Download Simplified Lake Eutrophication Modeling

Simplified Lake Eutrophication
Modeling: Using the Ecoregionbased model MINLEAP
Steven Heiskary,
Environmental Research Scientist
Minnesota Pollution Control Agency
wq-ppt1-01
MINLEAP - Development and
Acknowledgements
◆
◆
◆
“Minnesota Lake Eutrophication Analysis
Procedure”
Developed by Bruce Wilson and Dr. William
Walker Jr. (1989), published in Lake and Reserv.
Manage. 5(2): 11-22 “Development of Lake
Assessment Methods Based on Aquatic Ecoregion
Concept”
BASIC originally (Wilson), recent Windows
version - (Wade Gillingham, MPCA)
Overview
Session I
◆
◆
◆
◆
General modeling concepts;
Ecoregion framework and MINLEAP;
MINLEAP background and subroutines;
Application of MINLEAP
Session II
◆
◆
Loading and using MINLEAP
Case Studies;
Lake Eutrophication Modeling
(Drawn from Wilson, 1990)
◆ Predictive techniques to assess common
lake problems;
◆ Foundation is reasonable estimation of
water and nutrient budgets or loading;
◆ Lakes often exhibit similarities in response
to nutrient loading;
◆ Most techniques based on cross-sectional
studies -- e.g., ecoregion reference lakes;
Typical Empirical Model
Network
Inflow TP
Mean Depth Residence
In-lake Total P
Chlorophyll-a
Secchi
MINLEAP Development
◆ Based on ecoregion framework -- considers
regional patterns in geomorphology, soils,
landuse, and climatic characteristics;
◆ Uses average precip., evap., and runoff
◆ Canfield and Bachmann (1981) is the P
sedmentation model used;
◆ TP, Chl-a, and Secchi models based on MN
reference lakes;
Ecoregion Reference Lakes
N orthern
Minnesota W etlands
N = 30
Z = 6.3 m
A = 3 18 ha
Red
R iver
Valley
N = 36
Z = 6.6 m
A = 364 h a
No rthern L akes
an d Fo rests
N orth C entral
Hardwood Forests
N orthern
N = 8
Glaciated
Plains Z = 1.6 m
A = 21 8 ha
N = 11
Z = 2 .5 m
A = 10 7 ha
W estern C orn B elt Plains
Driftless
Area
Distribution of Phosphorus by Lake Mixing
Status and Ecoregion
D = Dimictic, I = Intermittent, P = Polymictic
NLF
NCHF
WCBP
Mixing Status:
Percentile value
for [TP]
90 %
D
I
P
D
I
P
D
37
53
57
104
263
344
--
I
--
P
284
75 %
29 35 39 58 100 161 101 195
50 %
20 26 29 39 62 89 69 135 141
25 %
13 19 19 25 38 50 39
10 %
# of obs.
211
58
97
25
--
69
257 87 199 152 71 145 4
3
38
9
13
12
19
21
32
TP and Chl-a
ECOREGION REFERENCE LAKES
Mean summer concentrations
Log Chlorophyll a ppb
2.5
2
1.5
1
0.5
0
0
0.5
1
1.5
Log Total Phosphorus ppb
2
2.5
Chlorophyll-a and Secchi
Summer-mean Secchi vs. Chlorophyll-a (log-log).
Based on ecoregion reference lakes.
1.00
y = -0.57x + 0.87
2
R = 0.82
0.80
0.60
Log Secchi meters
4. 0 m
0.40
0.20
0.00
1. 0 m
-0.20
-0.40
1 0 ppb
-0.60
0.00
0.50
1.00
1 0 0 ppb
1.50
Log Chl-a ppb
2.00
2.50
Algal Bloom Frequency
C hlorophyll-a B loom Frequency as a Function of Summer Mean. B ased on a single season of
data (W alker, 1985).
100
90
80
Chl-a Frequency (%)
70
60
50
40
30
>10
>20
>30
>60
20
10
ppb
ppb
ppb
ppb
0
10
20
30
40
S um m e r M e a n ppb
50
60
70
MN Lake P Criteria
Heiskary and Wilson, 1988
Ecoregion
Most Sensitive Use
P Criteria
Northern Lakes
and Forests
drinking water supply
cold water fishery
primary contact recreation
and aethetics
drinking water supply
primary contact recreation
and aesthetics
drinking water supply
primary contact recreation
(full support)
(partial support)
primary contact recreation
and aesthetics (partial
support)
< 15 µg/L
< 15 µg/L
< 30 µg/L
North Central
Hardwood
Forests
Western Corn
Belt Plains
Northern
Glaciated Plains
< 30 µg/L
< 40 µg/L
< 40 µg/L
< 40 µg/L
< 90 µg/L
< 90 µg/L
Model Statistics
◆ Error:
Difference between observed and
predicted mean value (e.g. standard error);
◆ Variability: Considers spatial and temporal
fluctuations in concentration about mean;
◆ T-test: If absolute value is < 2 the obs. mean
is not significantly different than predicted
(95 % confidence). Useful for identifying
problem lakes.
Vighi and Chiaudani
“Background P”
◆
◆
◆
◆
◆
Based on MEI -- ratio between TDS and Z, used
to estimate fishery yields;
Regression based on 53 lakes with negligible
anthropogenic P loading;
MEI = alk (meq/L)/ Z;
Lakes within the predicted [P] +/- C.I. approach
background P;
May not work well on ‘naturally eutrophic”
Applications
◆ How is lake doing given its ecoregion and
morphometric considerations;
◆ Estimates water and nutrient budget;
◆ Assesses inter-relationship of TP, Chl-a,
and Secchi -- compared to reference;
◆ Estimates background P (chla and Secchi);
◆ One basis for goal setting;
Goal Setting
1995
Longm ea n term
24 ± 3 40 ± 4
M IN - V ig h i LEAP
BATHTUB
31
26
25
L a k e M in n ew a sk a S u m m e r -M e a n P (p p b )
• =recen t su m m er-m ea n T P co m p a ra b le to
p re d ic te d - a t o r n ea r “ b a ck g ro u n d T P ”
• =G o a l - T P < 3 0 µ g /L (criteria fo r N C H F
eco reg io n < 4 0 µ g /L )
• =B elo w 3 0 µ g /L n u isa n ce b lo o m s < 5 % o f
su m m er; 4 0 µ g /L T P - b lo o m s ~ 3 0 %
=
Model Performance
◆
◆
◆
Good
Dimictic - NLF, NCHF;
No large upstream lakes;
Watershed char. similar
to norm for ecoregion;
◆
◆
◆
◆
Poor
Polymictic WCBP, NGP
high internal recycle,
high turbidity;
Seepage lakes
Chains of lakes -upstream sedimentation;
MINLEAP Predicted vs Observed
Confidence interval = mean std. error of observed data.
80
70
Wi l d Ri c e
Observed TP ppb
60
Tr o u t
I ndi a n
50
40
30
20
Mi d d l e
Co r mo r a n t
10
Gi r l &
Ch i l d
0
0
10
20
30
40
Predicted T P ppb
50
60
70
80
Conclusions and
Recommendations
◆
◆
◆
◆
◆
Simple model - first cut analysis;
Rough estimate of nutrient and water budgets;
Tool to flag lakes that may deserve additional
study -- relative to others in ecoregion;
Basis for communicating lake response and
estimating response to change in load;
One basis for goal setting -- along with Vighi, P
criteria, observed data, and related;