Download Gambusia, Gambusia holbrooki

Document related concepts
Transcript
Assessing the recovery of fish communities
following removal of the introduced Eastern
Gambusia, Gambusia holbrooki
Z. Tonkin, J. Macdonald, D. Ramsey, A. Kaus, F. Hames, D. Crook,
and A. King
2011
Arthur Rylah Institute for Environmental Research
Technical Report Series No. 232
Arthur Rylah Institute for Environmental Research Technical Series No. 232
Assessing the recovery of fish communities following removal of the introduced
Eastern Gambusia, Gambusia holbrooki
Zeb Tonkin, Jed Macdonald, David Ramsey, Andrew Kaus, Fern Hames, David
Crook and Alison King
Arthur Rylah Institute for Environmental Research
123 Brown Street, Heidelberg, Victoria 3084
June 2011
In partnership with the Murray–Darling Basin Authority
Native fish recovery following Eastern Gambusia removal
Report produced by:
Arthur Rylah Institute for Environmental Research
Department of Sustainability and Environment
PO Box 137
Heidelberg, Victoria 3084
Phone (03) 9450 8600
Website: www.dse.vic.gov.au/ari
© Murray-Darling Basin Authority; State of Victoria, Department of Sustainability and Environment 2012
Citation: Tonkin, Z., Macdonald, J., Ramsey, D., Kaus, A., Hames, F., Crook, D. and King, A. (2012). Assessing the
recovery of fish communities following removal of the introduced eastern gambusia, Gambusia holbrooki. Arthur Rylah
Institute for Environmental Research Technical Report Series No. 232. Department of Sustainability and Environment,
Heidelberg, Victoria
ISSN 1835-3827 (print)
ISSN 1835-3835 (online)
ISBN ISBN 978-1-74287-410-4 (Print)
Disclaimer: This publication may be of assistance to you but the State of Victoria and its employees do not guarantee
that the publication is without flaw of any kind or is wholly appropriate for your particular purposes and therefore
disclaims all liability for any error, loss or other consequence which may arise from you relying on any information in
this publication. This work is shared copyright between Murray–Darling Basin Authority and Department of
Sustainability and Environment. Graphical and textual information in the work (with the exception of photographs and
the MDBA logo) may be stored, retrieved and reproduced in whole or in part, provided the information is not sold or
used for commercial benefit and its source (Murray–Darling Basin Authority and Department of Sustainability and
Environment) is acknowledged. Such reproduction includes fair dealing for the purpose of private study, research,
criticism or review as permitted under the Copyright Act 1968. Reproduction for other purposes is prohibited without
prior permission of the Murray–Darling Basin Authority or the individual photographers and artists with whom
copyright applies. To the extent permitted by law, the copyright holders (including its employees and consultants)
exclude all liability to any person for any consequences, including but not limited to all losses, damages, costs, expenses
and any other compensation, arising directly or indirectly from using this report (in part or in whole) and any
information or material contained in it.
The contents of this publication do not purport to represent the position of the Murray–Darling Basin Authority. They
are presented to inform discussion for improved management of the Basin's natural resources.
Funding for this project was provided by the Murray-Darling Basin Authority’s Native Fish Strategy program
Accessibility:
If you would like to receive this publication in an accessible format, such as large print or audio, please telephone
136 186, or through the National Relay Service (NRS) using a modem or textphone/teletypewriter (TTY) by dialling
1800 555 677, or email [email protected]
Front cover photo: Ovens River wetland (background), Seine net haul of Eastern Gambusia and adult female Carp
Gudgeon (inset: L-R).
Authorised by: Victorian Government, Melbourne
iii
Native fish recovery following Eastern Gambusia removal
Contents
List of tables and figures................................................................................................................. vi
Acknowledgements........................................................................................................................... x
Non-technical summary ................................................................................................................. 11
Project background and objectives................................................................................................... 11
Literature review .............................................................................................................................. 11
Cross sectional study ........................................................................................................................ 12
Trial of Eastern Gambusia removal.................................................................................................. 13
Cost-effectiveness and logistics of Eastern Gambusia removal ....................................................... 14
Conclusions, future research and management recommendations ................................................... 15
1
General Introduction and project objectives..................................................................... 16
2
A review of the impact of Eastern Gambusia on native fishes of the Murray–Darling
Basin 18
2.1 Eastern Gambusia: biology, ecology and distribution ........................................................... 18
2.2
2.1.1
Taxonomy and morphology.................................................................................... 18
2.1.2
Biology and ecology ............................................................................................... 19
2.1.3
Global distribution and spread into Australia and the MDB .................................. 22
Impacts of Gambusia as alien species.................................................................................... 23
2.2.1
Impacts on native fishes.......................................................................................... 23
2.2.2
Impacts on native fishes of the MDB ..................................................................... 26
2.3
Mitigating the impacts of Eastern Gambusia ......................................................................... 29
2.4
Knowledge gaps and the current project ................................................................................ 31
3
Phase 1: Cross-sectional study of wetland fish communities across the mid-Murray
region of the MDB .......................................................................................................................... 32
3.1 Introduction............................................................................................................................ 32
3.2
3.3
3.4
iv
Methods.................................................................................................................................. 33
3.2.1
Study area and site selection................................................................................... 33
3.2.2
Water quality and habitat assessment ..................................................................... 34
3.2.3
Fish sampling and processing................................................................................. 34
3.2.4
Statistical analysis and development of correlative models ................................... 37
Results.................................................................................................................................... 39
3.3.1
Assessment of wetland fish communities of the mid-Murray River ...................... 39
3.3.2
The influence of Eastern Gambusia on species occupancy and abundance ........... 40
3.3.3
The influence of Eastern Gambusia on juvenile condition of native species ......... 46
Discussion .............................................................................................................................. 49
3.4.1
Assessment of wetland fish communities of the mid-Murray River ...................... 49
3.4.2
The influence of Eastern Gambusia on species occupancy, abundance and
condition ................................................................................................................. 54
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
3.4.3
Hypotheses on the influence of Eastern Gambusia on wetland fish communities ..59
4
4.1
Phase 2: Trial of Eastern Gambusia control ......................................................................60
Introduction.............................................................................................................................60
4.2
Methodology...........................................................................................................................60
4.3
4.4
4.2.1
Site description and treatments................................................................................61
4.2.2
Removal of Eastern Gambusia ................................................................................64
4.2.3
Monitoring of fish biota...........................................................................................69
4.2.4
Mark-recapture assessment .....................................................................................69
4.2.5
Analysis ...................................................................................................................72
Results.....................................................................................................................................76
4.3.1
Site parameters ........................................................................................................76
4.3.2
Methodology assessment.........................................................................................79
4.3.3
Eastern Gambusia population parameters and effectiveness of removal ................84
4.3.4
Fish community monitoring and condition indices .................................................90
Discussion.............................................................................................................................103
4.4.1
Physical control of Eastern Gambusia...................................................................103
4.4.2
Eastern Gambusia as an invasive species: colonisation and population dynamics107
4.4.3
Fish community response to Eastern Gambusia removal......................................108
4.4.4
Conclusions ...........................................................................................................111
5
5.1
Phase 3: Cost-effectiveness and logistics of Eastern Gambusia removal.......................112
Introduction...........................................................................................................................112
5.2
Cost-effectiveness and logistics of Eastern Gambusia removal using physical control
programs ...............................................................................................................................113
5.3
Social benefits of Eastern Gambusia removal: participation and education.........................116
5.4
Evaluating the benefits to native fish and cost-effectiveness of controlling other potentially
harmful alien species across the MDB: applicability of the processes used in the current
study......................................................................................................................................118
6
6.1
Conclusion ...........................................................................................................................121
Management / Research Recommendations .........................................................................121
7
References............................................................................................................................123
Appendices .....................................................................................................................................137
Appendix 1: Eastern Gambusia marking trial .................................................................................137
Appendix 2: Project communications .............................................................................................144
v
Native fish recovery following Eastern Gambusia removal
List of tables and figures
List of tables
Table 2.1 Potential macro-habitat, micro-habitat and dietary niche overlaps between Eastern
Gambusia and native freshwater fish of the MDB................................................................. 27
Table 3.1 Small native fish species and all alien species recorded in floodplain environments
(billabongs, anabranches and inundated floodplain) of the mid-Murray River and tributaries
(Albury – Barham). ................................................................................................................ 35
Table 3.2 Raw abundances of each species collected during the 2009/2010 cross-sectional study
from each of the four regions surveyed (Gunbower, Lower Ovens, Albury (Wonga
wetlands) and Barmah-Millewa/Goulburn).. ......................................................................... 39
Table 3.3 The relative influence of each predictor variable on abundances of Eastern Gambusia,
Australian Smelt, Carp Gudgeon, Flat-headed Gudgeon and carp in the final BRT models
for each species. ..................................................................................................................... 43
Table 4.1. Treatment type, habitat and water quality parameters for each of experimental sites
recorded at the commencement of the experiment in 2009/2010 and 2010/2011 seasons.. .. 78
Table 4.2. Raw abundances of each species collected from removal and control sites, for each trip
using all removal methodologies during the field depletion experiment in season one......... 79
Table 4.4 Eastern Gambusia mark-recapture parameters collected during the start of the
experiment (spring) during each of the two years (two sites in each year) and end of the
experiment in the first year only (end of summer)................................................................. 86
Table 4.5 Parameter estimates from the Bayesian SSM for Eastern Gambusia (Gh), Carp Gudgeon
(Hsp), Australian Smelt and carp (Cc)................................................................................... 87
Table 4.6 Total numbers of each species collected from all sites during each year of the removal
trial. ........................................................................................................................................ 91
Table 4.7 Total numbers of each species collected from each site for all trips and during both
seasons of the experiment.. .................................................................................................... 93
List of figures
Figure 2.1 Mature female Eastern Gambusia. Photo: Tarmo Raadik.............................................. 18
Figure 2.2 High densities of Eastern Gambusia in a shallow backwater environment. .................. 20
Figure 2.3 Fin-nipping and weight loss to dwarf galaxias as a result of Eastern Gambusia
aggression and competition.................................................................................................... 25
Figure 3.2.1 Cross-sectional study site locations (red squares) sampled during February / March
2009/ 2010 including detailed locations of those sites sampled in the Lower Ovens River
floodplain. .............................................................................................................................. 33
Figure 3.2.2 Example of (a) low; (b) Medium; and (c) High levels of snags and/or debris in site
littoral zones........................................................................................................................... 34
Figure 3.2.3 Sweep net electrofishing snag / debris habitat in the littoral zone of a billabong on the
Ovens River floodplain. ......................................................................................................... 36
Figure 3.2.4. Examples of different levels of caudal fin damage including fins scored as ‘1’ (top),
‘2’ (middle) and ‘3’ (bottom) for (a) Carp Gudgeon; and (b) Australian Smelt. ................... 37
vi
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Figure 3.3.1 Co-occurrence of sampled species with the presence or absence of Eastern Gambusia
(Gh) and aquatic vegetation (Veg) in the wetlands. ...............................................................40
Figure 3.3.2 Probability of Eastern Gambusia occupying a wetland when aquatic vegetation is
present or absent. ....................................................................................................................41
Figure 3.3.3 The influence of different environmental covariates: wetland size (size), water
temperature (temp), conductivity (cond), dissolved oxygen (DO) and turbidity (turb) on the
probability of occupancy by Eastern Gambusia. ....................................................................41
Figure 3.3.4 Fitted functions describing relationships between Eastern Gambusia abundance and
all predictor variables for the final BRT model. ....................................................................42
Figure 3.3.5 Fitted functions describing relationships between Australian Smelt abundance and all
predictor variables for the final BRT model. ..........................................................................43
Figure 3.3.6 Fitted functions describing relationships between Carp Gudgeon abundance and all
predictor variables for the final BRT model. ..........................................................................44
Figure 3.3.7 Fitted functions describing relationships between Flat-headed Gudgeon abundance
and all predictor variables for the final BRT model. ..............................................................45
Figure 3.3.8 Fitted functions describing relationships between carp abundance and all predictor
variables for the final BRT model. .........................................................................................46
Figure 3.3.9 Juvenile condition of Australian Smelt, Carp Gudgeon and Flat-headed Gudgeon
plotted against Eastern Gambusia abundance. .......................................................................47
Figure 3.3.10 Plots showing the probability of fin damage for Australian Smelt, Carp Gudgeon,
and Flat-headed Gudgeon in relation to Eastern Gambusia abundance, the extent of aquatic
vegetation and the debris load across all sampled wetlands. ..................................................48
Figure 4.1.1 Examples of removal sites undertaken in the field depletion experiment ....................63
Figure 4.2.1 General timeline of the experiment for each of the two years, including exercises
carried out for each of the treatment (repeated and single removal) and control types (control
and reference).. .......................................................................................................................64
Figure 4.2.2 (a) Standard collapsible bait trap containing a solar light which automatically operates
between dusk and dawn, and recharges throughout the day; (b) Single-wing fine-mesh fyke
net. ..........................................................................................................................................66
Figure 4.2.3 (a) Adult Eastern Gambusia collected during a targeted seine shot; and (b) Trial of
artificial light and heat using 100 watt, halogen globe spotlights as a means of attracting
Eastern Gambusia.. .................................................................................................................68
Figure 4.2.4 Picture of adult southern pygmy-perch. Photographs were taken of rarer species for
assessment of caudal fin condition .........................................................................................69
Figure 4.2.5 (a) Field set-up for Eastern Gambusia osmotic induction marking with calcein
solution with example of resulting marked fish; and (b) field-based detection of marked
Eastern Gambusia with example of a marked fish alongside an unmarked individual...........71
Figure 4.3.1 Ovens river height at Peechelba bridge between July 2009 and March 2010. The red
dashed line represents the approximate river height at which experimental sites connect to
adjoining river or anabranch habitats. Grey blocks represent the removal and monitoring
periods of the first year of the field-depletion experiment......................................................76
Figure 4.3.2 Example of site variability during the course of the project. (a) Site 2 in late October,
and 3 months later at the end of January during the 2009/2010 season. (b) Hillgrove’s site at
vii
Native fish recovery following Eastern Gambusia removal
the commencement of the experiment, and following localised flooding in the area resulting
in site connection with adjoining creek and control site. ....................................................... 77
Figure 4.3.3 Eastern Gambusia catch per unit effort (CPUE) predictions across the four methods
utilised in the study for (a) all trips and; (b) trip one only. ................................................... 80
Figure 4.3.4 Catch per unit effort (CPUE) predictions from the four methods utilised in the study
across all trips for (a) Carp Gudgeon; (b) Australian Smelt; and (c) southern pygmy-perch. 81
Figure 4.3.5 Catch per unit effort (CPUE + SE; fish per hour) predictions for Eastern Gambusia
from each of the trapping variables applied for (a) bait traps and; (b) fyke netting (number of
observations = 33) across all trips. ........................................................................................ 82
Figure 4.3.6 Bait trap (left column) and fyke net (right column) catch per unit effort (CPUE+ SE;
fish per hour) predictions for (a) Carp Gudgeon; (b) Southern pygmy-perch and; (c)
Australian Smelt for each of the trapping variables across all trips....................................... 83
Figure 4.3.7 Mean (± SE) number of Eastern Gambusia extracted from removal sites each
consecutive day of trip one during the 2009/10 and 2010/11 seasons. .................................. 85
Table 4.3 Raw numbers of Eastern Gambusia collected from each of the removal and control sites
using physical removal techniques during the first trip (Spring) of each of the two years.... 86
Figure 4.3.8 Predicted trajectories for Eastern Gambusia abundance at ‘control’ (no removal) and
‘treatment’ (removal) sites from November – March (time 1-5) predicted for 1000 simulated
‘sites’ from the SSM model (includes predictions for one additional time period). ............. 87
Figure 4.3.9 (a) Female Eastern Gambusia collected during the first trip of the experiment (before
the onset of spawning). ......................................................................................................... 88
Figure 4.3.10 Length frequency histograms (% frequency) of Eastern Gambusia collected for each
trip (October/November; December; January and February/March) for both the 2009/10 and
2010/11 seasons of the experiment. ....................................................................................... 89
Figure 4.3.11 Total raw number of fish extracted (bars) and the cumulative percentage of
population removed (line) based on mark-recapture data from Colbinabbin wetland over
eight days of removal exercises before the onset of spawning (trip 1, season 2)................... 89
Figure 4.3.12 Examples of species collected at sites during the field removal experiment. Native
species (LHS) from top to bottom – Murray–Darling rainbowfish, Carp Gudgeon, southern
pygmy-perch and flat-headed galaxias; and alien species (RHS – carp, Eastern Gambusia,
goldfish and oriental weatherloach). ...................................................................................... 92
Figure 4.3.13 Length frequency histograms (% frequency) for the five most common native
species collected for each trip (October/November; December; January and
February/March) of the experiment (both seasons combined)............................................... 94
Figure 4.3.14 Length frequency histograms (% frequency) of carp and goldfish collected for each
trip (October/November; December; January and February/March) of the experiment (both
seasons combined). ................................................................................................................ 95
Figure 4.3.14 The observed abundance of Eastern Gambusia (Gh), Carp Gudgeon (Hsp),
Australian Smelt (Rs) and carp (Cc) at each of the 4 time periods (November – February). 96
Figure 4.3.15 Trajectories of Eastern Gambusia (Gh), Carp Gudgeon (Hsp), Australian Smelt (Rs)
and carp (Cc) abundance from November – February (trip 1-4) predicted by the SSM
(dashed lines and small solid circles) overlaid on the observed trajectories (solid lines and
open circles). .......................................................................................................................... 97
Figure 4.3.16 Mean trajectories from 1000 simulated ‘sites’ for Eastern Gambusia (Gh), Carp
Gudgeon (Hsp), Australian Smelt (Rs) and carp (Cc) abundance from November – March
viii
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
(trip 1-5) predicted by the SSM including (includes predictions for one additional time
period).....................................................................................................................................98
Figure 4.3.17 Predicted effects of different abundances of Eastern Gambusia on the population
growth of Carp Gudgeon (Hsp), Australian Smelt (Rs) and carp (Cc) based on 1000
simulated ‘sites’. .....................................................................................................................99
Figure 4.3.18 General linear model outputs (± 95% confidence intervals) examining the
probability of fin damage (Top) and; relative condition (Krel; Bottom) of juvenile Australian
Smelt, Carp Gudgeon and Flat-headed Gudgeon in relation to Eastern Gambusia abundance
(log10abundance) at all sites from December - February. .....................................................101
Figure 4.3.19 Juvenile southern pygmy-perch collected in January 2010. Note the damage of the
caudal fin...............................................................................................................................102
Figure 4.3.19 Individual southern pygmy-perch lengths (mm) collected in each month (trips 1-4)
from Site 19 during the first year of the study. Fish < 30mm represent cohort of juvenile
fish. .......................................................................................................................................102
Figure 5.1 Decision support tool aimed at prioritising sites for physical control or Eastern
Gambusia on the basis of maximising the ecological benefits per dollar invested. .............114
Figure 5.2 Demonstration and information on Eastern Gambusia and wetland fish communities
presented to Colbinabbin primary school. ...........................................................................118
Figure 5.3 Template of the processes utilised in the current study to maximise the benefits to
native fish arising from an alien fish removal program. .......................................................120
ix
Native fish recovery following Eastern Gambusia removal
Acknowledgements
The project team thanks members of the steering committee, Heleena Bamford (MDBA), Renae
Ayres, John Koehn (DSE), Dale McNeil (SARDI), Mark Lintermans (UC), Jamie Knight (DII),
Dave Maynard (AMC Tasmania), Wayne Fulton and Kylie Hall (Vic DPI) for their valuable
guidance of the project. Thanks also to Justin O’Mahoney, Scott Raymond, Dean Hartwell, Jason
Lieschke, Joanne Kearns and David Semmens for contribution to field work; Thanks to John and
Glenys Avard, for access to private land and their generous hospitality. Thanks also to Steve
Saddlier and Wayne Koster for internal review of this document. This work was funded by the
Native Fish Strategy section of the Murray-Darling Basin Authority, MDBA Project No. MD1043.
This study was conducted under the following permits:
• DSE Animal Care and Ethics Approval (Permit No. AEC 08/07)
• NSW DPI Scientific Collection Permit (Permit No. P07/0115-1.0)
• Vic DPI Fisheries Collection Permit (Permit No. RP827)
x
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Summary
Project background and objectives
•
Alien fish species have been recognised as one of eight major threats to native fish in the
Murray–Darling Basin (MDB), and the control of these species is one of the key drivers of
the Native Fish Strategy. There is growing evidence of detrimental impacts of Eastern
Gambusia Gambusia holbrooki on native fish fauna globally, and it has been identified as
one of the key alien species contributing to the decline of a number of native fish within
the MDB, where it is widespread.
•
Understanding the ecological impacts of an alien species is an essential component of
vertebrate pest management. Unfortunately, the detrimental ecological impacts of the
Eastern Gambusia in the MDB remain uncertain. If we also consider the present lack of
effective control options, there is an urgent need for research into the feasibility of
reducing Eastern Gambusia populations to densities where measurable improvements to
native fish communities can be detected.
•
This project addressed these research needs by integrating surveys and quantitative
experimental work in natural billabong systems throughout the MDB. The specific
objectives of the project were to:
-
review current knowledge of the impacts of Eastern Gambusia on native fishes of
the MDB,
-
provide information on the response of native fish communities following the
reduction of Eastern Gambusia populations, and
-
provide a framework to evaluate the feasibility and effectiveness of such control
actions, and form a template for evaluating control options for other alien fishes
across the MDB.
Literature review
•
The literature review was compiled by searches of the Web of Science, Scopus and
Aquatic Sciences and Fisheries Abstract databases, the FishBase database, and from the
knowledge of experts in government and non-government organisations throughout
Australia. These searches found several hundred published scientific articles and
unpublished documents pertaining to the process of invasion by alien fishes, the biology,
ecology, distribution and ecological impacts of Gambusia species around the world, in
Australia and in the MDB, and options for, and effects of mitigating such impacts. The
detailed literature review has been published as MDBA Publication No. 38/09 (Macdonald
and Tonkin 2008). Three of the most significant finsings are outlined here.
•
Sixteen of the 37 native freshwater species have major niche overlaps (habitat, diet, or
both) with the Eastern Gambusia, and thus are at greatest risk. The review highlighted a
number of key families, in particular the ambassids (glassfish), nannopercids (pygmyperches), melanotaenids (rainbowfishes), atherinids (hardyheads), eleotrids (gudgeons)
and retropinnids (smelt), which make up the majority of the MDB’s wetland fish
communities.
•
The Eastern Gambusia is likely to have contributed to the decline in the distribution or
abundance (or both) of the olive perchlet, southern pygmy-perch, Murray–Darling
rainbowfish and purple-spotted gudgeon. Conversely, some species that have major niche
Arthur Rylah Institute for Environmental Research
11
Native fish recovery following Eastern Gambusia removal
overlaps with Eastern Gambusia, such as Australian Smelt Retropinna semoni and Carp
Gudgeons Hypseleotris spp., remain relatively widespread throughout the MDB, most
likely because of their highly flexible trophic niches that buffer them from the impacts of
competition. Such species are still likely to suffer localised impacts, particularly in areas
of limiting resources such as habitat availability.
•
In addition to the direct impacts that Eastern Gambusia is likely to have on the MDB’s
smaller species (occupying, at least in part, slow or still water habitats), there could be
indirect impacts on larger riverine species, particularly if Eastern Gambusia creates
energetic blocks within aquatic food webs.
Cross-sectional study
•
12
Accurately defining the level of risk (by quantifying the potential detrimental, neutral or
even positive impacts of invasion on ecosystem processes or particular species) can
provide a strong platform for generating and testing hypotheses about the level of control
(if any) required to achieve an acceptable level of risk. Therefore, the first phase of the
project involved a broad-scale, cross-sectional assessment of wetland fish communities
using standardised electrofishing and netting surveys throughout the mid-Murray region of
the MDB. This information was used to:
-
examine how the presence or absence of Eastern Gambusia influences the
probability of occupancy of native and alien fish species in wetlands across the
MDB
-
explore patterns of co-occurrence of native fish species and Eastern Gambusia in
these systems, and
-
quantify the relationships between Eastern Gambusia abundance and the
abundance and condition of sympatric fish species in differing habitat types and
with variability in environmental covariates.
•
A total of 93 sites were sampled over a two-year period. Carp Gudgeon (native species)
and Eastern Gambusia were the dominant species recorded, comprising of 43 and 44% of
the catch and occupying 90 and 83% of the sites respectively. Some shifts in fish
community composition were observed when compared with past surveys, including the
absence of flat-headed galaxias Galaxias rostratus at all sites, the restriction of southern
pygmy-perch Nannoperca australis to the lower Ovens River floodplain, and the presence
(and dominance at some sites) of Flat-headed Gudgeon Philypnodon grandiceps at
particular Ovens River sites.
•
The results suggest that Eastern Gambusia does not have a negative influence on the
occurrence or abundance of common species such as Carp Gudgeon and Flat-headed
Gudgeon, even where Eastern Gambusia density is high. Increasing Eastern Gambusia
abundance does, however, appear to decrease the body condition of juveniles of these
species. There is also some indication that greater habitat complexity (in the form of
aquatic vegetation and debris) limits the number of attacks of Eastern Gambusia on Carp
Gudgeon, as measured by the degree of fin damage in that species.
•
There was a slight negative relationship between the abundances of Australian Smelt and
Eastern Gambusia. In addition, an assessment of the condition of recruits of the common
native species (using two indices: length–weight relationships and fin condition) indicates
that there is a negative relationship between each condition index for juvenile Australian
Smelt and the abundance of Eastern Gambusia. The effects on rarer species such as
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
pygmy-perch and rainbowfish could not be assessed because too few of these species were
collected.
•
One of the more interesting results from the study was that juvenile Common Carp
Cyprinus carpio occurred in low numbers with Eastern Gambusia in the sampled
wetlands. We suggest that this may be due to predatory or competitive interactions
between the species, which would peak during late spring when early-stage carp larvae
would be exposed to pressures from an increasing density of Eastern Gambusia.
•
The results of this exploratory study enabled us to develop hypotheses about how native
fish might respond if Eastern Gambusia densities were reduced in wetland systems. These
hypotheses were explored in greater detail in a field removal trial. Specifically, we tested
the following hypotheses:
-
Removing Eastern Gambusia (to any level) will benefit Australian Smelt
populations, as measured by (a) increases in relative abundance, and (b)
improvement in morphometric condition.
-
Removing Eastern Gambusia (to any level) will impart minimal changes to Carp
Gudgeon and Flat-headed Gudgeon abundances.
-
Removing Eastern Gambusia (to any level) will improve morphometric condition
for carp and Flat-headed Gudgeon.
-
Removing Eastern Gambusia (to any level) will impart minor increases to carp
abundances.
-
Removing Eastern Gambusia (to any level) will benefit rarer wetland species such
as southern pygmy-perch, flat-headed galaxias and rainbowfish populations, as
measured by (a) increases in relative abundance, and (b) improvement in
morphometric condition.
Trial of Eastern Gambusia removal
•
In the second phase of the project, Eastern Gambusias were removed from small isolated
billabongs to test the hypotheses derived from the cross-sectional study, and to provide
information on control options and Eastern Gambusia population dynamics.
Unfortunately, extremes in environmental variables limited the analysis and conclusions
that could be drawn from the field removal trial, but it still provided important information
on Eastern Gambusia removal, population dynamics and native fish responses.
•
Eastern Gambusias were removed over a 5–10 day period (before the onset of the Eastern
Gambusia spawning season), resulting in major reductions in abundance (generally over
40%), and even eradication at several sites. This indicates that, under certain conditions,
direct removal could achieve major reductions in Eastern Gambusia populations.
However, determining the likelihood of success would require a thorough consideration of
various aspects of a site’s hydrology, climate, habitat and size.
•
During this trial, Eastern Gambusia displayed an astonishing capacity to rapidly colonise
habitats; a few individuals were able to re-establish populations of thousands in three or
four months. The fact that the intrinsic rate of increase of Eastern Gambusia populations
was far higher than even the most common native species in the region emphasises the
species’ ability to out-compete native species.
13
Native fish recovery following Eastern Gambusia removal
•
Most importantly, the results indicate that reductions of Eastern Gambusia abundances
will result in improvements to small-bodied native fish populations. The negative impacts
on the more common generalist species were relatively minimal and were probably a
result of the intact nature of the floodplain wetland sites used. There was some indication
that negative impacts may be far greater on species with limited trophic niches. (A large
proportion of these species have already suffered major reductions in range and
abundance.) This suggests that, in the short term, management and control of Eastern
Gambusia should focus on sites containing these native species, or sites with highly
uniform habitats (which are often degraded sites). Further field studies assessing the
response of the rarer specialist species and longer-term monitoring of fish communities at
more degraded sites after Eastern Gambusia removal is required to determine whether
these predictions are valid.
•
The results also suggest that Eastern Gambusia removal may result in small increases in
Common Carp populations, and it is possible that removal may have unexpected benefits
for other alien species. Site-specific ecosystem function in the absence of Eastern
Gambusia must therefore also be considered before undertaking a removal program.
Cost-effectiveness and logistics of Eastern Gambusia removal
The third phase identified strategies to maximise the improvement to native fish
communities through Eastern Gambusia control, given a fixed budget (benefit
maximisation), and to minimise the cost of achieving a defined significant improvement in
the native fish community (cost minimisation). The factors influencing overall costs and
the effectiveness of a physical control program at a specific site could be generalised to
four simple variables relating to a site’s hydrological connectivity, ecological value,
habitat complexity and size. Considering these factors in a methodical manner enables
managers to identify sites where the control of Eastern Gambusia would achieve the
maximum ecological benefit per dollar invested, or would minimise the cost of achieving
a defined ecological benefit.
All sites
Does the site facilitate permanent immigration and emigration?
Isolated sites
Is the site of high ecological value
(species / habitat)?
Low ecological value
High frequency
of connection
How frequently does the site connect to
adjoining waterbodies?
Low frequency
of connection
High ecological
value
How frequently does the site connect
to adjoining waterbodies?
Low frequency of
connection
How much structural habitat
does the site contain?
High frequency of
connection
How much structural habitat
does the site contain?
High structural
complexity
High structural
complexity
Low structural
complexity
Large surface
area
What is the size of the site?
Small surface
area
Low structural
complexity
Large surface
area
What is the size of the site?
Small surface
area
14
Very Low
Permanently connected
sites facilitating constant
immigration of pest fish
Permanently
connected
Arthur Rylah Institute for Environmental Research
Benefits per $ invested
•
Low
Isolated sites of low ecological
value and frequent connection
to adjoining habitats facilitates
frequent immigration of pest
fish
Medium
Isolated sites of high ecological
value but frequent connection
to adjoining habitats. Value of
investment increases with
reduced structural habitat
complexity and surface area
High
Isolated sites of high ecological
value and infrequent
connection to adjoining
habitats. Value of investment
increases with reduced
structural habitat complexity
and surface area due to a
reduction in required effort;
increased ability to undertake
active netting methods and
increased negative interaction
between pest and native
species due to a reduction in
habitat niche partitioning.
Native fish recovery following Eastern Gambusia removal
•
We developed a simple decision support tool for managers who might be considering
investing in a control program in their jurisdiction. By considering a few primary factors,
the tool enables managers to assign individual sites a general rating (from very low to
high) based on ecological benefit per dollar invested, ultimately enabling a decision on
whether or not to employ a control program, and the sites that a program should target. If
sites are likely to experience very low or low ecological benefits for a given investment,
alternative mitigation activities such as habitat restoration should still be investigated.
•
We found that there is an alarming lack of community awareness about Eastern Gambusia.
We strongly recommend that community awareness and education is incuded in all
Eastern Gambusia management programs. Community participation is a possibility
because the methods for removing Eastern Gambusia could easily be utilised by
community bodies such as Landcare and school groups. Even if this did not result directly
in immediate ecological benefits, it would still result in substantial social benefits,
particularly in light of the lack of education on the threats posed by Eastern Gambusia.
Conclusions, future research and management recommendations
•
Understanding the ecological impacts of an alien species is an essential component of pest
management, based on the concept of managing impacts rather than simply reducing
numbers. This project has taken a step forward in researching the feasibility of controlling
Eastern Gambusia populations to achieve measurable improvements to native fish
communities. The project provided important information on Eastern Gambusia removal,
population dynamics and native fish responses to the removal.
•
Reductions of Eastern Gambusia will result in improvements to native fish populations,
but management actions should focus on sites containing native species with limited
trophic niches, or sites containing highly uniform habitats, to maximise the ecological
benefits.
•
The removal of Eastern Gambusia, if conducted under certain conditions, can be used as
an effective management tool to achieve major reductions in Eastern Gambusia
populations, but the degree of success in reducing the density of Eastern Gambusia and
maximising the ecological response requires a thorough consideration of various aspects
of a site, hydrology, ecological value, climate, habitat and size.
•
Further field studies to assess the response of rarer specialist species, and longer-term
monitoring of fish communities across different habitat conditions after Eastern Gambusia
removal, are required to refine these predictions.
•
Because the removal of Eastern Gambusia may have unexpected benefits for other alien
fish species, ecosystem function in the absence of Eastern Gambusia must also be
considered before implementing removal programs.
•
Including community awareness, education and participation in Eastern Gambusia
management programs will result in substantial social benefits, particularly in light of the
lack of general public awareness about this alien fish species.
•
We suggest that the simple process used in the study could also prove valuable if applied
to other established alien fish species, particularly those with limited socio-economic
value.
15
Native fish recovery following Eastern Gambusia removal
1
General introduction and project objectives
Invasion by alien species is a primary threat to global biodiversity (Clavero and Garcia-Berthou
2005; Leprieur et al. 2008). Growth in international trade and concurrent increases in transport
capacity have accelerated the rate of introduction of alien species worldwide, with freshwater
ecosystems and their native fish communities being particularly susceptible (Vitousek et al. 1997;
Sala et al. 2000; Marchetti et al. 2004; Gozlan 2008). Alien fishes are implicated in the
displacement, reductions in abundance, distributional range and condition, local extirpation and
extinction of many native fish species worldwide (see Moyle and Light 1996; Amundsen et al.
1999; Irons et al. 2007). They can alter ecosystem function, affect genetic integrity (Weigel et al.
2002) and transmit parasites and disease (Gozlan et al. 2005), potentially resulting in high
ecological, economic and social damage (Pimentel et al. 2000, 2005, see Rowe et al. 2008).
In the last 30 years the number of freshwater fish species translocated outside their natural range
has more than doubled (see Williamson 1996), with the increase in freshwater aquaculture
production identified as a major cause (De Silva 2006; Gozlan 2008). In an analysis of the Food
and Agriculture Organization’s database and FishBase, Gozlan (2008) reported that 624 species of
fish throughout the world have now been introduced outside their natural range. Fifty-one per cent
of these introductions were for aquaculture production, 21% for the ornamental market, 12% for
angling or sport fishing, and the remainder for biological control or to fill an ecological niche.
Gozlan (2008) estimated the risk of ecological impact from the introduction of a freshwater fish
species is less than 10% for 84% of introduced fishes. This estimate is quite similar to that
provided by Simberloff (2007) or the ‘tens rule’ of Williamson (1996), which predicts that 10% of
introductions will become established, and 10% of these in turn will have an ecological impact on
the recipient ecosystem (but see Gherardi 2007).
In their synthesis of global patterns of freshwater fish invasion, Leprieur et al. (2008) identified
Australian freshwater systems as one of six major invasion hotspots where alien fish species
represent more than a quarter of the total number of species present. Australian freshwater systems
currently harbour 34 known alien fish species (Lintermans 2004), 11 of which are established in
the Murray–Darling Basin (MDB) (Lintermans 2007). Four of these 11 (including eastern
gambusia, Gambusia holbrooki) are classified in the top eight ‘worst’ invasive fish taxa by the
International Union for Conservation of Nature (Lowe et al. 2000; see also Koehn and Mackenzie
2004; Fausch 2007).
The ecological consequences of invasion by these alien fishes have been a major focus of research
in Australian freshwater systems (e.g. Roberts et al. 1995; Arthington and McKenzie 1997; King et
al. 1997; Howe et al. 1997; Jackson et al. 2004). Yet until quite recently the potential economic
and social costs and benefits of such invasions have been mostly overlooked (but see Bomford and
Hart 2002; McLeod 2004; West et al. 2007). In a recent comprehensive review, Rowe et al. (2008)
used a triple-bottom-line approach to assess the environmental, economic and social impacts of six
species of alien fish present in Australian freshwaters, including Eastern Gambusia.
The Murray–Darling Basin Authority’s Native Fish Strategy promotes the principles of integrated
pest management to mitigate such threats, and recently funded a three-year study (Project
MD1043: Native fish recovery following the removal of alien fish species) to explore the relative
economic costs of alien species control actions in relation to the ecological benefits to native fish
communities.
The Arthur Rylah Institute for Environmental Research was commissioned to undertake this
project, with the following objectives:
1. Review current knowledge of the impacts of Eastern Gambusia on native fishes of the MDB.
16
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
2. Provide information on the response of native fish communities following the reduction of
Eastern Gambusia populations.
3. Provide a framework to evaluate the feasibility and effectiveness of such control actions and,
form a template for evaluating control options for other alien fishes across the MDB.
17
Native fish recovery following Eastern Gambusia removal
2
A review of the impact of Eastern Gambusia on native
fishes of the Murray–Darling Basin
This chapter is a shortened version of the comprehensive review of the impact of Eastern
Gambusia on native fishes of the Murray–Darling Basin that was undertaken to address the first
objective of the project (Macdonald and Tonkin 2008).
2.1 Eastern Gambusia: biology, ecology and distribution
There is an extensive array of literature available on Eastern Gambusia and Western Gambusia
(Gambusia affinis) because of their widespread distribution, high abundance, ease of capture and
maintenance, and mixed attitudes towards them (Pyke 2005). McKay et al. (2001), Pyke (2005,
2008), Maynard et al. (2008), and Rowe et al. (2008) have presented extensive reviews on Eastern
Gambusia, including its taxonomy, biology and ecology. These reviews complement numerous
other publications about this species (e.g. McKay 1984; Meffe and Snelson 1989; Lintermans
2007). The following is a brief summary of this information to help understand the subsequent
chapters of the report.
2.1.1
Taxonomy and morphology
Family:
Poeciliidae
Scientific name:
Gambusia holbrooki (Girard, 1859)
Common names:
Eastern Gambusia, gambusia, mosquito fish, plague minnow, top minnow
Eastern Gambusia is a small (< 60 mm) poeciliid fish distinguished by a stout body, large cycloid
scales and a flattened upper head with a small up-turned mouth (Karolak 2006). It has strong
conical teeth and a shortened oesophagus and intestine, which are typical traits of predatory fish
(Pyke 2005). The colour is usually olive-brown on the back, blue-grey on the sides and whitesilver on the underside (Lintermans 2007). It is a sexually dimorphic species, with females larger
and much deeper bodied than males and have a large dark spot near the vent (Figure 1). Males,
which stop growing after reaching maturity, are generally slimmer and have a slender, elongated
anal fin that is used in copulation (McKay et al. 2001; Karolak 2006).
Figure 2.1 Mature female Eastern Gambusia. Photo: Tarmo Raadik.
18
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
2.1.2
Biology and ecology
Eastern Gambusia generally live for only a few months and die in the same season in which they
mature (McKay et al. 2001). However, females maturing towards the end of the season have been
recorded living up to 15 months (Cadwallader and Backhouse 1983, McKay et al. 2001), and
Karolack (2006) reported a lifespan of up to three years for the species. Males and females
generally mature when the body length is 17–20 mm, at one to two months of age (Pyke 2005).
Growth and maturation rates are highest when water temperatures are 25–30 °C (Pyke 2005).
Like most poeciliines, fertilisation in Eastern Gambusia is internal and young develop inside the
mother until they are born as free swimming fish (Parenti and Rauchenberger 1989; Pyke 2005).
Both male and female fish have an annual reproductive cycle with a distinct breeding season from
spring until autumn, peaking during the warmest times of the year. Females can store viable sperm
in their oviducts for several months, giving them considerable flexibility in the timing of egg
fertilisation (Pyke 2005). A single previously fertilised female can colonise a new site, but
seasonal timing and duration of spawning is strongly governed by water temperatures and day
length: temperatures over 16 °C and day lengths over 12–13 hours are required to initiate
spawning (Pyke 2005).
Female Eastern Gambusia can have several broods in a single breeding season; large females are
known to have up to nine broods per season in the wild (Milton and Arthington 1983), although 2–
5 broods is more typical (Pyke 2005). The clutch size of each brood is extremely variable and
depends on age, time of season, food availability, female size, and geographic location (Pyke
2005). Clutches are typically around 50, but clutches as large as 375 and as small as one have been
reported (Cadwallader and Backhouse 1983; Milton and Arthington 1983; Rowe et al. 2008). The
gestation period for each brood can range from 15 to 50 days, depending on the water temperature,
but is usually around 22–25 days (Milton and Arthington 1983). Because females cannot be
fertilised until after a litter is released and there is a delay of 2–14 days between birth and
fertilisation, broods can be produced every three to four weeks (Pyke 2005). Maglio and Rosen
(1969) reported that on average, females produced a brood every 25 days. The average number of
broods and clutch size of a population can be used to estimate its total life-time fecundity (Rowe et
al. 2008). Ignoring factors such as predation and resource availability, Maglio and Rosen (1969)
calculated that 10 adult females could produce 5 million individuals in six months.
