This program terminated
September, 2006, but the resources available on this web site will
remain available for some time. The Program's final reports
have been filed, and are accessible through the links below:
The objectives of this research are to develop approaches for spatio-temporal design and modeling in order to further our understanding of aquatic resources. We will extend current design and analysis methodology in order to incorporate data from different scales and different sources including data from EMAP's probability based sampling design, intensive site studies, and remote sensing platforms.
Our research will focus on three major objectives:
In meeting these objectives we will focus on specific EMAP needs.
The first objective of the proposed research is to develop spatio-temporal models for a continuous response to model EMAP data. Considerable work has been done in the area of modeling spatially correlated data, but many problems still need to be addressed. In particular, we propose to extend Bayesian hierarchical models to provide better predictions for spatially and possibly temporally correlated data. We also propose to address the issue of providing accurate assessments of uncertainty about these predictions via methodology called Bayesian model averaging. Finally, we propose to develop better methodology for model selection for these models. All of these new developments should lead to more accurate predictions and improved maps of aquatic based resources.
In the EMAP context, one may wish to model species counts, species presence/absence, or contaminant levels above or below a regulatory threshold. The second objective of the proposed research addresses these types of problems. The goal here is to extend and develop spatio-temporal models for count and/or categorical responses. The proposed methodology will extend existing methodology based on a model for binary response data called the autologistic model. We also propose to develop new methods to analyze spatially and temporally correlated discrete response data including development of a spatio-temporal autologistic model and models for latent processes. These models can be used to produce maps such as a map showing probability that a threshold for a particular contaminant has been exceeded over an area of interest. The methods also can be used to assess which indicators are relevant for predicting whether or not the threshold level has been exceeded.
The final objective of the proposed research is to address an on-going challenge in the modeling of EMAP and other aquatic resource monitoring data: how do we combine data at different scales? The proposed research will develop new guidelines for designing ecological monitoring studies aimed at understanding and assessing key aspects of aquatic systems. In addition to improving design of such studies, we also propose to develop new methodology to analyze data collected at different scales. The goal here will be to both understand and assess key processes in aquatic systems as well as to be able to produce estimates for all scales of interest.
A second benefit of this program will be the training of new statisticians in the area of environmental statistics. There is currently a severe shortage of statisticians with extensive knowledge and experience in solving environmental problems and analyzing environmental data. One goal of this project is to increase the number of statisticians who meet such a need. Graduate students and post-doctoral researchers will participate in all research and dissemination activities proposed in this grant, gaining valuable experience in environmental statistical problems. Our project will thus integrate research and education and partly address the chronic shortage of statisticians with expertise in environmental statistics.
|Home • Program • Opportunities • Learning Materials • Publications • Personnel • Related Links • Agencies • Email us •|