Third Annual Conference: Statistical Survey Design and Analysis for Aquatic Resources, Study Guides, Projects, Research of School management&administration

Download Third Annual Conference: Statistical Survey Design and Analysis for Aquatic Resources and more Study Guides, Projects, Research School management&administration in PDF only on Docsity! THIRD ANNUAL CONFERENCE: STATISTICAL SURVEY DESIGN AND ANALYSIS FOR AQUATIC RESOURCES Department of Statistics, Colorado State University September 10 - 11, 2004 SPONSORED BY: STARMAP in collaboration with DAMARS STARMAP = Space-time Aquatic Resources Modeling and Analysis Program DAMARS = Design and Model for Aquatic Resource Surveys Both funded by EPA STAR Cooperative Agreements (CR-829095 & 829096) ABSTRACTS in alphabetical order by first author, as available {w ith l ink s to presenta tion ma terials ; presented by first author , except w here noted by *} Nonparametric Survey Regression Estimation Using Penalized Splines F. Jay Breidt, Department of Statistics, Colorado State University, and Jean Opsomer, Department of Statistics, Iowa State University The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/breidt.2004.pps Model- vs Design-Based Sampling and Variance Estimation on Continuous Domains Cynthia Cooper, Department of Statistics, Oregon State University The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/cooper.2004.pps Sampling Strategies for Chinook-Salmon Spawning Populations Jean-Yves Pip Courbois, Steve Katz, Chris Jordan, Michelle Rub, and Ashley Steel NOAA-Fisheries, Northwest Fisheries Sciences Center, Seattle, WA, and Russel F. Thuron and Daniel J Isaak US Forest Service, Rocky Mountain Research Station We compare several sampling strategies to estimate the number of redds in the Middle Fork Salmon River Drainage. To evaluate the strategies the precisions, accuracies, and costs are compared in the form of the relative standard error, confidence interval coverage, and a cost function based on travel through the river network respectively. We find that systematic designs work best in terms of the most precise estimators. However the confidence intervals for these designs can be misleading if naive estimators for the estimates’ standard errors are used. Instead estimators based on a neighborhood should be used. These designs however are most expensive because the entire watershed needs to be visited for each sample. The cost precision trade-off needs to be made for different objectives and budget. The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/courbois.2004.pps Bayesian Models for Radio Telemetry Habitat Data Megan Dailey, Department of Statistics, Colorado State University, Alix I. Gitelman and Fred L. Ramsey, Department of Statistics, Oregon State University, and Steve Starcevich, Oregon Department of Fish and Wildlife Radio telemetry data used for habitat selection studies typically consists of a sequence of habitat types for each individual indicating habitat use over time. Existing models for estimating habitat selection probabilities have incorporated covariates in an independent multinomial selections (IMS) model (McCracken et al., 1998) and an extension of the IMS to include a persistence parameter (Ramsey and Usner, 2003). These models assume that all parameters are fixed through time. However, this may not be a realistic assumption in radio telemetry studies that run through multiple seasons. We extend the IMS and persistence models using a hierarchical Bayesian approach that allows for the selection probabilities, the persistence parameter, or both, to change with season. These extensions are particularly important when movement patterns are expected to be different between seasons, or when availability of a habitat changes throughout the study period due to weather or migration. The models are motivated by radio telemetry data for fish in which seasonal differences are expected and evident in the data. The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/dailey.2004.pps the model parameters correspond to conditional independence relationships. Using a Gibbs sampling approach, we illustrate application of the model on a data set of fish species richness in the mid-Atlantic region of the U.S. Keywords: compositional data; graphical models; logistic normal; random effects; species assemblage; conditional independence The materials used in this presentation are available as a pdf file at: http://www.stat.colostate.edu/starmap/pps/johnson.2004.pdf Increasing the Role of Statistics in Water Quality Management Decisions Dan McKenzie, WED, NHEERL, ORD, USEPA, Corvallis, OR The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/mckenzie.2004.pps Model Selection for Geostatistical Models Andrew A. Merton, Jennifer A. Hoeting, and Richard A. Davis Department of Statistics, Colorado State University We consider the problem of model selection for geospatial data. The importance of accounting for spatial correlation has been discussed in other contexts, but the effect of spatial correlation on the choice of covariates in the model has not been fully explored. We consider kriging for geostatistical models to predict a response at unobserved locations, which involves the fitting of explanatory variables and an autocorrelation function. Spatial correlation is typically ignored in the selection of explanatory variables and this can influence model selection results. To address this problem, we consider the Akaike Information