Eastern Gambusia often assumes plague proportions in favourable habitats because of its rapid
breeding ability (Lintermans 2007). Rowe et al. (2008) reported large aggregations of hundreds of
fish per square metre can occur in surface waters of lakes and ponds during summer (Figure 2).
Much lower densities are observed in the winter, although it is unclear whether this is a result of
reduced population sizes caused by mortality, or sheltering behaviour that makes the fish less
observable (Pyke 2005).
Like may successful alien fish species, Eastern Gambusia can exist in a range of habitat types,
including large rivers, creeks, wetlands, lakes, channels and bores. It is a poor swimmer and
prefers still waters to flowing waters (Rowe et al. 2008), and therefore is most commonly found in
areas such as wetlands, weir pools, lakes and backwaters (Figure 3). Eastern Gambusia tends to
prefer shallow areas (often less than 15 cm deep) within these macrohabitats, mostly around the
littoral margins, in surface waters or among freshwater plants (Karolak 2006; Lintermans 2007;
Rowe et al. 2008). Stoffels and Humphries (2003) reported that larger fish preferred the benthic
areas around macrophyte beds. Water velocity barriers such as those formed by rapids, chutes and
falls limit its upstream penetration because it is not able to tolerate fast-flowing areas (Rowe et al.
2008).
19
Native fish recovery following Eastern Gambusia removal
Figure 2.2 High density of Eastern Gambusia in a shallow backwater environment. Photo: Tarmo
Raadik.
Eastern Gambusia is not known to undertake active migrations; individuals generally remain
within relatively small areas (Pyke 2005; Rowe et al. 2008). Exceptions include downstream
displacement during flooding, and perhaps seasonal movement to deeper water before the onset of
winter (Pyke 2005). However, Lyon et al. (2010) reported movement of Eastern Gambusias
between the main river channel and off-channel environments, with the highest numbers of fish
moving during the day. It is not known whether this was an active migration or, given their diurnal
feeding pattern (see below), just part of daily feeding behaviour.
Eastern Gambusia have a remarkable ability to withstand adverse conditions (McKay et al. 2001).
The species is extremely tolerant of poor water quality, particularly high turbidity, extremes of
temperature and salinity ranges, and low dissolved oxygen (Karolak 2006), reflecting its success as
an invasive species in Australian floodplain habitats. While Eastern Gambusia can tolerate a wide
range of temperatures (1.8–38 °C), its preference and reproductive requirement for high
temperatures (see above) indicate it is a warm water species (Pyke 2005; Rowe et al. 2008).
McNeil and Closs (2007) reported Eastern Gambusia’s extreme tolerance of low dissolved oxygen
levels in both controlled conditions and MDB floodplain habitats, demonstrating the species’
ability to comfortably utilise aquatic surface respiration in conditions ranging from severe hypoxia
to anoxia. Those authors suggested that the species’ flattened head and upturned mouth indicate
Eastern Gambusia is morphologically adapted for this aquatic surface respiration.
Disturbed habitats are particularly susceptible to Eastern Gambusia invasion because of their
extreme tolerance of poor water quality (Arthington et al. 1983; Kennard et al. 2005; King and
Warburton 2007). This is evident in many urban systems that have suffered extreme habitat
alteration. For example, habitat alteration and subsequent water pollution have contributed to the
decline of native fishes and successful establishment of Eastern Gambusia in urban Brisbane
waterways (Arthington et al. 1983). The construction of dams and weirs reduces water discharge
and subsequent flow velocities, thus also creating additional favourable habitat for Eastern
20
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Gambusia (McKay et al. 2001). On the other hand, undisturbed lotic systems with naturally
variable discharge regimes are not favoured by Eastern Gambusia, and higher river discharges
reduce their populations and may even almost eliminate them (Arthington and Lloyd 1989;
Arthington et al. 1990; Chapman and Warburton 2006).
Eastern Gambusia exhibits both social and anti-social behaviour (Pyke 2005). On one hand it is a
schooling fish, often occurring in large aggregations. On the other hand it is well known for both
intraspecific and interspecific aggression towards other fish, often chasing and nipping the fins of
fish much larger than itself (Lintermans 2007). The effect of this behaviour is discussed in section
3.3.
Eastern Gambusia feeds during daylight hours and relies on sight to detect and attack prey
(Swanson et al. 1996; McKay et al. 2001). It is primarily carnivorous (as indicated by the dentition
and digestive organs), with a generalist diet that includes a range of aquatic macro-invertebrates,
terrestrial insects and arachnids, and the early life stages of fish and anurans (Arthington and
Marshall 1999; Ivantsoff and Aarn 1999; Stoffels and Humphries 2003; Pyke 2005). Cannabalism
of young is also known to occur (Maglio and Rosen 1969; Maynard et al. 2008). Eastern
Gambusia is best described as an opportunistic or generalist omnivore, as the species has also been
reported to consume filamentous algae, fragments of fruit and other plant tissue (Arthington and
Marshall 1999; McKay et al. 2001; Maynard et al. 2008). McDowall (1996) described Eastern
Gambusia as an adaptable generalist predator than is able to vary its diet according to the available
prey. For example, researchers conducting seine netting surveys in a floodplain billabong observed
large female Eastern Gambusia attacking and then ingesting juvenile Carp Gudgeon that were
recovering after being released after sampling (Tonkin pers. obs.). This generalist or opportunistic
nature of Eastern Gambusia’s feeding habits is another reason for its success as an invasive
species.
The flattened head and upturned mouth of Eastern Gambusia, coupled with its tendency to occupy
surface waters, indicates that a large proportion of its diet is sourced from or near the surface.
Because of its relatively small mouth, its prey is typically small, but prey size increases with
increasing fish size (Pyke 2005) and so larger females can be expected to have a wider diet.
Although mosquito larvae are consumed by Eastern Gambusia, they make up a very small
proportion of their overall diet (Arthington and Lloyd 1989). Furthermore, although Eastern
Gambusia is used extensively throughout the world as a mosquito control agent, there are no
striking examples of its effectiveness in reducing mosquito populations (Lloyd 1990).
In its native range Eastern Gambusia is susceptible to numerous predators, parasites and diseases
(Swanson et al. 1996). In Australia, however, it appears there are very few control agents. Other
than the introduced copepod Lernia, few species parasitise Eastern Gambusia in Australia
compared to native species, which may carry at least two or three parasite species. This relatively
light parasite burden may have also contributed to its success as an invader in Australia (Lloyd
1990).
There has been little work on predation of Eastern Gambusia in Australia, but predators are likely
to include birds such as cormorants and egrets (e.g. Boulton and Brock 1999, cited in Rowe et al.
2008), aquatic mammals (e.g. Lloyd 1987), invertebrates such as crayfish (e.g. Beatty 2006) and
fish. In particular, another introduced fish, redfin perch (Perca fluviatilis), is thought to prey
heavily on Eastern Gambusia, and Stoffels and Humphries (2003) and McNeil (2004) suggested
that the presence of redfin perch may govern the densities of Eastern Gambusia in Australian
floodplain billabong habitats. The level of predation on Eastern Gambusia by these various
organisms is largely unknown. For example, while the introduced redfin perch has been
documented to influence Eastern Gambusia numbers by direct predation, both native and alien fish
21
Native fish recovery following Eastern Gambusia removal
predators prefer other prey when given a choice, so that Eastern Gambusia could still establish
large populations in the presence of large predator populations (Lloyd 1990).
2.1.3
Global distribution and spread into Australia and the MDB
Eastern Gambusia is native to south-eastern United States and northern Mexico (McKay et al.
2001), but it was not until the genus was recognised as a potential biological control agent that it
became one of the most widespread in the world. The genus is now present on all continents
except Antarctica (Courtenay and Meffe 1989).
After the discovery that mosquitoes were responsible for the transmission of diseases such as
malaria and yellow fever, health authorities began searching for mosquito control techniques,
including the new concept of biological control (Lloyd 1990; McKay et al. 2001). During the early
1900s, scattered anecdotal evidence of Eastern Gambusia successfully controlling mosquito
populations led to their widespread introduction as a biological control, despite limited research
into their capabilities or the utility of native species as mosquito control agents (Lloyd 1990;
McKay et al. 2001). This resulted in the spread of Eastern Gambusia into freshwater environments
throughout the world, particularly from 1920 to 1940, including the first introduction into
Australia in 1925 (Rowe et al. 2008).
Shortly after the initial introduction of Eastern Gambusia into the Botanic Gardens, Sydney, the
species spread throughout New South Wales to the point that it was established in most of the state
by the early 1940s (McKay et al. 2001). At this time it was released in northern Queensland, parts
of the Northern Territory, South Australia, Victoria and Western Australia, all for the purpose of
mosquito control (McKay 1984; Maynard et al. 2008). Eastern Gambusia was recorded in
Tasmania in 1992, believed to be a result of an unauthorised release into a private dam (Maynard
et al. 2008). Arthington and McKenzie (1997) reported that the species was still being spread
about Australia for mosquito control during the late 1990s, despite substantial evidence that native
species were more effective at mosquito reduction than Eastern Gambusia (Lloyd 1990). Most
states have since adopted legislation that has meant the species is no longer advocated as a
mosquito control agent (see Chapter 5.1) and therefore unlikely to be intentionally distributed.
Further spread of the species is most likely to occur during widespread flooding, dispersing
individuals from existing populations (e.g. McKay 1984) and from populations in irrigation
channels and bore drains (McKay et al. 2001).
Eastern Gambusia is now present in all states and territories, having established populations
throughout most of the major drainage divisions in the country (Rowe et al. 2008). This includes
most of the eastern drainage division, from Port Douglas in northern Queensland to Adelaide in
South Australia (McDowall 1996; Arthington and McKenzie 1997; Rowe et al. 2008). In Western
Australia they are extremely abundant in the south-west and southern Pilbara, and a small
population has been discovered in an isolated region of the Kimberley (Morgan et al. 2004).
Several populations have been discovered in Darwin, several eastern catchments draining into the
Gulf of Carpentaria, and in central Australia’s Lake Eyre drainage region (Rowe et al. 2008).
Isolated populations have also become established in northern Tasmania’s Tamar basin (Maynard
et al. 2008). Eastern Gambusia is now present in all river basins of the MDB (MDBC 2008) and is
common and frequently extremely abundant in farm dams, slowly flowing waters and shallow
wetlands (Lintermans 2007). They have been recorded at altitudes from 20 to 1120 m, although
most populations have been recorded below 300 m (Faragher and Lintermans 1997; McKay et al.
2001), which is not surprising given their preference for warm water and slow flows. The only
areas where they have not been recorded are the higher altitudes in the south-eastern alpine areas,
and the upper reaches of parts of the Condamine and Warrego basins (Lintermans 2007).
22
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
2.2 Impacts of Gambusia as an alien species
Eastern Gambusia’s aggressive nature, high reproductive potential, fast maturation rate, flexible
behaviour and broad environmental tolerances have contributed to its success as an invader, and
the species poses a serious threat to native fishes in Australia and overseas (Courtney and Meffe
1989; Howe et al. 1997; Rowe et al. 2008). It can also have a detrimental effect frogs (see Webb
and Joss 1997; Gillespie and Hero 1999; Komak and Crossland 2000), influence
macroninvertebrate, zooplankton and phytoplankton communities (Hurlbert et al. 1972;
Margaritora et al. 2001; Angeler et al. 2007) and enhance primary productivity by increasing
allochthonous nutrient loads (Hargrave 2006). Such impacts can also indirectly affect native fishes
by reducing or removing available prey resources and altering physicochemical properties such as
turbidity, water temperature and phosphorus cycles (Hurlbert et al. 1972). Changes to these
variables may interfere with normal feeding, sheltering or reproductive strategies of native fishes
and alter ecosystem level processes (Cardona 2006). Recently, Pyke (2008) completed a thorough
review of the impacts of eastern and Western Gambusias on aquatic systems, giving particular
attention to the documented impacts of both species on mosquitos, other invertebrates, amphibians,
planktonic communities and ecosystem functions. Accordingly, we have decided to focus the
following discussion on the impacts of Gambusia species on native fishes alone, and draw
attention to Pyke (2008) for further information on indirect effects and ecosystem level responses,
and to Gillespie and Hero (1999) and Rowe et al. (2008) for reviews on the effects of Eastern
Gambusia on Australian native amphibians.
A substantial portion of the literature devoted to the impacts of Gambusia species has been based
on studies conducted in the USA and Australasia on Western Gambusia (e.g. Laha and Mattingly
2007; Rowe et al. 2007). Because the two species are similar in body size, diet, feeding rates,
reproductive capacity and habitat requirements (Rehage et al. 2005), the mechanisms of impacts
may be similar, notwithstanding the potential for such mechanisms to be strongly mediated by the
local physical environment, and the diversity and abundances of fish species present (Courtney
and Meffe 1989; Rowe et al. 2008). The following sections provide a global perspective on the
impacts of Gambusia species on native fishes.
2.2.1
Impacts on native fishes
Western and Eastern Gambusias have been introduced to more than 60 countries (Garcia-Berthou
et al. 2005), and populations of Eastern Gambusia are now established in 21 countries (Froese and
Pauly 2007; Fishbase). Direct predation on native fishes (Myers 1965; Lloyd et al. 1986; Laha and
Mattingly 2007), competitive exclusion from food resources and habitat (Arthington et al. 1983;
Howe et al. 1997; Caiola and de Sostoa 2005) and aggressive interactions in confined
environments (Howe et al. 1997; Knight 1999; Laha and Mattingly 2007) have been well
documented. The implications of such interactions range from reduced condition of native fishes,
such as stunted growth, reduced ovarian weight and low fecundity (Howe et al. 1997; Mills et al.
2004), and increased susceptibility of individuals to secondary infection and damage to skin and
fins (i.e. via fin-nipping) (Meffe et al. 1983), to mortality or more or less competitive interferencedriven reductions in population size and distribution (see Galat and Roberston 1992).
Combinations of these mechanisms may be acting simultaneously, and the magnitude of impacts
on a given species may vary between day and night, seasonally, and with ontogeny (Mills et al.
2004; Ayala et al. 2007). The nature of interactions between Gambusia species and sympatric fish
species may also depend strongly on the relative densities, and on the behaviour of the species
present (e.g. Knight, 1999; Breen, 2000; Conte 2001), the availability of vacant trophic or habitat
niches (e.g. Lloyd and Walker 1986; Lloyd 1987, 1990; Keller and Brown 2008), variation in
physicochemical and environmental parameters (e.g. McNeil 2004; Rincón et al. 2002), and
seasonal alteration to natural or managed flow regimes and associated changes to aquatic
ecosystem processes (e.g. McNeil 2004; Fairfax et al. 2007).
23
Native fish recovery following Eastern Gambusia removal
Also of concern is the potential for Eastern Gambusia to reduce or fragment populations of native
species to levels at which negative factors associated with small population size, such as
inbreeding depression, and loss of allelic diversity and heterozygosity, exist (Rowe et al. 2008).
The negative effects of inbreeding are well documented (Gall 1987; Rowe et al. 2008) and include
decreases in fitness and increases in deformed offspring (Kincaid 1976), and extinction probability
(Saccheri et al. 1998).
In Australia, a large number of studies have indicated that Eastern Gambusia has negative
ecological impacts on small-bodied native fish species, but others have documented no detrimental
effects; see Macdonald and Tonkin (2008) for list of references. Rowe et al. (2008) identified 23
native species (for which published information was available) that were adversely impacted.
These species included several members of the families Galaxiidae, Gobiidae, Eleotridae,
Melanotaenidae as well as an ambassid (Olive Perchlet) and a retropinnid (Australian Smelt)
among others (see section 4.4). Rowe et al. (2008) pointed out that much of this evidence is based
solely on correlative field data (11 of 23 species) or on controlled aquarium experiments in which
predatory or competitive interactions may be intensified (4 of 23). Evidence from both field
studies and aquarium experiments was available for only eight species. The mechanisms and
consequences of impacts documented in these studies reflect those reported in stuides overseas on
for small-bodied native species (mechanisms — direct predation, competitive exclusion from
essential food and habitat resources, aggressive interactions; consequences — increased mortality
rates, decreased growth, condition, reproduction, population declines and population
fragmentation) and are most likely driven by large overlaps in habitat use and trophic niches,
mediated by the local physical environment. Quantitative information regarding the magnitude of
impacts in Australia is still lacking, and caution is warranted in extrapolating aquarium
experiments to natural systems (Ling 2004).
Eastern Gambusia is one of the most widely distributed freshwater fishes in Australia, and its
range is still expanding. Because of its broad physiological tolerances, high reproductive capacity
and proven ability to rapidly colonise degraded freshwater environments (Arthington et al. 1983;
Kennard et al. 2005; King and Warburton 2007), the impact on Australian native fishes may not
have peaked and could increase (Rowe et al. 2008). Because invasions by alien species are
generally population-level processes, the most reliable evidence for impacts is likely to come from
density manipulations under natural conditions, whereby Eastern Gambusia populations are
depleted or removed entirely and the recovery of native species is then monitored and compared to
control treatments (e.g. Peterson and Fausch 2003). Such experiments have never been attempted
in Australia (Rowe et al. 2008).
24
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Figure 2.3 Fin-nipping and weight loss in Dwarf Galaxias as a result of Eastern Gambusia
aggression and competition. The fish below and on the left (respectively) was collected from an area with
high densities of Eastern Gambusia. The other fish was collected from a healthy population with low densities
of Eastern Gambusia. Source: Pitman and Tinkler (2007).
25
Native fish recovery following Eastern Gambusia removal
2.2.2
Impacts on native fishes of the MDB
The native fish fauna of the MDB comprises approximately 46 species. This low species richness
is typical of Australia’s depauperate freshwater fish fauna, which is a legacy of the continent’s
long isolation, low rainfall and high proportion of arid areas (Unmack 2001; Lintermans 2007).
Nevertheless, the MDB contains a variety of aquatic systems, including upland and lowland rivers,
wetlands, billabongs and lakes, each of which contain their own array of habitat types and
associated fish communities.
Over half of the native species in the MDB are listed as rare or threatened in state, territory or
national listings (Lintermans 2007), and overall abundances are believed to be around 10% of
those prior to European settlement (MDBC 2004). Alien species make up one quarter of fish
species diversity and in many areas account to 80–90% of the fish biomass, so it is not surprising
that alien species have been listed as one of the key threatening processes to the native fish fauna
of the MDB (MDBC 2004). Because of the growing evidence of its detrimental impacts on native
fish fauna globally and its widespread distribution throughout the MDB, Eastern Gambusia has
been identified as one of the key alien species contributing to the decline of a number of native
fish within the MDB (MDBC 2004).
Despite the common view that interspecific interaction between Eastern Gambusia and MDB
native fish is detrimental to the native species, the evidence is largely circumstantial and, like the
Australia-wide evidence (see Rowe et al. 2008), is based largely on a limited number of
speculative correlative studies and aquarium experiments. Although evidence from individual
cases is not strong, particularly for the development of control strategies (see chapter 4.2), the
collation and combination of evidence with biological information can provide a better
understanding of the potential impacts of an alien species. Therefore we must not only understand
the biology of the alien species in question (identifying the reasons for its success as an invader)
but also have an equally thorough knowledge of the biology and ecology of the native fish
community it has invaded (Rowe et al. 2008). This enables the identification of mechanisms by
which native species may or may not be directly affected, and the recognition of potential indirect
impacts on individual species and on the fish community as a whole. In the following sections we
combine the documented evidence with biological information for each native fish species in the
MDB to assess the impact of Eastern Gambusia on the fish fauna of the MDB.
The major mechanisms through which Eastern Gambusia directly affect native fish around the
world involves interspecific competition for resources such as food and habitat, aggression, and
predation of eggs, larvae and juveniles. For Eastern Gambusia to directly affect native species by
one of these mechanisms, there must be some overlap in ecological niches. We know that Eastern
Gambusia prefer still and slow-flowing waters such as wetlands and backwaters and
predominantly occupy the upper water column and littoral zone, although large individuals are
associated with benthic areas of macrophyte beds (Stoffels and Humphries 2003). We also jnow
that they have a broad, omnivorous diet that includes detritus, aquatic and terrestrial invertebrates,
and the eggs and larvae of fish and amphibians. By overlaying this information about diet and
macrohabitat and microhabitat niches with those of each native species (using the available
knowledge of each developmental stages of each native species), we found that 16 of the 37 native
freshwater species of the MDB may have major niche overlaps in a number of developmental
stages with Eastern Gambusia (Table 2.1). They included, in particular, ambassids (glassfish),
nannopercids (pygmy-perches), melanotaenids (rainbowfishes), atherinids (hardyheads), eleotrids
(gudgeons), and retropinnids (smelt), which occupy (at least in part) slowly flowing or still water
and together make up the majority of the MDB’s wetland fish communities.
26
Arthur Rylah Institute for Environmental Research
Table 2.1 Potential macrohabitat, microhabitat and trophic niche overlaps between Eastern Gambusia and native freshwater fish in the MDB.
Macrohabitat niche = still and slow flowing areas such as back waters, weir pools and wetlands; Microhabitat niche = aquatic vegetation, upper water column, and
shallow littoral areas;Trophic overlap includes detritus, microinvertebrates and macroinvertebrates, terrestrial invertebrates and fish eggs and larvae; E = eggs, L =
larvae, J = juveniles, A = adult developmental stages, - = minimal overlap of all developmental stages; * Carp Gudgeon species complex. Grey shading indicates species
at high risk, defined as species for which there is overlap with Eastern Gambusia for all four developmental stages of the native species in at least one niche, and
overlap in at least three developmental stages of the native species in at least one other niche.
Family
Species name
Common name
Macrohabitat niche
overlap
Microhabitat niche
overlap
Trophic niche
overlap
Ambassidae
Ambassis agassizii
Olive Perchlet
E,L,J,A4,18
E,L,J4,18
L,J,A4,15
Anguillidae
Anguilla australis
Short-finned Eel
J,A6,7,18
-
-
6,7,18
-
Anguilla reinhardtii
Atherinidae
Eleotridae
Bovichthyidae
Long-finned Eel
J,A
1,5,6,18
-
Craterocephalus stercusmuscarum fulvus
Unspecked Hardyhead
E,L,J,A
E,L,J,A
L,J,A6
Craterocephalus fluviatilis
Murray Hardyhead
E,L,J,A5,6,18
E,L,J,A5,6,18
L,J,A6
Craterocephalus amniculus
Darling River Hardyhead
E,L,J,A6,18
E,L,J,A6,18
L,J,Aan.
Hypseleotris spp.*
Carp Gudgeon*
E,L,J,A1,2,8,9,18
E,L,J,A1,2,8,9,18,34
L,J,A8,9,10,15,34
Mogurnda adspersa
Southern Purple-spotted Gudgeon
E,L,J,A6,16
E,L,J,A6,16
L,J16
Philypnodon grandiceps
Flat-headed Gudgeon
E,L,J,A1,2,10,18
E,L,J,A1,2,10,18
L,J1,2,10
Philypnodon macrostomus
Dwarf Flat-headed Gudgeon
E,L,J,A1,3,18
E,L,J,A1,3,18
L,Jan.
Pseudaphritis urvillii
Tupong, Congolli
-
-
-
Clupeidae
Nematalosa erebi
Bony Herring
E,L,J,A
E,L,J,A
J,A6,15
Gadopsidae
Gadopsis marmoratus
River Blackfish
J,A32
-
Lan
Gadopsis bispinosus
Two-spined Blackfish
J,A32
-
Lan
Galaxias rostratus
Flat-headed Galaxias
E,L,J,A6,17,18
E,L,J,A6,17,18
L,J,A6,17
Galaxias olidus
Mountain galaxias
-
-
-
Galaxias brevipinnis
Climbing galaxias
-
-
-
Galaxias truttaceus
Spotted Galaxias
-
-
-
Galaxiidae
11,6
1,5,6,18
11,6
(continued on next page)
Arthur Rylah Institute for Environmental Research
27
Native fish recovery following Eastern Gambusia removal
Table 2.1 (continued)
Family
Geotriidae
Nannopercidae
Melanotaeniidae
Mordaciidae
Percichthyidae
Species name
Common name
Macrohabitat niche
overlap
Microhabitat niche
overlap
Trophic niche
overlap
Galaxias maculatus
Common galaxias
J,A6,17,18,33
J,A6,17,18
J,A6,17
Geotria australis
Pouched lamprey
L6,21,22
-
-
1,6,19,20,24
1,6,19,20,24
Nannoperca australis
Southern pygmy-perch
E,L,J,A
E,L,J,A
L,J,A6,19,24
Nannoperca obscura
Yarra pygmy-perch
E,L,J,A6,24
E,L,J,A6,24
L,J,A6,24
Melanotaenia splendida tatei
Desert rainbowfish
E,L,J,A6,15
E,L,J,A6
L,J,A6,15
Melanotaenia fluviatilis
Murray–Darling rainbowfish
E,L,J,A1,2,25,37
E,L,J,A1,2,25,37
L,J,A1,2,9,25,37
Mordacia mordax
Shortheaded lamprey
L6,23
-
-
1,2
Maccullochella peelii peelii
Murray cod
L
-
L9,30,31
Maccullochella macquariensis
Trout cod
L1,2
-
L13
Macquaria ambigua
Golden perch
E,L,J1,2,26
L,J18,27
L13,14,30
Macquaria australasica
Macquarie perch
L28
L28
L,J6,27
Macquaria colonorum
Estuary perch
-
-
-
Tandanus tandanus
Freshwater catfish
E,L,J,A
-
L,J6,29
Neosilurus hyrtlii
Hyrtl's tandan
E,L,J,A6
-
L,J6,an.
Porochilus rendahli
Rendahl's tandan
E,L,J,A6
-
L,J6,an.
Retropinnidae
Retropinna semoni
Australian Smelt
E,L,J,A1,2
E,L,J,A1,2
L,J,A9,15
Terapontidae
Bidyanus bidyanus
Silver perch
L,J1,26
L,1,18,26
L,J12,13,30
Leiopotherapon unicolor
Spangled perch
E,L,J,A6,18,32
E,L,J,A6,32
L,J,A6,15,32
Plotosidae
6,18,29
1
King et al. 2007, 2Humphries et al. 2002, 3Tonkin and Rourke 2008, 4Allen 1996a, 5Wedderburn et al. 2007, 6Lintermans 2007, 7Koehn and O'Connor 1990, 8Stoffels and Humphries 2003,
9
King 2005, 10Gehrke 1992, 11Puckridge and Walker 1990, 12Warburton et al. 1998, 13Lake 1967, 14Arumugam and Geddes 1987, 15Balcome et al. 2005, 16Larson and Hoese 1996, 17McDowall
and Fulton 1996, 18Treadwell and Hardwick 2004, 19Llewellyn 1974, 20Tonkin et al. 2008., 21Potter 1996a, 22Potter et al. 1986, 23Potter 1996b, 24Kuiter et al. 1996, 25Allen 1996b, 26Geddes and
Puckridge 1989, 27Cadwallader and Douglas 1986, 28Jason Thiem personal communication, 29Clunie and Koehn 2001, 30Tonkin et al. 2006, 31Rowland 1992, 32Llewellyn 1973, 33SRA 2007,
34
Balcome and Humphries 2006. 35Arthington et al. 1983, 36Lloyd and Walker 1986, 37Lukies 2004, anAnecdotal from similar species.
28
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Although the direct detrimental impacts of Eastern Gambusia are likely to be less severe in
riverine environments, there still may be an impact on larger riverine species. For example,
Eastern Gambusia may interact directly with the early developmental stages of these species,
which often occupy slackwater habitats (e.g. Barlow and Bock 1981, cited in Howe et al.
1997) or there may be indirect interactions. The indirect interactions of Eastern Gambusia on
the MDB’s native fish community have been largely overlooked, mostly because of our poor
understanding of the species’ overall ecological impact on Australian environments. Fletcher
(1986) suggested that large populations of Eastern Gambusia are likely to alter invertebrate
communities through predation. This was documented by Hurlbert et al. (1972), who found
algal densities in experimental ponds increased as a result of Eastern Gambusia feeding
selectively on zooplankton. Localised reductions in zooplankton can directly limit the
availability of food and subsequent survivorship of larval stages and in situ small native fish
species (Wilson 2006), and is likely to also reduce the amount of food that is transported to
other environments, such as from wetland to river channel (e.g. the flood recruitment model:
see Harris and Gehrke 1994).
Large populations of small native fish are also likely to affect zooplankton communities,
particularly in wetlands. High native fish densities are a natural occurrence, acting as a key
source of dispersive offspring for the broader catchment (Wilson 2006; Stuart and Jones
2002). For example, Lyon et al. (2010) reported large numbers of Carp Gudgeon (> 175 000)
as well as Australian Smelt, Flat-headed Gudgeon and Eastern Gambusia, moving between
off-channel sites and the main river channel. The smaller native species are important forage
fish for larger riverine species such as Murray Cod and Golden Perch, which actively avoid
Eastern Gambusia as a prey item (Lloyd 1990). Wilson (2006) suggested that alien species
could create energetic blocks within aquatic food webs, which would be pertinent for Eastern
Gambusia if it resulted in a reduction or replacement of native fish or their prey. Such
energetic blocks could have an indirect impact on larger riverine species of recreational
significance such as Murray Cod and Golden Perch by reducing the amount of food (such as
zooplankton for larvae, and forage fish for adults) entering the broader environment. Because
these species are the basis of a large inland recreational fishery, Eastern Gambusia could have
both social and economic impacts, along with their environmental impacts.
2.3 Mitigating the impacts of Eastern Gambusia
If we consider the extensive literature relating to Eastern Gambusia invasion and global
recognition of the threats that Gambusia species pose to freshwater ecosystem function
outside their native range, it is surprising that there is very little information available on
mitigating their impacts. Pyke (2008) suggests there are two ways to reduce the negative
impacts of invasive Gambusia species on native species: lowering their numbers (control),
and reducing the impact per individual.
Wilson (2006) found that virtually no information was available to guide managers in
choosing the most appropriate control strategies in Australia, largely because traditional
techniques such as poisoning, exclusion, egg dehydration, direct removal, commercial harvest
and habitat restoration have little chance of success in controlling smaller species such as
Eastern Gambusia, which occupy more cryptic habitats. Nevertheless, there is some
information about Gambusia control from both overseas and in Australia. This control
focuses predominantly on chemical techniques and drying of habitats, and always requires
extremely thorough treatment and an extensive knowledge of the hydrology of the area.
In Australia there have been few attempts to control Eastern Gambusia; most activities on pest
fishes, particularly in the MDB, have focused on Common Carp (Wilson 2006). Most states
and territories have listed Eastern Gambusia as a noxious or controlled species, so that it is
illegal for Eastern Gambusia to be kept, returned to the water or translocated (Queensland
29
Native fish recovery following Eastern Gambusia removal
Fisheries Management Act 1995; New South Wales Fisheries Management Regulation 2002;
Victorian Fisheries Regulations 1998; Northern Territory of Australia Fisheries Act 2005;
Tasmania Inland Fisheries Act 1995; ACT Fisheries Act 2000; South Australian Fisheries
Management Act 2007). New South Wales also developed a threat abatement plan for Eastern
Gambusia that proposed further research, a review of existing legislation, and control actions
at feasible sites (NSW 2003). While such legislation and planning is an important step in
Eastern Gambusia control, particularly in slowing the establishment of new populations, there
is little, if any, on-ground activities occurring in areas of established populations, particularly
in the MDB. Those activities that have occurred have focused on two relatively new
invasions, one in the Northern Territory and the other in Tasmania. Although the success of
these programs is dependent on the results of ongoing monitoring (and perhaps further
treatment), Macdonald and Tonkin (2008) suggested that, with extensive biological and
hydrological knowledge of an area and a coordinated approach, the control of Eastern
Gambusia within closed systems or in areas of new invasion is possible.
Successful eradication of Eastern Gambusia using current methods is not likely to be feasible
in larger open systems, such as many of the areas of the MDB. Minimising the impact of
Eastern Gambusia on native fish may be an important strategy, particularly in areas
containing threatened fish fauna, until new control strategies such as harvesting techniques
and daughterless technology are developed. Pyke (2008) suggested that reducing any negative
impacts of Gambusia species on native species could be achieved by a reduction in their
numbers, and by reducing the impacts per individual. We know that lentic habitats that
contain Eastern Gambusia and provide cover such as aquatic vegetation and snags, contain
more native fish than areas without cover and Eastern Gambusia (Morgan et al. 2004).
Furthermore, it appears that Eastern Gambusia has the greatest impact on native fish during
times of low water levels (e.g. Fairfax et al. 2007). Therefore, habitat maintenance such as
enhancing thick aquatic vegetation (Pyke 2008) and watering during times of extreme low
levels may reduce the impact of Eastern Gambusia on rare native fish fauna until improved
control techniques are available.
The key objective of an alien species removal program is to reverse the negative impacts that
the species has had on native biota, to benefit native biological diversity (Zavaleta et al.
2001). One question that is often not considered in pest species management programs is
‘What negative effects will a reduction or complete removal of the target alien species have
on ecosystem function?’ There is evidence that successful eradications of alien species can
have unexpected and undesirable impacts on native species and ecosystems, particularly in
areas that the alien species has occupied for a long time and is an established species in the
food chain (e.g. Murphy et al. 1998). Maezono and Miyashita (2004) suggested two ways in
which such undesirable impacts may occur. First, the removal of an alien species can enhance
secondary establishment, or increase the impact of other alien species. Secondly, negative
impacts to native biota may occur if the alien species performs functions similar to those of
native species that are no longer in the system. Ultimately, the type of species being removed,
the degree to which it has replaced native taxa, and the presence of other non-native species
can all affect the eventual impacts of removal of an alien species (Zavaleta et al. 2001).
Thus, to restore native biodiversity it seems essential to clarify the community-wide impacts
of alien species, including interactions with other alien species, prior to eradication (Maezono
and Miyashita 2004). Zavaleta et al. (2001) suggested a pre-eradication assessment, including
qualitative evaluation of trophic interactions between alien and native species to anticipate the
need for special planning, and post-eradication monitoring. They concluded that, while
invasive species eradication is an increasingly important component of conservation
management, in natural systems a shift in focus is required from pure alien species control
towards broader ecosystem restoration goals.
30
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
2.4 Knowledge gaps and the current project
A central tenet of integrated pest management is the reduction of impacts on ecosystem
processes, not simply a reduction in numbers of the target species (Lodge and SchraederFrechette 2003; Koehn and Mackenzie 2004). Because of the ongoing threat Eastern
Gambusia pose to native fish communities (Lintermans 2007), the lack of current effective
control options (see McKay et al. 2001), and the uncertainty about the long-term detrimental
impacts of the species on ecosystem function in the MDB, research into the feasibility of
controlling Eastern Gambusia populations to densities where measurable improvements to
native fish communities can be detected should be a priority.
The core component of this project (Project MD1043) addressed these research needs by
taking a holistic approach. It integrated surveys and quantitative experimental work in natural
billabong systems throughout the MDB with cost-effectiveness approaches, to:
•
provide information on the response of native fish communities following the
reduction of Eastern Gambusia populations, and
•
provide a framework to evaluate the feasibility and effectiveness of such control
actions, and form a template for evaluating control options for other alien fishes
across the MDB.
To achieve these objectives the project was divided into three phases. The first phase involved
a broad-scale, cross-sectional study of wetland fish communities to develop hypotheses about
the effect of Eastern Gambusia on native fish communities in these enclosed systems. The
second phase was a field trial of Eastern Gambusia control in small isolated billabongs, to test
the hypotheses through density manipulation experiments and to provide information on
control options and Eastern Gambusia population dynamics. The third phase identified
strategies to maximise the level of improvement to the native fish community through Eastern
Gambusia control given a fixed budget (benefit maximisation), and to minimise the cost of
achieving a defined significant improvement in the native fish community (cost
minimisation). Finally, the project provided a template for evaluating control options for other
alien fishes across the MDB.
31
Native fish recovery following Eastern Gambusia removal
3
Phase 1: Cross-sectional study of wetland fish
communities across the mid-Murray region of the MDB
3.1 Introduction
The invasion, establishment and successful integration of an alien species into an ecosystem
often presents some form of ecological risk to that system (Copp et al. 2005). Accurately
defining the level of risk, by quantifying the potential detrimental, neutral or even positive
impacts of invasion on ecosystem processes or particular species, can provide a strong
platform for generating and testing hypotheses regarding what level of control (if any) is
required to achieve an acceptable level of risk; see Copp et al (2005) and Gozlan et al. (2010)
for discussion. But what constitutes a detrimental impact, and how do we measure it?
A recent paper by Gozlan et al. (2010) suggested that the focus should not be on the
ecological changes that inevitably occur after invasion, but on whether such changes can be
correlated with measurable reductions in native species diversity, abundance, reproduction,
condition or shifts in ecosystem function that compromise the long-term integrity of native
species populations. In Chapter 2 we detailed the mechanisms by which Eastern Gambusia
can impact Australian native fish species, the potential consequences of these mechanisms,
and how the nature of the physical environment can mediate the magnitude of such impacts at
multiple spatial and temporal scales (Kennard et al. 2005; Pyke 2005; Fairfax et al. 2007;
Rowe et al. 2008; Costelloe et al. 2010). The collation of biological information on wetland
fishes in the MDB, which was a key component of our literature review, suggests that Eastern
Gambusia are likely to have the greatest direct impact on smaller species, particularly the
ambassids (glassfish), nannopercids (pygmy-perches), melanotaenids (rainbowfishes),
atherinids (hardyheads), eleotrids (gudgeons) and retropinnids (smelt) that comprise the
majority of the MDB’s wetland and billabong fish communities (Macdonald and Tonkin
2008). The review found that 16 of the 37 native freshwater species present exhibit major
habitat or trophic niche overlaps with Eastern Gambusia, and we postulate that these 16
species are at greatest risk of negative impacts in systems where they coexist with Eastern
Gambusia.
In order to quantify the current level of impact, we designed a broad-scale, cross-sectional
study of wetland fish communities throughout the mid-Murray region of the MDB. The
overarching objective of the study was to examine the current influence of Eastern Gambusia
on native fish abundance, diversity and condition in ‘closed’ wetland or refugia systems,
where dispersal opportunities are limited and any impacts on native species are likely to be
magnified. In situations where baseline data are limited, correlative or species distribution
models (SDMs) can enable an investigation of variables that cannot be directly manipulated,
and reveal patterns between a species’ occurrence or abundance and environmental
characteristics (Elith and Leathwick 2009; Capinha and Anastácio 2011), although they
cannot identify the causes of such patterns. We collected correlative data on fish species
composition, abundance and several environmental covariates across a large spatial scale and
constructed such models to:
• examine how the presence or absence of Eastern Gambusia influences the probability of
occupancy of native and alien fish species in wetlands across the MDB
• explore patterns of co-occurrence of native fish species and Eastern Gambusia in these
systems
• quantify the relationships between Eastern Gambusia abundance and the abundance of
other sympatric fish species in differing habitat types and with variability in environmental
covariates.
32
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
We also examined relationships between Eastern Gambusia abundance and morphometric
condition in juveniles of the three most common native fish species captured during the study:
Australian Smelt, Carp Gudgeon and Flat-headed Gudgeon.
These models were exploratory and were designed to inform the development of hypotheses
on how native fish respond when Eastern Gambusia populations are reduced in wetland
systems. These hypotheses are introduced here, and the ecological basis for each is discussed.
Phase 2 of the project (see Chapter 4) provided a direct test of each hypothesis in a controlled
field removal experiment at selected wetlands in the Ovens River, Victoria.
3.2
Methods
3.2.1
Study area and site selection
The cross-sectional study focused on off-channel habitats such as billabongs and anabranches
in the mid-Murray region of the MDB. Because of the prolonged drought conditions in the
region, which persisted into summer 2009, these habitats were restricted to the Ovens River
and Reedy Creek floodplain, Gunbower Forest, Wonga Wetlands (Albury) and the Barmah–
Millewa Forest (Figure 3.2.1).
146ºE
New South Wales
36ºS
Yarrawonga
Lake Mulwala
Murray River
#
###
#
# #
## #
##
## ##
## #
#
#
##
N
W
## #
E
Ovens River
###
#####
#
##
#
#
#
##
#
S
#
#
##
##
#
0
10
20 km
Victoria
#
Wangaratta
Gunbower Forest
#
##
Wonga Wetlands
#
Barmah-Millewa/Goulburn
##
#
#
#
#
##
#
#
Ovens River floodplain
Melbourne
Figure 3.2.1 Cross-sectional study site locations (red squares) sampled during February–
March 2009 and 2010, including detailed locations of sites sampled in the Ovens River
floodplain.
33
Native fish recovery following Eastern Gambusia removal
A total of 93 sites were sampled in February–March (later summer – early autumn) 2009 and
2010. To ensure that the assessment of each site’s fish community was not influenced by
recent immigrations or emigrations, sites were selected on the basis that they were isolated
from the nearby main river channel or other waterbodies for several months prior to sampling.
