Download The influence of rock, forest community, and topographic position on

The influence of rock type, forest community, and topographic position on soil nutrients
in the Luquillo Mountains of Puerto Rico
Steven T. Goldsmith1, Susanna Mage1, and Stephen Porder1
1 Environmental
1. Introduction
Change Initiative, Brown University, Providence, RI 02912, USA (*correspondence: [email protected])
3. Methods
There are five state factors that define an ecosystem’s function and properties:
climate, biota, topography, parent material, and time (Amundson and Jenny
1997). While numerous studies have isolated one of these variables and
explored its effect on ecosystem properties (e.g. Vitousek, 2004, Chadwick et al,
2003, Walker and Syers, 1976), very few studies have systematically tested the
effects of several factors in the same experimental design. For example, we
know that soil parent material (Kitayama et al, 2000 ), climate (Chadwick et al,
2003) and slope position (Porder et al., 2005) can influence soil nutrient
availability. But it is not clear whether, in a system with two rock types and steep
slopes, if topographic position or parent material is more important for structuring
ecosystem properties.
In this context, we explored the influence of parent material, forest community
composition, and topographic position on several soil properties relevant to
ecosystem functioning and our understanding of silicate weathering at the
Luquillo LTER and CZO in Puerto Rico. We sampled soils across a full factorial
combination of two parent materials (granodiorite and volcaniclastics), two forest
types (Tabonuco and Colorado) and three topographic positions (ridge, midslope,
valley). Here we present three types of dependent variables: soil elemental
chemistry, soil elemental losses relative to parent material, and the relative
importance of atmospheric inputs to the soil exchangeable cation pools (as
measured by 87Sr/86Sr). We find that the relative importance of each state factor
varies with dependent variable. While some factors (e.g. soil P or Si content) are
driven by differences in parent material, topographic position is more important for
other properties (e.g. loss of P relative to parent material), and catena location,
rather than rock type, topographic position, or forest type, is most important for
driving the fraction of atmospherically-derived cations held on the exchange
pools.
Parameters exhibiting
statistical difference
• We sampled 16 sites across the Luquillo landscape (two rock types by two forest types with
four replicates of each combination). At each site, we collected soils from 3 catenas (ridge,
slope, valley). Mineral soils were sampled from 0-20, 20-50, and 50-80cm (Pic 2). The
original 432 soil samples were composited by depth and topographic position for any given
site (e.g. the three 0-20cm samples from a ridge at a particular site were combined), for a
total of 144 soil samples. We report here depth weighted averages of the top 50cm for all
data except Sr isotopes, which are reported for the top 20cm.
%P ( I < O, p = 0.0048)
%Si (I > O, p = 0.0004)
%Ca (I > O, p = 0.0025)
Tau Si (I > O, p = 0.001)
Tau Ca (I < O, p = 0.0001)
• We measured soil loss or gain of elements relative to parent material by indexing to an
immobile element (niobium, Nb; Brimhall and Dietrich, 1987). This value () accounts for soil
collapse or dilation, and allows a robust estimation of the degree of weathering a soil has
undergone. We calculate  as:
where C is concentration, w and P denote the concentration in soil and parent material,
respectively.
•
Volcaniclastics
Inceptisols
Tau P (R/S > V, p= 0.004)
Tau Ca (R/S > V, p= 0.02)
Oxisols
Forest Type
N/A
Tabonuco
Ridge
Slope
Topographic
Position
Valley
Ridge
Colorado
Slope
Valley
87/86Sr
isotopic analysis of soil ammonium acetate extractions was conducted on a Finnegan
Model 261 Thermal Ionization Mass Spectrometer (TIMS) at Brown University. Data was
normalized to 88Sr/86Sr=8.7352. The average for NBS-987 during the time when our
samples were run was 0.71002±0.00006.
Fig 2: Statistical relationships for total element concentrations
and Tau with selected state factors for Puerto Rico soils
We hypothesize that the soil-exchangeable Sr would be derived
primarily from atmospheric sources in these highly weathered soils,
and rock-Sr would be rejuvenated at lower slope positions (Porder et
al., 2005). Surprisingly, soil 87Sr/86Sr was determined by which catena
the samples came from, rock type, forest type, and slope position had
no statistically significant effect. As yet we have no explanation for
this.