Criterion (AIC) as applied to a geostatistical model. We offer a heuristic derivation of the AIC in this context and provide simulation results that show that using AIC for a geostatistical model is superior to the traditional approach of ignoring spatial correlation in the selection of explanatory variables. An example further demonstrates these ideas. Keywords: AIC ; kriging ; autocorrelation function The materials used in this presentation are available as a pdf file at: http://www.stat.colostate.edu/starmap/pps/merton.2004.pdf Adjustment Procedures to Account for Nonignorable Missing Data in Environmental Surveys Breda Munoz and Virginia Lesser* Department of Statistics, Oregon State University Methods for non response are well-known techniques used in survey practice for handling missing data. In this approach, sampling units (respondents and non-respondents) are classified in weighting classes, and the sampling weight for each respondent unit is weighted by the inverse of an estimate of its response probability (also known as propensity score). Optimal weighting classes are selected using variables associated with the response but uncorrelated with the response indicator. We explore the assumptions needed to construct optimal adjustment classes in the case of non-ignorable missing data in environmental surveys. We propose a modified Horvitz- Thompson non response estimator for the population total of the spatial random process of interest and study some of its properties. By using the weighting class adjustment, we will account for the non-ignorable missing data. The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/munoz.2004.adjustment.pps A Weighting Class Adjustment Estimator for the Total Under a Stratified Sampling Design in a Continuous Domain Breda Munoz and Virginia Lesser Department of Statistics, Oregon State University A weighting class adjustment is a common technique used by survey analysts for missing data. In complex surveys, this technique has been used for finite population sampling data. However, the weighting class adjustment has not been used for continuous populations. We explore in this paper some properties of the weighting class adjustment estimator under a stratified sampling design in a continuous domain and we develop an expression for the variance within the same framework. The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/munoz.2004.weighing.pps Predicting the Likelihood of Water Quality Impaired Stream Reaches Using Landscape Scale Data and a Hierarchical Methodology Erin E. Peterson Department of Geosciences, Colorado State University The Clean Water Act (CWA) (1972) requires states and tribes to identify water quality impaired stream segments, to create a priority ranking of those segments, and to calculate the Total Maximum Daily Load (TMDL) for each impaired segment based upon chemical and physical water quality standards. However, it is impossible to physically sample every stream within a large area due to the immense number of segments, limited personnel, and the cost associated with sampling. The purpose of this study is to develop methods that can be used by states and tribes to identify areas that have a higher probability of water quality impairment. A geographical information system (GIS) will be used to extract coarse scale landscape attributes related to sample points in two spatially dense datasets: one collected by the Maryland Department of Natural Resources and the other by Teck Cominco Alaska Incorporated, which is a mining company located in Alaska. All of the data will be used as input into two spatial models, which will then be used to predict reach scale water quality conditions. These models will differ from previous landscape models because they will be produced using newly developed methods of asymmetric kriging and geostatical model selection. The results of this study will provide a better understanding of the relationship between landscape and reach scale conditions. They will also provide managers with landscape indicators, which are crucial for the rapid and cost efficient monitoring of large areas. Most importantly, the methodologies used in this study will provide a cost effective tool that can be used by states and tribes to help fulfill the requirements of the CWA. The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/peterson.2004.pps Multi-Lag Cluster Enhancement of Fixed Grids for Variogram Estimation in Near Coastal Systems Kerry J. Ritter and Molly K. Leecaster Southern California Coastal Water Research Project, Westminster, CA N. Scott Urquhart* Department of Statistics, Colorado State University Kenneth C. Schiff Southern California Coastal Water Research Project, Westminster, CA Dawn Olsen and Tim Stebbins City of San Diego, CA Maps are useful tools for understanding, managing and protecting our marine environment. Despite the benefits, there has been little success in developing useful and statistically defensible maps of environmental quality and aquatic resources in the coastal regions. Heterogeneous SPACE-TIME AQUATIC RESOURCES M ODELING AND ANALYSIS PROGRAM (STARMAP): YEAR 3 N. Scott Urquhart Department of Statistics, Colorado State University The STARMAP program has made progress on each of its projects and responsibilities. Project 1 on Combining Environmental Data Sets has made progress on the analysis of complex contingency tables and on the selection of spatial models; Jennifer Hoeting, its principal