This was achieved by ground-truthing potential sites in December 2008 and 2009, coupled
with close monitoring of flow data in the associated rivers throughout spring, summer and
autumn each year, and knowledge of ‘commence-to-fill’ levels of anabranches and wetlands
in the region. All sites were sampled on a single occasion during February or March, timed to
coincide with the peak abundance of Eastern Gambusia and the conclusion of spawning of
most native species. We anticipated that any impacts of Eastern Gambusia on the native fish
community would be highest during this period (i.e. high gambusia densities coupled with
high densities of larval or juvenile native species in restricted habitats).
3.2.2
Water quality and habitat assessment
Prior to the commencement of fish sampling, measures of the wetland size (length × width, to
the nearest metre) and water quality parameters (pH, conductivity, turbidity, dissolved oxygen
and temperature) at a depth of approximately 30 cm were recorded. The mean and maximum
depth of the site as a whole, and of each replicate sampling unit, was also recorded. Habitat
complexity was also broadly assessed by an estimation of (a) the amount of aquatic vegetation
in the littoral zone (none, low, high); and (b) the amount of snags and other debris in the
littoral zone (none, low, medium or high) (Figure 3.2.2). Habitat complexity was only
assessed for the littoral zone, as all sampling was undertaken within 7 m of the water’s edge.
(a)
(b)
(c)
Figure 3.2.2 Example of (a) low (b) medium, and (c) high levels of snags and other debris in
site littoral zones.
3.2.3
Fish sampling and processing
Recent surveys throughout the study region recorded 10 small-bodied native species as well
as five alien species (including Eastern Gambusia) in floodplain environments (Table 3.1).
This assemblage composition, in conjunction with behavioural traits and habitat preferences
of these species, informed the development of a fish sampling protocol which aimed to
provide the most accurate representation possible of the small fish community at a given site.
34
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Table 3.1 Small native fish species and all alien species recorded in floodplain environments
(billabongs, anabranches and inundated floodplains) of the mid-Murray River and
tributaries between Albury and Barham.
Common name
Scientific name
Reference
Australian Smelt
Retropinna semoni
1,2,3,4,5,6,7,8
Carp Gudgeon
Hypseleotris spp.
1,2,3,4,5,6,7,8
Flat-headed Gudgeon
Philypnodon grandiceps
2,3,4,5,6,7,8
Un-specked Hardyhead
Craterocephalus stercusmuscarum
fulvus
3,4,6,7,8
Southern Pygmy-perch
Nannoperca australis
1,2,3,4
Murray–Darling
Rainbowfish
Melanotaenia fluviatilis
3,4,6,7
Dwarf flat-headed
Gudgeon
Philypnodon macrostomus
3,4,6,7
Flat-headed Galaxias
Galaxias rostratus
1,2
Mountain Galaxias
Galaxias olidus
2
Bony Bream
Nematalosa erebi
5,6,8
Eastern Gambusia
Gambusia holbrooki
1,2,3,4,5,6,7,8
Carp
Cyprinus carpio
1,2,3,4,5,6,7,8
Goldfish
Carassius auratus
1,2,3,4,5,6,7,8
Redfin Perch
Perca fluviatilis
1,2,3,4,5,6,7,8
Oriental Weatherloach
Misgurnus anguillicaudatus
1,2,3,4,5,6,7,8
Native
Exotic
1
McNeil 2004; 2 King et al. 2003; 3 King et al. 2007; 4Tonkin and Rourke 2008; 5 McKinnon 1997; 6 Rehwinkel and
Sharp 2009; 7 Richardson et al. 2005; 8 Jones and Stuart 2004.
To allow a rapid yet comprehensive assessment of the fish community at each site, only active
sampling methods were used. Passive methods such as fyke netting and bait traps which rely
on fish movement and give high interspecies and intraspecies variability in catch data were
deemed unsuitable. We restricted the sampling to the littoral zone in each wetland, defined
here as the area extending from the water’s edge to not more than 7 m from the bank. We
stratified the sampling by habitat type, using standardised fine-mesh seine netting (7 m length
× 1.5 m drop; 1 mm mesh diameter; 500 mm square purse) in open water littoral habitats and
sweep net electrofishing (SNE) in complex littoral habitats. ‘Open water’ habitats were
defined as containing no aquatic vegetation and no small or large woody debris. Habitats in
which aquatic vegetation or woody debris were present were classified as ‘complex’. Seine
netting and SNE were both effective at capturing fish and larvae as small as 6 mm in length,
but seine netting was more effective at capturing mobile species such as Australian Smelt and
Eastern Gambusia, and SNE was better at capturing species associated with complex habitats
(e.g. gudgeons and Southern Pygmy-perch).
35
Native fish recovery following Eastern Gambusia removal
The SNE method (King and Crook 2002) used a modified standard backpack electrofishing
unit (Smith-Root Model 12b) fitted with an anode ring 15 cm in diameter. A moulded plastic
rectangular frame (25 × 30 × 2 cm) holding a 250 µm mesh sampling net was attached near
the anode ring, with the net tapering to a removable collection jar. Each SNE sample was
obtained by approaching the selected complex habitat, activating the anode and moving at a
constant speed throughout the habitat for 20 seconds of electrofishing time (Figure 3.2.3).
Fish were then concentrated into the collection jar and emptied into an aerated 30 L bucket for
processing. Each seine sample was obtained using a standardised point–point method, in
which one end of the seine was held in a fixed position at the water’s edge while the other
was walked at a steady pace in a tear-drop shape, returning back to the same point; in effect,
covering an area of approximately 3 m2. All fish captured were then concentrated into the
square purse and quickly transferred to a 300 L aerated bin for processing.
Figure 3.2.3 Sweep net electrofishing snag / debris habitat in the littoral zone of a
billabong on the Ovens River floodplain.
The number of replications required to achieve an accurate representation of the fish
community at a given site was determined by conducting a power analysis, based on data
collected from three sites for each sampling method. This involved sampling the entire littoral
zone, with complex habitats sampled by SNE and open water habitats sampled by seine
netting (usually up to 12 samples for each site and method).
The results indicated that six replicate SNE samples in complex habitat and six seine samples
in open water habitat would provide an adequate assessment of species diversity and
abundance in the littoral zone of larger sites. Therefore, following the initial habitat
assessment at each site, any complex habitats in the littoral zone were sampled with a
maximum of six SNE passes, and all other open water areas of the littoral zone were sampled
with a maximum of six seine passes. At sites that were too small to achieve such replication
we made as many passes as possible without overlapping the areas sampled. In addition, if a
site contained no complex habitat in its littoral zone, only seine netting was undertaken; and
conversely, if a site only contained complex habitat, only SNE was undertaken.
At the completion of each replicate SNE and seine shot, all fish were identified and counted,
and twenty individuals of each species were randomly selected and measured for total length
(TL; nearest mm). Additionally, the first 20 individuals of each species measured at each site
were euthanised in an overdose of anaesthetic (alfaxalone 40 mg/L) for 10 minutes, and then
preserved in 95% ethanol to enable measures of condition (i.e. Krel, fin damage) and ageing to
be undertaken in the laboratory.
36
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
To assess morphometric condition, preserved juvenile Australian Smelt, Carp Gudgeon and
Flat-headed Gudgeon collected during the sampling events were measured for standard length
(nearest 0.1 mm) and weight (nearest 0.001 g). These measurements were made after a
minimum of 10 days to allow shrinkage associated with preservation to stabilise (Fey and
Hare 2005). Fish were then positioned on their right side and examined for fin condition. The
caudal, dorsal, anal and left hand pelvic and pectoral fins were given a score from 1 to 3,
where:
1 = no or minimal damage — fin has no sign of damage or only minor splitting of rays
2 = moderate damage — part of fin is missing, but more that 50% remains
3 = major damage — less than 50% of the fin remains.
Fin scores were subsequently converted to either 0 (scores of 1), or 1 (scores of 2 or 3) i.e. ‘no
damage’ or ‘damaged’ (Figure 3.2.4).
(a)
(b)
Figure 3.2.4. Examples of different levels of caudal fin damage including fins scored as ‘1’
(top), ‘2’ (middle) and ‘3’ (bottom) for (a) Carp Gudgeon; and (b) Australian Smelt.
3.2.4
3.2.4.1
Statistical analysis and development of correlative models
Species occupancy models
Models were fitted in a Bayesian framework in WinBUGS (Lunn et al. 2000) to explore how
the presence or absence of Eastern Gambusia or other covariates — presence or absence of
aquatic vegetation (veg), wetland size (size), water temperature (temp), conductivity (cond),
DO, pH, turbidity (turb) — can influence the probability of a site being occupied by native
and other alien fish species. Analyses was performed for 81 wetland sites using five samples
from each site (combined SNE and seine passes). The data consisted of the presence or
absence of seven species — Eastern Gambusia (Gh), Carp Gudgeon (Hsp), Flat-headed
Gudgeon (Pg), Australian Smelt (Rs), Unspecked Hardyhead (Cs), Common Carp (Cc) and
Goldfish (Ca) in each of the samples per site. The low detection of other native species (e.g.
Southern Pygmy-perch, Murray–Darling Rainbowfish) precluded their inclusion in the
models. These models estimate the true occupancy probability of a given species, accounting
for errors in detection from sampling methods. In addition, we used them to estimate the
patterns of co-occurrence of native species and Eastern Gambusia.
37
Native fish recovery following Eastern Gambusia removal
3.2.4.2
Relationships between species’ abundances and environmental covariates
The purpose of this analysis was primarily exploratory, in that we were attempting to
determine whether there were any relationships (whether linear or non-linear) between the
abundance of Eastern Gambusia, the abundance of native and other alien fish species, and
several environmental covariates. We used a relatively new machine-learning technique
called boosted regression trees (BRT) to explore the nature of these relationships in 58 of the
wetlands sampled in the cross-sectional study. Conventional regression models produce a
single ‘best’ model, whereas BRT combines large numbers of relatively simple tree models
adaptively to optimise the predictive performance (Elith et al. 2008). The optimal number of
trees in the final model is estimated using cross-validation techniques.
For each wetland, the data used was the average counts of fish species from six samples
(combined SNE and seine net passes). These averages were log(x + 1) transformed before
analysis to stabilise the variance. An analysis was undertaken for the same seven species used
in the occupancy modelling analysis (see section 1.2.4.1), but here results are presented for
Eastern Gambusia and the three most abundant native fish species based on our survey data:
Australian Smelt, Carp Gudgeon and Flat-headed Gudgeon. Abundances of species that were
not subject to analysis (i.e. response variable) were treated as potential explanatory variables.
Other explanatory variables included in the models were environmental parameters —
wetland size (size), aquatic vegetation cover (veg), debris load (debris), pH, dissolved oxygen
(DO), conductivity (cond), turbidity (turb) and water temperature (temp).
3.2.4.3
Condition indices for juvenile native fishes
An exploratory assessment of body and fin condition was undertaken for the juveniles of
Australian Smelt, Carp Gudgeon and Flat-headed Gudgeon. The body condition index for
juvenile fish was estimated by applying the length/weight polynomial model for the juveniles
of these species. The residuals from each model were used as an index of condition for each
individual.
A second measure of condition was undertaken using the relative condition factor (see Froese
2006). Length and weight of juvenile fish were combined to give an indicator of general body
form of juvenile Australian Smelt, Carp Gudgeon and Flat-headed Gudgeon, using the
equation
W = aLb
where W is the weight in grams, L is the TL in mm, and a and b are constants. Length–weight
relationships could then be used to compare body condition using the relative condition factor
Krel, using the formula
Krel = W / aLb
The relationship between juvenile condition of each species and the Eastern Gambusia
abundance (measured as catch per unit effort, CPUE) at each sampling site was then assessed
using simple linear regression.
An initial examination of the fin condition data for the three aforementioned native species
indicated that the caudal fin sustained by far the most damage of any fin type. As a result we
decided to limit our analysis to this fin type. We constructed logistic regression models in R
(version 2.12.2, The R Foundation for Statistical Computing) to examine the impact of
Eastern Gambusia CPUE on caudal fin condition in juveniles of each of the three species. We
were also interested in the potential influence of habitat complexity and its interaction with
Eastern Gambusia CPUE in mediating the prevalence of aggressive encounters and
subsequent effects on the degree of caudal fin damage. Aquatic vegetation cover and the
38
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
debris load at each site were added as categorical variables in the models, and fin condition
was assessed in relation to changes in these variables.
3.3 Results
3.3.1
Assessment of wetland fish communities of the mid Murray River
A total of 61 575 fish comprising six native and five alien species were captured during the
cross-sectional study (Table 3.2). The most abundant and frequently occurring species were
Carp Gudgeon (27 346; 44% of abundance and occupying 90% of sites), Eastern Gambusia
(26 932; 43% of abundance and occupying 83% of sites), Australian Smelt (4040, 7% of
abundance and occupying 49% of sites) and Flat-headed Gudgeon (2186, 4% of abundance
and occupying 48% of sites) respectively. Four of the ten small-bodied native species
previously recorded in off-channel habitats in these regions — Dwarf Flat-headed Gudgeon,
Flat-headed Galaxias, Mountain Galaxias and Bony Bream — were not recorded during the
study: see Table 3.1). Although three of the four species occur in higher abundances outside
the study region or are more riverine species, the absence of the Flat-headed Galaxias is of
concern. Very few Southern Pygmy-perches were captured in the Lower Ovens, and none
from other sites previously known to harbour healthy populations (McNeil 2004, Tonkin et al.
2008). Small numbers of Unspecked Hardyhead and Murray–Darling Rainbowfish were
captured, but both these species are known to persist in good numbers in the middle and lower
reaches of the MDB (Lintermans 2007). All of the alien species previously recorded in offchannel habitats in the regions were collected in the surveys.
Table 3.2 Raw abundances of each species collected during the 2009–2010 cross-sectional
study from each of the four regions surveyed (Gunbower, Lower Ovens, Albury (Wonga
wetlands) and Barmah–Millewa/Goulburn). The number of sites surveyed within each
region is shown in parentheses.
Scientific name
Gunbower
(n = 5)
Ovens
(n = 83)
Albury
(n = 2)
Australian Smelt
Retropinna semoni
1234
3170
48
–
4452
Carp Gudgeon
Hypseleotris spp.
2100
23985
974
287
27346
Flat-headed Gudgeon
Philypnodon
grandiceps
–
2208
55
28
2291
Unspecked Hardyhead
Craterocephalus
stercusmuscarum
fulvus
–
18
–
1
19
Southern Pygmy-perch
Nannoperca
australis
–
16
–
–
16
Murray–Darling
rainbowfish
Melanotaenia
fluviatilis
3
–
–
3
6
1266
25042
455
169
26932
Common name
B–M/G
(n = 3)
Total
Native
Alien
Eastern Gambusia
Gambusia holbrooki
Carp
Cyprinus carpio
–
194
1
–
195
Goldfish
Carassius auratus
–
293
1
–
294
Redfin perch
Perca fluviatilis
–
9
–
–
9
Oriental weatherloach
Misgurnus
anguillicaudatus
–
15
–
–
15
4603
54950
1534
488
61575
Total
39
Native fish recovery following Eastern Gambusia removal
3.3.2
The influence of Eastern Gambusia on species occupancy and abundance
3.3.2.1
Species occupancy models
1.0
The output from the occupancy modelling suggests that the presence of Eastern Gambusia
and the presence or absence of aquatic vegetation at a particular wetland can interact to
influence the probability of occupancy by several sympatric fish species (Figure 3.3.1). The
three most common native fish species captured in the surveys — Carp Gudgeon, Australian
Smelt and Flat-headed Gudgeon — were most likely to coexist with Eastern Gambusia in
wetlands lacking aquatic vegetation, with the probability of occupancy for these species
highest when Eastern Gambusia were present. Unspecked Hardyhead and Common Carp
were least likely to occur in wetlands containing Eastern Gambusia and no vegetation.
Moreover, Common Carp were significantly more likely to coexist with Eastern Gambusia in
wetlands containing aquatic vegetation, or in the absence of Eastern Gambusia. The
likelihood of Goldfish being present at a particular wetland appears to be largely unaffected
by the presence or absence of Eastern Gambusia and aquatic vegetation. At sites where
Eastern Gambusia were not captured, the presence of aquatic vegetation had minimal
influence on the probability of that wetland being occupied by every other species examined
(Figure 3.3.1).
0.8
0.6
0.4
0.0
0.2
Occupancy probability
Veg present: Gh present
Veg present: Gh absent
Veg absent: Gh present
Veg absent: Gh absent
Hsp
Pg
Rs
Cs
Cc
Ca
Species
Figure 3.3.1 Co-occurrence of sampled species with the presence or absence of Eastern
Gambusia (Gh) and aquatic vegetation (Veg) in the wetlands. Species codes are Carp
Gudgeon (Hsp), Flat-headed Gudgeon (Pg), Australian Smelt (Rs), Unspecked Hardyhead
(Cs), carp (Cc) and Goldfish (Ca).
Eastern Gambusia exhibited substantial flexibility in the types of habitats in which they
persisted. The probability of occupancy was greater than 0.75 regardless of the presence of
aquatic vegetation, although the species was more likely to occupy wetlands devoid of aquatic
vegetation (Figure 3.3.2). In addition, larger wetlands sampled were more likely to be
occupied by Eastern Gambusia than smaller ones, and the odds of occupancy were decreased
as temperature and turbidity increased (Figure 3.3.3).
40
Arthur Rylah Institute for Environmental Research
0.8
0.6
0.4
0.2
0.0
Occupancy probability
1.0
Native fish recovery following Eastern Gambusia removal
Veg present
Veg absent
Aquatic vegetation
2
0
-2
-4
-6
Log (odds of occupancy)
4
6
Figure 3.3.2 Probability of Eastern Gambusia occupying a wetland when aquatic vegetation
was present or absent.
size
temp
cond
DO
turb
Environmental variable
Figure 3.3.3 The influence of different environmental covariates: wetland size (size), water
temperature (temp), conductivity (cond), dissolved oxygen (DO) and turbidity (turb) on the
probability of occupancy by Eastern Gambusia.
41
Native fish recovery following Eastern Gambusia removal
3.3.2.2
Relationships between species’ abundances and environmental covariates
The primary aim of the BRT analysis was to explore the nature of any relationships existing
between the abundances of Eastern Gambusia and other sympatric species, and the potential
influence of environmental covariates in sampled wetlands. As stated in the Methods (section
3.2.4.2), we ran separate BRT models on all seven species (Gh, Hsp, Pg, Rs, Cs, Cc, Ca)
examined in the occupancy modelling section, but here present results only for Eastern
Gambusia, Australian Smelt, Carp Gudgeon, Flat-headed Gudgeon and Common Carp. The
output from the BRT models reports the relative contribution of each potential predictor
variable for the final model of each species (Table 3.3) and provides fitted functions of each
predictor in relation to the abundance of each species.
Eastern Gambusia abundance
M
N
200
4
5
H
L
N
veg (1%)
0.5
0.0
fitted function
9.0
-0.5
0.5
8.5
9.5
15
20
1
2
3
35
0.5
0.0
20000
2
4
6
8
10
12
DO (4.9%)
fitted function
0
30
-0.5
fitted function
5000 10000
25
Temp (6.9%)
0.5
0
fitted function
0.5
0.0
fitted function
3
8.0
size (4.9%)
-0.5
0.0
2
Rs (2.9%)
600
Cond (5.7%)
-0.5
1
400
7.5
0.0
fitted function
0
debris (6.8%)
0
7.0
-0.5
0.5
0.0
fitted function
L
6.5
pH (7.1%)
-0.5
0.5
0.0
-0.5
H
300
4
Pg (0.2%)
0.5
200
Turb (8.5%)
0.0
100
0.0
fitted function
0
-0.5
0.5
6
-0.5
5
0.5
4
0.0
3
Hsp (51.1%)
-0.5
2
0.0
fitted function
0.0
-0.5
1
0.5
Eastern gambusia abundance
0
-0.5
0.5
The final BRT model fitted to the combined SNE and seine netting data had 1750 trees, as
estimated by cross-validation, and explained 51% of the deviance. The most influential
predictor variable was the abundance of Carp Gudgeon, which contributed 51% of the
variance of the final BRT (Table 3.3). The fitted functions of each predictor variable relative
to Eastern Gambusia abundance are shown in Figure 3.3.4. The plots illustrate the strong
positive relationship between Eastern Gambusia and Carp Gudgeon abundance, and very
weak negative relationships with turbidity, pH and Australian Smelt abundance.
0.0
0.5
1.0
1.5
2.0
Cc (0%)
Predictor variable
Figure 3.3.4 Fitted functions describing relationships between Eastern Gambusia
abundance and all predictor variables for the final BRT model. Refer to Figure 6 for species
codes. The value in brackets after each predictor variable shows the relative influence of
that predictor in the final model (see also Table 3.3). For variables ‘debris’ and ‘veg’,
N = none, L = low, M = medium, H = high levels.
42
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Table 3.3 Relative influence (I) of each predictor variable (PV) on abundances of Eastern
Gambusia, Australian Smelt, Carp Gudgeon, Flat-headed Gudgeon and Common Carp in the
final BRT models for each species.
Eastern
Gambusia
Australian
Smelt
Carp
Gudgeon
Flat-headed
Gudgeon
Common
Carp
PV
I (%)
PV
I (%)
PV
I (%)
PV
I (%)
PV
I (%)
Hsp
Turb
pH
temp
debris
cond
size
DO
Rs
veg
Pg
Cc
Ca
51.1
8.5
7.1
6.9
6.8
5.7
4.9
4.9
2.9
1.0
0.2
0.0
0.0
Hsp
size
turb
cond
Gh
pH
temp
DO
Cc
Pg
debris
Veg
Ca
28.5
25.0
9.6
8.8
7.1
5.8
5.3
4.5
2.9
1.1
0.8
0.6
0.1
Gh
Rs
DO
turb
size
Cc
temp
Pg
cond
pH
veg
debris
Ca
36.8
19.2
16.4
10.2
4.1
3.7
2.7
2.2
2.1
1.5
0.6
0.3
0.0
Hsp
DO
pH
size
cond
turb
Gh
temp
veg
Rs
debris
Cc
Ca
39.9
20.9
10.9
7.8
6.0
4.8
4.4
4.0
0.7
0.3
0.3
0.0
0.0
Hsp
pH
size
cond
Gh
turb
temp
Rs
DO
Ca
veg
debris
Pg
29.1
18.9
12.5
10.2
7.6
6.1
4.9
3.7
2.1
1.6
1.6
1.0
0.5
Australian Smelt abundance
20000
5
6
8.0
8.5
9.0
9.5
2.0
2
3
Pg (1.1%)
0.4
0.2
0.0
-0.2
fitted function
30
0.0
35
2
4
4
6
8
10
12
0.4
DO (4.5%)
fitted function
-0.4
fitted function
1
0.2
0.4
25
-0.4
0
600
-0.2
fitted function
20
0.4
0.4
0.2
0.0
fitted function
-0.4
1.5
400
-0.4
15
Temp (5.3%)
-0.2
0.4
0.0
-0.2
1.0
Cc (2.9%)
0.2
0.4
7.5
200
Cond (8.8%)
0.0
fitted function
7.0
pH (5.8%)
-0.4
0.5
0
-0.4
6.5
Gh (7.1%)
0.0
300
0.2
4
200
-0.2
0.4
0.2
0.0
fitted function
3
100
Turb (9.6%)
-0.4
2
-0.4
0
-0.2
0.4
0.2
0.0
-0.2
1
0.2
0.4
5000 10000
size (25%)
-0.4
0
0.0
fitted function
0
0.2
6
0.0
5
-0.2
4
0.0
3
Hsp (28.5%)
-0.2
2
-0.4
-0.4
1
-0.2
0.4
0.2
0.0
fitted function
0.0
-0.2
-0.4
0
0.2
Australian smelt abundance
-0.2
0.2
0.4
The optimal BRT model contained 1900 trees and explained 50% of the deviance. Carp
Gudgeon abundance was the most influential variable, followed by wetland size, with positive
relationships noted between these two variables and Australian Smelt abundance (Figure
3.3.5). Australian Smelt abundance was found to increase slightly at more turbid and saline
wetlands, and there was a weak negative relationship with Eastern Gambusia abundance.
H
L
M
debris (0.8%)
N
H
L
N
veg (0.6%)
Predictor variable
Figure 3.3.5 Fitted functions describing relationships between Australian Smelt abundance
and all predictor variables for the final BRT model.
43
Native fish recovery following Eastern Gambusia removal
Carp Gudgeon abundance
3
4
20000
1.0
1.5
2.0
0
7.0
7.5
8.0
pH (1.5%)
100
8.5
9.0
9.5
200
300
20
25
30
0.0
0.5
Turb (10.2%)
35
0
1
2
3
4
fitted function
0.5
Pg (2.2%)
0.5
fitted function
6.5
0.0
12
fitted function
15
0.5
0.0
fitted function
600
10
Temp (2.7%)
-0.5
0.0
400
Cond (2.1%)
8
0.5
0.5
Cc (3.7%)
-0.5
200
6
0.0
fitted function
0.0
size (4.1%)
0
4
-0.5
0.5
0.0
fitted function
5000 10000
0.5
2
DO (16.4%)
-0.5
0.5
0.0
-0.5
0
fitted function
5
-0.5
2
Rs (19.2%)
-0.5
0.5
1
0.0
fitted function
0
H
L
N
veg (0.6%)
0.0
6
-0.5
5
0.0
4
-0.5
3
Gh (36.8%)
-0.5
0.5
2
0.0
fitted function
0.0
-0.5
1
0.5
Carp gudgeon abundance
0
-0.5
0.5
The final model contained 2800 trees and explained 73% of the deviance. The most important
variables in the model were Eastern Gambusia abundance, which contributed 37% of the
variance, followed by Australian Smelt abundance, DO and turbidity. Carp Gudgeon
abundance was positively related to each of these variables (Figure 3.3.6).
H
L
M
N
debris (0.3%)
Predictor variable
Figure 3.3.6 Fitted functions describing relationships between Carp Gudgeon abundance
and all predictor variables for the final BRT model.
44
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Flat-headed Gudgeon abundance
4
6
8
10
12
600
100
200
300
9.5
0
2
3
Rs (0.3%)
4
5
20000
0.10
fitted function
1
2
3
4
5
6
15
20
25
30
35
1.5
2.0
0.10
Temp (4%)
-0.10
1
5000 10000
size (7.8%)
0.10
fitted function
0.10
-0.10
0
0.00
0.10
9.0
Gh (4.4%)
0.00
fitted function
0.10
N
8.5
-0.10
0
Turb (4.8%)
-0.10
L
veg (0.7%)
8.0
-0.10
0
Cond (6%)
H
7.5
0.10
fitted function
0.10
-0.10
400
7.0
pH (10.9%)
0.00
fitted function
0.10
0.00
200
-0.10
6.5
DO (20.9%)
0.00
2
-0.10
0
fitted function
0.10
-0.10
6
0.00
5
fitted function
4
-0.10
3
Hsp (39.9%)
0.00
2
0.00
1
0.00
fitted function
0.10
0.00
-0.10
-0.10
0
0.00
Flat-headed gudgeon abundnace
0.00
fitted function
0.10
The final BRT model contained 500 trees and explained 15% of the deviance. Carp Gudgeon
abundance was by far the most influential variable, evidenced by a strong positive
relationship with Flat-headed Gudgeon abundance. DO and pH were also positively related to
Flat-headed Gudgeon abundance, whereas Eastern Gambusia abundance had minimal little
influence (Figure 3.3.7).
H
L
M
debris (0.3%)
N
0.0
0.5
1.0
Cc (0%)
Predictor variable
Figure 3.3.7 Fitted functions describing relationships between Flat-headed Gudgeon
abundance and all predictor variables for the final BRT model.
45
Native fish recovery following Eastern Gambusia removal
Common Carp abundance
7.5
8.0
8.5
9.0
9.5
5
6
100
200
300
fitted function
0.10
8
10
12
DO (2.8%)
25
30
0.0
0.5
1.0
1.5
Ca (1.7%)
0.10
0.00
35
2.0
400
600
0
1
2
3
4
5
Rs (4.1%)
-0.10
0.00
fitted function
6
20
Temp (4.8%)
-0.10
0.00
4
fitted function
15
Turb (5.6%)
-0.10
2
fitted function
-0.10
0
Gh (7.5%)
200
Cond (9.7%)
fitted function
4
0
-0.10
3
20000
-0.10
fitted function
0.10
-0.10
2
5000 10000
size (12.5%)
0.00
fitted function
0.10
0.00
-0.10
1
-0.10
0
0.10
7.0
0.00
6.5
pH (17.3%)
0.10
Carp abundance
0
0.10
-0.10
6
0.10
5
0.00
4
0.10
3
Hsp (30.5%)
0.00
2
0.10
1
0.00
0
0.00
fitted function
0.10
0.00
-0.10
-0.10
0.00
fitted function
0.10
The optimal BRT model contained 200 trees and explained 43% of the deviance. The most
influential variable in determining Common Carp abundance was Carp Gudgeon abundance,
with a clear negative association between the two variables. The pH of the wetland and its
size were the next most important predictor variables. Common Carp abundance showed a
slight negative relationship with Eastern Gambusia abundance (Figure 3.3.8).
H
L
N
veg (1.7%)
H
L
M
N
debris (1.3%)
Predictor variable
Figure 3.3.8 Fitted functions describing relationships between carp abundance and all
predictor variables for the final BRT model.
3.3.3
3.3.3.1
The influence of Eastern Gambusia on juvenile condition of native species
Body condition
The assessment of body condition in relation to Eastern Gambusia abundance at particular
sites was undertaken for 223 juvenile Australian Smelts, 336 juvenile Carp Gudgeons and 106
juvenile Flat-headed Gudgeons. Condition was expressed in two forms: the residuals derived
from the population’s length–weight relationship and the relative condition factor (Krel). The
linear regression models showed significant negative relationships between both indices of
juvenile condition of all three species as Eastern Gambusia abundance increased (all
p < 0.001: Figure 3.3.9). Although marked variation in condition scores for each of the
species was apparent at some sites, resulting in relatively low R2 values for these relationships,
the data suggest a general decline in body condition for the three native species in sites
containing high Eastern Gambusia densities.
46
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Australian smelt (n=223)
p < 0.001
R2 = 0.1856
L-W residuals
80
Carp gudgeon (n=336)
p < 0.001
R2 = 0.114
30
Flat-headed gudgeon (n=107)
p < 0.001
R2 = 0.239
60
60
20
40
20
10
20
0
0
40
-20
0
-10
d
u
le
a
rW
si
-
-20
-40
L
-20
-60
p < 0.001
R2 = 0.074
1.2
K rel
p < 0.001
R2 = 0.199
1.4
1.2
1.0
le
r
p < 0.001
R2 = 0.116
1.4
1.0
1.2
0.8
1.0
0.8
K
0.6
0.6
0.4
0.8
0.4
0
10
20
30
40
0
10
20
30
40
0
5
10
15
20
Eastern gambusia CPUE
Figure 3.3.9 Juvenile condition of Australian Smelt, Carp Gudgeon and Flat-headed
Gudgeon plotted against Eastern Gambusia abundance. Condition indices are expressed as
residuals of each species’ length–weight (L–W) relationship (top), where 0 = average
condition, and the relative condition factor (Krel; bottom). Fitted and observed relationships
of the regression analysis are presented with 95% confidence intervals.
3.3.3.2
Fin condition
The probability of caudal fin damage in juvenile Australian Smelt rose significantly as
Eastern Gambusia abundance increased: Australian Smelt were 1.4 times more likely to
sustain damage with every one unit increase in Eastern Gambusia CPUE (logistic regression:
odds ratio = 1.407, p = 0.029) (Figure 3.3.10). The extent of aquatic vegetation cover and
debris load had little effect on the likelihood of fin damage in juvenile Australian Smelt, and
the interactions between these variables and Eastern Gambusia CPUE were also not
significant (all p > 0.100).
Carp Gudgeon fin condition was largely unaffected by increases in Eastern Gambusia
abundance (Figure 3.3.10). The logistic model actually predicted a slight decrease in
likelihood of fin damage as Eastern Gambusia CPUE increased, although this effect was not
significant. Similarly, a negative relationship was observed between the extent of aquatic
vegetation cover and the probability of fin damage, but again this relationship was relatively
weak. Higher debris loads significantly reduced the probability of fin damage across our sites
however, with fin damage estimated to be almost four times less likely in sites with medium
versus low debris, and high versus medium debris loads (odds ratio = 0.283, p = 0.029).
Interactions between Eastern Gambusia CPUE, and both aquatic vegetation cover and debris
load were not significant (all p > 0.100). Flat-headed Gudgeon showed minimal fin damage
across all sites, and the probability of damage remained low regardless of increases in Eastern
Gambusia CPUE or variation in aquatic vegetation cover and debris load.
47
Native fish recovery following Eastern Gambusia removal
Carp gudgeon (n=173)
1.0
p = 0.029
p = 0.128
p = 0.279
0.8
Value
DAMAGE
Value
DAMAGE
0.8
0.6
0.6
0.4
0.4
0.6
0.4
0.4
0.2
0.2
0.2
2
3
p = 0.336
4
0
0.4
3
High
0.0
1
None
2
Low
AQUAVEG
3
High
1
None
p = 0.029
Value
DAMAGE
Value
DAMAGE
0.6
0.6
0.4
0.2
0.0
0.0
4
High
2
Low
p = 0.235
0.8
0.4
0.2
0.2
0.0
3
High
0.6
0.4
0.4
0.2
2
Low
AQUAVEG
1.0
0.8
0.8
0.6
3
Med
DEBRIS
0.4
0.2
1.0
1.0
p = 0.545
4
0.6
Aquatic vegetation cover
1.0
3
p = 0.807
0.8
Value
DAMAGE
0.6
0.0
2
Low
AQUAVEG
2
LOGCPUE
p = 0.377
0.2
0.0
1
1.0
0.8
0.2
2
Low
3
1.0
0.4
0.8
2
Log (eastern gambusia CPUE+1)
0.6
1
None
1
LOGCPUE
DAMAGE
Value
0.8
0.0
0
4
Value
DAMAGE
1
1.0
DAMAGE
Value
0.2
0.0
0.0
LOGCPUE
Probability of fin damage
1.0
0.8
0.6
0
Probability of fin damage
Flat-headed gudgeon (n=69)
1.0
1.0
0.8
Value
DAMAGE
Probability of fin damage
Australian smelt (n=183)
0.0
3
Med
DEBRIS
4
High
Low
2
Debris load
Med
3
DEBRIS
High
4
Figure 3.3.10 Plots showing the probability of fin damage for Australian Smelt, Carp
Gudgeon, and Flat-headed Gudgeon in relation to Eastern Gambusia abundance, the extent
of aquatic vegetation and the debris load across all sampled wetlands. Fitted curves from
the logistic regression analyses are shown with 95% confidence intervals. Data points
represent observations of damage (coded as 1) or no damage (coded as 0).
48
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
3.4 Discussion
3.4.1
Assessment of wetland fish communities of the mid-Murray River
The cross-sectional study provided a snapshot of wetland fish community structure and
associated environmental variables across a large portion of the southern MDB. The
dominance of Carp Gudgeon, Eastern Gambusia and (to a lesser extent) Australian Smelt
across the four regions surveyed was expected, given the timing of sampling (late summer –
early autumn), and the results from previous work (e.g. Humphries et al. 2002; McNeil 2004;
McMaster and Bond 2008). Yet there were some differences in fish community composition
when compared with past surveys in the MDB (e.g. King et al. 2003; McNeil 2004; King et
al. 2007; Tonkin and Rourke 2008; Rehwinkel and Sharp 2009), namely the total absence of
Flat-headed Galaxias from any of our sites, the restriction of Southern Pygmy-perch to the
lower Ovens River floodplain, and in very low numbers, and the presence and in some cases
dominance of Flat-headed Gudgeon at particular Ovens River sites.
In designing this phase of the study, we hypothesised that the control exerted by Eastern
Gambusia on the diversity, abundance and condition of sympatric fishes would be maximised
during late summer and early autumn in MDB wetland systems. We predicted that peak
Eastern Gambusia densities would coincide with low water levels at this time, subsequently
increasing ecological niche overlap with sympatric species and raising both the intensity and
frequency of negative interactions. However, the interplay of many biotic and abiotic forces
could shape fish community structure in freshwater ecosystems (Winemiller et al. 2000;
Matthews and Marsh-Matthews 2003; Magoulick and Kobza 2003; and Prince 2010). In
hydrologically variable environments such as floodplain wetlands, the timing, magnitude and
duration of flooding and drying plays a pivotal role in mediating dispersal, immigration and
reproductive opportunities and controlling physicochemical conditions, resource availability
and habitat use — factors that can ultimately define fish species diversity, abundance, and the
extent and nature of intraspecific and interspecific interactions (e.g. Jackson et al. 2001;
McNeil 2004, Arthington et al. 2005; Beesley and Prince 2010).
Factors associated with dispersal and colonisation may be more influential during or shortly
after inundation, with predation, competition and environmental stressors becoming more
important as drying proceeds (Matthews and Marsh Matthews 2003; Magoulick and Kobza
2003, McNeil 2004). During the survey the wetlands we sampled were rapidly contracting,
exposing the local fish communities to increasingly harsh physicochemical conditions (e.g
low DO and high water temperatures) and at the same time intensifying potential competition
for ever-diminishing resources. Such conditions are likely to favour the persistence of species
with high physiological tolerances and morphological or behavioural adaptations that enable
them to exploit these situations, species that can successfully buffer competition effects by
occupying different ecological niches, and ‘generalist’ species that are highly flexible in
relation to resource use (see Poff and Allan 1995; Magoulick and Kobza 2003; McNeil and
Closs 2007).
Our data support this prediction. The two most abundant species captured during the surveys,
Eastern Gambusia and Carp Gudgeon, have high tolerances to physiological stress (McNeil
and Closs 2007; McMaster and Bond 2008), high reproductive capacities, and protracted
spawning periods over the summer months. These species coexisted in over 80% of wetlands
sampled and appeared to occupy similar habitats in the littoral zone, often in extremely high
densities. Low capture rates of certain key wetland species (Southern Pygmy-perch and
Murray–Darling Rainbowfish) and the total absence of Flat-headed Galaxias during the study
unfortunately precluded their inclusion in the correlative modelling component. In the
following discussion we focus on these rarer species and discuss the mechanisms that may
have contributed to their low abundances in our surveys.
49
Native fish recovery following Eastern Gambusia removal
Adult Carp Gudgeon collected from lower Goulburn wetland.
50
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Southern Pygmy-perch
The southern pygmy-perch is patchily distributed in rivers and wetlands in northern Victoria
and has undergone a major decline in distribution and abundance throughout the MDB,
particularly throughout New South Wales, where it is endangered (Tonkin et al. 2008). Loss
of habitat and the effects of river regulation have been cited as the primary reasons for the
observed declines (Arthington et al. 1983; Lintermans 2007; Tonkin et al. 2008). Although
only very few fish were captured in our study, the off-channel habitats within Barmah–
Millewa Forest (BMF) and adjacent to the unregulated lower Ovens River, where most our
effort was focused, have supported healthy Southern Pygmy-perch populations in the past,
particularly during years of substantial floodplain inundation (McNeil 2004; Tonkin et al.
2008). A study examining trends in abundance and recruitment of the species in wetlands of
the BMF over five years and under varying hydrological conditions found that dispersal and
recruitment was enhanced in a year of prolonged spring–summer floodplain inundation
(associated with a major environmental watering event in 2005–6), coinciding with the
highest abundances (Tonkin et al. 2008). The authors concluded that reduced flooding
frequency associated with river regulation may be an important driver of the species’ decline
in the MDB. In light of these findings, it is possible the lack of significant flooding in BMF in
the years preceding our cross-sectional study may have diminished the Southern Pygmy-perch
population in the area, resulting in the lack of detection in BMF.
Our sampling was limited to late summer – early autumn in 2008–9 and 2009–10, a time
when we predicted that overlap in trophic niches and the intensity of antagonistic interactions
between Eastern Gambsuia and native fishes would be greatest. Southern Pygmy-perch is
known to occupy habitat and trophic niches that are very similar to those of Eastern Gambusia
(see Table 2.1), and despite its apparently high tolerance to physicochemical extremes
(McNeil and Closs 2007; McMaster and Bond 2008; but see Morrongiello et al. in press) the
species may be particularly susceptible in shrinking wetlands during late summer. Following
large-scale floods that inundated the entire lower Ovens floodplain in 1996, McNeil (2004)
documented substantial shifts in wetland fish assemblages as drying progressed from spring to
late summer. During spring sampling in November 1996, he noted that Carp Gudgeon
dominated floodplain fish communities, Southern Pygmy-perch was abundant and widely
distributed, and Eastern Gambusia was found in large numbers only in a few wetlands. As
wetlands contracted through summer, numerical dominance switched from Carp Gudgeon to
Eastern Gambusia. By the late dry season (February–March 1997), captures of Southern
Pygmy-perch were rare and all billabongs were arranged along a gradient based solely on the
relative dominance of Carp Gudgeons and Eastern Gambusia.