1.00
Fraction of 87/86Sr
from the atmosphere
Fig. 1. Map of the Luquillo National Forest showing
both the two major rock and forest types. Samples
sites are denoted with a ( ).
Pic. 2. View of the sampling
methodology for the 0-20cm interval.
Granodiorite
Soils
5. 87/86Sr Isotopes to Discern End-Member
Inputs
The study site, Luquillo National Forest, is located within the
Luquillo Mountain system in eastern Puerto Rico (Fig. 1).
Elevations range for our soil sampling spanned two major
forest types: Colorado and Tabonuco (United States
Department of Agriculture, 2002). Soils in this region are
derived from one of two parent materials. volcaniclastic and
granodiorite (Guariguata, 1990), each of which has given rise
to a definitive soil type (Oxisol and Inceptisol, respectively).
Inceptisol soils have been found to have higher clay content,
soil cohesive strength and exhibit less frequent landslides
compared to Oxisols.
Lithology
%Si ( G > V, p= 0.002)
%P (G < V, p= 0.0014)
%Ca (G ≈ V)
• Major elemental analysis of air-dried soil was conducted after lithium metaborate fusion and
analysis by x-ray flouresence spectrometry (XRF) by ALS Chemex (Reno, NV).
2. Study Site
Pic. 1. View of the Luquillo National
Forest.
4. Major Elemental Concentrations and Open System Mass Loss
0.80
Acknowledgements
• Si and P concentrations differ significantly between both parent
materials and their respective soil types. Surprisingly, there is a
difference in Ca concentrations between the two soil types
despite similar concentrations in the two parent materials, which
must reflect mineralogical differences in loss rate, and may reflect
a mineralogical difference between the rock types.
We thank Art Johnson and Fred Scatena of
the University of Pennsylvania for their
logistical assistance and helping in collecting
samples and the U.S. Forest Service for
providing access to El Yunque National Park I
Puerto Rico. Financial support was provided
to SP from NSF DEB-0918387 and the
Andrew Mellon Foundation.
• The loss of P from the different soils proceeds at a similar rate
despite differences in mineralogy, concentrations of other Pbinding elements such as iron and aluminum, and soil texture. P
is heavily bio-cycled, and relatively immobile in soils, so this
result is not surprising, but it indicates that in a given landscape
parent material may play a substantial role in driving the p status
of ecosystems.
References
Amundson, R., and Jenny, H., Bioscience 47, 536-543 (1997).
Brimhall, G.H. Jr., and Dietrich, W.E., Geochim. Cosmochim. Acta.
51, 567–587 (1987).
Chadwick, O.A., Gavenda, R.T., Kelly, E.F., and Zeigler, K., et al.
Chem. Geol. 202, 195–203 (2003).
Guariguata, M.R., Ecology 78, 814–832 (1990).
0.60
• Lower P and Ca values in valleys reflect the cumulative input of
less weathered material from landslides/erosive transport
supports the hypothesis that erosion can provide rejuvenation of
rock-derived material in tropical landscapes with highly
weathered soils.
R² = 0.37
0.40
0.20
0.00
-1.00
Pic. 3. View of soil sampling
equipment.
6. Conclusions
• Forest type does not appear to play a controlling role on either
soil elemental concentrations or weathering processes.
-0.50
P
Fig 3: P versus 8786Sr from 16 catenas.
0.00
Kitayama, K., Majalap-Lee, N., and Aiba, S, Oecologica 123: 342-249
(2003).
Porder, S., Paytan, A., and Vitousek, P.M., Ecosystem Ecology 142,
440–449 (2005).
United States Department of Agriculture, Soil Survey of Caribbean
National Forest and Luquillo Experimental Forest, Commonwealth of
Puerto Rico (2002)
Vitousek, P., Chadwick, O., Matson, P., Allison, S., Derry, L., Kettley,
L., Luers, A., Mecking, E., Monastra, V., Porder, S. Ecosystems 6,
762–772 (2003).
Walker, L.R., Zarin, D.J., Fetcher, N., Myster, R.W., Johnson, A.H.,
Biotropica 28, 566–576 (1996).