investigator, will describe this in more detail; several investigators will report specific results. Project 2 on Local Inference has made progress on models for smoothing environmental responses along stream networks. Jay Breidt, its principal investigator, will describe this work and the work of several collaborators, including Jean Opsomer of Iowa State University. Project 3 on Indicator Development supports work in Geographic Information Systems (GIS), directed toward statistics. David Theobald of the Natural Resources Ecology Laboratory, its principal investigator, will describe GIS tools relevant to the statistical analysis of aquatic systems. Project 4 on learning materials continues to expand and test learning materials about statistics and sampling relevant to water quality scientists. STARMAP has participated substantially in several professional conferences as part of its outreach responsibilities and to provide training opportunities for developing environmental statisticians. The materials used in this presentation are available as a PowerPoint Presentation file at: http://www.stat.colostate.edu/starmap/pps/urquhart.2004.pps POSTERS Sampling Perennial Streams: An Application in Model Based Optimal Design William J. Coar and F. Jay Breidt Department of Statistics, Colorado State University One-dimensional Point Processes in Ecology Jean-Yves Pip Courbois NOAA-Fisheries, Northwest Fisheries Sciences Center, Seattle, WA We show two examples of one-dimensional point processes in Ecology, turtle nests along a beach and chinook salmon nests along a river. Our ultimate goal is to describe the distributions of these events in order to optimize sampling designs for enumerating these events. Although technically two- dimensional the long-skinny nature of the event locations necessitates analysis in one dimension. This dimension reduction is necessary because edge effects in two dimensions prevents the analysis of long-range associations, isotropy is likely violated as both the turtles and fish may be choosing locations for nests differently along the beach/river than across the beach/river, and when sampling nests we are likely to sample across the entire beach/river. The turtle-nest distribution is well approximated by the Neyman-Scott spatial cluster process implying that the turtle-nests are clustered along the beach. One reason for this clustering is that turtles build more than one nest and tagging studies have shown that they do not stray far between nests. The chinook redd (nest) distribution is more complex. It demonstrates both clustering at a small scale, a possible regular pattern, as well as non-stationarity caused by large-scale environmental factors. As in much of statistics, one wishes to model both the first and second moments of this distribution but techniques to model both simultaneously are lacking in the point process field. Also displayed at The International Environmetrics Conference, Portland, ME, June 28 - July 1, 2004 Hierarchical Bayesian Models for Seasonal Radio Telemetry Habitat Data Megan C. Dailey Department of Statistics, Colorado State University Alix I. Gitelman Department of Statistics, Oregon State University Also displayed at the Graybill Conference, Colorado State University, June 16 - 18, 2004 and The International Environmetrics Conference, Portland, ME, June 28 - July 1, 2004 Distribution Function Estimation in Small Areas for Aquatic Resources Mark J. Delorey Department of Statistics, Colorado State University Also displayed at the Graybill Conference, Colorado State University, June 16 - 18, 2004 Nonparametric, Model-Assisted Estimation for a Two-Stage Sampling Design Mark Delorey and F. Jay Breidt Department of Statistics, Colorado State University What Is a Multi-scale Analysis? Implications for Modeling Presence/Absence of Bird Species Kathi Georgitis, Alix Gitelman, and Don L. Stevens Jr. Department of Statistics, Oregon State University Nick P. Danz, and JoAnn M. Hanowski Natural Resources Research Center, Univ of Minnesota Duluth Also displayed at The International Environmetrics Conference, Portland, ME, June 28 - July 1, 2004 Predicting the Likelihood of Water Quality Impaired Stream Reaches Using Landscape Scale Data and a Hierarchical Methodology: a Case Study in the Southern Rocky Mountains Erin E. Poston Department of Geosciences David M. Theobald Natural Resources Ecology Laboratory and Department of Natural Resources Recreation & Tourism Melinda J. Laituri Department of Forest, Rangeland, and Watershed Stewardship N. Scott Urquhart, Department of Statistics, all of Colorado State University Related posters also have been displayed at the American Water Resources Association GIS Specialty Conference, Nashville, TN, May 17-19, 2004, and the Second Annual Conference: Statistical Survey Design and Analysis for Aquatic Resources, Corvallis, OR, August 11 - 13, 2003 Two-stage Sampling Designs for Birds in Great Lakes Wetlands Ron Regal Department of Mathematics and Statistics, University of Minnesota Duluth Don L. Stevens, Jr. Department of Statistics, Oregon State University, Nick P. Danz and JoAnn M. Hanowski Natural Resources Research Center, Univ of Minnesota Duluth Robert W. Howe, Department of Natural and Applied Sciences, University of Wisconsin-Green Bay Learning Materials for Surface Water Monitoring N. Scott Urquhart, Department of Statistics, Colorado State University Also displayed at the EMAP Symposium 2004, Newport, RI, May 4 - 7, 2004