Notwithstanding the strong evidence relating altered flow regimes to declines in populations
across the MDB (e.g. Tonkin et al. 2008), seasonal decreases in population size in relatively
unmodified systems, as reported by McNeil (2004), point more to interactions of local
physical and biological factors associated with the progression of drying in these habitats
(Matthews and Marsh-Matthews 2003; Morrongiello et al. in press). In addition, the weight of
correlative evidence from our study and that of McNeil (2004) and others (e.g. Lloyd and
Walker 1986; Koster 1997; SRA 2007) suggests that Eastern Gambusia has at least some
level of negative impact on Southern Pygmy-perch throughout the MDB. Clearly, further data
is need to define the level of impact and assess the relative risk to the species at key sites. We
discuss these issues further in relation to Phase 3 of the project (see Chapter 5).
51
Native fish recovery following Eastern Gambusia removal
Flat-headed Galaxias
The Flat-headed Galaxias has undergone substantial range contraction across much of the
southern MDB in recent decades (Morris et al. 2001; Lintermans 2007; Fisheries Scientific
Committee 2008), and in 2008 the species was listed as critically endangered in New South
Wales under the state’s Fisheries Management Act 1994. Apart from aspects of its
reproductive and larval biology (see Llewellyn 2005), little information is available on the life
history of the species or the causes of its decline. Habitat degradation and limitations on
dispersal associated with altered flow regimes, and recruitment failure caused by releases of
cold water from dams in addition to predatory and competitive interactions with alien (e.g.
redfin, Common Carp and Eastern Gambusia) and native fish species (e.g. climbing galaxias
Galaxias brevipinnis) have been suggested as possible factors (McNeil 2004; Lintermans
2007).
Billabongs and wetlands associated with north-central Victorian rivers have harboured
healthy populations of the species in the recent past (McNeil 2004; Lintermans 2007; McNeil
and Closs 2007). McNeil (2004) captured large numbers of Flat-headed Galaxias on the lower
Ovens floodplain following widespread flooding of the region in winter–spring 1996. In
explaining the declines in abundance observed from spring to late summer, McNeil (2004)
suggested that reduced DO in small wetlands and predation by Redfin in larger ones may
constrain the species distribution across the floodplain (see also Stoffels and Humphries 2003;
McNeil and Closs 2007). He attributed the persistence of Flat-headed Galaxias in relatively
marginal habitats to a trade-off between environmental harshness and predation pressure from
Redfin, and suggested that in the absence of Redfin, the distribution of Flat-headed Galaxias
would expand and its and recruitment potential would increase. Because we captured just nine
small Redfin from 6 of the 83 sites sampled in the Lower Ovens region, we suggest that
alternative factors must have contributed to the absence of flat-headed galaxias on the
floodplain across the two years of the study. Opportunities for colonisation of our sites from
the Ovens River main channel did exist in both years, as was the case during the 1999 and
2000 surveys by King et al. (2003) who captured only two individuals across both years, so
the extent of connectivity does not appear be a limiting factor. Similarly, anthropogenic
disturbance effects and habitat degradation are likely to be minimal at the majority of these
sites, as the unregulated Ovens River provides a relatively natural cycle of flooding and
drying to the floodplain, which is in state forest dominated by River Red Gum (Eucalyptus
camaldulensis). Although we did not sample these wetlands in spring when physicochemical
and other environmental stressors are likely to be less severe, factors associated with wetland
drying over summer (see section 3.4.1) and concurrent peaks in (1) abundance of Eastern
Gambusia and (2) competitive pressure from juvenile Common Carp, Carp Gudgeons and
Australian Smelt, which occupy similar dietary niches, may interact to suppress spawning and
increase mortality in these closed systems. Further targeted sampling is required to assess the
current status of Victorian Flat-headed Galaxias populations, but our data suggest that the
species may be at risk of local extinction in a region that once supported healthy populations.
Murray–Darling Rainbowfish
The absence of Murray–Darling Rainbowfish from the lower Ovens River wetlands in our
study was not unexpected, because previous studies in the region had failed to detect the
species (King et al. 2003; McNeil 2004). The few that we did capture in the remaining
wetlands in Gunbower and BMF were at least 42 mm long (TL) and most likely in their
second year of life (Milton and Arthington 1984). It is significant that we did not capture
juveniles at any of our sites, as this suggests that: (a) local spawning did not occur during the
previous spring – early summer, which is the known spawning period of the species
52
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
(Backhouse and Frusher 1980; Humphries et al. 2002), (b) we were unable to detect it with
our sampling techniques, or (c) that biotic or physical factors (or both) increased mortality
during the early life stages. Murray–Darling Rainbowfish are still collected within creek and
main river channel environments throughout the Barmah–Millewa and lower Goulburn region
(e.g. King et 2007; Tonkin and Rourke 2008), indicating this species has a much broader
trophic niche than the two species previously mentioned.
In terms of biological factors, the threats posed by Eastern Gambusia to melanotaeniids have
been well documented (e.g. Arthington 1991; Arthington and Marshall 1991). Predation on
eggs and larvae has been reported (Ivanstoff and Aarn 1999) and suggested (Lukies 2004),
and — because of the substantial overlap in dietary and habitat niches identified for Eastern
Gambusia and Murray–Darling Rainbowfish across all life stages (see Table 2.1; Lukies
2004; Kellaway et al. 2010) — we suggest that competitive and predatory impacts of Eastern
Gambusia may go some way to explaining the increasing patchiness of the species across the
southern MDB. If this is the case, then the late and restricted spawning time of species such as
the Murray–Darling Rainbowfish may magnify the impacts (Pen et al. 1993). Spawning
during October, November and December (Backhouse and Frusher 1980; Humphries et al.
2002) exposes newly hatched larvae and juvenile stages to rapidly expanding Eastern
Gambusia populations, which thus could exert increasing control where emigration
opportunities are limited; for example, in rapidly drying wetlands such as those sampled
during our study.
Little information exists on the effects of environmental factors on rainbowfish abundance.
Flooding does not appear to affect spawning time or larval production in the species (King et
al. 2010), and larvae rarely drift (Humphries et al. 2002). The species was never collected
during a seven-year period in the highly regulated Campaspe River in central Victoria, yet it
was captured in the mildly regulated Broken River (Humphries et al. 2002). But because
temperature — the suggested spawning stimulus (Backhouse and Frusher 1980; Milton and
Arthington 1984) — is often independent of flow, and egg development requires aquatic
vegetation (Milton and Arthington 1984), the absence of the species from the Campaspe
likely reflects other factors such as habitat quality and availability rather than river regulation.
Because of small samples sizes in the present study it is not possible to accurately define the
relative importance of each of these processes in structuring populations of Murray–Darling
rainbowfish. It may be that the simple absence of the species from so many sites containing
high densities of Eastern Gambusia is enough to infer that there is a negative impact (without
identifying the mechanism), but clearly more data is needed to validate this inference.
Murray–Darling Rainbowfish collected from a wetland in the Gunbower Forest.
53
Native fish recovery following Eastern Gambusia removal
3.4.2
The influence of Eastern Gambusia on species occupancy, abundance and
condition
Identifying the factors that currently influence fish community composition across southern
MDB wetlands, and defining the nature and extent of the impact of Eastern Gambusia, were
key objectives of the cross-sectional study. If the wetland fish community is structured by
local abiotic factors, then we would expect strong relationships to exist between
physicochemical parameters or habitat variables and species abundances. Conversely, if
competitive or predatory forces dominate, clear patterns in co-occurrence or relationships
between relative abundances of different species are predicted.
The use of correlative or species distribution models (SDMs) is an intuitive way to explore
and visualise such relationships and develop hypotheses about the mechanisms underlying the
patterns observed (Elith and Leathwick 2009; Capinha and Anastácio 2011). Such models can
also be used in a predictive capacity to make forecasts on likely changes in species
distributions and abundances under a range of future scenarios of ecological change. Boosted
regression trees (BRTs) are a relatively new type of SDM, based on traditional regression and
more modern machine learning techniques (Elith et al. 2008). Their use is increasing among
freshwater and marine fish ecologists who require robust tools to explain and predict patterns
of occurrence at numerous spatial and temporal scales (see Elith et al. 2008; Leathwick et al.
2006, 2008).
The correlative models we constructed were based on this framework. They were designed to
gain an insight into relationships the presence and abundance of Eastern Gambusia and the
presence and abundance of other species in the sampled wetlands, and to examine the
influence of physicochemical and other environmental parameters on these relationships. The
results from the occupancy modelling clearly suggests that the occurrence of Eastern
Gambusia in a wetland has a strong effect on the probability of the four most common native
species captured in our surveys (Carp Gudgeon, Australian Smelt, Flat-headed Gudgeon and
un-specked hardyhead) being present. The magnitude of the effect for these species appears to
depend on the amount of aquatic vegetation, although it is notable that all four species were
most likely to occur in the presence of Eastern Gambusia, regardless of the presence or
absence of aquatic vegetation. The high probability of co-occurrence of these species with
Eastern Gambusia is not unexpected, since Eastern Gambusia occupied 83% of all sites
sampled. The results, in effect, highlight the success of Eastern Gambusia as an invader and
coloniser of these wetland systems. Its broad tolerance of environmental extremes (Karolak
2006; McNeil and Closs 2007), high reproductive capacity (Milton and Arthington 1983) and
ability to thrive in disturbed habitats (Arthington et al. 1983; Kennard et al. 2005) is reflected
in its wide distribution in the MDB and beyond, and its extraordinary capacity to rapidly
dominate fish communities following the onset of spawning (see Chapter 4). The output from
the BRT models provides a deeper insight into the relationships between Eastern Gambusia
abundance and potential predictor variables in the sampled wetlands. The results suggest that
environmental variables have little influence on Eastern Gambusia abundance — a result
which aligns well with the characteristics of a broadly distributed, highly tolerant alien
species. By far the most significant predictor of Eastern Gambusia abundance was the
abundance of Carp Gudgeon, with a strong positive relationship evident. From the data we
have it is not possible to say whether or not these species show partial segregation in
microhabitat niches at scales undetectable by our sampling, or employ other behavioural traits
that limit competition between them at particular life stages (but see Stoffels and Humphries
2003). However, it is clear that both species are thriving in most wetlands across the southern
MDB.
Our findings also illustrate the resilience of the most common native species to high Eastern
Gambusia densities. At sites where Eastern Gambusia was present during late February –
54
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
March, it was almost always very abundant. When we predicted the likely outcomes of the
study we expected that such high densities combined with the increasingly harsh
physicochemical conditions and other stressors associated with wetland drying would act to
decrease the abundance of many common wetland species. This may have been the case with
some rarer species (e.g. Southern Pygmy-perch, Murray–Darling Rainbowfish, Flat-headed
Galaxias), although the small numbers of individuals collected during our study limit the
conclusions we can reach for these species. In any case, the observation that Carp Gudgeon,
Australian Smelt and Flat-headed Gudgeon were able to persist at many sites under such
conditions (albeit at lower abundances or decreased morphometric condition at some sites)
reflects the ability of these species to adapt quickly to changes in ecosystem processes
associated with rapid increases in Eastern Gambusia abundance, and to resist some of the
effects of drought disturbance (Morrongiello et al. 2006; McNeil and Closs 2007; McMaster
and Bond 2008; Crook et al. 2010).
Carp Gudgeon typically inhabited wetlands containing high abundances of Eastern Gambusia
and Australian Smelt, and tended to prefer those with higher DO and moderate to high
turbidity. As noted above, the positive relationships observed with Eastern Gambusia and
Australian Smelt may reflect differential use of habitat and dietary resources by one or more
of these species at different developmental stages (e.g. Stoffels and Humphries 2003). Any
such separation of habitat or trophic niche would ultimately decrease the frequency of
interactions between species, and hence competition with Carp Gudgeon. Alternatively, such
resources may not be limited in some wetlands, enabling these species to persist even when
ecological niches need to be shared (Stoffels and Humphries 2003; Kellaway al. 2010). In
analysing species abundance data from the lower Ovens, McNeil (2004) reported that,
although Carp Gudgeon and Eastern Gambusia distributions overlapped substantially on the
floodplain, higher Carp Gudgeon abundances were generally associated with bigger, deeper,
more permanent wetlands, lower macrophyte cover and higher pH. Eastern Gambusia
abundance was, by contrast, associated with smaller, shallower, ephemeral habitats with
higher macrophyte cover and lower pH. McNeil (2004) also showed that Carp Gudgeon
abundance was positively correlated with minimum oxygen concentrations in these wetlands,
a finding corroborated by our data. However, apart from DO and a weak positive relationship
with turbidity, none of the other environmental parameters that we measured appeared to
influence abundances for this species at the time of our sampling. Relative body condition in
Carp Gudgeons was shown to decrease at higher Eastern Gambusia densities, but there was
no such relationship with fin condition. However, caudal fin damage was considerably less at
sites with higher debris loads. We conclude from this data and from previous work that a mix
of biotic (i.e. competitive interactions, or their absence) and abiotic controls (i.e. DO levels,
habitat complexity) have contributed to the abundance patterns and condition scores observed
for Carp Gudgeon.
Australian Smelt was widely distributed across our sites and was collected in moderate
numbers over the two years of the study. In each year the catch was dominated by recruits
spawned in the previous spring, although there were small numbers of adult fish that may
have been more than one year old. At sites where Eastern Gambusia abundance was high,
many Australian Smelt captured showed visible signs of caudal fin damage and were
generally in poor condition. This species is known to occupy trophic and habitat niches that
are similar to those of Eastern Gambusia. Combining that with our knowledge of their
behaviour and physiological tolerances (see McNeil 2004; McNeil and Closs 2007), we
hypothesised that Eastern Gambusia would impart a strong influence on the abundance and
body and fin condition indices at sites where the two species coexisted, particularly where
habitat complexity was low. The output from the BRT models revealed a weaker negative
relationship between Eastern Gambusia and Australian Smelt abundance than we expected.
55
Native fish recovery following Eastern Gambusia removal
Rather, the abundance of Carp Gudgeon and wetland size were most important in explaining
the abundance of Australian Smelt, which was more likely to occupy larger wetlands
containing a large number of Carp Gudgeons. Interestingly, habitat complexity and
physicochemical predictors had very little influence on Australian Smelt abundance, which
again suggests a high tolerance to a range of environmental conditions associated with
wetland drying. As predicted, the assessment of morphometric condition revealed a
significant decline in both body and fin condition with increasing Eastern Gambusia
abundances. Our data suggest that these effects may be exacerbated in smaller wetlands,
despite the extent of habitat complexity having little effect. The positive correlation between
Australian Smelt abundance and wetland size suggests that the permanence of refuge habitats
may influence recruitment success and persistence of the species under conditions of
environmental stress, and that larger wetlands may provide some release from competitive
pressures conferred by other sympatric species that occupy similar trophic niches (Kellaway
et al. 2010).
Flat-headed Gudgeon was prevalent in the lower Ovens sites during our surveys, a result that
contrasts markedly with previous work in region. King et al. (2003) sampled the area over
two years, incorporating a year of minimal floodplain inundation during spring and summer
(1999–2000) and a year of repeated flooding (2000–1). Although only small numbers of Flatheaded Gudgeon were captured in both years, a slight increase in numbers was recorded in
2000–1 coinciding with regular floods during the sampling period and prolonged periods of
floodplain inundation. In late 1996, following widespread floods that inundated all floodplain
habitats, McNeil (2004) failed to capture a single individual while sampling shortly after
floodwaters had subsided and again later that summer. Our sampling comprised a large range
of size-classes and a high proportion of juveniles, suggesting that either spawning had
occurred locally or that a large influx of early-stage fish had colonised these sites during a
recent connection to the main stem (see Lyon et al. 2010). This is noteworthy because the
species does not typically use the floodplain for larval development (Lintermans 2007). Large
adult Flat-headed Gudgeon appeared to completely dominate the fish community at some
sites, which were generally characterised by low species diversity and very low abundances of
other species, including Eastern Gambusia. Although we predicted a high risk of negative
impacts from Eastern Gambusia, based on habitat and dietary niche overlaps, our data,
combined with knowledge of the species utilisation of benthic habitats (Kilsby and Walker in
press) suggests that impacts would most likely be on juvenile stages in situations where
habitat and food resources were limited.
At the time of our sampling in the lower Ovens region, a positive relationship was found
between Flat-headed Gudgeon and Carp Gudgeon abundance, and Flat-headed Gudgeon were
more likely to inhabit wetlands with relatively high DO and pH. The relative abundance of
Eastern Gambusia or any other habitat or environmental variables that we measured did not
substantially affect Flat-headed Gudgeon densities. These findings, in addition to the lack of a
relationship between fin damage in the species and Eastern Gambusia abundance at our sites,
points to a partitioning of resources between these two species, which is not unexpected given
their natural benthic versus pelagic behavioural traits and habitat preferences. Whether
resource use is more partitioned at sites with higher Eastern Gambusia abundances we are
unable to say from our data, but clearly factors other than Eastern Gambusia abundance are
the major determinants of Flat-headed Gudgeon abundance in these systems. Flat-headed
Gudgeon is considered highly tolerant to low hypoxic conditions (Gee and Gee 1991; McNeil
2004), so although higher DO habitats were preferred in our study, DO is not likely to
constrain the distribution or abundance of the species. McNeil (2004) reported that the sudden
appearance of Flat-headed Gudgeon on the lower Ovens floodplain during 2000 (a year of
repeated and substantial spring flooding) coincided with a large influx of Common Carp into
56
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
the wetlands, and suggested that the same mechanisms that drive carp recruitment and use of
off-channel habitats in the region (i.e. flows) may be also determine the distribution of Flatheaded Gudgeon. We found no correlation with carp abundances in our surveys, but the
importance of connectivity with the Ovens main channel for dispersal and colonisation
opportunities is obvious (Lyon et al. 2010).
One of the more interesting results from the study was that juvenile Common Carp, a highly
successful and pervasive alien species that commonly makes use of floodplain habitats in the
MDB for spawning and larval development (King et al. 2003; Koehn et al. 2004; Macdonald
and Crook 2006; Stuart and Jones 2006), occurred in low numbers with Eastern Gambusia in
the sampled wetlands. These species were significantly more likely to coexist in wetlands
containing aquatic vegetation, and substantial numbers of Common Carp were collected only
in the complete absence of Eastern Gambusia. Common Carp is highly resilient and resistant
to the impacts of drought (Crook et al. 2010) and, like Eastern Gambusia, has an extreme
tolerance of poor water quality and the effects of anthropogenic disturbances (Kennard et al.
2005), a very high fecundity (Sivakumaran et al. 2003) and a flexible diet (Crivelli 1981).
Whether carp and Eastern Gambusia populations exert any simultaneous control upon each
other in these wetlands is unknown. However, because the vast majority of carp collected
were at least a few months of age, having been spawned during the previous spring, with a
small number of fish probably older than one year also captured (Brown et al. 2005; Smith
and Walker 2004), it is likely that any impacts associated with predatory or competitive
interactions between the species would have peaked during late spring (well before our
sampling), so that early-stage carp larvae would have been exposed to pressures from everincreasing densities of Eastern Gambusia. It is important to note, however, that our sampling
techniques would not have captured large adult Common Carp effectively even if they were
present at our sites.
The BRT models indicate that Common Carp abundance was strongly negatively correlated
with Carp Gudgeon abundance. This suggests that either Carp Gudgeon are having a direct or
indirect impact on carp populations (most likely through resource competition between early
life stages) or that sites with high Carp Gudgeon abundances are in some way less prone to
invasion by Common Carp. Our sampling was restricted to higher quality sites on the lower
Ovens floodplain, but it is likely that carp may be advantaged in more degraded systems
where human disturbance can act to reduce the biological resistance of the native fish fauna
(Kennard et al. 2005; Crook et al. 2010). We predict that in those circumstances the influence
of Carp Gudgeon and local physical factors such as pH would be lessened, and that other
landscape factors such as the frequency and timing of flooding, and access to spawning and
nursery habitat on the floodplain would be major factors (Crook and Gillanders 2006;
Macdonald and Crook 2006; Stuart and Jones 2006).
57
Native fish recovery following Eastern Gambusia removal
Juvenile carp collected from lower Ovens wetland.
Billabong on the lower Ovens floodplain with healthy riparian and aquatic vegetation
and woody debris. Does this diverse habitat availability buffer some native species from
negative alien species interactions?
58
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
3.4.3
Hypotheses for the influence of Eastern Gambusia on wetland fish
communities
Correlative models provide a platform to examine current trends in species diversity,
abundance and environmental parameters in MDB wetlands. The models we used were
essentially hypothesis-development tools that could allow predictions to be made on the likely
response of wetland fish communities to different levels of Eastern Gambusia reduction in
these systems, and then tested in a field-based control program.
Using the results from the models, we developed four hypotheses relating to the outcomes of
Eastern Gambusia removal in disconnected wetlands and refugia for the most common
species captured. These predictions were developed under a scenario in which there was no
dispersal away from the site and no immigration into it over the course of the removal
program. Because of the low numbers of rarer species collected and the dominance of Eastern
Gambusia, and information in current literature, an additional hypothesis was included to
allow for the absence of model predictions for these species. The hypotheses were as follows:
1. Removal of Eastern Gambusia (to any level) will benefit Australian Smelt populations, as
measured by (i) increases in relative abundance and (ii) improvement in morphometric
condition.
4. Removal of Eastern Gambusia (to any level) will impart minimal changes to Carp
Gudgeon and Flat-headed Gudgeon abundances.
5. Removal of Eastern Gambusia (to any level) will improve morphometric condition for
carp and Flat-headed Gudgeon.
6. Removal of Eastern Gambusia (to any level) will impart minor increases to carp
abundances.
7. Removal of Eastern Gambusia (to any level) will benefit rarer wetland species such as
Southern Pygmy-perch, Flat-headed Galaxias and Murray–Darling Rainbowfish
populations, as measured by (i) increases in relative abundance and (ii) improvement in
morphometric condition.
Our testing the validity of these hypotheses in a field-based removal experiment is the focus
of the next chapter.
59
Native fish recovery following Eastern Gambusia removal
4
4.1
Phase 2: Trial of Eastern Gambusia control
Introduction
Despite the vast amount of literature devoted to the Gambusia species, there is a surprising
lack of knowledge about factors controlling their abundance, impacts on native ecosystems,
and the mitigation of these impacts (Pyke 2008). Studies examining trophic interactions (as
suggested by Zavaleta et al. 2001) and subsequent ecological effects of small alien species
such as Eastern Gambusia have received very little attention, particularly in comparison to
larger species such as Common Carp (e.g. Wilson 2006). Whilse we have documented a
number of controlled experimental studies in this review, rigorous comparisons of control
success under field conditions have not been undertaken. Rowe et al. (2008) recommended
full BACI (before/after control/impact) or manipulation studies (or both) for the assessment of
impacts of alien species on indigenous species. Such experiments have not yet been
undertaken on Eastern Gambusia. Furthermore, information regarding how native species
respond to reductions of an alien species in the wild is necessary before large amounts of
money are spent on large-scale control strategies. Rowe et al. (2008) noted that ‘scientific
proof of the impact of gambusia on indigenous biodiversity is likely to be required in the
future as management efforts to control gambusia increase in number and size and therefore
attract closer public scrutiny of cost’. Field-based manipulation experiments can provide
scientifically defensible information on an alien species’ impact in the natural environment
(Rowe et al. 2008) and also valuable information on controlling populations of alien species
to a level that results in a measurable improvement in native fish communities.
Although there is global recognition of the threats Gambusia pose to freshwater ecosystem
function outside their native range, there is very little information available on mitigating the
impacts. This is largely due to traditional techniques including poisoning, exclusion, egg
dehydration, direct removal, commercial harvest and habitat restoration having minimal
chance of success for smaller species such as Eastern Gambusia which occupy more cryptic
habitats (Wilson 2006). The few documentations of Gambusia species control both
internationally and within Australia focus predominantly on chemical techniques and drying
of habitats, (see McKay et al. 2001). More recently, some consideration has also been given
to management options in systems where total eradication using chemical treatments is
undesirable because of the presence of threatened species or fragile ecosystems, e.g. Maynard
et al. (2008) and Brookhouse and Coughran (2010). These techniques all call for extremely
thorough treatment and extensive knowledge of the hydrology of the area. With such limited
information available for managers, research should be directed towards the feasibility of
existing and new techniques for mitigating the impacts of Eastern Gambusia.
This chapter reports on Phase 2 of the project — a field-based trial of Eastern Gambusia
control in small isolated billabongs. The trial had three broad objectives:
1. Investigate aspects of Eastern Gambusia population dynamics to inform pest management
practices for the species.
2. Investigate the effectiveness of a variety of physical removal techniques.
3. Assess the response of native fish populations following reductions of Eastern Gambusia
populations, by determining the level of support for each hypothesis (model validation)
derived from the predictive models developed in Phase 1 of the project.
4.2
Methodology
The trial involved two major components. The first was the removal of Eastern Gambusia at
specific sites, ultimately creating a range of abundances across sites. The second was the
60
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
monitoring of the fish community to investigate the impacts of different abundances of
Eastern Gambusia and ultimately assess what responses we might expect from native fish
communities following Eastern Gambusia removal.
Because of the high variability in fish assemblages at small off-channel sites, both between
sites and across years within sites (derived from the results presented in Phase 1 of the
project; see also King et al. 2007), and, more importantly, a lack of information on methods
for physically removing Eastern Gambusia, the trial was undertaken over two seasons in an
attempt to minimise the risk of factors such as flooding affecting the experiment, while
maximising the project outputs that could be achieved within the budget and time-frame.
4.2.1
Site description and treatments
The trial was undertaken in wetland habitats on the lower Ovens River floodplain and (in the
second year only) on private land on the lower Goulburn – Campaspe catchments. These
areas were also where much of the cross-sectional data was sourced (see chapter 3.2.1). Site
selection was highly dependent on winter–spring hydrology (site connection, accessibility,
etc.) as well as the predictions of the correlative models, especially in relation to species
composition.
During the first season of the experiment (2009–10), all sites were located on the lower Ovens
River floodplain. The area experienced some minor flooding during early spring, filling
floodplain habitats such as anabranches and billabongs which had been dry for several years.
The majority of sites during the second year of the experiment (2010–011) were shifted to the
lower Goulburn and Campaspe region given the major flooding of the Ovens river sites in
spring and long range forecasts predicting further flooding in the region. The new sites were
all on private land, where there was a much lower risk of reconnecting to adjoining habitats
during late spring and summer compared to sites on the Ovens River. Several sites on the
Ovens River used in the first season were used as reference sites in 2010–11 (see below).
The population dynamics of Gambusia species involve a population increase, peak, decline,
and low phase. Based on pest management theory for organisms with similar population
synamics (e.g. rodents), we hypothesised that the best time to undertake control operations
would be during the low phase or early in the increase phase (Ramsey and Wilson 2000). In
the case of Eastern Gambusia this is in late winter – spring (spawning begins in early summer
when water temperatures exceed 16°C and day length exceeds 12–13 hours). Unfortunately,
late winter and early spring is also when the probability of flooding is highest, which would
make long-term removal difficult and risk confounding results of any removal that may
precede reconnection to other aquatic habitats. Therefore, to minimise this risk, and still
maintain removal exercises commencing during the most effective period, the field depletion
experiment commenced in late October just prior to the onset of gambusia spawning, and
following flooding and connectivity to adjoining aquatic habitats.
Based on the hypotheses derived from Phase 1, the project steering committee agreed that,
ideally, sites used in the field depletion experiment should be selected on the basis that they:
• contain a high Eastern Gambusia density
• contain at least two native species across all sites
• are smallish in size (not more than 400 m2 in area) yet should still retain water throughout
summer
• have an accessible perimeter
• have similar habitat (density snags and aquatic vegetation)
• support few other alien species (particularly Redfin).
61
Native fish recovery following Eastern Gambusia removal
Finding sites that met all of these criteria was difficult, mainly because of flooding of the
region in previous months. Nevertheless, the proposed site selection guidelines were followed
as closely as possible using data from fish surveys in adjoining waterbodies in recent years
(e.g. if large numbers of Eastern Gambusia occupied an adjoining waterbody, it was assumed
that they had migrated into a newly inundated site) as well as random seine netting and staff
experience of sampling in the region (e.g. sites that were known to retain water throughout the
summer period). This resulted in the selection of 13 sites in both the 2009/10 and 2010/11
seasons (some sites were repeated in the 2010/2011 season). Sites were assigned to one of two
treatments (repeated removal or single removal) or one of two control types (control or
reference).
The treatments and controls were as follows:
•
Repeated removal — 6 sites (3 sites in each season) × removal of Eastern Gambusia
for a period of 5 days each month + monitoring from November until February (i.e.
monthly removal).
•
Single removal — 4 sites (2 sites in each season) × removal of Eastern Gambusia for
a period of 5 days during November only (release thereafter) + monitoring from
November until February (i.e. a single spring removal).
•
Control (no removal) — 6 sites (3 sites in each season) × release of all Eastern
Gambusia + monthly monitoring from November until February (i.e. no removal).
•
Reference — 4 sites (4 sites in each season) × monthly monitoring from November
until February.
Examples of the sites used in the study and the experimental timeline are presented in Figures
4.1.1 and 4.2.1 respectively. The conclusion of the removal experiment also coincided with
the time when the cross-sectional study was undertaken (February – March), enabling a direct
comparison.
62
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Figure 4.1.1 Examples of removal sites used in the field trial.
63
Native fish recovery following Eastern Gambusia removal
4.2.2
Removal of Eastern Gambusia
Eastern Gambusia removal was carried out each month, from late October – early November
until late February, at each of the repeated removal, single removal and control sites (Figure
4.2.1). Each removal was generally carried out over a 5-day period, but the first removal in
each season was extended over a 10-day period because of the extra effort required to
undertake the mark–recapture component (see below). Eastern Gambusia removal was carried
out at both treatment and control sites, the difference between each treatment being whether
Eastern Gambusia were permanently removed and euthanised, or released back into the site
after capture. For example, the same trapping regime was carried out at all sites, along with an
equivalent effort of seine netting (based on the number of shots per square meter of site),
during each sampling trip. This standardised any influence the removal exercise may have had
on the native fish population. Reference sites were not subject to any removal exercises.
Treatment
Repeated
M,R, E
M,R, E
M,R, E
M,R, E
Single
M,R, E
M,R
M,R
M,R
Control
M,R
M,R
M,R
M,R
M
M
Reference
M
Oct
Nov
Dec
Jan
M
Feb
Mar
Figure 4.2.1 General timeline of the experiment for each of the two years, including
exercises carried out for each of the treatments (repeated removal and single removal) and
controls (control and reference). M = monitoring; R = removal; E = Eastern Gambusia
permanently removed and euthanised.
At removal sites we attempted to reduce fish numbers to as close to zero as possible. We used
targeted seine netting and a variety of trapping techniques to maximise the efficiency of
removal of Eastern Gambusia while minimising the impact of the removal activities on
sympatric native species. The specific techniques described below were based largely on
work presented by Maynard et al. (2008) in an attempt to exploit Eastern Gambusia
behaviour; specifically, its attraction to light and heat, and other microhabitat preferences (see
chapter 2.3).
Standard collapsible bait traps
It was decided that standard collapsible bait traps would be the main passive removal
technique used in the experiment because of the results obtained by Maynard et al. (2008). By
applying a ranking system to compare catch data for four trap designs, they concluded that the
standard collapsible bait trap containing a light source and set in a manner where the
inception area (entrance) was near the water’s surface was the best design to employ in an
64
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Eastern Gambusia control program. This was because it has a high CPUE during winter and is
selective for large breeding females. It is also readily available, cheap, and easy to deploy.
Therefore, at each site and for each trip, up to ten 10 bait traps were deployed. Although all
traps were set so that the inception area was near the surface of the water, several variations in
setting were applied in the first season to enable an assessment of the best technique for
maximising Eastern Gambusia collection using the standard collapsible bait traps.
Specifically, for the first season of the field depletion experiment (2009/2010), the traps:
• were set for a minimum of three daylight sets (approximately 0900–1700 h) and three
night-time sets (approximately 1800–0800 h) for each sample trip
• contained three illuminated traps (either solar lights which operate between dusk and
dawn, or a 12-hour chemical light stick placed in the trap at dusk) and a minimum of three
(maximum of seven) unilluminated traps during night-time sets for each sample trip
(Figure 4.2.2a)
• contained three traps baited with dry pet food for a minimum of three sets during daylight
hours (approximately 0900–1700 h) and three night-time sets (approximately 1800–0800
h) for the first sample trip.
At the completion of each set, the number of each species of fish captured was recorded. All
fish were then released unless the site was a designated removal site, in which case all Eastern
Gambusia were retained and euthanised.
Therefore, in addition to the primary purpose of collecting Eastern Gambusia for the field
removal experiment, the traps set in the first season provided data that was analysed to
determine whether the addition of light or bait, or being set during the day or night, improved
the catch rate of Eastern Gambusia. An assessment of trapping variables on the catch rates of
native species was also undertaken because these traps are commonly employed for general
fish surveys (e.g. in the Murray–Darling Basin Authority’s Sustainable Rivers Audit). The
second season of trapping involved the same number of replicates, but the variations were
refined in accordance with the results of the first season.
Fyke netting
To complement the bait traps used during the first season, we set a fine-mesh, single-wing
fyke net at each site for one daylight set (approximately 0900–1700 h) and one night-time set
(approximately 1800–0800 h). As with the collapsible bait traps, the inception area was set at
the water’s surface (Figure 4.2.2b). For the first removal in the first season, a 12-volt
fluorescent light was positioned in the rear chamber of the net during night-time sets in an
attempt to increase the attraction of Eastern Gambusia. This was discontinued after the first
trip. For each set, catch data (CPUE) was standardised to the number of fish collected per
hour to allow for the shorter soak time employed during the daylight set (approximately 8
hours compared to 14 hours soak time).
65
Native fish recovery following Eastern Gambusia removal
(a)
(b)
Figure 4.2.2 (a) Standard collapsible bait trap containing a solar light which automatically
operated between dusk and dawn, and recharged throughout the day, and (b) single-wing
fine-mesh fyke net. Note that both traps are set so the inception area is at the water
surface.
66
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Targeted seine netting
Because Eastern Gambusia is generally an open water species, swimming at or near the
surface, seine netting is a very reliable method for their capture. However, this method can be
difficult where there is complex habitat or the banks are steep, and it is potentially destructive
to habitat such as aquatic vegetation and fragile organisms such as larval fish. To overcome
these disadvantages, the lightweight seine nets that were used in the cross-sectional study
(7 m long × 1.5 m drop, 1 mm mesh diameter, 500 mm square purse) were employed in a
manner that specifically targeted Eastern Gambusia. This was achieved by exploiting two
aspects of Eastern Gambusia behaviour:
1.
Eastern Gambusia generally occur at or near the surface throughout spring and summer,
making them relatively easy to locate (using polarised sunglasses).
2. Eastern Gambusia’s affinity to warmth and light (see literature review) makes the areas it
occupies within specific sites very predictable, particularly during times when water
temperatures are low. For example, during late afternoon the fish will be concentrated in
in areas that receive the last remaining sunlight.
Subsequently, seine netting was employed only when Eastern Gambusia were directly
observed occupying an area, or to target an area where they were likely to be concentrated.
Once fish or such an area was located, the seine net was deployed to encircle the fish or the
designated area, and then retrieved (Figure 4.2.3a). At the completion of each targeted seine
netting the numbers of Eastern Gambusia were recorded, along with the presence of any other
native species (which were then released). The number of seine nettings was also recorded for
each site to document the effort of the removal exercises. A single seine netting took
anywhere from 30 seconds to 15 minutes, as it was highly dependent on the abundance of
both Eastern Gambusia and native species.
An attempt was also made to mimic Eastern Gambusia’s preferred areas using 100 watt,
halogen globe spotlights powered by deep-cycle, heavy duty 12-volt batteries. These lights
not only produced an immense amount of light, but also heat. During the first sample trip for
each site, one light was erected in the area receiving the last sunlight. Each was positioned at
a 45° angle, 30 cm above the water and was triggered approximately one hour before darkness
(Figure 4.2.3b). Each light was left on for approximately two hours after dark, upon which
time, a single seine netting circling the illuminated area was made, and the catch assessed.
Although the technique concentrated large numbers of zooplankton and native species, not a
single Eastern Gambusia was collected using this technique at any of the sites so it was not
repeated after the first sampling in the first season.
Sweep-net electrofishing
Although sweep-net electrofishing is used only as a general monitoring method, any Eastern
Gambusia collected within removal sites using this technique were euthanised. (See Chapter 3
for a description of the technique.) The data generated from this method were analysed in the
methods assessment.
67
Native fish recovery following Eastern Gambusia removal
(a)
(b)
Figure 4.2.3 (a) Adult Eastern Gambusia collected during a targeted seine netting, and (b)
trial of artificial light and heat using 100 watt halogen spotlights to attract Eastern
Gambusia.
68
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
4.2.3
Monitoring fish biota
All sites were monitored for fish biota to test the hypotheses generated in Phase 1 of the
project. This enabled us to assess Eastern Gambusia population trends (including the
effectiveness of removal and subsequent population growth) as well as any response of native
fish to the removal. During each trip (monthly from November to February), the fish biota of
each site was assessed by a combination of:
• Catch data generated from the removal exercises over the first four days (see chapter 4.2.2)
given the methodologies used for the Eastern Gambusia removal (bait trap, fyke nets and
seine nets) are also efficient at capturing native species; and
• Sweep net electrofishing (SNE; employed in the cross-sectional surveys) to enable a direct
comparison with results obtained in phase 1 (see chapter 3.2.2).
A minimum of three 20-second SNE passes were made before removing any Eastern
Gambusia, within complex habitats in the littoral zone (aquatic vegetation, snags/debris). As
with removal methodologies, at the completion of each SNE pass all fish were identified and
counted. For each site, a minimum of twenty individuals of each species were randomly
selected and measured for total length (TL; nearest mm). This enabled an assessment of
recruitment, and subsequent tracking of a cohort’s growth throughout the study period.
Additionally, samples of young-of-year Carp Gudgeon, Flat-headed Gudgeon, Australian
Smelt and Murray–Darling Rainbowfish collected from some sites in January and February
were euthanised in an overdose of anaesthetic (immersion for 10 mintes in 40 mg/L
alfaxalone) and then preserved in 95% ethanol to enable further measures on condition to be
made in the laboratory. Only fish collected in January and February were assessed given this
was the period where young-of-year native species and high abundances of Eastern Gambusia
coincided. All fish were measured for standard length (Ls - nearest 0.1mm) and weight
(nearest 0.001g) after a minimum of 10 days to allow shrinkage associated with preservation
to stabilise (Fey & Hare 2005). Individual fish were also examined for fin condition using the
same assessment protocol as described in Chapter 3.
Any Southern Pygmy-perch collected were released because of their rarity in the region, but
photographs of juvenile and adult fish were taken for assessing fin damage (Figure 4.2.4).
Figure 4.2.4 Adult Southern Pygmy-perch. Photographs were taken of rarer species for
assessment of caudal fin condition.
4.2.4
Mark-recapture assessment
A mark-recapture component was undertaken at selected sites to compliment estimates of
Eastern Gambusia removal efficiency and population growth derived from the monitoring
69
Native fish recovery following Eastern Gambusia removal
data. In the first season this was undertaken at two of the sites during the first removal (in
spring) and at another two sites in February. It was envisaged that this would occur at all sites,
but the extremely low numbers of Eastern Gambusia i spring 2009 and the subsequent
difficulty with their capture meant that only the two sites with the highest numbers of Eastern
Gambusia present were used. In the second season the mark–recapture assessment was
undertaken at three sites during the first trip of the season, but was not repeated in February as
a result of flooding throughout the entire experimental period, which compromised any
interpretation of population change (because of influences of immigrating and emigrating
fish).
For each of the sites where mark–recapture was used, all Eastern Gambusia collected on the
first day were marked and released back into the site. Marking was by osmotic induction,
which involves a chemical dye (calcein) that creates a fluorescent mark on bony structures of
the fish, such as fin rays and otoliths (see Appendix 1 for details). Marked fish can be rapidly
identified without being euthanised, using a handheld blue light. The procedure was favoured
for this project as it can mark a large number of fish and minimise handling before release,
compared to other techniques such as fin clipping and tagging.
In this procedure a 5% saline solution was prepared by dissolving 100 g of commercially
available natural salt in 2 L of water. A 0.5% solution of calcein was prepared by adding 10 g
of powdered calcein (2,4-bis-[N,N0-fdicarbo methylg-aminomethyl] fluorescein) to 2 L of
water. The consequential decrease in the pH of the solution was counteracted by gradually
adding sodium hydroxide until a pH of 7.0 was obtained. The salt and dye solutions were
aerated by bubbling air through the solutions during the marking procedures (Figure 4.2.5a).
Up to 20 Eastern Gambusia were placed together in a 2 L plastic container having a mesh
bottom and immersed in the salt solution for 3 minutes. After immersion in the salt solution
the fish were rinsed in fresh water and then immersed in the calcein dye solution for 5
minutes. They were then placed in a 300 L bin of aerated fresh water and assessed for
condition. When their condition was judged to be satisfactory they were released where they
were collected.
