Grantee Research Project Results
Final Report: Effects of Interacting Stressors in Agricultural Ecosystems: Mesocosm and Field Evaluation of Multi-level Indicators of Wetland Responses
EPA Grant Number: R826595Title: Effects of Interacting Stressors in Agricultural Ecosystems: Mesocosm and Field Evaluation of Multi-level Indicators of Wetland Responses
Investigators: Threlkeld, Stephen , Benson, William H. , Crain, Andrew , D'Surney, Stephen , Easson, Greg , Ochs, Clifford , Schlenk, Daniel , Slattery, Marc
Institution: University of Mississippi
EPA Project Officer: Packard, Benjamin H
Project Period: October 1, 1998 through September 30, 2001 (Extended to September 30, 2002)
Project Amount: $897,634
RFA: Ecological Indicators (1998) RFA Text | Recipients Lists
Research Category: Ecological Indicators/Assessment/Restoration , Aquatic Ecosystems
Objective:
The primary objective of this research project was to evaluate indicators of molecular, cellular, population, community, and ecosystem responses to multiple, potentially interacting, natural, and anthropogenic stressors that vary at different spatial and temporal scales in agricultural wetlands. The indicators were chosen to represent a selection of mechanism-based, and system-level integrative characteristics that would be amenable to cost-effective routine monitoring. Our null hypothesis was that indicators that effectively characterize ecosystem responses to single stressors also are scale- and interaction-independent (i.e., useful even when there are multiple, interacting stressors with diverse operational scales). Our alternative hypothesis was that when multiple, interacting stressors are present, responses are not well characterized by indicators that are useful for monitoring the effects of single stressors. This outcome would require the use of either a different set of indicators, or a different spatial or temporal scale of resolution for evaluating the indicators.
Summary/Accomplishments (Outputs/Outcomes):
Significant Mesocosm Effort Findings. Our approach used a wetland mesocosm experiment (27-1 center point enhanced fractional factorial design with sampling on days 2, 12, 24, and 48 post-exposure) to evaluate indicator sensitivity to single and multiple stressors. We then incorporated field tests of these indicators in comparable wetlands in the northern Mississippi agricultural landscape. Included in the treatment structure were three pesticides (atrazine, chlorpyrifos, and MSMA), nutrient fertilizers (N, P), ultraviolet radiation, suspended solids, and presence of a predator. The calculation of principal components (all days combined) for the mesocosm effort yielded groupings of responses by periphyton nutrients (PC1), periphyton pigments (PC2), and phytoplankton pigments and dissolved nutrients (PC3). The effects of the treatment structure on the added eigenvector values for each variable (e.g., a score for PC1 for each mesocosm, a score for PC2, etc.) are seen by significant effects of atrazine*nutrients, MSMA*UV, and fish predators*UV on PC1, 2, and 3. The level of significance was p0.05 for all tests.
Significant Field Effort Findings. We measured responses at several levels of biological organization, which evaluated how damage to components of agricultural wetlands is first promulgated, and the extent to which the damage is indirectly transmitted to other levels of biological organization, including those more amenable to routine monitoring and assessment.
A total of 101 ponds were visited during the sampling season (see Figure 1). The samples that were collected and the endpoints that were measured were developed to correlate with endpoints measured from our mesocosm experiment. A Principal Components Analysis of the pond data revealed a grouping of response variables by spectral signatures (bands 1-3) and phytoplankton pigments (PC1); turtle diversity and phytoplankton pigments (PC2); spectral signatures (bands 4-6) and dissolved organic carbon (DOC) (PC3); and several turtle size and quantity variables (PC4).
Regression of PC scores (by pond) against the organizing features of crop type and pesticide use, using a stepwise procedure, on the square mile grid, 3 x 3 square mile grid, and 5 x 5 square mile grid scales revealed soybean acreage as a dominant predictive variable across several scales. This relationship also is apparent in correlation matrices between the principal components and crop/pesticide use across the different spatial scales. For example, PC1 and PC2 were significantly correlated (negative) with soybean acres on the single square mile [p <0.0001 (PC1); p = 0.0104 (PC2)], 3 x 3 square mile grid [p = 0.0016 (PC1); p = 0.0079 (PC2)], and 5 x 5 square mile grid scales [p = 0.0062 (PC1); p = 0.0215 (PC2)]. Each principal component was analyzed separately in the correlation matrices.
