Grantee Research Project Results
Final Report: A Statistical Modeling Methodology for the Detection, Quantification, and Prediction of Ecological Thresholds
EPA Grant Number: R832447Title: A Statistical Modeling Methodology for the Detection, Quantification, and Prediction of Ecological Thresholds
Investigators: Richardson, Curtis J. , Qian, Song S.
Institution: Duke University
EPA Project Officer: Packard, Benjamin H
Project Period: February 1, 2005 through June 30, 2007 (Extended to August 31, 2008)
Project Amount: $278,876
RFA: Exploratory Research: Understanding Ecological Thresholds In Aquatic Systems Through Retrospective Analysis (2004) RFA Text | Recipients Lists
Research Category: Aquatic Ecosystems , Water
Objective:
The primary objective was to develop a statistical modeling approach for the detection and quantification of ecological thresholds that can be used in aquatic ecosystems worldwide. Specifically, we used a Bayesian hierarchical modeling approach to integrate the single-species changepoint analysis method developed by Qian et al. (2003) to study the interactions among fast and slow responding species or ecological metrics indicating alternative stable states. Under this structural approach, the resulting species/metric thresholds were integrated to form an ecosystem-level threshold distribution. A risk-based criteria approach was developed to assess ecological resilience and predict the threshold for alternative state changes.
Summary/Accomplishments (Outputs/Outcomes):
In the past three years, we focused our work in several areas:
- Utilized our 15 year Everglades database to develop a new integrated hierarchical statistical method for threshold analysis.
- Wrote and published several manuscripts on key indicator species and thresholds.
- Presented our results at two workshops on ecological thresholds. Papers were also given on these topics at the annual ESA and SWS meetings in 2007 and 2008.
We refined our data files related to the threshold studies that were done in the Everglades and then tested alternative threshold models. Our analysis concluded that the traditional confidence interval is not a suitable tool for summarizing uncertainty in ecological threshold. Our Bayesian approach resulted in a more relevant way for summarizing uncertainty and linking the estimated ecological threshold to environmental standards. In a separate study, we investigated the possible applications of a multilevel analysis of variance (ANOVA) method. The multilevel ANOVA model is a potentially a valuable tool for ecological threshold study. We wrote a manuscript on multilevel analysis, which was published in 2007. A paper on our threshold work in the Everglades was also published in 2007. A new paper entitled A Hierarchical Modeling approach for Estimating Ecological thresholds” is under final development. Key outcomes from this paper concern the development of a new hierarchical approach to assessing ecological thresholds, which are increasingly used in studies of ecological and environmental management. Quantitative methods for estimating ecological thresholds are specific to the particular response variables of interest, while changes in the ecosystem or environment are usually defined at a broader scale which must include multiple responses. This paper discusses the Bayesian hierarchical modeling approach for integrating thresholds of individual responses and developing an integrated ecosystem model. This approach has numerous applications to the field of ecological threshold analyses. Key findings from our earlier published Ecology and ES&T papers are as follows: (Copies of these papers are provided in our final report)
Ecology 88 (10):2489-2495.A Bayesian representation of the analysis of variance by A. Gelman is introduced with ecological examples. These examples demonstrate typical situations encountered in ecological studies. Compared to conventional methods, the multilevel approach is more flexible in model formulation, easier to set up, and easier to present. Because the emphasis is on estimation, multilevel models are more informative than the results from a significance test. The improved capacity is largely due to the changed computation methods. In our examples, we show that (1) the multilevel model is able to discern a treatment effect that is smaller than the conventional approach can detect, (2) the graphical presentation associated with the multilevel method is more informative, and (3) the multilevel model can incorporate all sources of uncertainty to accurately describe the true relationship between the outcome and potential predictors.
Environmental Science and Technology 41 (23):8084-8091.The Florida Everglades, a wetland of international importance, has been undergoing a significant shift in its native flora and fauna due to excessive total phosphorus (TP) loadings (an average of 147 metric tons per annum from 1995 to 2004) and an elevated mean TP concentrations (69 mg L-1 of TP in 2004) from agricultural runoff and Lake Okeechobee outflow despite the use of 16,000 hectares of stormwater treatment areas. We utilized the data base from 15 years of field research and developed a Bayesian changepoint analysis of long-term experimental research which showed that exceeding a surface water geometric mean TP threshold concentration of 15 mg L-1 causes an ecological imbalance in algal, macrophyte, and macroinvertebrate assemblages as well as slough community structure. A Phosphorus threshold for all trophic levels was also developed, which may be more realistic and protective when presented as a threshold zone (12-15 mg L-1) because estimates of uncertainty must be utilized to accurately define TP thresholds, which change with seasons and water depths. Most interior areas of the Everglades are currently at or below this threshold zone, but the exterior areas near inflow structures (except for the Everglades National Park) are presently receiving double or triple the proposed threshold. Our Bayesian modeling approach, used here to address ecological imbalance along nutrient gradients, is applicable to determining thresholds and stable states in other aquatic ecosystems.
Journal Articles on this Report : 2 Displayed | Download in RIS Format
Other project views: | All 8 publications | 2 publications in selected types | All 2 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Qian SS, Shen Z. Ecological applications of multilevel analysis of variance. Ecology 2007;88(10):2489-2495. |
R832447 (2006) R832447 (2007) R832447 (Final) |
Exit |
|
Richardson CJ, King RS, Qian SS, Vaithiyanathan P, Qualls RG, Stow CA. Estimating ecological thresholds for phosphorus in the Everglades. Environmental Science & Technology 2007;41(23):8084-8091. |
R832447 (2006) R832447 (2007) R832447 (Final) |
Exit Exit |
Supplemental Keywords:
Total phosphorus; soluble reactive phosphorus; ecological thresholds, Bayesian changepoint analysis, credible interval, RFA, Scientific Discipline, Ecosystem Protection/Environmental Exposure & Risk, Aquatic Ecosystems & Estuarine Research, Aquatic Ecosystem, Environmental Monitoring, Ecology and Ecosystems, species interaction, ecosystem monitoring, statistical modeling, aquatic ecosystems, ecological classification, aquatic indicatorsRelevant Websites:
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.