Removal of Eastern Gambusia at these sites commenced two days later to ensure that marked
fish regained normal behaviour. Eastern Gambusia collected afterwards were exposed to a
handheld blue light, which caused bony structures such as the spines, fins and head on any
fish that had been subjected to the calcein dye to fluoresce, thus identifying it as an individual
from the initial capture (Figure 4.2.5b).
The overall number of marked fish collected during each visit gave an estimation of the
population size of Eastern Gambusia, as well as an estimate of the proportion of the
population removed in the experiment.
70
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
(a)
(a)
(b)
Figure 4.2.5 (a) Field set-up for Eastern Gambusia osmotic induction marking with calcein,
with examples of marked fish. (b) Detection of marked Eastern Gambusia, with an example
of a marked fish alongside an unmarked individual.
71
Native fish recovery following Eastern Gambusia removal
4.2.5
4.2.5.1
Analysis
Assessment of Eastern Gambusia removal methodologies
Because of the large amount of data, we assessed the methodologies for the first season only,
allowing us to refine the techniques for the second season of removal. To assess the catch
rates of the different methods used (and subsequent ranking of each method as a removal
method) we analysed the standardised catch data using an unbalanced Analysis of Variance
(ANOVA), using sample trip and method as the factors and blocking for site. The response
variable was Eastern Gambusia abundance, but species-specific abundances were also tested.
This enabled a comparison between bait trapping, fyke netting, sweep net electrofishing and
seine netting CPUEs.
All trapping data was standardised to the number of fish per hour soak time, and targeted
seine netting data was standardised to the number of fish per pass. The sweep net
electrofishing data used for the monitoring was kept standard as the number of fish per 20
second pass. We considered a single targeted seine pass to take a similar amount of time as an
electrofishing pass, however, processing times (often a large by-catch of native species) and
observation time (time it would take to observe fish) is highly variable, and as a result, is
explored in the Discussion (section 4.4).
To assess the variables incorporated within each trapping regime, the standardised catch data
for bait trap and fyke netting was analysed (separately) using an unbalanced Analysis of
Variance (ANOVA), and using site, sample trip, and trap variable as the factors (along with
their interactions with trip), blocking for site. The primary response variable tested for was
Eastern Gambusia abundance, but species specific abundances were also tested. This enabled
an assessment of catch data in relation to:
•
•
•
•
whether the trap was set during the day or night
whether the trap had a light or not
whether the trap was baited or unbaited
whether any of the above was influenced by trip (interactions).
The analysis of methods did not consider the size of Eastern Gambusia collected. This was
mainly because only mature fish were present during the first removal trip, but also because
we felt that targeting for a specific size range of fish would be futile given the rapid growth
rates and subsequent maturation the species exhibits. As removal success depends largely on
the removal of fish prior to the onset of the spawning season, we predicted that maximising
the effectiveness of removal strategies during the first trip would be of the greatest
importance. Therefore, an assessment of Eastern Gambusia catch for each of the methods
during the first trip was also undertaken by analysing the standardised catch data using an
unbalanced ANOVA for the first trip, using method as the factor, blocking for site.
4.2.5.2
Population growth and impacts of Eastern Gambusia
The results from the monitoring data provided information on population dynamics for
several species and, importantly, how it relates to the abundance of Eastern Gambusia.
Unfortunately the analysis was restricted to the first season of the experiment, since any
information used to inform population growth was compromised by frequent flooding (and
thus frequent immigration and emigration). There were also numerous missing values within
the time series database (given the absence or omission of data caused b y drying or flooding
of sites) as well as small numbers of other species collected in the experiment. The analysis
was therefore restricted to assessments of Eastern Gambusia (Gh), Carp Gudgeon (Hsp),
Australian Smelt (Rs) and Common Carp (Cc).
72
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
The data consisted of a time series of observations of the four species over four time periods
from 13 sites between October 2009 and February 2010, with the sampling periods
approximately 1.3 months apart. Each species was sampled using three methods (SNE, bait
traps and seine nets), and the abundance of each species was expressed as the mean count
averaged over all sampling methods.
To monitor population change we used a non-linear state-space model (SSM) to fit a
mathematical model of population growth to the observed counts to estimate the trajectory of
each species through time. In addition to the population growth rate, the SSM also included
parameters that governed interactions between species. The SSM considers that the observed
data arise from two distinct processes: the system process, which models the underlying
biological process of population growth, and the observation process, which takes into
account the error associated with the sampling process (King et al. 2010).
The underlying model we used for the system process to model population growth and species
interactions was a discrete time Gompertz model (e.g. Dennis et al 2006). The system process
equation is given by
N t = N t −1 exp( a + b ln N t −1 + E t )
(equation 1)
where Nt is the population abundance at time t, a and b are constants, and E is the process
error or stochastic variation around the true biological process. E is assumed to be normally
distributed with mean 0 and variance σ2. On a logarithmic scale, the Gompertz model is a
linear autoregressive time series model of order 1. Thus
X t = X t −1 + a + bX t −1 + E t
(equation 2)
where Xt = log(Nt), a represents the intrinsic rate of increase and b governs the strength of
density dependence. When b = 0 the model corresponds to one with no density-dependence
(i.e. density independent growth).
However, the true (log) population abundances are not observed directly because of errors
induced from the observation or sampling process. Hence the observed counts Yt are given by
the observation process
Yt = X t + U t
where Yt are the observed (log) abundance data and Ut is the random variability that results
from the sampling process assumed to also be normally distributed with mean 0 and variance
τ2. Hence the full SSM model is specified as
X t = X t −1 + a + bX t −1 + E t
Yt = X t + U t
Et ~ N (0, σ 2 )
U t ~ N (0, τ 2 )
(equation 3)
We can also generalise equation 3 to multiple species to estimate population growth of each
species as well as interactions between species. If there are p interacting species, a
multivariate version of equation 3 is given by
X t = X t −1 + A + BX t −1 + E t
(equation 4)
Where Xt is a vector of size p of (log) abundances of each species, A is a vector of growth
rates, B is a p × p matrix whose elements bij give the effect of the abundance of species j on
the per-capita growth rate of species i, and E is a vector of process errors (Ives et al. 2003).
73
Native fish recovery following Eastern Gambusia removal
A useful addition to equation 4 occurs when there are other covariates that could be used to
model population growth. This is easily accommodated in the model by adding a term for
covariate effects:
X t = X t −1 + A + BX t −1 + CZ t + E t
(equation 5)
Where Zt is a vector of covariate values and C is a p × p matrix of the effects of covariate j on
species i. For the present analysis, we had a single covariate describing whether Gambusia
removal took place at a site or not.
Estimation of equation 4 is usually undertaken using maximum likelihood via either the EM
(expectation maximisation) or Kalman filter algorithm. Here we take a Bayesian approach
and fit equation 4 using MCMC techniques constructed in OpenBUGS software.
Initial attempts at fitting equation 5 ran into problems because of the short nature of each time
series (four observations) and the large number of missing observations (30% of observations
were missing in the Eastern Gambusia time series). As a result there was poor mixing of the
chains, and the model failed to converge. To remedy this we fitted a simpler version of
equation 5 that excluded all density-dependent terms, so that it assumed that each species was
subject to exponential growth modified by potential interactions with other species. Hence,
the density-dependent regulation terms in the model were set to 0. We also assumed that the
only species interactions were the effects of Eastern Gambusia on native species, and that the
treatment effect (Eastern Gambusia removal) applied only to Eastern Gambusia population
growth.
Parameter estimates for the fit of this simplified version of equation 5 were sampled from 4
chains following a burn-in of 10 000 samples, thereafter a further 40 000 samples were taken
with a thinning rate of 10, leaving a sample of 4000 to give the posterior distribution of each
parameter. Convergence was assessed from the Gelman/Rubin diagnostic, with a values less
than 1.05 used as the criteria for convergence of the chains.
4.2.5.3
Impacts of Eastern Gambusia on morphometric condition of juvenile native
species
To test the hypotheses based on Eastern Gambusia abundance influencing the condition of
native species, the juveniles of common native species collected in December, January and
February were assessed for both fin condition and body condition.
As in the cross-sectional component, body condition was measured using the relative
condition factor (Froese 2006). Length (TL) and weight (W) data of juvenile fish were used as
an assessment of general body form of juvenile Australian Smelt, Carp Gudgeon and Flatheaded Gudgeon, using the equation
W = aLb
where W is the weight in grams, L is the TL in mm, and a and b are constants. Length–weight
relationships could then be used to compare body condition using the relative condition
factor, where
Krel = W/aLb
General linear models were used to assess the influence of Eastern Gambusia abundance on
the condition of juvenile Carp Gudgeon, Flat-headed Gudgeon and Australian Smelt. Site
specific Eastern Gambusia abundance (derived from monitoring data) were log10 transformed
for normality and used as the fixed factor, with relative condition of juvenile fish used as the
dependent variable for each of the species.
74
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Each of the individual juvenile fish subject to the body condition assessment was also
assessed for fin damage. Using the same protocol as that in cross-sectional analysis, a
categorical assessment was made based on the degree of damage to an individual’s caudal fin.
The scoring system was as follows:
1 = no or minimal damage —no sign of damage to fin, or only minor splitting of fin rays.
2 = moderate damage — part of the fin was missing, but more than 50% remained.
3 = major damage — less than 50% of the caudal fin remained.
Fin scores were subsequently converted to either 0 (scores of 1), or 1 (scores of 2 or 3), i.e.
‘no damage’ or ‘damaged’. Logistic regressions (modelling of binomial proportions) were
used to assess the relationship between Eastern Gambusia abundance and the probability of
juvenile fin damage for each of the three native species, using the log10 Eastern Gambusia
abundance as the fixed factor and fin condition score as the dependant variable. A separate
species-specific analysis on relative condition was conducted for each of the years because of
the possibility of morphological differences between populations specific to the two regions
of the study (i.e. year one was conducted in the lower Ovens, but year two predominantly the
lower Goulburn – Campaspe).
4.2.5.4
Mark-recapture component
An estimate of the Eastern Gambusia population size was generated using the Chapman
estimator (Seber 1982) closed population mark – recapture method for each of the sites where
mark–recapture was used. This method was applied because of its widespread use and
theoretical basis, as it assumes that:
1. the population is geographically closed
2. the population is demographically closed with no births or deaths
3. no marks are lost or missed
4. marking does not change fish behaviour or vulnerability to capture
5. marked fish mix at random with unmarked fish
6. all animals have equal probability of capture (Seber 1982; Hayes et al. 2007).
The Chapman estimator for population size is given by
^
N =
( n1 + 1)(n2 + 1)
(m2 + 1)
where n1 = number caught and marked during the first sampling period; n2 = number caught
in second sampling period; and m2 = number of marked animals in second sampling period.
After the second season of marking we suspected that the protocol might have caused an
increased mortality of the species, which is not desirable for subsequent population measures
derived from recapture data. A laboratory study to assess the marking protocol’s impact on
the short-term mortality of Eastern Gambusia confirmed these concerns (see Appendix 1).
Therefore, to improve estimates of population size and capture rates, the initial number of fish
marked included only female fish (n1 in the Chapman estimator equation). The revised n1 was
also reduced by 25% and the equation was re-run, giving an estimated population range rather
than a single value. This was based on the mortality figures generated in the laboratory trial
(see Appendix 1). The population estimates were then used to generate removal efficiencies.
75
Native fish recovery following Eastern Gambusia removal
4.3 Results
4.3.1
Site parameters
Season one
Removal and subsequent monitoring commenced during the week of 26 October 2009.
Monitoring of river levels and temperatures during this period confirmed the experiment
commenced after the last connection to adjoining aquatic habitats as a result of flooding
(Figure 4.3.1), and before Eastern Gambusia had initiated spawning. In general, sites were of
similar size and had similar water quality and habitat complexity, all being within the largely
natural river redgum floodplain of the Ovens River (Table 4.1).
During the final trips in summer, several sites either dried completely, or had drawn down to
extremely low levels, most likely as a result of lowered groundwater levels and/or disturbed
substrates during the drought (Figure 4.3.2). In these circumstances, monitoring data that was
collected during this period was excluded from the analysis.
132
Gauge height (m)
130
Site connection height
128
126
Mar
Feb
Jan
Dec
Nov
Oct
Sep
Aug
Jul
124
Figure 4.3.1 Ovens river height at Peechelba bridge between July 2009 and March 2010.
The red dashed line represents the approximate river height at which experimental sites
connect to adjoining river or anabranch habitats. Grey blocks represent the removal and
monitoring periods of the first year of the field-depletion experiment.
Season two
In contrast to the previous year, the 2010–11 season experienced major flooding before the
trial and also several weeks after initial removals. Even sites that did not frequently connect
directly to major water courses (and thus perceived to be ‘low risk’ of flooding) connected to
adjoining treatments as a result of both localised flash flooding and widespread regional
flooding. As a result our experimental treatments were compromised, with major fish
immigration and emigration between sites post-treatment. Nevertheless, sites were still
subject to the monthly monitoring protocol, providing information on recolonisation of
Eastern Gambusia after control actions along with information on juvenile condition in
relation to Eastern Gambusia abundance. Habitat parameters collected for each site at the
commencement of the experiment are presented in Table 4.1. Being largely located on private
land, the majority of sites (other than the Ovens river sites from the previous season) had
some degree of disturbance including access by domestic stock and lower structural habitat
complexity.
76
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
(a)
(b)
Figure 4.3.2 Example of site variability during the course of the project. (a) Site 2 in late
October, and three months later at the end of January during the 2009–10 season. (b)
Hillgrove’s site at the commencement of the experiment, and following localised flooding in
the area that resulted in connection with the adjoining creek and control site.
77
Native fish recovery following Eastern Gambusia removal
Table 4.1. Treatment type, habitat and water quality parameters for each site, recorded at the commencement of trials in 2009–10 and 2010–11.
Categories H = high; M = medium and; L = low. Note that all sites in 2010–11 were subject to frequent or ongoing connection to adjoining habitats after
the trial had commenced.
Site
2009–10
1
2
19
20
21
22
8
9
ST
2010–11
Willow 1
Willow 2
Hillgrove’s 1
Hillgrove’s 2
Browns 1
Browns 2
Cornella Ck
Colbo wetland
1
2
19
ST
78
Size (m2)
Aquatic
Veg. (H,M,L)
Snag/Debris
(H,M,L)
pH
Repeated
Single
Control
Repeated
Single
Control
Repeated
Control
Ref
225
600
900
180
180
120
180
176
625
H
L
H
H
H
H
L
L
L
M
M
M
M
M
M
M
M
M
6.48
6.01
5.94
6.26
6.24
6.17
6.89
6.76
6.22
49.8
45.8
38.8
15.9
18.2
62.2
47.5
464
83
2.9
8.5
10.9
20.6
33
9.4
13.1
25.3
27
3.55
2.3
3.73
2.4
2.24
1.8
2.32
2.63
1.24
29
26.6
32.4
23.3
23.3
22.5
26.2
24.8
21
1.8
3
2.5
2
1.3
2.3
1.5
1.8
2
Repeated
Control
Single
Control
Repeated
Control
Single
Repeated
Ref
Ref
Ref
Ref
300
320
400
250
150
400
1000
1225
244
600
900
600
M
M
M
M
L
L
L
L
M
M
M
M
L
L
L
L
M
M
M
M
M
M
M
M
7.2
7.2
6.82
6.39
6.5
6.5
7.08
7.37
6.54
7.78
7.34
6.9
131
142
126
126
164
164
450
169
5.12
43.2
50
53.5
36
104
48
19
8
8
24
141
20.1
40.9
10.1
11.2
8.55
8.5
6.77
5.08
6.74
6.74
6.48
6.8
8.1
7.57
11.3
10.48
20.5
21.6
21
19
15.2
15.2
18
21.5
18.6
15.2
21.2
19.5
1.5
1.5
1.8
1.3
2
2
3
1.2
1.8
3
2.5
2
Treatment
Arthur Rylah Institute for Environmental Research
Conductivity
(µS/cm)
Turbidity
(NTU)
DO (mg / L)
Surface
temp. (°C)
Max. depth
(m)
Native fish recovery following Eastern Gambusia removal
4.3.2
Methodology assessment
In the first season, 11 397 fish were collected from the eight sites (excluding reference sites)
(Table 4.2). Eastern Gambusia was the most abundant species collected (n = 5165), followed
by Common Carp (n = 3039) and Carp Gudgeon (n = 2872).
The analysis of the standardised catch data indicated that the number of Eastern Gambusia
collected was significantly influenced by both trip number (p < 0.001), and the method of
collection (p < 0.001). Specifically, the targeted seine netting was by far the most effective
method for collecting Eastern Gambusia when comparing standardised efforts across all trips
(Figure 4.3.3a), and more importantly, for the first trip (Figure 4.3.3b), which only collected
mature fish prior to spawning. Method type also had a significant influence on collection of
each of the native species (all p < 0.001). The seine netting was also significantly more
efficient at collecting the two most abundant native species, Carp Gudgeon and Australian
Smelt, than the trapping methods (Figure 4.3.4). CPUE of southern pygmy-perch were highest
for fyke nets.
The different variables employed in the trapping methods (bait traps and fyke nets) were also
assessed for differences in CPUE of Eastern Gambusia and native species. The analysis of the
standardised catch data indicated that the number of Eastern Gambusia collected by both bait
traps and fyke nets was significantly influenced by trip number (p < 0.001 and p < 0.01
respectively), but not whether the trap was set during the day or night (both p > 0.05). Eastern
Gambusia CPUE for bait traps was higher during the day than night (Figure 4.3.5a), but this
difference was not statistically significant (p = 0.07). For traps set at night, there was no
significant difference in Eastern Gambusia CPUE between traps that did and did not contain a
light (both p > 0.05). There was also no significant difference between baited and unbaited
traps during trip one (p > 0.05). The pattern in trapping variables and CPUE for native species
was similar to those for Eastern Gambusia, in that the CPUE for Carp Gudgeon, Australian
Smelt and Southern Pygmy-perch was not significantly influenced by whether traps were set
during the day or night, contained a light or not, or were baited or unbaited (all p > 0.05).
Table 4.2. Raw abundances of each species collected from removal and control sites, for
each trip using all removal methodologies during the field depletion experiment in season
one.
Species
Trip
1
2
52
225
1718
3170
5165
196
228
1073
1375
2872
1
3
1
0
5
16
51
122
9
198
3
20
4
5
32
Eastern Gambusia
Carp Gudgeon
Flat-headed Gudgeon
Australian Smelt
Southern pygmy-perch
Common Carp
3
4
Total
506
1444
786
303
3039
Goldfish
4
0
34
4
42
Oriental weatherloach
0
5
2
13
20
Redfin
1
22
1
0
24
Total
779
1998
3741
4879
11397
79
Native fish recovery following Eastern Gambusia removal
30
(a)
20
10
CPUE
0
1.2
(b)
0.8
0.4
0
BT
Fyke
Seine
Per hour
SNE
Per shot
Figure 4.3.3 Eastern Gambusia catch per unit effort (CPUE) predictions for the four
methods used in the study for (a) all trips, and (b) the first trip only. Analysis based on
catch per hour for a single bait trap (BT) and fyke net, and fish per pass for targeted seine
net and sweep net electrofishing (SNE). Number of observations = 1698.
80
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
18
(a) Carp gudgeon
12
6
0
2.5
(b) Australian smelt
CPUE
2
1.5
1
0.5
0
0.025
(c)Southern pygmy perch
0.02
0.015
0.01
0.005
0
BT
Fyke
Per hour
Seine
SNE
Per shot
Figure 4.3.4 Catch per unit effort (CPUE) predictions for the four methods utilised in the
study across all trips for (a) Carp Gudgeon, (b) Australian Smelt, and (c) Southern Pygmyperch. Analysis based on catch per hour for a single bait trap (BT) and fyke net, and fish per
shot for seine net and sweep net electrofishing (SNE). Number of observations = 1698.
81
Native fish recovery following Eastern Gambusia removal
0.3
(a) Bait traps
0.2
0.1
0
CPUE
Day
4
Night
Light
No light
Bait
No bait
(b) Fyke netting
3
2
1
0
Day
Night
Light
No light
Figure 4.3.5 Catch per unit effort (CPUE + SE; fish per hour) predictions for Eastern
Gambusia for each of the trapping variables applied for (a) bait traps, and (b) fyke netting
for all trips (number of observations = 33). Variables included in the analysis are whether
traps were set during the day or night, contained a light or not (for night sets only; fyke
nets during first trip only), and were baited or unbaited (bait traps during first trip only). p
> 0.05 for all comparisons.
82
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Bait traps
Fyke nets
(a) Carp gudgeon
0.15
2.5
2
0.1
1.5
1
0.05
0.5
0
0
CPUE
(b) Pygmy perch
0.005
0.04
0.004
0.03
0.003
0.02
0.002
0.01
0.001
0
0
(c) Australian smelt
0.006
0.15
0.004
0.1
0.002
0.05
0
0
Day
Night
Light
No light
Bait
Unbaited
Day
Night
Figure 4.3.6 Bait trap (left column) and fyke net (right column) catch per unit effort
(CPUE+ SE; fish per hour) predictions for (a) Carp Gudgeon, (b) Southern Pygmy-perch,
and (c) Australian Smelt for each of the trapping variables for all trips. Variables included in
the analysis are whether traps were set during the day or night, contained a light or not,
and were baited or unbaited (bait traps only). p > 0.05 for all comparisons.
83
Native fish recovery following Eastern Gambusia removal
4.3.3
Eastern Gambusia population parameters and effectiveness of removal
Season 1
In general, all sites during the first season contained very low numbers of Eastern Gambusia
at the onset of the trial (late October – early November). During the first trip the raw numbers
of Eastern Gambusia collected at each site ranged from 0–21 individuals; at six of the eight
sites, fewer than six were captured (Table 4.3). Mark–recapture data collected from the two
sites that were subject to the protocol during the first trip also suggested small Eastern
Gambusia population sizes at the start of the trial (Table 4.4), estimating a maximum
population size of 25 and 41 individuals respectively (sites 1 and 19; Table 4.4).
Despite the low numbers of Eastern Gambusia present at the onset of the trial (and subsequent
low numbers of fish removed), the initial removals substantially reduced population sizes.
Although limited, mark–recapture data suggest that, despite minimal effort, the physical
removals over five to seven days collected a large proportion of the population (e.g. the initial
number of fish removed from site 1 represented between 84 and 100% removal; Table 4.4).
The most encouraging result was that Eastern Gambusia were not recorded in follow-up
monitoring at two of the four sites where they were removed (Table 4.3). The success of
removal exercises did rely on consecutive visits to sites within this early trip, both to detect
and ultimately capture fish (Figure 4.3.7). Each of these site visits were, however, relatively
brief (generally less than one hour). Unfortunately it was evident that, despite a similar effort,
repeated removals during the following trips (which were also after the onset of spawning)
did not result in complete eradication of Eastern Gambusia from the remaining removal sites.
Nevertheless, the rate of increase of Eastern Gambusia population growth was far less within
removal sites when compared to control sites. The value for the effect of the removal
treatment was negative, indicating that the removal treatment resulted in a lower growth rate
of Eastern Gambusia at removal sites, although the 95% CL just overlapped 0, indicating a
small amount of uncertainty around this estimate (Table 5; Figure 4.3.8). Hence, the growth
rate of Eastern Gambusia at control sites was estimated to be 1.19 while the growth rate at
removal treatment sites was estimated to be 0.39 (i.e. 1.189 – 0.796).
Populations of Eastern Gambusia at the onset of the removal experiment comprised entirely
adult fish, particularly females (Figures 4.3.9, 4.3.10). Furthermore, female fish collected
during the initial removal period were all in gravid condition. Consequently, the sites still
containing Eastern Gambusia during the second trip all had an influx of the first cohorts for
the season (Figure 4.3.10). From this point onwards, Eastern Gambusia population sizes
rapidly increased (Figure 4.3.8) despite the extremely low abundances of Eastern Gambusia at
the onset of the experiment or a removal efficiency greater than 80%. For example, mark–
recapture at control site 19 gave population estimates of Eastern Gambusia ranging from 38 to
41 in November, but this increased to 19 030 to 25 365 at the end of February. Similarly, at
control site 2, where only two Eastern Gambusia were collected at the start of the experiment,
had an estimated population size (derived from the mark–recapture data) of between 6520 and
8646 four months later. Further detailed assessment of population trajectories are presented
below in the monitoring results.
Season 2
Eastern Gambusia were much more abundant at the sites at the onset of the experiment during
the second season (Table 4.3). Unlike the first season, raw numbers of fish collected in
removals and mark–recaptures suggested (with the exception of two sites) populations of
hundreds of fish.
84
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
As in season 1, mark–recaptures at the onset of the trial still suggested that a large proportion
of the population were being collected (Table 4.4) and that the initial removal exercises
substantially reduced population sizes. This was confirmed when Eastern Gambusia were not
recorded during several site visits (which included observations and seine netting) over three
consecutive weeks following the initial removal. Again, the success of removals relied on
consecutive visits to sites during this early trip, although each visit was brief. For example,
eight days of removal exercises undertaken at Colbinabbin wetland resulted in a gradual
reduction amounting to 75% of the population, based on mark–recapture information (Figure
4.3.11). Unfortunately, all sites were completely inundated shortly afterwards, and large
numbers of adult and juvenile fish were consequently reintroduced to the sites. Because of the
constant flooding, which conrinued until late February, no further removals or monitoring of
Eastern Gambusia population change could be conducted after the initial removals.
Nevertheless, these results provided important information about Eastern Gambusia removal
and reinvasion.
5
2009/10
4
No. Fish removed
3
2
1
n/a
n/a
n/a
7
8
9
n/a
0
160
2010/11
120
80
40
0
1
2
3
4
5
6
10
Day
Figure 4.3.7 Mean (± SE) number of Eastern Gambusia extracted from removal sites each
consecutive day of trip one during the 2009–10 and 2010–11 seasons. n/a indicates no
removals were undertaken.
85
Native fish recovery following Eastern Gambusia removal
Table 4.3 Raw numbers of Eastern Gambusia collected from each of the removal and
control sites using physical removal techniques during the first trip (spring) of each of the
two years. The detection of Eastern Gambusia during follow-up assessments is also
presented.
Year
2009–
10
2010–
11
Fish detected in
follow-up
assessment
Site
Treatment
Total fish collected
1
Repeated
21
Yes
19
Control
15
Yes
2
Single
2
No
8
Repeated
0
No
9
Control
2
Yes
20
Repeated
5
Yes
21
Single
3
No
22
Control
2
Yes
Colbo
Repeated
Willow 1
Willow 2
555
Yes*
Repeated
195
No*
Control
186
Yes*
Hillgroves 1
Single
269
No*
Hillgroves 2
Control
23
Yes*
Browns 1
Repeated
3
No*
Browns 2
Control
1
Yes*
Cornella Ck
Single
220
Yes*
Table 4.4 Eastern Gambusia mark-recapture parameters collected during the start of the
experiment (spring) during each of the two years (two sites in each year) and end of the
experiment in the first year only (end of summer). Population estimates are presented as a
range allowing for a maximum mortality of marked female fish of 30% (see Appendix 1).
Year
2009–
Period
Site
Treatment
No. fish
Total fish
No.
Population
Recapture / removal
marked
collected
recap’s
estimate
efficiency (%)
Spring
1
R. removal
11
21
9
21 – 25
82 – 100
Spring
19
Control
13
15
5
36 – 41
41 – 47
Summer
9
Control
0
2
0
Summer
19
Control
1000
1038
40
19030 – 25366
4–6
Summer
9
Control
60
3
6520 – 8646
7–9
Spring
Colbo
R. removal
25
555
14
731 – 963
58 – 76
Spring
Willow 1
R. removal
35
195
11
444 – 587
32 – 44
10
2010–
< 10
n.a.
11
86
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Table 4.5 Parameter estimates from the Bayesian SSM for Eastern Gambusia (Gh), Carp
Gudgeon (Hsp), Australian Smelt and carp (Cc). a – intrinsic rate of increase of; b – The
effect of Eastern Gambusia (Gh) abundance; c – effect of the removal treatment on the rate
of increase of Gh; τ – observation error; σ – process error.
Species
Gh
Parameter
Hsp
Rs
Cc
x
lwr
upr
x
lwr
upr
x
lwr
upr
1.189
0.633
1.755
0.250
-0.042
0.547
0.063
-0.080
0.206
a
b
x
lwr
-
-
0.008
0.262
-
-
upr
0.257
-0.016
-0.316
0.290
-0.075
-0.222
0.078
0.061
0.332
0.200
c
-0.796
-1.664
0.034
τ
0.510
0.029
1.038
0.153
0.005
0.438
0.127
0.008
0.259
0.346
0.067
0.549
σ
0.777
0.073
1.421
0.470
0.191
0.707
0.184
0.014
0.346
0.246
0.007
0.652
150
0
50
100
Abundance
150
100
0
50
Abundance
200
Treatment sites
200
Control sites
1
2
3
Time
4
5
1
2
3
4
5
Time
Figure 4.3.8 Predicted trajectories for Eastern Gambusia abundance at control (no removal)
and treatment (removal) sites from November to March (trips 1–5) predicted for 1000
simulated sites from the SSM model (includes predictions for one additional time period).
Dashed lines indicate the 95% credible interval. Predictions exclude observation error.
87
Native fish recovery following Eastern Gambusia removal
(a)
(b)
Figure 4.3.9 (a) Female Eastern Gambusia collected during the first trip of the experiment
(before the onset of spawning). Note that all fish are in gravid condition. (b) Eastern
Gambusia collected in a single targeted seine shot at the end of the experiment.
88
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
2009/10
30
2010/11
20
n = 22
Trip 1
n = 206
10
0
30
20
Trip 2
Frequency (%)
n = 197
n = 263
10
0
30
20
n = 44
Trip 3
n = 545
10
0
30
20
Trip 4
n = 290
n = 339
10
0
6
10
14
18
22
26
30
34
38
42
6
10
14
TL (mm)
18
22
26
30
34
38
TL (mm)
200
100
160
80
120
60
80
40
40
20
0
Cum. % removed
No. fish removed
Figure 4.3.10 Length frequency histograms (% frequency) of Eastern Gambusia collected
on each trip (October–November, December, January and February–March) for both the
2009–10 and 2010–11 seasons of the trial.
0
1
2
3
4
5
6
7
8
Day
Figure 4.3.11 Total raw number of fish extracted (bars) and the cumulative percentage of
population removed (line) based on mark–recapture data from Colbinabbin wetland over
eight days of removal exercises before the onset of spawning (trip 1, season 2). Note day
one involved the setting of traps only, and therefore no fish were extracted on this day.
89
42
Native fish recovery following Eastern Gambusia removal
4.3.4
4.3.4.1
Fish community monitoring and condition indices
Fish community monitoring
A total of 44 482 fish were collected during the field removal experiment (14 621 and 29 600
in 2009–10 and 2010–11 respectively), comprising seven native and five alien species (Table
4.6, Figure 4.3.12). There was considerable variation in species composition and abundances
across sites, although Carp Gudgeon were present at all sites and both Eastern Gambusia and
Goldfish were present at 19 of the 20 sites (Table 4.7). In the 2009–10 season, Eastern
Gambusia were the dominant species, followed by Carp Gudgeon, Common Carp and
Australian Smelt respectively. Very few individuals of other species were collected, so they
could not be used in the population modelling exercise. In the 2010–11 season, Eastern
Gambusia and Carp Gudgeon were again the dominant species, followed by Flat-headed
Gudgeon, Murray–Darling Rainbowfish and Australian Smelt. Three Flat-headed Galaxias
were collected following the frequent flooding in the second season of the experiment — the
first time this species had been recorded in the study (including the cross-sectional study).
An assessment of length frequency data gave some indication of the degree of recruitment
into the populations. Of the native species, only Carp Gudgeon appeared to spawn in late
summer, with the arrival of early cohorts (i.e. fish under 10 mm TL: Figure 4.3.13).
Australian Smelt, Flat-headed Gudgeon, Murray–Darling Rainbowfish and Southern Pygmyperch, apparently spawned only in the early months of the trial, or just prior to the start of the
trial (although data is limited for the last two species). This pattern was also evident for the
alien species: Common Carp and Goldfish spawned in the months prior to the trial, or in the
first month of the trial (Figure 4.3.14). Recruitment patterns of Eastern Gambusia were
discussed in section 4.3.3.
Given the high variability in species composition and abundance between sites, an assessment
of population trajectories of the most common species collected in the first year of the
experiment was undertaken using the SSM approach. The intrinsic rate of increase was
positive for Eastern Gambusia, Carp Gudgeon and Australian Smelt, and slightly negative for
Common Carp (Table 4.5). Eastern Gambusia had by far the greatest rate of population
increase (using control site data), being almost five times greater than the highest native
species, Carp Gudgeon (Figure 4.3.16). The populations of Australian Smelt and Common
Carp were relatively stable because the project commenced after the major spawning period
(as indicated by the length frequency data). Observed population trajectories for each site are
shown in Figure 4.3.14, and predicted trajectories from the SSM are shown in Figure 4.3.15.
Predicted simulated mean trajectories (+95% CL) over five time periods are shown in Figures
4.3.8, 4.3.16 and 4.3.17.
The results suggest that Eastern Gambusia had a negative influence on the population growth
of native species and Common Carp, so that removing Eastern Gambusia would result in
slightly increased abundances of other fish. The values for the interaction coefficient b were
all negative, indicating that the presence of Eastern Gambusia had a negative impact on the
per capita growth rate of all three species, with the largest estimated impact (–0.75) occurring
for Australian Smelt (Table 4.5, Figure 4.3.17). The predicted effects of different abundances
of Eastern Gambusia on the population growth of Carp Gudgeon, Australian Smelt and
Common Carp indicates an increasing negative impact with increasing abundance of Eastern
Gambusia (Figure 4.3.17). It must be noted, however, that the predicted negative effects of
Eastern Gambusia on population growth was extremely low, even when Eastern Gambusia
abundances were extremely high, and especially for Carp Gudgeon. Again, the 95% CL for
these estimates all overlapped 0, indicating a degree of uncertainty around these effects.
90
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Table 4.6 Total numbers of each species collected from all sites during each year of the
removal trial.
Month
Species
Nov
Dec
197
233
1
Jan
Feb
Total
1078
1851
3359
3
1
40
45
16
51
122
151
340
Flat-headed galaxias
0
0
0
0
0
Murray–Darling rainbowfish
0
0
0
4
4
Southern pygmy-perch
3
20
4
5
32
Unspecked hardyhead
0
0
0
0
0
505
1444
787
341
3077
Oriental weatherloach
0
5
2
13
20
Redfin
1
22
1
1
25
43
243
1739
5694
7719
4
0
35
40
79
766
2021
3734
8100
14700
1711
569
1042
526
3848
161
52
74
74
361
65
29
22
35
151
1
1
1
0
3
85
63
30
27
205
Southern pygmy-perch
5
1
1
0
7
Unspecked hardyhead
0
0
5
0
5
Carp
3
49
75
8
135
11
0
23
0
34
4
7
3
1
15
2202
4287
5344
13003
24836
24
19
111
28
182
4248
5058
6620
13674
29782
2009–10
Carp Gudgeon
Flat-headed Gudgeon
Australian Smelt
Carp
Eastern Gambusia
Goldfish
Total
2010–11
Carp Gudgeon
Flat-headed Gudgeon
Australian Smelt
Flat-headed galaxias
Murray–Darling rainbowfish
Oriental weatherloach
Redfin
Eastern Gambusia
Goldfish
Total
91
Native fish recovery following Eastern Gambusia removal
Figure 4.3.12 Examples of species collected at sites during the field removal experiment.
Left, top to bottom: native species Murray–Darling Rainbowfish, Carp Gudgeon, Southern
Pygmy-perch and Flat-headed Galaxias; and alien species. (Right, top to bottom: Common
Carp, Eastern Gambusia, Goldfish and Oriental Weatherloach).
92
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Table 4.7 Total numbers of each species collected from each site for all trips and during both seasons of the experiment. Note that the sampling effort was not equal at each site
because of the effects of drought and flood.
Site
9
19
20
21
22
Colbo
Corn
. Ck
Hill.
1
Hill.
2
PB
8
Br’ns
Rd 2
PB
2
Br’ns
Rd 1
PB
1
17b
1b
2b
ST
Will.
Rd 1
Will.
Rd 2
132
763
45
404
2273
5
3
3
5
9
622
152
444
7
212
162
87
294
326
1259
Flat-headed Gudgeon
1
8
1
0
5
0
0
0
13
8
271
76
0
0
12
9
0
2
0
0
Australian Smelt
0
133
12
41
4
9
1
3
8
0
2
128
0
0
16
82
44
7
1
0
Flat-headed Galaxias
0
0
0
0
0
0
0
0
0
0
2
1
0
0
0
0
0
0
0
0
Murr.–Darl. Rainbowfish
0
0
0
0
0
0
0
0
0
0
209
0
0
0
0
0
0
0
0
0
Southern Pygmy-perch
0
0
0
0
34
0
0
0
0
0
0
4
0
0
0
0
0
0
0
0
Unspecked Hardyhead
0
0
0
0
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2306
291
57
84
32
126
57
128
52
8
0
0
1
0
17
0
21
32
0
0
10
12
0
0
21
0
0
0
0
0
0
0
0
0
0
0
0
11
0
0
3
1
0
0
0
5
3
12
0
1
2
12
0
0
0
1
0
0
0
0
1709
847
0
938
3237
372
3
1146
3167
522
2004
557
5320
1479
270
172
307
1810
7241
1454
24
26
9
10
41
5
0
2
13
3
3
8
10
1
1
2
33
19
20
31
4185
2081
124
1477
5652
522
67
1294
3258
551
3115
938
5775
1487
528
428
492
2175
7588
2744
Species
Carp Gudgeon
Common Carp
Oriental Weatherloach
Redfin
Eastern Gambusia
Goldfish
Total
93
Native fish recovery following Eastern Gambusia removal
November
January
December
February
12
n = 272
Carp
gudgeon
n = 450
n = 185
8
n = 319
4
0
35
30
25
20
15
10
5
0
Australian
smelt
n = 37
n = 39
n = 24
n = 71
Flat-headed
gudgeon
% frequency
80
60
n = 23
n = 72
n = 30
n = 84
40
20
0
100
80
Murray River
rainbowfish
n=7
n = 30
n=7
n=4
n = 30
n = 63
60
40
20
0
80
Southern
Pygmy perch
60
n=4
n=4
40
20
0
<5
20
30
40
>50
<5
20
30
40
>50
<5
20
30
40
>50
<5
20
30
40
>50
TL (mm)
Figure 4.3.13 Length frequency histograms (% frequency) for the five most common native species collected for each trip (October–November,
December, January, February–March) of the trial (both seasons combined).
94
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
November
January
December
February
80
60
Carp
% frequency
Goldfish
n = 183
n = 84
n = 64
n = 24
n = 97
n = 26
40
20
0
80
n=3
60
n = 10
40
20
0
<5
20
30
40
>50
<5
20
30
40
>50
<5
20
30
40
>50
<5
20
30
40
>50
TL (mm)
Figure 4.3.14 Length frequency histograms (% frequency) of Common Carp and Goldfish collected for each trip (October–November, December, January,
February–March) of the trial (both seasons combined).
95
Native fish recovery following Eastern Gambusia removal
Hsp
50
40
30
10
0
0
2
3
4
1
2
3
Time
Rs
Cc
4
15
10
5
0
0
5
10
Abundance
15
20
Time
20
1
Abundance
20
Abundance
150
100
50
Abundance
200
Gh
1
2
3
4
1
Time
2
3
4
Time
Figure 4.3.14 Observed abundance of Eastern Gambusia (Gh), Carp Gudgeon (Hsp),
Australian Smelt (Rs) and Common Carp (Cc) from November to February (trips 1–4). Lines
connecting points indicate data from the same site. Red lines indicate Eastern Gambusia
removal sites.
96
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Hsp
50
40
30
10
0
0
2
3
4
1
2
3
Time
Time
Rs
Cc
4
10
0
0
10
5
20
30
Abundance
15
40
50
20
1
Abundance
20
Abundance
150
100
50
Abundance
200
Gh
1
2
3
Time
4
1
2
3
4
Time
Figure 4.3.15 Trajectories of Eastern Gambusia (Gh), Carp Gudgeon (Hsp), Australian
Smelt (Rs) and carp (Cc) abundance from November to February (trips 1–4) predicted by
the SSM (dashed lines and small solid circles) overlaid on the observed trajectories (solid
lines and open circles).
97
Native fish recovery following Eastern Gambusia removal
Hsp
20
5
0
0
3
4
5
1
2
3
Time
Time
Rs
Cc
4
5
4
5
4
3
2
1
0
0
1
2
3
Abundance
4
5
2
5
1
Abundance
10
Abundance
15
150
100
50
Abundance
200
Gh
1
2
3
4
5
1
2
Time
3
Time
Figure 4.3.16 Mean trajectories from 1000 simulated sites for Eastern Gambusia (Gh), Carp
Gudgeon (Hsp), Australian Smelt (Rs) and Common Carp (Cc) abundance from November to
March (trips 1–5) predicted by the SSM including (includes predictions for one additional
time period). Predictions include process error but exclude observation error. Dashed lines
indicate the upper and lower 95% credible interval.