Figure 1. The Little Tallachtchie Drainage Basin Agrichemical Loading Combinations
One of our concerns about this research project has centered on whether or not the "square mile" level of spatial resolution has been adequate to objectively identify a random set of potential sampling sites. We were constrained in the use of this unit because the Farm Service Agency offices maintain their records according to the Public Land Survey System. Essentially, our question was: Are the land and pesticide use practices that are present in a square mile similar to practices within a 1- or 2-mile radius? For this reason, we resampled our crop and pesticide data to obtain a running average of crop coverage and pesticide use across increasing spatial scales (i.e., 3 x 3 and 5 x 5 square mile grids, 9 and 25 total square miles, respectively). This procedure allowed us to conclude that the square mile is an adequate unit as square mile practices are similar to those in surrounding square miles.
A promising line of future research to be developed from this project is the potential for using LANDSAT spectral signatures as indicators of changing conditions in aquatic habitats. The quantitative values for bands 1-3 are highly correlated with high-performance liquid chromatography (HPLC) analysis of phytoplankton pigment concentrations, as well as significantly contributing to the calculation of Principal Component 1. For example, bands 1, 2, and 3 are positively and significantly correlated with beta-carotene (p = 0.0115, p = 0.0332, and p = 0.0289, respectively) and chlorophyll a (p = 0.0218, p = 0.0650, and p = 0.0485, respectively). These contributions are consistent in this data set, even when corrected for excessively turbid waters. Although the significance of the regression relationship between the principal components and agricultural practices may drop across increasing spatial scales, the relationship is consistent as evidenced by significant correlations between the principal components and agricultural variables. The promise of this line of investigation also is highlighted by the groupings of periphyton pigments and phytoplankton pigments on PC2 and PC3, respectively, from the mesocosm study. The next step in this area would be to obtain spectral signatures for the 8,700+ ponds in our study basin by resampling our imagery data and obtaining additional LANDSAT images. Two passes (images) of the satellite provide data for the entire basin. These values would then be analyzed against the crop and pesticide use data, per square mile (2,100+), that we have already developed. The correlation matrices for our present data and resampled data then will be evaluated to determine the consistency of the relationships between the data sets. A high degree of consistency would support the potential use of spectral data, as related to biological activity, to be an effective ecosystem indicator, both financially and biologically, in monitoring the effects of multiple stressors in wetlands of agricultural ecosystems.
Journal Articles on this Report : 2 Displayed | Download in RIS Format
| Other project views: | All 33 publications | 3 publications in selected types | All 2 journal articles |
|---|
| Type | Citation | ||
|---|---|---|---|
|
|
Easson GL, Robinson H. Assembling DEM files for watershed analysis. ArcUser July-September 2001. |
R826595 (Final) |
not available |
|
|
Easson GL, Robinson H. Using 30-meter resolution digital elevation data for basin analysis-a practical utilization of USGS 24K digital elevation data-complications and solutions. ArcUser 2001;4(3):26-28. |
R826595 (2001) R826595 (Final) |
not available |
Supplemental Keywords:
agricultural watershed, agriculture ecosystems, agriculture, agronomy, anthropogenic stresses, aquatic ecosystems, cellular, chlorpyrifos, ecological risk assessment, ecosystem indicators, environmental stress, enzymes, exploratory research environmental biology, factorial experiment, field validation, genetics, interactive stressors, landscape, mesocosm, metals, Mississippi, MS, molecular biology, monitoring, multiple spatial scales, multiple stressors, multiscale assessment, organics, organism, pesticides, population, remote sensing, scaling, UV effects, water, watershed, scaling, organics, genetics, ecology, limnology, population., RFA, Scientific Discipline, Water, Ecosystem Protection/Environmental Exposure & Risk, Water & Watershed, Ecology, Limnology, Ecosystem/Assessment/Indicators, Environmental Chemistry, Ecological Effects - Environmental Exposure & Risk, Agronomy, Watersheds, Ecological Indicators, anthropogenic stresses, ecological risk assessment, interactive stressors, remote sensing, UV effects, agricultural watershed, chlorpyrifos, enzymes, metal release, multiple spatial scales, multiple stressors, ecosystem indicators, field validation, mesocosm, aquatic ecosystems, environmental stress, water quality, stress responses, multiscale assessment, agriculture ecosystemsRelevant Websites:
Effects of interacting stressors in agricultural ecosystem wetlands Exit
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.