98
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
1
2
3
4
20
15
10
Abundance
10
0
5
5
Gh = 148
5
20
15
Gh = 7
0
5
10
Abundance
15
Gh = 0
0
Abundance
20
Hsp
1
Time
2
3
4
5
1
Time
2
3
4
5
4
5
4
5
Time
5
5
5
Rs
1
2
3
4
5
4
3
0
1
2
Abundance
3
0
1
2
Abundance
3
2
0
1
Abundance
Gh = 148
4
Gh = 7
4
Gh = 0
1
Time
2
3
4
5
1
Time
2
3
Time
5
5
5
Cc
1
2
3
Time
4
5
4
3
0
1
2
Abundance
3
0
1
2
Abundance
3
2
0
1
Abundance
Gh = 148
4
Gh = 7
4
Gh = 0
1
2
3
Time
4
5
1
2
3
Time
Figure 4.3.17 Predicted effects of different abundances of Eastern Gambusia on the
population growth of Carp Gudgeon (Hsp), Australian Smelt (Rs) and Common Carp (Cc)
based on 1000 simulated sites. Dashed lines are the upper and lower 95% credible
intervals. Predictions exclude observation error.
99
Native fish recovery following Eastern Gambusia removal
Condition indices of juvenile native species
A total of 860 juvenile native species were collected for the body and fin condition
assessment — predominantly Carp Gudgeon, Australian Smelt and Flat-headed Gudgeon (n =
580, 170 and 83 respectively).
There was a general trend that increasing Eastern Gambusia abundance resulted in an
increased likelihood of fin damage and decreased body condition of juvenile fish, although
the degree of this influence varied between species. The assessment of all available data (trips
2–4) indicated that, for all three species, the relative condition of juvenile fish decreased with
increasing Eastern Gambusia abundance (Figure 4.3.18). While there is some uncertainty in
these relationships for Australian Smelt and Flat-headed Gudgeon (both p > 0.05), the
predictions for Carp Gudgeon are significant (p < 0.001). The probability of fin damage
increased with increasing Eastern Gambusia abundance for Australian Smelt (p < 0.05) and
Flat-headed Gudgeon but was negligible for Carp Gudgeon, although there was some
uncertainty in the last two relationships (all p > 0.05).
The numbers of juveniles of rarer species collected, such as Murray–Darling Rainbowfish,
Flat-headed Galaxias and Southern Pygmy-perch, were insufficient for a statistical analysis of
body and fin condition. All of these species were collected from sites containing high
abundances of Eastern Gambusia. Although adults of these rarer species did not have any fin
damage, 50% of juvenile Murray–Darling Rainbowfish and all juvenile Southern Pygmyperch collected in January and February had fin damage (Figure 4.3.19). On the other hand,
none of the juvenile Southern Pygmy-perch collected in December displayed fin damage. A
detailed examination of individual juvenile Southern Pygmy-perch collected in December,
January and February suggested that there was very little growth in this cohort after
December (Figure 4.3.20), but as very few individuals were collected this suggestion should
be treated with caution.
100
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Australian smelt
Carp gudgeon
Flat-headed gudgeon
n = 170
n = 580
n = 83
1.0
0.8
0.8
0.6
Value
FINSCORE
Prob. fin damage
1.0
0.6
0.4
0.4
0.2
0.2
P < 0.05
0.0
0.0
0.0
0.0
0.5
0.5 1.0
P = 0.667
0.0
2.0
2.0 1.5 2.5
1.5
1.0
0.5
1.5
2.5
2.0
LOGCPUE
LOGCPUE
LOGCPUE
Relative condition (Krel)
1.0
P = 0.129
0.9
0.0
0.5
1.0
1.5
2.0
2.5
LOGCPUE
3.0
6
0.8
2.5
5
0.7
2.0
4
0.6
3
1.5
2
1.0
0.5
0.4
P = 0.391
0.0
0.4
0.8
1.2
P < 0.001
1
1.6
0.50
1.00
1.50
2.00
P = 0.465
0.5
0.6
0.8
1.0
1.2
1.4
1.6
1.8
Log_Gambusia abundance
Figure 4.3.18 General linear model outputs (± 95% confidence intervals) examining (top) the probability of fin damage, and (bottom) relative condition
(Krel) of juvenile Australian Smelt, Carp Gudgeon and Flat-headed Gudgeon in relation to Eastern Gambusia abundance (log10 abundance) at all sites from
December to February.
101
Figure 4.3.19 Juvenile Southern Pygmy-perch collected in January 2010. Note the damage of
the caudal fin.
Pygmy perch TL (mm)
70
60
50
40
30
20
10
0
Oct/Nov
Dec
Jan
Feb
Month
Figure 4.3.19 Individual Southern Pygmy-perch lengths (mm) collected in each month (trips 1–
4) from Site 19 during the first year of the study. Fish < 30 mm represent the cohort of juvenile
fish.
Arthur Rylah Institute for Environmental Research
102
Native fish recovery following Eastern Gambusia removal
4.4
Discussion
4.4.1
Physical control of Eastern Gambusia
One of the most encouraging outcomes of the study was demonstrating that, under certain
situations, physical removal can be effective in reducing Eastern Gambusia abundance. During
both years of the study, physical removal conducted over a 5–10 day period (before the onset of
the Eastern Gambusia spawning season), resulted in major reductions in Eastern Gambusia
abundance (generally over 40%), and even complete eradication at several sites. Furthermore, the
modelling exercises indicated the removal substantially reduced the rate of population increase.
Physical removal techniques have rarely been considered as a management tool for Eastern
Gambusia, the focus instead being on unselective draining and chemical poisoning (Maynard et al.
2008). While the effectiveness of these two techniques for removing alien species cannot be
questioned, there will always be regions or sites where these methods are not feasible. For
example, where threatened species are present or the hydrology is or unfavourable, draining or
chemical treatment may have a greater impact on the site’s biota than the impact of Eastern
Gambusia themselves.
Three recent studies have documented the potential of physical removal to reduce gambusia
populations in the wild (Maynard et al. 2008; Kerezsy 2009; Brookhouse and Coughran 2010).
Maynard et al. (2008) and Brookhouse and Coughran (2010) demonstrated the potential utility of
trapping methods that exploit specific behavioural responses of Eastern Gambusia to maximise the
efficiency as a removal tool and, just as importantly, minimise their impact on non-target biota.
For example, Maynard et al. (2008) established that Eastern Gambusia exhibit a positive
phototactic response (attraction to light) and positive thermotactic response to heat (attracted to
warm water). The authors then exploited this behaviour by incorporating heat or light sources (or
both) into a variety of trap designs, to ultimately develop a trap and setting methodology that
maximises the capture of Eastern Gambusia while having a minimal effect on native fish.
Similarly, Brookhouse and Coughran (2010) established that setting bait traps in such a manner
that their inception area is at the surface, as well as positioned in unshaded areas during the day,
vastly increases catch rates of Eastern Gambusia and minimises catch rates of native species.
The success of the Eastern Gambusia reductions in our study was a result of a combination of
repeated site visits on each trip and the effectiveness of the seine netting method, which employed
lightweight fine mesh nets to specifically target Eastern Gambusia. As in the examples mentioned
above, seine netting exploits aspects of Eastern Gambusia behaviour, in particular their occurrence
at or near the surface throughout spring and summer and their attraction to warmth and sunlight.
For example, Maglio and Rosen (1969) found that the the major concentration of Gambusia affinis
roughly corresponded to the areas of a pond receiving the most direct sunlight. This makes the
areas they occupy within specific sites very predictable, particularly during times when water
temperatures are low so that they are relatively easy to locate using polarised sunglasses. This
behaviour also reduces any limitations that site depth had on the removal technique, as areas that
were originally perceived being too deep to seine were generally the non-preferred areas for
Eastern Gambusia (particularly during spring). For example, during spring the adult fish were
constantly trying to access the shallowest, brightest area of each site (presumably to initiate
reproduction). These areas would be targeted with seine netting several times a day for a 5–10 day
period. Ultimately this method involved considerably less effort (each site visit was less than one
hour) and created far less site disturbance than traditional random netting of an entire site.
Understanding and exploiting the target species behaviour is not new in pest management
activities, including those designed for fish. The management of carp in Australia is a prime
example, with the development and optimisation of exclusion screens for wetlands (e.g. Hillyard et
103
Native fish recovery following Eastern Gambusia removal
al. 2010), separation cages on fishways (e.g. Stuart et al. 2006), and general targeted removal and
exclusion (e.g. Stuart and Jones 2002), all being based on thoroughly understanding carp
behaviour and applying it to the region in question (e.g. Stuart and Jones 2006).
Recently Kerezsy (2009) also investigated physical control of Eastern Gambusia using manual
dip-netting in Australian artesian springs. He reported some success with the method, suggesting
that sustained physical removal using dip nets in small springs may be an effective means of
reduction or control. It must be noted that both our study and that of Kerezsy (2009) relied on
active targeted removal techniques rather than trapping. Maynard et al. (2008) observed a
reduction in CPUE over time, but our study and that of Brookhouse and Coughran (2010) did not
assess the overall efficiency or effectiveness of trapping techniques in reducing population sizes,
so the efficiency of trapping techniques alone is still in question. Although we included the
trapping methods used in these earlier studies (using key findings of light and net position) in our
the methodology assessment, we demonstrated that targeted seine netting was by far the most
effective physical method for collecting Eastern Gambusia across all trips, and more importantly,
for the first trip. While it is difficult to compare the efforts across different methods, the difference
in catch rates between the targeted seine netting and other methods was so great that an equivalent
catch of Eastern Gambusia using any of the other methods would require a huge increase in
replication to achieve the removal outcomes of a single targeted seine netting.
For example, during the trial the bait trap equivalent to the catch rate of a single targeted seine
netting was approximately 171 hours immersion time (e.g. 10 traps immersed for 17 hours). This
difference was even greater when comparing methods during the first trip (over 1000 times
difference between CPUEs). Maynard et al. (2008) also compared CPUEs of dip netting with a
variety of traps. They found that dip netting had the overall highest CPUE, although they stated
that it was hard to compare CPUEs of trapping (trap hour) and active netting. Furthermore,
damage to the aquatic environment, labour costs, and size and sex selectivity must also be
considered when comparing methods. Nevertheless, this suggests that the use of currently
available trapping techniques (e.g. bait traps) alone would require extremely high trap replication,
or would have be undertaken for a much longer period of time before the onset of spawning if it
were to achieve similar removal results to a control strategy which also employed active removal
techniques such as seine netting.
The active physical removal techniques employed in both this study and that of Kerezsy (2009) did
have limitations. While depth did not seem to influence removal efficiency because Eastern
Gambusia were not using deeper areas when they were targeted, the amount of structural habitat
such as dense macrophytes and large woody debris that was present at a site did make these
techniques difficult to use. The targeted techniques described in both studies overcome the
problems of habitat to some degree, although high densities of this habitat does reduce the
efficiency of the technique (particularly if eradication is the desired outcome). Although the
targeted netting is far less damaging to habitat and non-target fauna than chemical applications or
drying (or conventional seine netting), it did collect more native species and appeared to damage
aquatic vegetation more than trapping methods. Therefore trapping is still a useful method in an
Eastern Gambusia control program, particularly in sites that are extraordinarily fragile or contain
dense structural complexity.
The variables tested within the trapping component of the study were used primarily to streamline
removal during the trial, but the information disseminated from results is applicable to future
control programs that use trapping for removal. The results of the trapping assessment indicated
that the number of Eastern Gambusia collected from daytime sets was higher than numbers
collected during night sets, although this difference was not statistically significant (p = 0.07).
104
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Furthermore, for traps set at night there was no significant difference in numbers of Eastern
Gambusia collected in traps with and without a light, and with and without bait. This suggests that
setting collapsible bait traps or fyke nets throughout the day without any additional light or bait
attraction will collect as many Eastern Gambusia as setting these nets with lights or bait overnight.
Because it is the most simplistic setting technique (both in time and investment) it is also the
cheapest option for setting traps.
Maynard et al. (2008) and Brookhouse and Coughran (2010) also assessed different trapping
variables. Maynard et al. (2008) concluded that the use of collapsible bait traps set with the
inception area at or near the surface was the best available option for Eastern Gambusia removal
(which is why we used this method in the current study). While that study did not assess the catch
rates of nets during the daytime, the authors concluded that having nets fitted with a portable light
source improved catches. The reasons for the differences between that conclusion and our findings
are unclear, but may be linked to the different climates where the studies were conducted
(Tasmania vs the mid-MDB) or the different light sources used in the studies. Our study also
recorded slightly higher numbers of Eastern Gambusia during the day, and the early trial using the
floodlight (generating immense amounts of light and heat) attracted no Eastern Gambusia. We
suggest that this is because Eastern Gambusia are largely inactive during the night, given that they
feed during daylight hours and rely on sight to detect and attack prey (Swanson et al. 1996;
McKay et al. 2001). Brookhouse and Coughran (2010) did not assess light attraction, but did
conclude that nets set during daylight hours collected far greater numbers of fish if they were set in
sunlight (as opposed to shade) and near the surface.
The patterns in trapping variables of native species were similar to that of Eastern Gambusia, in
that numbers of Carp Gudgeon, Australian Smelt and southern pygmy-perch collected were not
influenced by whether traps were set during the day or night, contained a light or not and, were
baited or unbaited. This information should be considered when designing survey regimes,
potentially saving costs that would otherwise have been spent on setting traps throughout the night
or investing in bait or light attraction sources. Furthermore, it supports the use of unbaited,
daytime sets of collapsible bait traps used in the MDBA sustainable rivers audit (SRA) fish
sampling program currently used throughout the MDB.
Irrespective of methodology, the success of physical removal was strongly influenced by a number
of other factors in the current study, most notably timing, hydrology (site connection) and size of
the site. Although replication was limited, the results suggested that the success of removal
undertaken before spawning had commenced largely governed the overall effectiveness of the
control strategy, whether repeated removal was undertaken after spawning had commenced or not.
This is because of the ability of Eastern Gambusia to rapidly colonise a site, as demonstrated by
this study and others (e.g. Milton and Arthington 1983) coupled with the much more intensive
effort required to physically remove post-larval and early juvenile Eastern Gambusia (e.g. Kerezsy
2009). In most cases this means that the window for physical removal of Eastern Gambusia is
during winter and early spring, when daylight is less than 12 hours and water temperatures are
below 16 °C (Pyke 2005). However, in many regions in Australia, including the northern regions
of the MDB, this window for Eastern Gambusia removal is severely reduced, and in tropical areas
even non-existent. To overcome this, Kerezsy (2009) suggested that removals be repeated over
several sampling occasions to allow all size classes of fish to be targeted. This allows juvenile
Eastern Gambusia to grow and then be subsequently targeted, thus overcoming the difficulties of
removing early life-stages. This has implications for the design and, in particular, the methodology
employed during control program. For example, if a strategy such as that proposed by Kereszy
(2009) were employed, a window of just a few weeks exists between early juveniles and spawning
105
Native fish recovery following Eastern Gambusia removal
adults. As a result, active removal techniques should be employed because a trapping program on
its own is highly unlikely to achieve successful removal outcomes in this small window of time.
Hydrology is another major factor governing the success, and more specifically, the duration of the
success of Eastern Gambusia control exercises. The second season of the removal experiment
provided a good example of how hydrology will largely govern the overall success of a control
program. After what appeared to be very successful removal efforts at numerous sites, late season
flooding and subsequent connection to adjoining habitats enabled both adult and early juvenile
Eastern Gambusia to recolonise these sites approximately two weeks after the initial removal.
Additionally, many sites in temperate regions, including those on the Ovens floodplain, generally
reconnect at predictable periods (e.g. every winter–spring), which further restricts the ideal
window for control actions. For this reason the removal exercises were conducted early enough to
be before the onset of Eastern Gambusia removal, but late enough to minimise the risk of
reconnection. Alternatively, measures could be taken to prevent reinvasion. For example, Kerezsy
(2009) erected polypropelene barrier fencing across inflow points following Eastern Gambusia
removal, to prevent reinvasion. However, the effectiveness of such measures is still uncertain.
Current control strategies require an extensive knowledge of the distribution of Eastern Gambusia
throughout a region and an equally detailed knowledge of the region’s hydrology, irrespective of
methodology (e.g. Freeman 2007; Scurr 2007; Kerezsy 2009).
Finally, the surface area of a site will also govern the effectiveness of control programs. Sites with
a large surface area and greater numbers of preferred habitats that need to be repeatedly targeted
require a far greater effort than small sites with a single target area. While eradication in the
current study was achieved in sites up to 600 m2, Kerezsy (2009) found that physical removal
could only be a successful strategy for complete eradication for sites of less than 3 m2. Of course
this is a limitation for all control strategies, whether they be physical, chemical or drying of sites.
Our study has highlighted the fact that that, unless removal exercises result in complete
eradication, the suppression of Eastern Gambusia population size lasts little more than a single
month. Lydeard and Belk (1993) considered two management strategies for Gambusia species:
partial removal by seining to reduce population size, and complete removal by poisoning and
subsequent reintroduction of native species. After conducting mesocosm experiments they found
that the presence of Eastern Gambusia at both high and low densities had a negative effect on
population growth of Least Killifish Heterandria formosa. The experiment involved an initial
stocking of mesocosms with zero, low and high densities of Eastern Gambusia, together with the
native species. Eastern Gambusia numbers in the low-density treatment increased to levels present
in the high-density treatment in less than a month, finishing in higher abundances at the
completion of the four-month trial.
Therefore, if managers are aiming for longer-term benefits from Eastern Gambusia removal, they
must either achieve complete eradication or undertake removal exercises at regular enough
intervals so that the rate of removal is greater or equal to the rate of population growth. For Eastern
Gambusia, the latter is extremely hard to achieve because of the remarkable ability of the species
to recolonise a site. Typically this would involve repeated site visits to undertake control actions,
and is thus one of the primary reasons why current management strategies focus on chemical and
drying techniques, as they are the most likely techniques that can result in complete eradication of
the species using fewer visits to a site. Even so, repeated treatments are still recommended in
chemical control strategies (e.g. Elkington and Maley 2005). Willis and Ling (2000) noted the
difficulties in using chemical treatment alone as a means of eradication, suggesting that a single
treatment is unlikely to completely eradicate a small livebearing fish such as Western Gambusia.
While our study and that of Kerezsy (2009) has demonstrated that eradication can be achieved by
106
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
physical removal, it is far less likely to achieve this result in much larger systems, or in warmer
regions where the the ideal removal period is severely reduced or is non-existent. This also
highlights the importance of repeated removal efforts, whatever method is used. This is not to say
that physical removal should not be trialled in larger enclosed systems, particularly in those within
temperate regions where the period of non-spawning activity is much longer and repeated site
visits can be undertaken.
Research should now be directed to new control techniques for Eastern Gambusia because of the
current lack of successful and feasible control techniques. Trap designs have already been
investigated in Tasmania (Maynard et al. 2008) and gene technology aimed at affecting the
reproductive capabilities of Eastern Gambusia has been trialled (Fairfax et al. 2007). Rigorous
investigation into possible biological controls needs to continue (or expand or commence) if
Eastern Gambusia control or eradication is to be successful. It is an encouraging sign that scientists
and managers are now considering alternative removal exercises beyond chemical applications and
drying.
4.4.2
Eastern Gambusia as an invasive species: colonisation and population dynamics
During the course of the trial, Eastern Gambusia displayed an astonishing capacity to rapidly
colonise habitats. By combining mark–recapture data with a population growth modelling
component, the trial revealed that just a few individuals of this alien species can colonise an area
over the course of just several months. At the onset of both years of the experiment (October –
November), Eastern Gambusia were found in very low abundances, often just a few individuals.
Where these fish persisted (i.e. in control sites) populations rapidly established, reaching
population sizes in their thousands in three to four months. This supports the general view that a
single gravid female can colonise a site (Pyke 2005; Rowe et al. 2008).
The state-space modeling approach used in this study demonstrated the alarming intrinsic rate of
increase. This is not surprising given the species generally has several broods in a single breeding
season, a gestation period usually around 22–25 days and maturation at around 4–6 weeks of age
(Milton and Arthington 1983; Pyke 2005; Rowe et al. 2008). While not accounting for factors such
as predation and resource availability, Maglio and Rosen (1969) calculated that 10 adult females
could produce a population of 5 million individuals in six months, such is the reproductive
potential of the species. Although not following this extreme trajectory, our study provides fieldbased evidence that astonishing population increases are still achieved even with resource
limitations and predation.
The modelling approach also indicated that the intrinsic rate of increase of Eastern Gambusia was
far higher than even the most common native species in the region. This emphasises the species’
potential to out-compete native species, particularly if we also consider that Eastern Gambusia is
an opportunistic or generalist omnivore (Arthington and Marshall 1999; McKay et al. 2001;
Maynard et al. 2008) and has an extreme tolerance of poor water quality (Arthington, et al. 1983;
Kennard et al. 2005; King and Warburton, 2007). Wide environmental tolerances, short generation
times, rapid growth, broad diet, early sexual maturity and a high reproductive capacity have all
been identified as being important attributes of successful invasive species (Koehn 2007). These
attributes make Eastern Gambusia particularly suited to establishing populations after major
disturbances of native fish populations, whether they are natural or human-induced. For example,
ihe first year of our trial the sites had only recently been refilled following several years of extreme
drought. Similar abundances of adult Eastern Gambusia and native species had recolonised these
habitats, yet it was Eastern Gambusia that quickly asserted dominance. This pattern is also evident
in many urban systems that have suffered extreme habitat alteration. For example, habitat
alteration and subsequent water pollution have contributed to the decline of native fishes and
107
Native fish recovery following Eastern Gambusia removal
subsequent establishment of Eastern Gambusia in urban Brisbane waterways (Arthington et al.
1983).
4.4.3
Fish community response to Eastern Gambusia removal
The ultimate aim of the field removal experiment was to assess the fish community response (in
particular native species) following the reduction in Eastern Gambusia, and subsequently, test the
assumptions generated in phase 1 of the experiment. Whilst the native fish response component of
the experiment was limited to a reduced number of sites and species (as a result of extremes in
environmental factors, namely drought followed by flood), the results gave some indication of the
responses we might expect to see by native small bodied fish communities in the Mid-Murray
region following reductions in Eastern Gambusia abundances. These results also showed some
alignment with the predictions of the cross-sectional component (phase 1 of the study).
The assessment of overall fish community responses to Eastern Gambusia removal (population
modeling) were limited to the first year of the experiment and could only focus on two common
native species and one alien species. Nevertheless, the overall results suggest Eastern Gambusia
had a slight negative impact on the short term intrinsic rate of population increase for Australian
Smelt, Carp Gudgeon and Common Carp, with this negative impact increasing with increasing
abundances of Eastern Gambusia. Additionally, the assessment of condition indices also suggested
that increasing Eastern Gambusia abundance will increase the likelihood of fin damage for
Australian Smelt and Flat-headed Gudgeon, and decrease the morphometric condition of all three
native species. Whilst there is still some uncertainty surrounding these predictions, the combined
results suggest that Eastern Gambusia removal resulted in an improvement in fish condition (both
increased and decreased fin damage) which is likely to have resulted in increased population sizes
for these species. These results provide field-based support for numerous predictions based on
theoretical considerations and aquarium trials regarding the likely impacts of Eastern Gambusia in
the natural environment. For example, Macdonald and Tonkin (2008) identified significant
ecological niche overlaps (habitat and/or dietary) between Eastern Gambusia and many smallbodied native species within the MDB (including those investigated in this study). Additionally,
aquarium experiments investigating behavioural interactions between Eastern Gambusia and
several small Australian native fishes demonstrated that interference competition and aggression
towards Pacific Blue-eye Pseudomugil signifier, Duboulay’s Rainbowfish Melanotaenia
duboulayi, ornate rainbow fish Rhadinocentrus ornatus and Firetail Gudgeon Hypseleotris galii
increased with relative densities of Eastern Gambusia (Knight 1999; Breen 2000; Conte 2001). All
of these studies concluded that the establishment of Eastern Gambusia is likely to have contributed
to reductions in natural populations of these species.
Although our modelling predictions indicated that Eastern Gambusia removal would result in a
positive response in condition and population size of the three co-occurring fish species, the degree
of improvement, as indicated by changes in population trajectory and condition, is likely to be
minimal, at least in the short term. Despite the uncertainty in the modelled predictions of
population growth and condition indices, the predictions for all of the species indicated that the
degree of severity of these negative effects was extremely low. Australian Smelt displayed the
largest estimated negative impact on population growth and condition indices in response to
increasing Eastern Gambusia abundance, most likely because of major trophic niche overlaps in all
life stages (see Macdonald and Tonkin 2008). Carp Gudgeon displayed a significant reduction in
morphometric condition with increasing Eastern Gambusia abundance, but only a slight negative
change in population growth. These results support the modelled predictions of the cross-sectional
analysis, in which Eastern Gambusia was predicted to have very little impact on the occupancy
and population size of the common native species in floodplain wetlands (the greatest negative
impact was predicted to be for Australian Smelt). As suggested in the chapter 3, it may be that the
108
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
minimal impact of Eastern Gambusia on these species is a major factor contributing towards them
remaining largely widespread and abundant throughout the MDB. There is therefore uncertainty as
to why the impact of Eastern Gambusia on these species is minimal, despite what appears to be
major ecological niche overlaps.
The Carp Gudgeon species complex displayed a significant decline in condition in relation to
increasing gambusia abundance, but overall population change was shown to be minimal. The
species is still widespread throughout the Basin (Lintermans 2007), despite major overlaps in
dietary and habitat niches with Eastern Gambusia. Stoffels and Humphries (2003) reported all size
classes of Carp Gudgeon and Eastern Gambusia showed significant partitioning of food resources,
despite small Carp Gudgeons exhibiting high spatial overlap with Eastern Gambusia within
surface habitats coupled with similar diets. This may suggest that while similar, the trophic niche
of smaller Carp Gudgeon is broad enough to accommodate Eastern Gambusia. In addition to the
Carp Gudgeon species complex, Flat-headed Gudgeon and Australian Smelt are all still relatively
widespread throughout the MDB and all also occupy a broad range of habitats (King et al. 2007).
Ling (2004) reported that many of New Zealand’s small native fish occupy relatively broad niches
and that some degree of niche contraction caused by Eastern Gambusia may be accommodated
without severe consequences. Similarly, Ayala et al. (2007) observed that the impacts of gambusia
on Least Chub Iotichthys phlegethontis may also be mitigated by changes in the seasonal and daily
habitat use of these two species. During the summer months, Least Chub exploit cooler, deeper
habitats that are free of Western Gambusia, or moved to shallow, warmer habitats at night when
Western Gambusia were less active. The availability of such habitats, even though suboptimal,
may have acted as a buffer, reducing the magnitude of impacts and promoting the coexistence of
the two species. The current study was conducted in largely intact floodplain wetlands, which
provides frequent connection to adjoining habitats, as well as a variety of microhabitat niches (e.g.
a variety of structural complexities in the form of aquatic vegetation and woody debris). The minor
impact of Eastern Gambusia we observed on these common species may be a reflection of the
complexity of microhabitats in natural systems, allowing these generalist native fishes to shift
niches to those not utilised by Eastern Gambusia and thereby potentially reducing resource overlap
and competitive interactions, and lessening the impacts of high Eastern Gambusia densities. A
recent study of rivers in the Lake Eyre Basin suggested that the naturally variable hydrological
regimes and native dominant fish assemblages of the area afford some resistance to the
establishment and proliferation of alien fish (Costelloe et al. 2010).
The negative impacts of Eastern Gambusia on native fish populations is likely to be far greater for
fish populations occupying highly degraded sites. Highly degraded sites typically exhibit a
reduction in connectivity and habitat uniformity resulting in limited opportunity for recolonisation
or niche contraction. Morgan et al. (2004) found that lentic habitats that contained Eastern
Gambusia, but provided cover such as aquatic vegetation and snags, contained more native fish
than areas without Eastern Gambusia where cover was lacking. This suggests that the minor
impacts of Eastern Gambusia on the common generalist native species that we have reported in
this study are likely to be exacerbated in more degraded sites. Pyke (2008) suggested that reducing
any negative impacts of Gambusia species on native species can be achieved not only by a
reducing their abundance but by reducing the impacts per individual. From a management
perspective, minimising the impact of Eastern Gambusia on common generalist species may be
achieved by habitat restoration, such as the introduction of woody debris, flow restoration and the
rehabilitation of riparian vegetation (e.g. Kennard et al. 2005). Further assessments of more
isolated or degraded sites, involving longer-term monitoring of the fish community following
Eastern Gambusia removal, is required to confirm this.
109
Native fish recovery following Eastern Gambusia removal
The patterns reported thus far in the study have focused on the minimal impact on the common
generalist species of the MDB, suggesting that the ability of these species to shift or contract their
trophic niche dampens the negative impacts of Eastern Gambusia. This also suggests that the
negative impacts of Eastern Gambusia are likely to be far greater on species which do not display
such a broad trophic niche. The low numbers of Murray–Darling rainbowfish, southern pygmyperch and flat-headed galaxias collected during the experiment did not allow the population
modeling to be undertaken, however, the limited information collected on fin condition for
southern pygmy-perch and Murray–Darling rainbowfish suggested that these species were much
more prone to fin damage than the generalist species. The extremely low numbers of these species
collected during the study, coupled with the dominance of Eastern Gambusia across sites may in
itself be an indicator of potentially greater negative impacts of Eastern Gambusia on these more
specialist species. A large proportion of these more specialist wetland and confined waters species
have undergone major reductions in range and abundance across the MDB. For example, Southern
Pygmy-perch, Southern Purple-spotted Gudgeon and Olive Perchlet are now classified as
threatened within all or certain areas of the MDB (Lintermans 2007). While habitat loss and
degradation are likely to be the major cause of the declines in such species, numerous studies have
suggested that the dominance of Eastern Gambusia within much of these species’ preferred habitat
is likely to have accelerated the declines (e.g. Arthington 1983; Lloyd and Walker 1986;
Lintermans 2007). For example, during surveys in the lower Murray River, Lloyd and Walker
(1986) found that Southern Pygmy-perch was absent from sites of suitable habitat but with an
abundance of Eastern Gambusia, suggesting that the interaction between these species could be
responsible. Similarly, recent surveys of the Wimmera River basin during the Victorian
Sustainable Rivers Audit showed that there were few or no Southern Pygmy-perch in areas where
Eastern Gambusia were collected (SRA 2007), although this may be the result of factors such as
habitat degradation. Collectively, the results of this and earlier studies suggest that, in the short
term, management and control of Eastern Gambusia should focus on sites containing species with
a narrow trophic niche, or on sites containing highly uniform habitats. Of course, this may also
include sites which connect to these sites (thus potentially act as a source of Eastern Gambusia
invasion), or where these species could be re-introduced. Further field based studies assessing the
response of these rarer specialist species following Eastern Gambusia removal is required to
ascertain these predictions.
The results of the trial also indicated a slight negative impact on another co-existing alien species,
Common Carp. This therefore suggests that removal of Eastern Gambusia may have unexpected
benefits to other alien species. One question that is often not considered in pest species
management programs is, ‘What negative effects will a reduction or complete removal of the
target alien species have on ecosystem function?’ There is other evidence that successful
eradications of alien species can have unexpected and undesired impacts on native species and
ecosystems, particularly in areas which have accommodated the alien species for long periods of
time and where they are an established species in the food chain (e.g. Murphy et al. 1998).
Maezono and Miyashita (2004) suggested two ways in which such undesired impacts may occur.
First, the removal of an alien species can enhance secondary establishment, or increase the impact
of other alien species. Secondly, negative impacts to native biota may occur if the alien species
performs functions similar to those of native species that are no longer in the system. An example
of the first mechanism was reported by Maezono and Miyashita (2004), who investigated the
removal of introduced Largemouth Bass Micropterus salmoides on native communities in farm
ponds. While removal of this alien species did result in an increase in native fish and shrimp, there
was also a substantial increase in alien crayfish, which resulted in a substantial reduction in
macrophytes and associated rare odonate species. Zavaleta et al. (2001) suggest eradication of the
alien prey species only can also cause problems by forcing the alien predator to switch to native
110
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
prey. Indeed, such alien predator and prey interactions may be relevant for areas containing both
Eastern Gambusia and Redfin. McNeil (2004) suggested that Eastern Gambusia may be subject to
selective predation by Redfin in billabong habitats. This alien species predator–prey interaction
was hypothesised as responsible for the dominance of small native species in habitats containing
Redfin, and the dominance of Eastern Gambusia in habitats without Redfin. This suggests that the
removal of Redfin alone may result in an increase in Eastern Gambusia dominance, or that the
removal of Eastern Gambusia alone may result in increased predation pressure on the existing
native fish community.
4.4.4
Conclusions
While the extremes in environmental variables have limited the analysis and conclusions that
could be drawn from the field removal trial, the experiment still provided important information on
Eastern Gambusia removal, population dynamics and native fish responses following such
intervention, as follows.
•
Under certain conditions, physical removal can be used as a management tool to achieve
major reductions in Eastern Gambusia populations, but the degree of success depends on a
thorough consideration of various aspects of a sites hydrology, climate, habitat and size.
•
Eastern Gambusia has a remarkable capacity to rapidly recolonise habitats: just a few
individuals rapidly established population sizes in their thousands in three to four months.
The intrinsic rate of increase of Eastern Gambusia populations was far higher than even
the most common native species in the region. This emphasises the species’ ability to
establish populations rapidly and out-compete native species.
•
Most importantly, reductions in Eastern Gambusia abundance benefit small-bodied native
fish populations. The negative impacts on the more common generalist species within
these intact floodplain wetlands were shown to be relatively minimal. However, negative
impacts might be far greater on species with narrow trophic niches. (A large proportion of
such species have already suffered major reductions in range and abundance.) This
suggests that, in the short term, management and control of Eastern Gambusia should
focus on sites containing these species, or on sites containing highly uniform habitats.
Further field studies to assess the response of these rarer specialist species, and longerterm monitoring of fish communities within more degraded sites following Eastern
Gambusia removal, are required to test these predictions.
•
The removal of Eastern Gambusia may have unexpected benefits to other alien species,
such as Common Carp, so site-specific ecosystem function in the absence of Eastern
Gambusia must also be considered before undertaking a removal program.
111
Native fish recovery following Eastern Gambusia removal
5
Phase 3: Cost-effectiveness and logistics of Eastern
Gambusia removal
5.1 Introduction
The effective management of alien fish relies largely on prevention and containment, with the
chances of success usually being directly proportional to the extent of dispersal (Raynor and
Creese 2006; Gozlan et al. 2010). Consequently, for established, widespread alien fish species the
options for management are restricted mainly to mitigation schemes that compensate for the
presence of the species (Gozlan et al. 2010).
Eastern Gambusia is now well established throughout the Murray–Darling Basin (MDBC 2008). If
we also consider its ability to rapidly colonise an area (the results of phase 2 of this project being a
prime example) and the absence of any effective large-scale control, it is clear that it is not feasible
at present to undertake large-scale Eastern Gambusia control or eradication across the MDB. This
does not, however, preclude management actions for this species. The NSW control plan for
Common Carp reported that there is anecdotal evidence of localised efforts to reduce carp numbers
helping to keep numbers under control, especially when repeated over a period of time (NSW
2010), despite this species being well established throughout NSW. Although direct ecological
benefits were the ultimate aim of these efforts, the plan also notes these actions have important
social benefits, particularly in community education, which in itself is an important management
tool for pest fish, e.g. preventing new incursions through deliberate introductions, illegal use as
live bait, and accidental translocation via gear (NSW 2010).
Even at a more local scale there is very little information available on mitigating the impacts of
Eastern Gambusia; the few documented cases of control focus predominantly on chemical
techniques and drying of habitats (see McKay et al. 2001). While managers await the development
of possible wide-ranging solutions such as large-scale harvesting techniques and daughterless
technology, there is an urgent need for control options at a local scale, where total eradication
using chemical treatments is undesirable because of the presence of threatened species or fragile
ecosystems. Consequently, minimising the ecological impact of Eastern Gambusia at a local scale
may be an important conservation strategy, particularly in areas containing threatened fish species,
until new control measures are developed.
The key objective of an alien species removal program is to reverse the negative impacts the
species has had on environmental and socio-economic values. An assessment of these values in
relation to invasive alien species is critical for sound environmental management and policy
development, but such assessments are rare, particularly at an economic level. In Australia for
example, economic assessments of freshwater fish invasions to our knowledge total one, for
Common Carp (McLeod 2004; Rowe et al. 2008). In the case of Eastern Gambusia control, the
benefits of controlling or eradicating populations do not have a quantifiable monetary value
(primarily because ecological benefits accrue in systems for which the economic values are
unknown), so they cannot be contrasted directly with the costs of control (e.g. research, sampling
time and effort, removal). This makes justifying management and environmental policy for the
species difficult because a true cost–benefit analysis cannot be undertaken.
Nevertheless, the results of the study have provided important information on the physical removal
of Eastern Gambusia that can be used to identify strategies that maximise the level of improvement
to the native fish community for a fixed budget (budget maximisation), or minimise the cost of
achieving a defined removal outcome (cost minimisation) (e.g. Choquenot et al. 2004; Koehn and
MacKenzie 2004).
112
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
5.2 Cost-effectiveness and logistics of Eastern Gambusia removal using
physical control programs
Chapter 4 provided a thorough assessment of the methodology and ecological responses associated
with the physical control of Eastern Gambusia, incorporating the results of the current project,
other studies and theoretical information. We concluded that physical control programs can be
used to substantially reduce Eastern Gambusia populations and ultimately provide benefits to
native fish communities, and provide an alternative to conventional non-specific methods such as
chemical treatments and drying of sites. The physical methodology used in this study involved a
combination of targeted trapping and seine netting. Aside from the advantages of these methods in
targeting Eastern Gambusia while having minimal impact on native fauna, these methods are also
relatively inexpensive in terms of equipment costs (approximate AUD$90 and AUD$5 for a lightweight seine net and bait trap respectively) and are easy to deploy and maintain (e.g. Brookhouse
and Coughran 2010). After brief training and supervision by an experienced zoologist (particularly
in fish identification), both techniques could be used easily by community groups such as
Landcare and schools.
Although the these techniques are low-cost and require little training, the costs associated with the
amount of effort, and their effectiveness in achieving the primary objective of substantially
reducing Eastern Gambusia abundances to benefit native fauna, is highly variable between sites,
and depends on a number of factors. As discussed in chapter 4, the factors influencing overall
costs and the effectiveness of a physical control program at a specific site can be generalised to
four simple variables relating to a site’s hydrological connectivity, ecological value, habitat
complexity and size. Considering these factors in a methodical manner would enable managers to
prioritise sites for the physical control of Eastern Gambusia in a way that maximises the ecological
benefits per dollar invested, or that minimises the costs to achieve a defined ecological benefit
(e.g. eradication from a certain number of sites). Subsequently, we have developed a simple
decision support tool for managers considering investing in a physical control program (Figure
5.1). By considering a few basic questions regarding the primary factors we have discussed, the
tool enables managers to rate individual sites (from ‘very low’ to ‘high’) according to the
ecological benefits per dollar invested. This then would enable managers to decide whether to
employ a physical control program and to prioritise the sites that the program should target.
113
All sites
Low ecological value
Is the site of high ecological value
(species / habitat)?
High frequency
of connection
How frequently does the site connect to
adjoining waterbodies?
Low frequency
of connection
High ecological
value
High frequency of
connection
How frequently does the site connect
to adjoining waterbodies?
Low frequency of
connection
How much structural habitat
does the site contain?
How much structural habitat
does the site contain?
High structural
complexity
High structural
complexity
Low structural
complexity
Large surface
area
What is the size of the site?
Small surface
area
Low structural
complexity
Large surface
area
What is the size of the site?
Benefits per $ invested
Permanently
connected
Does the site facilitate permanent immigration?
Isolated sites
Very Low
Permanently connected
sites facilitating constant
immigration of pest fish
Low
Isolated sites of low ecological
value and frequent connection
to adjoining habitats facilitates
frequent immigration of pest
fish
Medium
Isolated sites of high ecological
value but frequent connection
to adjoining habitats. Value of
investment increases with
reduced structural habitat
complexity and surface area
High
Isolated sites of high ecological
value and infrequent
connection to adjoining
habitats. Value of investment
increases with reduced
structural habitat complexity
and surface area due to a
reduction in required effort;
increased ability to undertake
active netting methods and
increased negative interaction
between pest and native
species due to a reduction in
habitat niche partitioning.
Small surface
area
Figure 5.1 Decision support tool for prioritising sites for physical control or Eastern Gambusia on the basis of maximising the ecological benefits per
dollar invested. Note that this does not factor in the socio-economic benefits of control programs.
Arthur Rylah Institute for Environmental Research
114
Irrespective of methodology, a fundamental knowledge of a site’s hydrology is required before
large amounts of time and money are invested in Eastern Gambusia removal (e.g. Kerezsy 2009)
— in particular, is a site isolated, or is it permanently connected to adjoining habitats so that a
continuing immigration or emigration of fish is possible? Clearly, if the latter is the case (for
example, if the site is a creek that connects directly to a river), any ecological benefits of Eastern
Gambusia removal will be short-lived (or non-existent) if fish simply recolonise the site
immediately after (or during) the removal program (particularly given the re-population rate
demonstrated for the species). For more isolated sites, consideration must also be given to how
often they are isolated. Many sites reconnect to adjoining waterbodies every year (e.g. many
wetlands on the lower Ovens River, Victoria), but others may only connect during a one-in-tenyear event. This implies that, for a similar investment, one site may get up to 10 years of
ecological benefits compared to less than 12 months at another site.
Prioritising sites based on their frequency of connection seems obvious, but the ecological value of
each sites must also be considered. Managers must question what ecological benefits may arise
from investing in an Eastern Gambusia removal program at a site which is not associated with high
ecological values (e.g. contains threatened fish species or frequently connects to areas which do),
compared to investing in those connecting to or containing threatened species or communities. For
example, if one site supports only Carp Gudgeon (which is widespread and, based on this study,
not drastically impacted by Eastern Gambusia abundance) and another site contains fish such as
pygmy-perch or glassfish that are more likely to be more heavily impacted, then the ecological
benefits achieved at the latter site will far outweigh those achieved at the former, irrespective of
differences between frequency of connection. For example, the basis behind employing Eastern
Gambusia control in a range of Australian artesian springs was the presence of two critically
endangered fish species: Red-finned Blue-eye Scaturiginichthys vermeilipinnis and Edgbaston
Goby Chlamydogobius squamigenus (Kerezsy 2009). Brookhouse and Coughran (2010) suggested
that trapping Eastern Gambusia could have particular use in short-term applications, such as
greatly reducing the abundance of fish in preparation for the spring and summer spawning periods
of rare native fishes and amphibians. In the Murray–Darling Basin there are several small-bodied
fish species occupying small isolated sites that are now severely reduced in range and abundance
including pygmy-perch (e.g. Nannoperca australis), galaxids (Galaxias rostratus), Murray
Hardyhead (Craterocephalus fluviatilis), Southern Purple-spotted Gudgeon (Morgurnda adspersa)
and Olive Perchlet (Ambassis agassizii). Managers need to be aware of such species, as well as
other rare aquatic fauna such as some amphibian species (e.g. Komak and Crossland 2000), if
planning Eastern Gambusia control.
Chapter 4 highlighted that the structural habitat complexity of a site will largely govern the
effectiveness of a removal strategy, in that sites containing large amounts of this habitat (such as
snags), will severely reduce the effectiveness of the most efficient physical removal method
(targeted active netting such as seine and dip-netting). This implies that the control program will
require a far greater effort, because of the reduced efficiency of targeted active methods and an
increase in the reliance on trapping techniques. Of course, sites containing greater amounts of
structural habitat may also be in less need of Eastern Gambusia control because there is likely to
be fewer aggressive interactions and more ecological niche separation with native species,
although this depends on the native species occupying the site. So for a fixed investment at sites
with similar ecological value and frequency of connection, the success of the removal program (in
terms of the proportion of fish removed along with the benefits to native species) will decrease as
the structural habitat complexity of a site increases.
Depth does not influence removal efficiency (as Eastern Gambusia is unlikely to be using deeper
water when they are targeted), but the surface area of a site does govern the effectiveness of
Arthur Rylah Institute for Environmental Research
115
Native fish recovery following Eastern Gambusia removal
control programs. The larger a site’s surface area, the greater number of preferred habitats that
need to be repeatedly targeted. Thus, larger sites require a far greater effort than small sites with a
single target area. This is a limitation for all control strategies, whether they are physical, chemical
or drying of sites.
While the decision support tool considers the primary factors governing the effectiveness of a
physical control program for Eastern Gambusia, there may be other factors specific to jurisdictions
which enable sites to be prioritised further. These factors range from detailed biological
information to legislation. For example, chapter 4 highlighted the importance of undertaking
removal before the onset of Eastern Gambusia spawning. Because spawning times of Eastern
Gambusia depend on the regional climate, the ideal window for removal in a particular area needs
to be considered when developing a physical control program (e.g. regions with a short removal
window may need a short, intensive program). Additionally, ecological knowledge of a region’s
native biota should also be considered. For example, removal activities should ideally be
undertaken outside the peak spawning periods of threatened native fish species present at a site.
Although not relevant specifically to this project, legislation in regard to the use of gear and animal
ethics may also affect how a control program is undertaken. For example, we recommend the use
of lightweight seine nets as an active removal technique, but their use in inland waterways is
restricted to nets less than 6 m long and restricted for use in nine inland waterways. Realistically,
this makes the technique illegal for the general public to use. If managers intend to involve the
community in a removal program they must therefore consider relevant fisheries and animal ethics
legislation beforehand.
Finally, this decision support tool relates to the logistics and cost-effectiveness of the physical
control of Eastern Gambusia, not other methods. If sites are characterised as being likely to
experience very low or low ecological benefits for a given investment, alternative mitigation
activities should still be investigated. For example, a site may be of low priority for physical
removal due to the high frequency of connection compared to other sites however, there may be
potential value in habitat restoration activities such as resnagging and replanting riparian and/or
aquatic vegetation (see chapter 4). Of course, these activities may also be used in conjunction with
physical removal programs.
The decision support tool and associated information on the physical removal of Eastern Gambusia
will also been presented in a ‘guidelines for managers’ brochure that will be available to CMAs
and Landcare groups (see appendix 2).
5.3 Social benefits of Eastern Gambusia removal: participation and
education
The major objectives of the current project focused on ecological issues with respect to Eastern
Gambusia impacts and control, but the social implications of alien fish management cannot be
ignored. While the difficulties of quantifying purely economic measures associated with Eastern
Gambusia impacts (and therefore benefits of control actions) have been discussed, the social
benefits of removal programs are even more difficult to quantify but are a vital component of any
alien fish management program.
Community education is a vital component of pest fish management for a number of reasons
(Koehn 2007). As is the case with Eastern Gambusia, people are the principal means by which
alien fish are spread (Wells 2007), yet there is a substantial lack of community education and
awareness of the threats posed by alien fish (Koehn et al. 2007). During the course of this project,
frequent discussions with landholders and, on one occasion, a demonstration and information
116
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
session presented to a local school group (Figure 5.2) revealed an alarming lack of awareness in
regard to Eastern Gambusia. For example, Eastern Gambusia were being ‘rescued’ by members of
the community from drying water bodies and re-released back into rivers and wetlands because
they were thought to be small-bodied native fish (Fern Hames pers. comm.). If a control program
can also raise community awareness of the issue then there is less potential for humans to continue
to spread this alien species through such measures as deliberate introductions and illegal use as
bait. The NSW control plan for carp has also reported that localised carp removal can have
important social benefits, particularly at an educational level, which in itself is an important
management too for controlling pest fish (NSW 2010).
Finally, Koehn and MacKenzie (2004) point out that funding for on-ground natural resource
management activities in Australia is largely directed through regional Natural Resource
Management (NRM) bodies, which determine NRM priorities for their region (funding is allocated
according to those priorities). These groups also coordinate and support other community-based
activities such as Landcare and integrated catchment management. Therefore funding and support
for on-ground alien species management can be accessed or enhanced if regional bodies are
provided with sufficient information about the significance of the alien species threat and how the
community can assist (Koehn and MacKenzie 2004).
A key aspect of an alien fish management plan should therefore include a community awareness
and education component. This project included a variety of ways of disseminating information on
the Eastern Gambusia problem and the objectives of the project (see appendix 2). Communication
activities and materials produced through the MDBA’s Native Fish Strategy have also made
excellent contributions in this area, including workshop reports (e.g. Ansell and Jackson 2007),
materials such as the Aliens in the Basin! brochure and Alien Fish in the Murray–Darling Basin
book, events associated with Demonstration Reaches and Native Fish Awareness Week, and many
activities delivered by the Native Fish Strategy Community Stakeholder Taskforce and local
coordinators. Such activities and materials are clearly important and should continue.
A range of other educational tools and resources have also been introduced which are beginning to
build community education around alien fish. These include resources such as the on-line River
Rescue scenarios developed by the University of Canberra, and the Sustaining River Life kit
established by WaterWatch, RiverSmart and the MDBA. In the case of Eastern Gambusia, one
area we wish to highlight is the potential for community involvement in control programs. As we
have discussed, the methodology used for the physical removal of Eastern Gambusia could easily
be used by community groups such as Landcare and schools. Aside from any direct ecological
benefits that may arise from these activities, involving the community in removal activities will
provide them with a sense of ownership of the pest fish issue and an improved understanding of
the complexities of managing alien fish species (Koehn and MacKenzie 2004). As the Victorian
Coordinator for the MDBA Native Fish Strategy and Community Engagement has pointed out at
community field days, ‘informing individuals that removing just one female Eastern Gambusia
will prevent tens of thousands of others in a few months time, is an extremely powerful message to
convey’. So even if community involvement in removal activities does not directly result in
immediate ecological benefits, these exercises will still result in substantial social benefits,
particularly in light of the evidently insufficient education on the threats of Eastern Gambusia.
117
Native fish recovery following Eastern Gambusia removal
Figure 5.2 Demonstration and information on Eastern Gambusia and wetland fish communities
presented to Colbinabbin primary school. (Photo: Glenace Avard)
5.4 Evaluating the benefits to native fish and cost-effectiveness of
controlling other potentially harmful alien species across the MDB:
applicability of the processes used in the current study
Because the control of numerous alien species (not just Eastern Gambusia) within the MDB is one
of the key driving actions of the Native Fish Strategy, it was relevant to consider whether the
processes undertaken in this study could be applied to evaluate the ecological benefits of
controlling other alien fish across the MDB. In this study we took a holistic approach to
investigating the feasibility of Eastern Gambusia control that would result in measurable
improvements to native fish communities. We did this by integrating surveys and quantitative
experimental work in natural billabong systems throughout the MDB and collating this
information with cost-effectiveness approaches, ultimately enabling managers to use resources in a
manner which maximises the potential ecological benefits of control strategies for a given
investment (i.e. the development of the decision support tool).
The cost-effectiveness assessment, and ultimately the decision support tool (section 5.3), was
derived from an assessment of the factors influencing the removal efficiency of Eastern Gambusia
— in turn derived from this study and results of research into control strategies, e.g. Maynard et al.
(2008) — and most importantly, the likely response of native fish following these reductions.
While the decision support tool is specifically for the physical removal of Eastern Gambusia
118
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
removal, the processes it was derived from have major relevance for controlling other established
alien species across the basin, particularly those which also have limited socio-economic value.
These processes have been generalised to form a template for the required approach to maximise
the benefits to native fish arising from an alien fish removal program (Figure 5.3).
The basic premise of this template is the development of a thorough understanding of the
ecological impacts of the alien species (and therefore the response following control measures) as
well as the factors influencing the removal activities themselves. This in turn will enable
maximum benefits to be achieved for native fish. Understanding the ecological impacts of an alien
species is an essential component of best practice vertebrate pest management, which is based on
the concept of managing impacts rather than numbers (Braysher 1993; Koehn and MacKenzie
2004). This together with the use of pest management principles (Braysher 1993; Bomford and
Tilzey 1997) and adoption of integrated approaches will give the best results by ultimately
providing a basis for benefit:cost analyses of management options (Choquenot et al. 2004; Koehn
and MacKenzie 2004).
Unfortunately, our understanding of the impacts of most alien species is still rudimentary, even for
direct impacts such as predation, let alone impacts at the community or ecosystem levels
(Townsend 2003). As a result, a thorough understanding of ecological impacts and therefore sound
predictions of ecological responses to removal is rarely utilised in the management of alien fish in
Australia. For example, despite millions of dollars being spent annually on carp control throughout
the MDB, the effects of carp on native biota remains largely speculative (NSW 2010) and is
limited to impacts on water quality (King et al. 1997; Koehn 2007). As a result, current costeffectiveness assessments and subsequent control measures for the species generally focus on
maximising the number of fish removed, as opposed to minimising the negative impacts of the
species on native biota. While the latter is generally assumed to be a product of the first (and this is
often the case), the ultimate goal of management for alien species is not a reduction in numbers
per se but a reduction in the impacts caused by each species (Lodge and Shrader-Frechette 2003;
Koehn 2007; Zavaletta 2010). Indeed, the current study has highlighted that, for established alien
fish species, an understanding of the response of native species to a control program under
different scenarios will enable the use of resources to maximise these ecological benefits,
assuming that ecological benefits are the primary objective of an alien fish removal program.
There is still the question of whether there is a need to definitively determine the specific impacts
of an alien species, or whether managers should simply apply the precautionary principle (based
on a reasonable suspicion of impacts) and commence management actions to control the threat
(Lintermans et al. 2007). The management of new incursions of an alien pest species is a case in
point: the high priority of preventing the spread and establishment of new populations (Koehn and
MacKenzie 2004) tends to override the lengthy processes involved in developing the predictive
capabilities (theoretical, proof-of-concept and cost-effectiveness stages in Figure 5.3), relying on a
rapid response to maximise the number of fish removed (e.g. Ayres and Clunie 2010). This is a
sensible approach for new incursions of alien fish species when they are generally still confined in
their distribution, but (as we have highlighted) undertaking large-scale control or eradication of
established alien species across the entire MDB is not feasible at present. For these species,
options for management are largely restricted to localised mitigation schemes that compensate for
the presence of the species (Gozlan et al. 2010). Therefore, for established species in the MDB
such as Common Carp, Eastern Gambusia, Redfin and Goldfish, management should be based on
an integrated and strategic approach that addresses the range of factors that threaten the health of
native fish populations and the general health of the MDB (Braysher 2007). Collectively, this
indicates a through understanding of the ecological impacts (and therefore sound predictions of the
responses following control measures) of specific alien species is essential to maximise the
119
Native fish recovery following Eastern Gambusia removal
ecological benefits (such as improvements in native fish fauna) of investment in the control of
established alien species within the MDB.
Again, this discussion and the template developed for prioritising control actions for established
alien fish is based purely on achieving ecological benefits. It has not considered socio-economic
factors, which are much more relevant for angling species such as Redfin and trouts. While this
assessment is outside the scope of the study, we suggest the simplistic process presented in the
study aimed purely at assessing the ecological benefits of alien fish control has relevance for other
established alien fish species with limited socio-economic value, such as Goldfish and Oriental
Weatherloach (for which the biological impacts on native biota is also largely unknown). Thus, we
suggest that a similar approach to that undertaken in the present study would prove valuable.
Thorough understanding of the life-history
and behavioural traits of the alien species
and co-occurring native fish species
Identify the species / communities which
are most impacted or at risk
Development of methodologies to
maximise the efficiency of removal
Undertake a field based proof of concept
approach testing predictions
Predictive capabilities on ecological
response
Identify factors influencing removal
measures
Cost-effectiveness assessment and
development of site prioritisation tool
Maximise the
ecological benefits
of alien fish
control program
Figure 5.3 Template for the processes used in the current study to maximise the benefits to
native fish arising from an alien fish removal program.
120
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
6
Conclusion
This project has taken a forward step into researching the feasibility of controlling Eastern
Gambusia populations to achieve measurable improvements to native fish communities. By
integrating surveys and quantitative experimental work in natural billabong systems throughout the
MDB, the project provided important information on Eastern Gambusia removal, population
dynamics, and the responses of native fish following such intervention.
Understanding the ecological impacts of an alien species is an essential component of pest
management, based on the concept of managing impacts rather than simply reducing numbers
(Braysher 1993; Koehn and MacKenzie 2004). The negative impacts of Eastern Gambusia on the
more common generalist species in these intact floodplain wetlands were minimal, but negative
impacts might be far greater on species that have narrow trophic niches. There was also some
indication that Eastern Gambusia may have unexpected impacts on other alien fish species.
While successful eradication of an established alien species such as Eastern Gambusia using
current methodologies is not likely to be feasible in larger open systems, minimising the impact of
Eastern Gambusia on native fish may be an important management strategy until new control
strategies such as harvesting techniques and daughterless technology are developed. Despite
Eastern Gambusia displaying an astonishing capacity to rapidly colonise habitats (with an intrinsic
rate of increase far higher than even the most common native species in the region), we found that
physical removal of Eastern Gambusia, when conducted under certain conditions, can achieve
major reductions in Eastern Gambusia populations.
6.1 Management / research recommendations
•
Although an assessment of control methods was not a major objective of this study, we
found that physical removal of Eastern Gambusia, when conducted under certain
conditions, can be used as an effective management tool to achieve major reductions in
Eastern Gambusia populations. However, the degree of success requires thorough
consideration of various aspects of a sites hydrology, climate, habitat and size.
•
Reductions of Eastern Gambusia will result in improvements to native fish populations,
but management and control actions should focus on sites containing species that have
narrow trophic niches, or on sites containing highly uniform habitats, to maximise the
ecological benefits of Eastern Gambusia reductions. A decision support tool has been
developed for managers to assess the feasibility and prioritise sites for Eastern Gambusia
control actions.
•
Further field studies to assess the response of rarer specialist native species and conduct
longer-term monitoring of fish communities across different habitat conditions following
Eastern Gambusia removal are required to refine these predictions.
•
Given the unexpected benefits to other alien fish species, ecosystem function in the
absence of Eastern Gambusia must also be considered before implementing a removal
program.
•
We found an alarming lack of community awareness about Eastern Gambusia. We
strongly recommend that community awareness and education exercises be included in
any Eastern Gambusia management program. Involving the community in removal
activities is a possibility because of the methods used, which could easily be utilised by
community groups such as Landcare and schools. Even if these activities do not directly
result in immediate ecological benefits, they will result in substantial social benefits,
particularly in light of the lack of education on the threats from this alien fish species.
121
Native fish recovery following Eastern Gambusia removal
•
122
We suggest that the simple process used in the study could also prove valuable if applied
to other established alien fish species, particularly those with limited socio-economic
value.
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
7
References
Allen, G. R. (1996a) Family Chandidae: Glassfishes, chanda perch. In Freshwater Fishes of SouthEastern Australia. (McDowall, R. M., Ed.), pp. 146-149. Reed Books, Sydney.
Allen, G. R. (1996b) Family Melanotaeniidae: Rainbowfishes. In Freshwater Fishes of Southeastern Australia. (McDowall, R. M., Ed.), pp. 134-140. Reed Books, Sydney.
Amundsen, P-A., Staldvik, F. J., Reshetnikov, Y. S., Kashulin, N., Lukin, A., Bøhn, T., Sandlund,
O. T. and Popova, O. A. (1999) Invasion of vendace Coregonus albula in a subarctic watercourse.
Biological Conservation 88, 405-413.
Angeler, D. G., Sanchez-Carrillo, S., Rodrigo, M. A., Alvarez-Cobelas, M. and Rojo, C. (2007)
Does seston size structure reflect fish-mediated effects on water quality in a degraded semiarid
wetland? Environmental Monitoring and Assessment 125, 9-17.
Ansell, D. and Jackson, P. (Eds). (2007) Emerging Issues in Alien Fish Management in the
Murray–Darling Basin: Statement, recommendations and supporting papers. Proceedings of a
workshop held in Brisbane QLD, 30-31 May 2006. Murray–Darling Basin Commission, Canberra.
Arthington, A. H. and Lloyd L. N. (1989) Introduced Poeciliids in Australia and New Zealand. In:
Ecology and evolution of livebearing fishes (Poeciliidae). (Meffe, G. K. and Snelson, F. F. Jnr.,
Eds.), pp. 333-348, Prentice-Hall, Inc., New Jersey.
Arthington, A. H. and Marshall, C. J. (1999) Diet of the exotic mosquitofish, Gambusia holbrooki,
in an Australian lake and potential for competition with indigenous fish species. Asian Fisheries
Science 12, 1-16.
Arthington, A. H. and McKenzie, F. (1997) Review of impacts of displaced/introduced fauna
associated with inland waters, Australia: State of the Environment Technical Paper Series (Inland
Waters), Department of the Environment, Canberra.
Arthington, A. H., Hamlet, S. and Bluhdorn, D. R. (1990) The role of habitat disturbance in the
establishment of introduced warm-water fishes in Australia. In: Introduced and translocated fishes
and their ecological effect. (Pollard, D. A., Ed.), pp. 61-66. Australian Government Publishing
Service, Canberra.
Arthington, A. H., Milton, D. A. and McKay, R. J. (1983) Effects of urban development and
habitat alteration on the distribution and abundance of native and exotic freshwater fish in the
Brisbane region, Queensland. Australian Journal of Ecology 8, 87-101.
Arumugam, P. T. and Geddes, M. C. (1987) Feeding and growth of golden perch larvae and fry
(Macquaria ambigua Richardson). Transactions of the Royal Society of South Australia 111, 5965.
Ayala, J. R., Rader, R. B., Belk, M. C. and Schaalje, G. B. (2007) Ground-truthing the impact of
invasive species: spatiotemporal overlap between native least chub and introduced western
mosquitofish. Biological Invasions 9, 857-869.
Ayres, R. and Clunie, P. (2010) Management of freshwater fish incursions: a review. Arthur Rylah
Institute for Environmental Research. Department of Sustainability and Environment, Heidelberg,
Victoria
Backhouse, G.N., and Frusher, D.J. (1980) The crimson-spotted rainbowfish, Melanotaenia
fluviatilis (Castelnau 1878). Victorian Naturalist 97(July/Aug): 144-148.
123
Native fish recovery following Eastern Gambusia removal
Balcome, S. and Humphries, P. (2006) Diet of the western Carp Gudgeon (Hypseleotris
klunzingeri Ogilby) in an Australian floodplain lake: the role of water level stability. Journal of
Fish Biology 68, 1484-1493.
Balcome, S., Bunn, S. E., Mckenzie-Smith, F. J. and Davies, P. M. (2005) Variability of fish diets
between dry and flood periods in an arid zone floodplain river. Journal of Fish Biology 67, 15521567.
Barlow, C. G. and Bock, K. (1981) Fish in farm dams: final report to the Reserve Bank’s Rural
Credits Development Fund. NSW Fisheries Research Institute, Cronulla, Australia.
Beesley, L.S., and Prince, J. (2010) Fish community structure in an intermittent river: the
importance of environmental stability, landscape factors and within-pool habitat descriptors.
Marine and Freshwater Research 61, 605-614.
Beatty, S.J. 2006. The diet and trophic positions of translocated, sympatric populations of Cherax
destructor and Cherax cainii in the Hut River, Western Australia: evidence of resource overlap.
Marine and Freshwater Research 57, 825-835.
Bomford, M. and Hart, Q. (2002) Non-indigenous vertebrates in Australia. In: Biological
Invasions - Economic and Environmental Costs of Alien Plant, Animal and Microbe Species.
(Pimental, D. Ed.), CRC Press, New York.
Boulton, A. J. and Brock, M. A. (1999) Australian freshwater wetlands: processes and
management. Gleneagles Publishing, Glen Osmond.
Bomford, M., and Tilzey, R. (1997) Pest management principles for European carp. Controlling
carp: exploring the options for Australia, Albury, pp. 9-22.
Braysher, M. (1993) Managing vertebrate pests: principles and strategies. Bureau of Resource
Sciences, Australian Government Publishing Service, Canberra.
Braysher, M. (2007) Management of pest fish. Emerging Issues in Alien Fish Management in the
Murray–Darling Basin: Statement, recommendation and supporting papers. Proceedings of a
workshop held in Brisbane QLD, 30-31 May 2006, Brisbane, pp. 33-39.
Breen, A. (2000) Density dependent interference competition between the exotic poeciliid
Gambusia holbrooki (Girard, 1859) and the Australian native melanotaeniid Rhadinocentrus
ornatus (Regan, 1914). Unpublished Honours thesis, Southern Cross University, Lismore.
Brookhouse, N., and Coughran, J. (2010) Exploring the potential for an ecology-specific, physical
control method of the exotic pest Mosquitofish, Gambusia holbrooki. Ecological Management and
Restoration 11, 226-228.
Brown, P., Sivakumaran, K.P., Stoessel, D., and Giles, A. (2005) Population biology of carp
(Cyprinus carpio L.) in the mid-Murray River and Barmah Forest wetlands, Australia. Marine and
Freshwater Research 56, 1151-1164.
Cadwallader, P. L. and Backhouse, G. N. (1983) A guide to the freshwater fish of Victoria.
Ministry for Conservation, Melbourne.
Cadwallader, P. L. and Douglas, J. (1986) Changing food habits of Macquarie Perch, Macquaria
australasica Cuvier (Pisces: Percichthyidae), during the initial filling phase of Lake Dartmouth,
Victoria. Australian Journal of Marine and Freshwater Research 37, 647-657.
Caiola, N. and de Sostoa, A. (2005) Possible reasons for the decline of two native toothcarps in the
Iberian Peninsula: evidence of competition with the introduced eastern mosquitofish. Journal of
Applied Ichthyology 21, 358-363.
124
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Capinha, C. and Anastácio, P. (2011) Assessing the environmental requirements of invaders using
ensembles of distribution models. Diversity and Distributions 17, 13-224.
Chapman, P. and Warburton, K. (2006) Postflood movements and population connectivity in
gambusia. Ecology of Freshwater Fish 15, 357-365.
Choquenot, D., Nicol, S. J. and Koehn, J. D. (2004) Bioeconomic modelling in the development of
invasive fish policy. New Zealand Journal of Marine and Freshwater Research 38, 419-428.
Clavero, M. and García-Berthou, E. (2005) Invasive species are a leading cause of animal
extinctions. Trends in Ecology and Evolution 20, 110.
Clunie, P. and Koehn, J. (2001) Freshwater Catfish: A Resource Document. Final report to
Murray–Darling Basin Commission, Canberra. Prepared by the Arthur Rylah Institute for
Environmental Research, Department of Sustainability and Environment, Heidelberg.
Conte, S. (2001) An investigation of density-dependant interference competition between the
exotic poeciliid Gambusia holbrooki (Girard, 1859) and the Australian native fish Hypseleotris
galii (Ogilby, 1903). Unpublished Honours thesis, Southern Cross University, Lismore.
Copp, G. H., Bianco, P. G., Bogutskaya, N. G., Eros, T., Falka, I., Ferreira, M. T., Fox, M. G.,
Freyhof, J., Gozlan, R. E., Grabowska, J., Kovac, V., Moreno-Amich, R., Naseka, A. M., Penaz,
M., Povz, M., Przybylski, M., Robillard, M., Russell, I. C., Stakenas, S., Sumer, S., Vila-Gispert,
A. and Wiesner, C. (2005) To be, or not to be, a non-native freshwater fish? Journal of Applied
Ichthyology 21, 242-262.
Costelloe, J.F., Reid, J.R.W., Pritchard, J.C., Puckridge, J.T., Bailey, V.E., and Hudson, P.J.
(2010) Are alien fish disadvantaged by extremely variable flow regimes in arid-zone rivers?
Marine and Freshwater Research 61, 857-863.
Courtenay, W. R. J. and Meffe, G. K. (1989) Small fishes in strange places: a review of introduced
poeciliids. In: Ecology and evolution of livebearing fishes (Poeciliidae). (Meffe, G. K. and
Snelson, F. F. Jnr., Eds.), pp. 319-332. Prentice-Hall, Inc., New Jersey.
Crivelli, A.J. (1981) The biology of the Common Carp, Cyprinus carpio L. in the Camargue,
southern France. Journal of Fish Biology 18, 271-290.
Crook, D.A., Reich, P., Bond, N.R., McMaster, D., Koehn, J.D. and Lake, P.S. (2010) Using
biological information to support proactive strategies for managing freshwater fish during drought.
Marine and Freshwater Research 61, 379-387.
Dennis, B., Ponciano, J.M., Lele, S.R., Taper, M.L., Staples, D.F. (2006) Estimating density
dependence, process noise, and observation error. Ecological Monographs, 76 (3), pp. 323-341.
De Silva, S. S., Nguyen, T. T. T., Abery, N. W. and Amarasinghe, U. S. (2006) An evaluation of
the role and impacts of alien finfish in Asian inland aquaculture. Aquaculture Research 37, 1-17.
Elith, J., Leathwick, J.R. and Hastie, T. (2008) A working guide to boosted regression trees.
Journal of Animal Ecology 77, 802-813.
Elith, J. and Leathwick, J.R. (2009) Species distribution models: ecological explanation and
prediction across space and time. Annual Review of Ecology, Evolution, and Systematics 40, 677697.
Elkington, S. (2004) Eradication of Gambusia (Gambusia affinis) and koi carp (Cyprinus carpio)
from the Nelson/Marlborough Conservancy. Department of Conservation Te Papa Atewhai,
Nelson.
125
Native fish recovery following Eastern Gambusia removal
Elkington, S. and Maley, R. (2005) Eradication of Gambusia (Gambusia affinis), koi carp
(Cyprinus carpio), and rudd (Scardinius erythrophthalmus) from the Nelson/Marlborough
Conservancy. Department of Conservation Te Papa Atewhai, Nelson.
Fairfax, A., Fensham, R., Wager, R., Brooks, S., Webb, A. and Unmack, P. (2007) Recovery of
the red-finned blue-eye: an endangered fish from springs of the Great Artesian Basin. Wildlife
Research 34, 156-166.
Faragher, R. A. and Lintermans, M. (1997) Alien fish species from the New South Wales River
Survey. In: Fish and Rivers in Stress: The NSW Rivers Survey. (Harris, J. H. and Gehrke, P. C.
Eds.), pp. 201-223. NSW Fisheries Office of Conservation and the CRC for Freshwater Ecology.
Cronulla.
Fausch, K. D. (2007) Introduction, establishment and effects of non-native salmonids: considering
the risk of rainbow trout invasion in the United Kingdom. Journal of Fish Biology 71, 1-32.
Fey, D.P., and Hare, J.A. (2005) Length correction for larval and early-juvenile Atlantic menhaden
(Brevoortia tyrannus) after preservation in alcohol. Fishery Bulletin 103, 725-727.
Fletcher, A. R. (1986) Effects of introduced fish in Australia. In Limnology in Australia (De
Deckker, P. and Williams, W. D., eds.). Melbourne and the Netherlands: CSIRO/Dr W. Junk
Publishers.
Freeman, R. (2007) Tamar Estuary, Gambusia Eradication. Internal Report, Inland Fisheries
Service, Hobart.
Froese, R. (2006) Cube law, condition factor and weight-length relationships: history, metaanalysis and recommendations. Journal of applied ichthyology 22, 241-253.
Froese,
R.
and
Pauly,
D.
http://www.fishbase.org/search.php
Eds.
(2007)
Fishbase
Version
07/2008.
Galat, D. L. and Robertson, B. (1992) Response of endangered Poeciliopsis occidentalis
sonoriensis in the Rio Yaqui drainage, Arizona, to introduced Gambusia affinis. Environmental
Biology of Fishes 33, 249-264.
Gall, G. A. (1987) Inbreeding. In Population Genetics and Fisheries Management (Ryman N. and
Utter, F. Eds.), pp. 47-87. University of Washington Press, Seattle.
García-Berthou, E., C. Alcaraz, Q. Pou-Rovira, L. Zamora, G. Coenders, and C. Feo (2005)
Introduction pathways and establishment rates of invasive aquatic species in Europe. Canadian
Journal of Fisheries and Aquatic Sciences 62, 453-463.
Geddes, M. C. and Puckridge, J. T. (1989) Survival and growth of larval and juvenile native fish.
The importance of the floodplain. In: Proceedings of the Native Fish Management Workshop
(Lawrence, B. Ed.), pp. 101-115. Murray Darling Basin Commission, Canberra.
Gee, J.H. and Gee, P.A. (1991) Reactions of gobioid fishes to hypoxia: buoyancy control and
aquatic surface respiration. Copeia 1, 17-28.
Gehrke, P. C. (1992) Diel abundance, migration and feeding of fish larvae (Eleotridae) in a
floodplain billabong. Journal of Fish Biology 40, 695-707.
Gherardi, F. (2007) Biological invasions in inland waters: an overview. In: Biological invaders in
inland waters: profiles, distribution and threats (Gherardi, F. Ed.), pp. 3-25. Springer, Dordrecht,
The Netherlands.
126
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Gillespie, G. and Hero, J. M. (1999) Potential impact of introduced fish and fish translocations on
Australian amphibians. In: Declines and Disappearances of Australian Frogs. (Campbell, A. Ed.),
pp. 131-144. Environment Australia, Canberra.
Gozlan, R. E. (2008) Introduction of non-native freshwater fish: is it all bad? Fish and Fisheries 9,
106-115.
Gozlan, R. E., St-Hilaire, S., Feist, S. W., Martin, P. and Kent, M. L. (2005) Disease threat to
European fish. Nature 435, 1046.
Gozlan, R.E., Britton, J.R., Cowx, I., and Copp, G.H. (2010) Current knowledge on non-native
freshwater fish introductions. Journal of Fish Biology 76, 751-786.
Hargrave, C. W. (2006) A test of three alternative pathways for consumer regulation of primary
productivity. Oecologia, 149, 223-232.
Harris, J. H. and Gehrke, P. C. (1994) Modelling the relationship between streamflow and
population recruitment to manage freshwater fisheries. Agricultural Systems and Information
Technology 6, 28-30.
Hayes, D.B., Ferreri, C.P., and Taylor, W.W. (1996) Active fish capture methods. In Fisheries
Techniques. Edited by B.R.a.W. Murphy, D.W. American Fisheries Society, Bethesda, Marysland,
USA. pp. pp 193-220.
Hillyard, K.A., Smith, B.B., Conallin, A.J., and Gillanders, B.M. (2010) Optimising exclusion
screens to control exotic carp in an Australian lowland river. 61, 418-429.
Howe, E., Howe, C., R., L. and Burchett, M. (1997) Impact of the introduced poeciliid Gambusia
holbrooki (Girard, 1859) on the growth and reproduction of Pseudomugil signifer (Kner, 1865) in
Australia. Marine and Freshwater Research 48, 425-434.
Humphries, P., Serafini, L. and King, A. J. (2002) River regulation and fish larvae: variations
through space and time. Freshwater Biology 47, 1307-1331.
Hurlbert, S. H., Zedler, J. and Fairbanks, D. (1972) Ecosystem alteration by Mosquitofish
(Gambusia affinis) predation. Science 175, 639-641.
Irons, K. S., Sass, G. G., McClelland, M. A. and Stafford, J. D. (2007) Reduced condition factor of
two native fish species coincident with invasion of non-native Asian carps in the Illinois River,
U.S.A. Is this evidence for competition and reduced fitness? Journal of Fish Biology 71, 258-273.
Ivantsoff, W. and Aarn (1999) Detection of predation on Australian native fishes by Gambusia
holbrooki. Marine and Freshwater Research 50, 467-478.
Ives, A.R., Dennis, B., Cottingham, K.L., Carpenter, S.R. (2003) Estimating community stability
and ecological interactions from time-series data. Ecological Monographs, 73, 301-330.
Jackson, D.A., Peres-Neto, P.R. and Olden, J.D. (2001) What controls who is where in freshwater
fish communities - the roles of biotic, abiotic, and spatial factors. Canadian Journal of Fisheries
and Aquatic Sciences 58, 157-170.
Jackson, J. E., Raadik, T. A., Lintermans, M. and Hammer, M. (2004) Alien salmonids in
Australia: impediments to effective impact management, and future directions. New Zealand
Journal of Marine and Freshwater Research 38, 447-455.
Jones, M., and Stuart, I.G. (2004) Impact of flow regulation structures on fish in the BarmahMillewa forest. Final report to Barmah-Millewa Forum, Arthur Rylah Institute for Environmental
Research, Heidelberg.
127
Native fish recovery following Eastern Gambusia removal
Karolak, S. (2006) Alien Fish in the Murray–Darling Basin. MDBC publication No. 03/06.
Murray–Darling Basin Commission, Canberra.
Kelleway, J., Mazumder, D., Wilson, G.G., Saintilan, N., Knowles, L., Iles, J. and Kobayashi, T.
(2010) Trophic structure of benthic resources and consumers varies across a regulated floodplain
wetland. Marine and Freshwater Research 61, 430-440.
Keller, K. and Brown, C. (2008) Behavioural interactions between the introduced plague minnow
Gambusia holbrooki and the vulnerable native Australian ornate rainbowfish Rhadinocentrus
ornatus, under experimental conditions. Journal of Fish Biology 73, 1714-1729.
Kennard, M. J., Arthington, A. H., Pusey, B. J. and Harch, B. D. (2005) Are alien fish a reliable
indicator of river health? Freshwater Biology 50, 174-193.
Kerezsy, A. (2009) Gambusia control in spring wetlands.Final report South Australian Arid Lands
Natural Resources Management Board.
Kilsby, N.N. and Walker, K.F. (in press) Behaviour of two small pelagic and demersal fish species
in diverse hydraulic environments. River Research and Applications.
Kincaid, H. L. (1976) Inbreeding depression in rainbow trout. Transactions of the American
Fisheries Society 105, 273-280.
King, A. J. (2005) Ontogenetic dietary shifts of fishes in an Australian floodplain river. Marine
and Freshwater Research 56, 215-225.
King, A. J., Robertson, A. I. and Healey, M. R. (1997) Experimental manipulations of the biomass
of introduced carp (Cyprinus carpio) in billabongs. I. Impacts on water-column properties. Marine
and Freshwater Research 48, 435-443.
King, A.J., and Crook, D.A. (2002) Evaluation of a sweep net electrofishing method for the
collection of small fish and shrimp in lotic freshwater environments. Hydrobiologia 472, 223-233.
King, A.J., Tonkin, Z. and Mahoney, J. (2007) Assessing the effectiveness of environmental flows
on fish recruitment in Barmah-Millewa Forest. Final report to Murray–Darling Basin Commission,
Canberra. Project No. BMF 2004.09. Prepared by the Arthur Rylah Institute for Environmental
Research, Department of Sustainability and Environment, Heidelberg.
King, A.J., Ward, K.A., O'Connor, P., Green, D., Tonkin, Z., and Mahoney, J. (2010) Adaptive
management of an environmental watering event to enhance native fish spawning and recruitment.
Freshwater Biology 55, 17-31.
King, S. and Warburton, K. (2007) The environmental preferences of three species of Australian
freshwater fish in relation to the effects of riparian degradation. Environmental Biology of Fishes
78, 307-316.
Knight, J. (1999) Density dependent interference competition between the Australian native fish
Pseudomugil signifer (Kner, 1865) and the introduced Poeciliid Gambusia holbrooki (Girard,
1859). Integrated project prepared as partial fulfilment of the requirements of the B. App. Sc.
(Fisheries and Aquaculture Management). Southern Cross University, Lismore.
Koehn, J. (2007) Managing pest fish in the Murray–Darling Basin: lessons learnt from previous
invaders. Emerging Issues in Alien Fish Management in the Murray–Darling Basin: Statement,
recommendation and supporting papers. Proceedings of a workshop held in Brisbane QLD, 30-31
May 2006, Brisbane, pp. 54-57.
Koehn, J. D. and MacKenzie, R. F. (2004) Priority management actions for alien freshwater fish
species in Australia. New Zealand Journal of Marine and Freshwater Research 38, 457-472.
128
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Koehn, J. D. and O'Connor, W. G. (1990) Biological Information for Management of Native
Freshwater Fish in Victoria. Victorian Government Printing Office, Melbourne.
Komak, S. and Crossland, M. R. (2000) An assessment of the introduced mosquitofish (Gambusia
affinis holbrooki) as a predator of eggs, hatchlings and tadpoles of native and non-native anurans.
Wildlife Research 27, 185-189.
Koster, W. M. (1997) A study of the interactions between dwarf galaxias (Galaxiella pusilla),
southern pygmy-perch (Nannoperca australis) and eastern gambusia (Gambusia holbrooki). B.Sc.
Honours thesis, Deakin University, Clayton.
Kuiter, R. H., Humphries, P. and Arthington, A. H. (1996) Family Nannoperca: Pygmy-perches.
In: Freshwater Fishes of South-Eastern Australia (McDowall, R. M., Ed.), pp. 168-175. Reed
Books, Sydney.
Laha, M. and Mattingly, H. T. (2007) Ex situ evaluation of impacts of invasive mosquitofish on
the imperilled Barrens topminnow. Environmental Biology of Fishes 78, 1-11.
Lake, J. S. (1967) Rearing experiments with five species of Australian freshwater fishes. II.
Morphogenesis and ontogeny. Australian Journal of Marine and Freshwater Research 18, 155-173.
Larson, H. K. and Hoese, D. F. (1996) Gudgeons. In Freshwater Fishes of South-Eastern Australia
(McDowall, R. M., ed.), pp. 200-228. Reed Books, Sydney.
Leathwick, J.R., Elith, J., Francis, M.P., Hastie, T. and Taylor, P. (2006) Variation in demersal fish
species richness in the oceans surrounding New Zealand: an analysis using boosted regression
trees. Marine Ecology Progress Series 321, 267-281.
Leathwick, J.R., Elith, J., Chadderton, L., Rowe, D. and Hastie, T. (2008) Dispersal, disturbance,
and the contrasting biogeographies of New Zealand’s diadromous and nondiadromous fish species.
Journal of Biogeography 35, 1481-1497.
Leprieur, F., Beauchard, O., Blanchet, S., Oberdorff, T. and Brosse, S. (2008) Fish Invasions in the
World’s River Systems: When Natural Processes Are Blurred by Human Activities. PLoS Biology
6, 0404-0410.
Ling, N. (2004) Gambusia in New Zealand: really bad or just misunderstood? New Zealand
Journal of Marine and Freshwater Research 38, 473-480.
Lintermans, M. (2004) Human-assisted dispersal of alien freshwater fish in Australia. New
Zealand Journal of Marine and Freshwater Research 38, 481-501.
Lintermans, M. (2007) Fishes of the Murray–Darling Basin: an introductory guide: Murray–
Darling Basin Commission, Canberra.
Llewellyn, L. C. (1973) Spawning, development and temperature tolerance of the spangled perch
Madigania unicolor (Günther) from inland waters in Australia. Australian Journal of Marine and
Freshwater Research 24, 73-94.
Llewellyn, L. C. (1974) Spawning, development and distribution of the southern pygmy-perch
Nannoperca australis australis Günther from inland waters in Eastern Australia. Australian
Journal of Marine and Freshwater Research 25, 121-129.
Lloyd, L. N. (1987) Ecology and distribution of the small native fish of the lower River Murray,
South Australia and their interactions with exotic mosquitofish, Gambusia affinis holbrooki
(Girard). MSc thesis, University of Adelaide, Adelaide.
129
Native fish recovery following Eastern Gambusia removal
Lloyd, L. N. (1990) Ecological interactions of Gambusia holbrooki with Australian native fishes.
In Introduced and translocated fishes and their ecological effect (Pollard, D. A., Ed.), pp. 94-97.
Australian Government Publishing Service, Canberra.
Lloyd, L. N., Arthington, A. H. and Milton, D. A. (1986) The mosquitofish – a valuable control
agent or a pest? In: The Ecology of Exotic Animals and Plants; some Australian Case Histories
(Kitching, R. L. Ed.), John Wiley and Sons, Brisbane.
Lloyd, L. and Walker, K. F. (1986) Distribution and conservation status of small freshwater fish in
the River Murray, South Australia. Transactions of the Royal Society of South Australia 110, 4957.
Lodge, D. M. and Schrader-Frechette, K. (2003) Nonindigenous species: ecological explanation,
environmental ethics, and public policy. Conservation Biology 17, 31-37.
Lowe, S. J., Browne, M. and Boudjelas, S. (2000) 100 of the world’s worst invasive alien species.
IUCN/SSC Invasive Species Specialist Group (ISSG), Auckland, New Zealand.
Lukies, S. (2004) Diet and foraging interactions between the native crimsonspotted rainbowfish
(Melanotaenia fluviatilis) and the introduced mosquitofish (Gambusia holbrooki) in the Broken
River of northeast Victoria. Unpublished Thesis, Department of Environmental Management and
Ecology. La Trobe University, Wodonga.
Lunn, D.J., Thomas, A., Best, N., Spiegelhalter, D. (2000) WinBUGS - A Bayesian modelling
framework: Concepts, structure, and extensibility. Statistics and Computing 10, 325-337.
Lydeard, C. and Belk, M. C. (1993) Management of indigenous fish species impacted by
introduced mosquitofish: An experimental approach. The Southwestern Naturalist 38, 370-373.
Lyon, J., Stuart, I., Ramsey, D., and O'Mahoney, J. (2010) The effect of water level on lateral
movements of fish between river and off-channel habitats and implications for management.
Marine and Freshwater Research 61, 271-278.
Macdonald, J.I. and Crook, D.A. (2006) Contribution of the Barmah-Millewa floodplain lakes to
carp populations in the mid-Murray River, 2005⁄06. Report to the Murray–Darling Basin
Commission. Arthur Rylah Institute for Environmental Research, Heidelberg, Victoria.
Macdonald, J., and Tonkin, Z. (2008) A review of the impacts of Eastern Gambusia on native
fishes of the Murray–Darling Basin. Arthur Rylah Institute for Environmental Research.
Department of Sustainability and Environment. Murray–Darling Basin Authority Publication No.
38/09
Maezono, Y. and Miyashita, T. (2004). Impact of exotic fish removal on native communities in
farm ponds. Ecological Research 19, 263-267.
Maglio, V. J. and Rosen, D. E. (1969) Changing preferences for substrate colour by reproductively
active mosquito fish Gambusia affinis (Baird and Girard) (Poeciliidae: Atherinoformes). American
Museum Novitates 2397, 1-39.
Magoulick, D.D. and Kobza, R.M. (2003) The role of refugia for fishes during drought: a review
and synthesis. Freshwater Biology 48, 1186-1198.
Marchetti, M. P., Moyle, P. B. and Levine, R. (2004) Invasive species profiling? Exploring the
characteristics of non-native fishes across invasion stages in California. Freshwater Biology 49,
646-661.
130
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Margaritora, F. G., Ferrara, O. and Vagaggini, D. (2001) Predatory impact of the mosquitofish
(Gambusia holbrooki Girard) on zooplanktonic populations in a pond at Tenuta di Castelporziano
(Rome, Central Italy). Journal of Limnology 60, 189-193.
Mathews, W.J. and Marsh-Mathews, E. (2003) Effects of drought on fish across axes of space,
time and ecological complexity. Freshwater Biology 48, 1232-1253.
Maynard, D., Gaston, T. and Henderson, A. (2008) Assessment of removal techniques for
Gambusia holbrooki, to assist in the eradication of an introduced pest from ecologically sensitive
environments. Milestone report submitted to the Tamar Region Natural Resource Management
Strategy Reference Group Inc. by the Australian Maritime College, University of Tasmania,
Hobart.
McDowall, R. M. (1996) Family Poeciliidae: Livebearers. In Freshwater Fishes of South-Eastern
Australia (McDowall, R. M., ed.), pp. 116-122. Reed Books, Sydney.
McDowall, R. M. and Fulton, W. (1996) Galaxiids. In Freshwater Fishes of South-Eastern
Australia (McDowall, R. M., ed.), pp. 52-77. Reed Books, Sydney.
McKay, R. J. (1984) Introductions of exotic fishes in Australia. In Distribution, Biology and
Management of Exotic Fishes (Courtenay, W. R. and Stauffer, J. R., eds.). John Hopkins
University Press, Baltimore.
McKay, S., Clunie, P., Gillespie, G., Raadik, T., Saddlier, S., O'Brien, T., Ryan, T. and Aland, G.
(2001) Predation by Gambusia holbrooki: A review of the literature. Arthur Rylah Institute for
Environmental Research, Department of Natural Resources and Environment, Heidelberg.
McKinnon, L.J. (1997) Monitoring of Fish Aspects of the Flooding of Barmah Forest. Final report
to Murray–Darling Basin Commission. Marine and Freshwater Resources Institute, Queenscliff,
Australia.
McLeod, R. (2004) Counting the cost: Impact of Invasive Animals in Australia 2004. Cooperative
Research Centre for Pest Animal Control, Canberra.
McMaster, D., and Bond, N. (2008) A field and experimental study on the tolerances of fish to
Eucalyptus camaldulensis leachate and low dissolved oxygen concentrations. Marine and
Freshwater Research 59, 177-185.
McNeil, D. G. (2004) Ecophysiology and behavior of Ovens River floodplain fish: Hypoxia
tolerance and the role of the physicochemical environment in structuring Australian billabong fish
communities. Unpublished PhD Thesis. Cooperative Research Centre for Freshwater Ecology,
LaTrobe University, Wodonga.
McNeil, D. G. and Closs, G. P. (2007) Behavioural responses of a south-east Australian floodplain
fish community to gradual hypoxia. Freshwater Biology 52, 412-420.
MDBC (2004) Native Fish Strategy for the Murray–Darling Basin 2003-2013. Murray–Darling
Basin Commission Publication No. 25/04. Canberra.
MDBC (2008) Murray–Darling Basin Rivers: Ecosystem Health Check, 2004-2007. A summary
report based on the independent Sustainable Rivers Audit Group’s SRA Report 1: A report on the
ecological report of rivers in the Murray–Darling Basin, 2004-2007, submitted to the Murray–
Darling Basin Ministerial Council in May 2008.
Meffe, G. K. and Snelson, F. F. Jnr. (1989) An ecological overview of poeciliid fishes. In Ecology
and evolution of livebearing fishes (Poecilidae) (Meffe G. K. and Snelson, F. F. Jnr. Eds.). pp. 1331, Prentice-Hall, Inc., New Jersey.
131
Native fish recovery following Eastern Gambusia removal
Meffe, G. K., Hendrickson, D. A. and Minckley, W. L. (1983) Factors resulting in decline of the
endangered Sonoran topminnow Poeciliopsis occidentalis (Atheriniformes: Poeciliidae) in the
United States. Biological Conservation 25, 135-159.
Merrick, J. R. and Schmida, G. E. (1984) Australian freshwater fishes: biology and management.
J. Merrick, North Ryde, New South Wales.
Mills, M. D., Radar, R. B. and Belk, M. C. (2004) Complex interactions between native and
invasive fish: the simultaneous effects of multiple negative interactions. Oecologia 141, 713-721.
Milton, D. A. and Arthington, A. H. (1983) Reproductive biology of Gambusia affinis holbrooki
Baird and Girard, Xiphophorus helleri (Gunther) and X. maculatus (Heckel) (Pisces; Poeciliidae)
in Queensland, Australia. Journal of Fish Biology 23, 23-41.
Morgan, D. L., Gill, H. S., Maddern, M. G. and Beatty, S. J. (2004) Distribution and impacts of
introduced freshwater fishes in Western Australia. New Zealand Journal of Marine and Freshwater
Research 38, 511-523.
Morris, S.A., Pollard, D.A., Gehrke, P.C. and Pogonoski, J.J. (2001) Threatened and potentially
threatened freshwater fishes of coastal New South Wales and the Murray–Darling Basin. NSW
Fisheries Final Report Series No. 33.
Morrongiello, J., Elith, J. and Crook, D. (2006) Impacts of drought on fish in Victorian rivers and
streams. Arthur Rylah Institute for Environmental Research, Heidelberg, Victoria.
Morrongiello, J.R., Bond, N.R., Crook, D.A. and Wong, B.B.M. (in press). Eucalyptus leachate
inhibits reproduction in a freshwater fish. Freshwater Biology.
Moyle, P. B. and Light, T. (1996) Biological invasions of fresh water: Empirical rules and
assembly theory. Biological Conservation 78, 149-161.
Murphy, E. C., Kay Clapperton, B., Bradfield, P. M. F. and Speed, H. J. (1998) Effects of ratpoisoning operations on abundance and diet of mustelids in New Zealand podocarp forests. New
Zealand Journal of Zoology 25, 315-328.
Myers, G. S. (1965) Gambusia, the fish destroyer. Australian Zoologist 13, 102.
NSW (2003) New South Wales threat abatement plan. Predation by Gambusia holbrooki - the
plague minnow. NSW, National Parks and Wildlife Service. Hurstville, NSW.
NSW (2010) NSW control plan for the noxious fish carp Cyprinus carpio. Aquatic Biosecurity &
Risk Managment Unit, Industry and Investment. Orange, NSW.
Parenti, L. R. and Rauchenberger, M. (1989) Systematic overview of the Poeciliines. In Ecology
and Evolution of Livebearing Fishes (Poeciliidae) (Meffe, G. K. and Snelson, F. F. Eds.), pp. 3-31.
Englewood Cliffs, Prentice Hall, New Jersey.
Pen, L. J. and Potter, I. C. (1991) Reproduction, growth and diet of Gambusia holbrooki (Girard)
in a temperate Australian river. Aquatic Conservation: Marine and Freshwater Ecosystems 1, 159172.
Pen, L. J., Potter, I. C. and Calver, M. C. (1993) Comparisons of the food niches of three native
and two introduced fish species in an Australian river. Environmental Biology of Fishes 36, 167182.
Peterson, D. P. and Fausch, K. D. (2003) Upstream movement nonnative brook trout (Salvelinus
fontinalis) promotes invasion of native cutthroat trout (Oncorhynchus clarki) habitat. Canadian
Journal of Fisheries and Aquatic Sciences 60, 1502-1516.
132
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Pimental, D., Lach, L., Zuniga, R. and Morrison, D. (2000) Environmental and economic costs of
nonindigenous species in the United States. BioScience 50, 53-65.
Pimentel, D., Zuniga, R. and Morrison, D. (2005) Update on the environmental and economic
costs associated with alien-invasive species in the United States, Ecological Economics 52, 273288.
Pitman, K. S. and Tinkler, P. (2007) Recovery planning survey for dwarf galaxias (Galaxias
pusilla) in the West Gippsland region, Victoria. Arthur Rylah Institute for Environmental
Research, Department of Sustainability and Environment, Heidelberg.
Poff, N.E., and Allan, J.D. (1995) Functional organisation of stream fish assemblages in relation to
hydrological variability. Ecology 76: 606-627.
Potter, I. C. (1996a) Family Geotriidae: pouched lampreys. In: Freshwater Fishes of South-Eastern
Australia (McDowall, R. M., Ed.), pp. 36-38. Reed Books, Sydney.
Potter, I. C. (1996b) Family Mordaciidae: shortheaded lampreys. In: Freshwater Fishes of SouthEastern Australia (McDowall, R. M., Ed.), pp. 32-35. Reed Books. Sydney.
Potter, I. C., Hilliard, R. W., Bradley, J. S. and McKay, R. J. (1986) The influence of
environmental variables on the density of larval lampreys in different seasons. Oecologia 70, 433440.
Puckridge, J. T. and Walker, K. F. (1990) Reproductive biology and larval development of a
gizzard shad, Nematalosa erebi (Gunther) (Dorosomatinae: Teleostei), in the River Murray, South
Australia. Australian Journal of Marine and Freshwater Research 41, 695-712.
Pyke, G. H. (2005) A review of the biology of Gambusia affinis and G. holbrooki. Reviews in Fish
Biology and Fisheries 15, 339-365.
Pyke, G. H. (2008) plague minnow or mosquito fish? A review of the biology and impacts of
introduced Gambusia species. Annual Review of Ecology and Systematics 39, 171-191.
Ramsey, D.S.L., and Wilson, J.C. (2000) Towards ecologically based baiting strategies for rodents
in agricultural systems. International Biodeterioration & Biodegradation 45, 183-197.
Raynor, T.S., and Creese, R.G. (2006) A review of rotenone use for the control of non-indigenous
fish in Australian fresh waters, and an attempted eradication of the noxious fish, Phalloceros
caudimaculatus. New Zealand Journal of Marine and Freshwater Research 40, 477-486.
Rehwinkel, R., and Sharpe, C. (2009) Gunbower Forest Fish Monitoring Surveys. Report prepared
for North Central Catchment Management Authority, January, 2009.
Richardson, A., Meredith, S., Conallin, A., and Sharpe, C. (2005) Status of the Gunbower Island
fish community, June 2005 - Including recommendations for future monitoring. MDFRC
Consultancy Report for the North Central Catchment Management Authority. Murray–Darling
Freshwater Research Centre, Wodonga.
Rincón, P. A., Correas, A. M., Morcillo, F., Risueño, P. and Lobón-Cerviá, J. (2002) Interaction
between the introduced eastern mosquitofish and two autochthonous Spanish toothcarps. Journal
of Fish Biology 61, 1560-1585.
Roberts, J., Chick, A., Oswald, L. and Thompson, P. (1995) Effect of carp, Cyprinus carpio L. an
exotic benthivorous fish, on aquatic plants and water quality in experimental ponds. Marine and
Freshwater Research 46, 1171-1180.
133
Native fish recovery following Eastern Gambusia removal
Rowe, D. K., Moore, A., Giorgetti, A., Maclean, C., Grace, P., Wadhwa, S. and Cooke, J. (2008)
Review of the impacts of gambusia, redfin perch, tench, roach, yellowfin goby and streaked goby
in Australia. Draft report prepared for the Australian Government Department of the Environment,
Water, Heritage and the Arts, Canberra.
Rowe, D. K., Smith, J. P. and Baker, C. (2007) Agonistic interactions between Gambusia affinis
and Galaxias maculatus: implications for whitebait fisheries in New Zealand rivers. Journal of
Applied Ichthyology 23, 668–674.
Rowland, S. J. (1992) Diet and feeding of Murray Cod (Maccullochella peelii) larvae. Proceedings
of the Linnean Society of New South Wales 113, 193-201.
Saccheri, I., Kuussaari, M., Kankare, M., Vikman, P., Fortelius W. and Hanski, I. (1998)
Inbreeding and extinction in a butterfly metapopulation. Nature 392, 491-493.
Sala, O. E., Chapin, F. S. III., Armesto, J. J., Berlow, E., Bloomfield, J., Dirzo, R., HuberSanwald, E., Huenneke, L. F., Jackson, R. B., Kinzig, A., Leemans, R., Lodge, D. M., Mooney, H.
A., Oesterheld, M., Poff, N. L., Sykes, M. T., Walker, B. H., Walker, M. and Wall, D. H. (2000)
Global biodiversity scenarios for the year 2100. Science 287, 1770-1774.
Scurr, G. (2007) Gambusia control update. Tasmanian Conservationist 311, 10-11.
Sivakumaran, K.P., Brown, P. Stoessel, D. and Giles, A. (2003) Maturation and reproductive
biology of female wild carp, Cyprinus carpio, in Victoria, Australia. Environmental Biology of
Fishes 68, 321-332.
Smith, B.B. and Walker K.F. (2004) Reproduction of Common Carp in South Australia, shown by
young-of-the-year samples, gonadosomatic index and the histological staging of ovaries.
Transactions of the Royal Society of South Australia 128, 249-257.
SRA (2007) Victorian Sustainable Rivers Audit Fish data summary. Arthur Rylah Institute for
Environmental Research, Department of Sustainability and Environment, Heidelberg.
Stoffels, R. J. and Humphries, P. (2003) Ontogenetic variation in the diurnal food and habitat
associations of an endemic and an exotic fish in floodplain ponds: consequences for niche
partitioning. Environmental Biology of Fishes 66, 293-305.
Stuart, I. G. and Jones, M. (2002) Ecology and management of Common Carp in the BarmahMillewa Forest. Final report to Agriculture, Fisheries and Forestry Australia. Arthur Rylah
Institute for Environmental Research. Department of Sustainability and Environment, Heidelberg.
Stuart, I.G., and Jones, M. (2006) Movement of Common Carp, Cyprinus carpio, in a regulated
lowland Australian river: implications for management. Fisheries Management and Ecology 2006,
231-219.
Stuart, I.G., Williams, A., McKenzie, J., and Holt, T. (2006) Managing a migratory pest species: a
selective trap for Common Carp. North American Journal of Fisheries Management 26, 888-893.
Swanson, C., Cech, J. J. J. and Piedrahita, R. H. (1996) Mosquitofish: Biology, Culture, and Use in
Mosquito Control. California, U.S.A.: Mosquito and Vector Control Association of California;
The University of California Mosquito Research Program. California.
Tonkin, Z. and Rourke, M. (2008) Barmah-Millewa fish condition monitoring - 2008 annual
summary and refuge habitat report submitted to the Murray–Darling Basin Commission. Arthur
Rylah Institute for Environmental Research (Department of Sustainability and Environment) and
NSW Department of Primary Industries.
134
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Tonkin, Z., King, A.J., and Mahoney, J. (2008) Effects of flooding on recruitment and dispersal of
the Southern Pygmy-perch (Nannoperca australis) at a Murray River floodplain wetland.
Ecological Management and Restoration 9, 196-201.
Tonkin, Z., Humphries, P. and Pridmore, P. A. (2006) Ontogeny of feeding in two native and one
alien fish species from the Murray–Darling Basin, Australia. Environmental Biology of Fishes 76,
303-315.
Townsend, C.R. (2003) Individual, population, community and ecosystem consequences of a fish
invader in New Zealand streams. Conservation Biology 17, 38-47.
Treadwell, S. and Hardwick, R. (2004) Review of Habitat Associations of Native Fish in the
Murray–Darling Basin. Murray–Darling Basin Commission project R2105. Sinclair Knight Merz,
Malvern.
Unmack, P. J. (2001) Biography of Australian freshwater fishes. Journal of Biogeography 28,
1053-1089.
Vitousek, P. M., Mooney, H. A., Lubchenco, J. and Melillo, J. M. (1997) Human domination of
earth's ecosystems. Science 277, 494-499.
Warburton, K., Retif, S. and Hume, D. (1998) Generalists as sequential specialists: diets and prey
switching in juvenile silver perch. Environmental Biology of Fishes 51, 445-454.
Webb, C. and Joss, J. (1997) Does predation by the fish Gambusia holbrooki (Atheriniformes:
Poeciliidae) contribute to declining frog populations. Australian Zoologist 30, 316-324.
Wedderburn, S. D., Walker, K. F. and Zampatti, B. P. (2007) Habitat separation of
Craterocephalus (Artherinidae) species and populations in off-channel areas of the lower River
Murray, Australia. Ecology of Freshwater Fish 16, 442-449.
Weigel, D. E., Peterson, J. T. and Spruell, P. (2002) A model using phenotypic characteristics to
detect introgressive hybridisation in wild Westslope cutthroat trout and rainbow trout.
Transactions of the American Fisheries Society 131, 389-403.
Wells, A. (2007) Community understanding and attitudes to alien fish. Emerging Issues in Alien
Fish Management in the Murray–Darling Basin: Statement, recommendation and supporting
papers. Proceedings of a workshop held in Brisbane QLD, 30-31 May 2006, Brisbane, pp. 69-72.
West, P., Brown, A. and Hall, K. (2007) Review of alien fish monitoring techniques, indicators
and protocols: implications for national monitoring of Australia’s inland river systems. Invasive
Animals Cooperative Research Centre, prepared for the National Land and Water Resources
Audit, Canberra.
Williamson, M. (1996) Biological Invasions. Chapman & Hall, London.
Willis, K. and Ling, N. (2000) Sensitivities of mosquitofish and black mudfish to a piscicide:
could rotenone be used to control mosquitofish in New Zealand wetlands? New Zealand Journal of
Zoology 27, 85-91.
Wilson, G. (2006) Impact of invasive exotic fishes on wetland ecosystems in the Murray–Darling
Basin. In: Native fish and wetlands in the Murray–Darling Basin: Action Plan, knowledge gaps
and supporting papers. (Phillip, B. Ed.). Murray–Darling Basin Commission, Canberra.
Winemiller, K.O., Tarim, S., Shoreman, D., and Cotner, J.B. (2000) Fish assemblage structure in
relation to environmental variation among Brazos river oxbow lakes. Transactions of the American
Fisheries Society 129, 451-468.
135
Native fish recovery following Eastern Gambusia removal
Zavaleta, E. S., Hobbs, R. J. and Mooney, H. A. (2001) Viewing invasive species removal in a
whole-ecosystem context. Trends in Ecology and Evolution 16, 454-459.
136
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Appendices
Appendix 1: Eastern Gambusia marking trial
An assessment of short term mortality of Eastern Gambusia Gambusia holbrooki after being
subject to an osmotic induction marking technique
Abstract
The use of chemical markers has become a common fish tagging technique for mark-recapture
studies, particularly in situations that involve either marking large numbers of small individuals
and/or where long-term marks are required. A recent project utilised an osmotic induction marking
protocol using calcein (adopted from Crook et al. (2009) to mark Eastern Gambusia Gambusia
holbrooki in the wild. However, preliminary observations suggested that the marking protocol may
have been causing increased mortality, which was not desirable for subsequent population
measures derived from recapture data. This study was therefore designed to assess whether the
marking protocol has any impact on short term mortality and growth.
The experiment indicated a significant increase in rates of short-term mortality of adult Eastern
Gambusia following the osmotic induction marking protocol. This mortality stabilised 3 – 6 days
following marking, with mortality of marked fish approximately 25% higher for female fish and
50% higher for male fish, than fish that were not subject to the marking procedure. There was no
influence of the marking protocol on overall growth of male or female fish. The results of the
study have demonstrated that the osmotic induction marking protocol will increase short term
mortality of Eastern Gambusia, and that any mark-recapture data derived from field based marking
using this protocol must account for these mortality rates to improve estimates of population size
and capture rates. This study highlights the need for developing species specific marking protocols
before being applied to field or hatchery based studies.
Introduction
Mark-recapture studies provide vital information to key aspects of fisheries and ecological
research including population, mortality and methodology assessments. The accuracy of these
assessments rely on sound marking techniques which ideally ensure marks (or tags) are retained by
individuals for the duration of the required detection period, do not increase individual mortality
(or if they do, the rate of mortality must be quantifiable) and do not have any significant effect on
the normal biological functioning or behaviour of the fish (either negative or positive). For this
reason, chemical markers such as calcein and oxytetracycline have become preferable for marking
large numbers of fish to traditional tagging methods (such as floy tags, pit tags etc), given that (i)
they can be applied to fish of all ages and developmental stages; (ii) they are not visible under
normal sunlight (so should not influence behaviour or predation) and; (iii) they take less time to
apply to large numbers of fish and so are more cost effective (e.g. Leips et al. 2001; Crook et al.
2009).
A recent project examining Eastern Gambusia Gambusia holbrooki populations in small wetlands
has utilised a mark-recapture method to assess the level of fish removal, as well as provide an
137
Native fish recovery following Eastern Gambusia removal
estimate of the abundance of Eastern Gambusia in selected sites. The study utilised an osmotic
induction marking protocol developed for use in fish hatcheries producing golden perch
Macquaria ambigua (see Crook et al. 2009), to mark large numbers of gambusia. The technique
uses the fluorescent dye, calcein, which binds to calcified tissue such as otoliths, scales and the
skeleton of the fish (see Mohler 2003; Crook et al. 2009). Marked fish fluoresce in particular zones
on their body when exposed to blue or ultraviolet light allowing external detection in a field
environment without sacrificing the individual (e.g. Bashey 2004; Crook et al 2009). The osmotic
induction technique has the advantage of marking large numbers of individuals in a very short
time, and thus is ideal for use on small bodied species in the field.
The detection period of the mark-recapture component in the study was undertaken for a
maximum of 14 days post-marking (generally less than 10 days). Given this technique has been
shown to produce detectable external marks for months or even years (see Negus and Tureson
2004; Bashey 2004) any influence on mark retention over a period of days is likely to be minimal.
This was confirmed by all individual fish recaptured displaying clear external marking,
particularly on ventrally located bony structures which are not as subject to photodegredation as
dorsal areas. Recapture rates however, were slightly lower than expected and several recaptured
individuals displayed what appeared to be a fungal growth on their caudal fin. This caused concern
that the technique may be influencing mortality in this species, which would have serious
implications for reliably estimating of removal efficiency and population size.
This study was used to assess any short-term influence on growth and mortality of adult Eastern
Gambusia after they have been subject to the osmotic induction marking protocol. This will
ultimately improve the accuracy of the field based mark-recapture population and removal
assessments which use this technique to mark adult fish.
Methods
Approximately 100 adult Eastern Gambusia (> 24 mm total length) were collected from a farm
dam in central Victoria during October 2010 using fine-mesh (1mm) seine netting and transported
to an aquarium facility. The following day, half of the fish received a calcein mark via the osmotic
induction method as described by Crook et al (2009). The technique involves a 5% salt solution
being prepared by dissolving 100 g of commercially available natural salt in 2L of water. A 0.5%
solution of calcein is prepared by adding 10 g of calcein powder (2,4-bis-[N,N0-fdicarbo methylgaminomethyl] fluorescein) to 2L of water. At this concentration, the calcein causes a decrease in
the pH of the solution, and is inturn adjusted back to 7.0 by gradually adding sodium hydroxide
(NaOH) to the solution. Both solutions were aerated by bubbling air into the solutions during the
marking procedures. Fifty Eastern Gambusia were placed in a mesh bottom vessel and immersed
in the salt solution for three minutes. After immersion in the salt solution, fish were rinsed in fresh
water for 5 seconds, and then immersed in the calcein dye solution for five minutes. Fish were then
placed into a 20L aerated container and assessed for signs of stress. Individual fish from both were
then assessed for gender and measured for total length (TL; nearest mm) and randomly assigned to
one of four 10L bucket replicates at a rate of six females and three male fish per replicate.
Additional control groups consisting of fish that were not exposed to the marking procedure were
also assessed for gender, measured for TL and assigned to at the same abundance and sex ratio as
the treatment tanks. This resulted in a total of eight 10 L vessels (four treatment and four controls),
each of which contained six female fish and three male fish. Each vessel had a flow through water
138
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
system (ambient water temperature ranging 17 – 22 °C) and aeration supplied for the duration of
the experiment. The experiment was run for a period of 15 days, where a daily routine occurred of
feeding of blood worms, mortality check and general aquaria maintenance. Any dead fish were
recorded and removed from the experiment. At the completion of the 15th day, all surviving
individuals were measured, assessed for an external mark with a handheld blue light, then
euthanized with anaesthetic solution (Alfaxalone – 40 mg/L for 10 minutes), and then preserved in
95% ethanol.
Mortality data from the experiment at day zero, three, six, nine, 12 and 15, was converted to
percentage of overall mortality and compared between treatment and controls using a repeated
measures Analysis of variance (ANOVA) with treatment, sex, time and their interaction as factors.
Total growth (mm) for all surviving individuals was calculated by using the mean initial and final
standard length (loge[mean SLfinal] - loge[mean SLinitial]). A two-way ANOVA was used to test
whether total growth differed due to treatment (marked or unmarked), sex and their interaction.
Results
At the completion of the experiment, fish mortality was found to be at a relatively high rate in both
treatment (marked) and control (unmarked) vessels. Total mortality for control and treatments
combined was 46 % for males and 44% for females. Some fish that were subject to the marking
procedure developed a fungus-type growth on their caudal fin, which rapidly developed over the
posterior of their body, followed by mortality (within two days). As a result, mortality rates were
significantly higher in fish that had been subject to the marking procedure than those that were not
marked (Figure 1; treatment, F = 21.85; p < 0.001). Mortality rates in marked fish were higher in
males than female fish, however, the differences were not significant (F = 4.07; p = 0.067).
Not surprisingly mortality rates were significantly influenced by the duration of the experiment
(time; F = 17.93; p < 0.001), where mortality was highest across all treatments at the end of the
experiment (Figure 1). Mortality rates of female fish in control and treatments also increased
during the final days of the experiment. Of most relevance was the interaction between duration
and treatment, with male and female fish having significantly higher mortalities 3-6 days
following marking (Figure 1; F = 6.23; p = 0.001). This mortality stabilised after this time, with
mortality of marked fish approximately 25% higher for female fish and 50% higher for male fish
than fish that were not subject to the marking procedure. There was, however, much higher
variability in male mortality (most likely due to the smaller sample size).
There was little variation in overall growth of surviving fish over the 15-day period. Overall,
growth was not significantly influenced by gender, whether it was marked or unmarked, or their
interaction (Figure 2; all p > 0.05). At the completion of the experiment, all fish in marked
treatments displayed clear fluorescent marking on scales and bony surfaces when viewed under a
handheld ultraviolet light (Figure 3).
139
Native fish recovery following Eastern Gambusia removal
100
Male
80
60
Cumulative % mortality
40
20
0
100
Female
80
60
40
20
0
0
3
6
9
12
15
Day
Figure 1. Mean ± SE cumulative percent mortality of fish subject to osmotic induction
marking (treatment; Pink line) and unmarked fish (Control; black line) for both adult males
(top; n = 12 marked and unmarked fish) and females (bottom; n = 24 marked and unmarked
fish) for every third day of the experiment.
140
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
MALE
FEMALE
Growth (mm)
8
6
4
2
0
Unmarked
Marked
Unmarked
Marked
Figure 2. Growth in total length (Mean ± SE) of female and male fish over the 15-day period
from unmarked (control) and marked (treatment) replicates.
Figure 3. Photograph of adult female gambusia 15 days post-marking displaying clear
fluorescent marking under ultraviolet light (viewed through an amber lense).
141
Native fish recovery following Eastern Gambusia removal
Discussion
The results of the experiment indicate a significant increase in mortality of adult Eastern Gambusia
following the osmotic induction marking procedure. This mortality stabilised 3 – 6 days following
marking after which time mortality of marked fish was approximately 25% higher for female fish
and 50% higher for male fish than fish that were not subject to the marking procedure. The slightly
higher mortality of marked male fish compared to marked female fish may be due to their smaller
body size, although results should be treated with some caution given the small sample size. Leips
et al. (2001) and Bashey (2004) reported calcein marking of other poeciliids (Heterandria formosa
and Poecilia reticulate) having no significant effect on growth or mortality after conducting long
and short term experiments (9 weeks and 14 days respectively). Both of these studies, however,
did not use the osmotic induction technique for marking (i.e. no salt bath before exposure to a
concentrated chemical). Crook et al. (2009) presented a protocol using the osmotic induction
technique with calcein, which produced no significant increase in mortality (reduction in growth)
when marking golden perch fingerlings. At fingerling stage, this species of percichthyid are much
larger, and perhaps more robust, when compared to adult Eastern Gambusia, particularly male
Eastern Gambusia. Therefore, larger fish may be more tolerant of the marking technique than
smaller individuals.
Mohler (2003), who developed the osmotic induction technique, reported no significant difference
in mortality between Salmo salar that had been marked with calcein through osmotic induction
and control fish, which had only been exposed to the salt bath. Mohler (2003) reported that after
47 days fish in both control and marked treatments had mortalities between 11-28%, which is
similar to mortality levels of female Eastern Gambusia in the present study. Furthermore, a 5% salt
solution was also used in his study. This may suggest that the salt bath is the process in the
technique that may influence mortality, not the calcein itself. Further experimentation varying
dosage rates and holding times used in the osmotic induction marking protocol are required to
further test this.
Similar to other studies (Negus and Tureson (2004) and Bashey (2004)), marking did not influence
the short term growth of surviving fish in this study. Crook et al. (2009) reported significantly
higher growth rates for calcein marked fish than unmarked fish during their study on golden perch
fingerlings, suggesting the salt exposure may have had a prophylactic effect given it is commonly
used in hatcheries for disease and parasite treatment.
All surviving fish showed clear fluorescent marks on hard surfaces such as scales and fins, fifteen
days after their initial marking. Whilst assessing mark duration was not an objective of this trial,
these results are supported by other similar studies which have shown external mark longevity in
numerous other species to greatly exceed the 15-days used in this study (e.g. Negus and Tureson
2004; Stubbing and Moss 2007; Crook et al. 2009). However, we suggest future trials of mark
longevity are conducted outdoors (subject to sunlight and ambient temperatures) for Eastern
Gambusia if external marks are required to be detected in the field for extended periods (> 15
days). Eastern Gambusia recaptured in the field have shown substantial photodegredation of marks
as little as 7-days after marking (particularly dorsal surfaces; pers. obs). This is not unexpected
given that the detectability of marked fish may decrease rapidly in fish that are more exposed to
sunlight and higher temperatures (Leips et al. (2001) and Bashey (2004)), coupled with Eastern
Gambusia’s tendency to occupy warm shallow / surface water (Karolak 2006).
The results of the trial have demonstrated that that the osmotic induction marking protocol adopted
from Crook et al. (2009), will increase short term mortality of Eastern Gambusia. Therefore, any
mark-recapture data derived from field based marking that uses this marking procedure must
account for the mortality rates presented in this study to improve estimates of population size and
142
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
capture rates. Whilst the protocol as it stands is indeed useful, the trial highlights a need for further
experimental trials of varied exposure rates to develop species specific marking protocols. The
trial has also demonstrated that mortality rates may also differ between genders and that this
should be considered when developing such marking protocols. This reinforces the
recommendation by Leips et al. (2001) whereby these techniques should be tested for adverse or
beneficial effects on the species in question.
References
Bashey, F. (2004). A comparison of the suitability of Alizarin Red S and Calcein for inducing a
nonlethally detectable mark in juvenile guppies. Transactions of the American Fisheries Society
133, 1516-23.
Crook, D., A., O’Mahony, D.,J., Sanger, A., C., Munro, A., R., Gillanders, B., M., and Thurstan.,
S. (2009) Development and Evaluation of Methods for Osmotic Induction Marking of Golden
Perch, Macquaria ambigua with Calcein and Alizarin Red S. Northern American Journal of
Fisheries Management 29: 279- 287.
Karolak, S. (2006) Alien Fish in the Murray–Darling Basin. MDBC publication No. 03/06,
Murray–Darling Basin Commission, Canberra.
Leips, J., Baril, C.T., Rodd, F.H., Reznick, D.N., Bashey, F., Visser, G.J. & Travis, J. (2001) The
suitability of calcein to mark poeciliid fish and a new method of detection. Transactions of the
American Fisheries Society, 130, 501-507.
Mohler, J.W. (2003) Producing Fluorescent Marks on Atlantic Salmon Fin Rays and Scales with
Calcein via Osmotic Induction. North American Journal of Fisheries Management, 23, 1108-1113.
Negus, M.T. and Tureson, F.T. (2004). Retention and nonlethal detection of Calcein marks in
rainbow and chinook salmon. North American Journal of Fisheries Management 24, 741-47.
Stubbing, D. N., and Moss, R. D. (2007). Success of calcein marking via osmotic induction in
brown trout fry, Salmo trutta. Fisheries Management and Ecology 14:231–233.
143
Native fish recovery following Eastern Gambusia removal
Appendix 2: Project communications
Publications
DSE (2008) “Recovery of Wetland Native Fish Communities Following Removal of Eastern
Gambusia”. Arthur Rylah Institute for Environmental Research Website Article.
http://www.dse.vic.gov.au/arthur-rylah-institute/research-themes/invasive-species
Macdonald, J., Tonkin, Z., Ramsey, D., Kaus, A., King, A. K. and Crook, D.A. (In Press) Do
invasive eastern gambusia (Gambusia holbrooki) shape wetland fish assemblage structure in southeastern Australia? Marine and Freshwater Research
Tonkin, Z., Macdonald, J., Ramsey, D., Kaus, A., Crook, D.A. and King, A.K. (In prep) Fish
community responses to reductions of an alien species. Applied Ecology
Tonkin, Z., Kaus, A., Ramsey, D., Macdonald, J., Crook, D., Hames, F., and King, A. (2011) A
field based assessment of native fish responses following Eastern Gambusia removal: putting
theory into practice. Conference proceedings from the MDBA Gambusia Forum, Melbourne,
1-2 June 2011.
Tonkin, Z., Kaus, A., Ramsey, D., Macdonald, J., Crook, D., Hames, F., and King, A. (2010)
Assessing the recovery of wetland native fish communities following removal of the introduced
Eastern Gambusia, Gambusia holbrooki. Conference proceedings from the 2010 Native Fish
Forum, National Museum of Australia, Canberra 15 – 16 September 2010.
Macdonald, J., and Tonkin, Z. (2009) Eastern Gambusia and native fish communities in wetlands –
assessing native fish recovery following removal of an alien invader. Article in the Australian
Society for Fish Biology Newsletter, 2009.
Macdonald, J., and Tonkin, Z. (2008) A review of the impacts of Eastern Gambusia on native
fishes of the Murray–Darling Basin. Arthur Rylah Institute for Environmental Research.
Department of Sustainability and Environment. Murray–Darling Basin Authority Publication No.
38/09
Macdonald, J., Tonkin, Z., Ramsey, J. and Jin, C. (2008) Native fish recovery following alien
species removal. Conference proceedings from the Native Fish Forum, Canberra 9 – 10 September
2008.
Tonkin, Z. and Macdonald, J (2008) “Alien invaders in our billabongs and wetlands”. DSE Inform
Article. Department of Sustainability and Environmental, August 2008.
Presentations
Tonkin, Z., Kaus, A., Ramsey, D., Macdonald, J., Crook, D., Hames, F., and King, A. (in prep) A
field based assessment of native fish responses following Eastern Gambusia removal: putting
theory into practice. Spoken presentation at the ASFB conference, Townsville, 22-23 July 2011.
Tonkin, Z., Kaus, A., Ramsey, D., Macdonald, J., Crook, D., Hames, F., and King, A. (2011) A
field based assessment of native fish responses following Eastern Gambusia removal: putting
theory into practice. Spoken presentation by Z. Tonkin at the MDBA Gambusia Forum,
Melbourne, 1-2 June 2011.
Tonkin, Z., Kaus, A., Ramsey, D., Macdonald, J., Crook, D., Hames, F., and King, A. (2010)
Assessing the recovery of wetland native fish communities following removal of the introduced
144
Arthur Rylah Institute for Environmental Research
Native fish recovery following Eastern Gambusia removal
Eastern Gambusia, Gambusia holbrooki. Spoken presentation by Z. Tonkin, Native Fish Forum
2010 National Museum of Australia, Canberra 15 – 16 September 2010.
Macdonald, J., Tonkin, Z., Ramsey, J. and Jin, C. (2008) Native fish recovery following alien
species removal. Spoken presentation by J. Macdonald, Native Fish Forum, Canberra 9 – 10
September 2008.
Other communications activities (see below)
DSE Media release, 31st May 2011
Project fact sheet
Removal guidelines for managers
Waranga news, 28th October 2010
DSE inform June 15 2011
145
Native fish recovery following Eastern Gambusia removal
ISSN 1835-3827 (print)
ISSN 1835-3835 (online)
ISBN 978-1-74287-410-4 (print)
146
Arthur Rylah Institute for Environmental Research