|
The Spatially Explicit Exposure Module
(SEEM): An Expanded Landscape For Wildlife Exposure Assessment Tools
|
|||||||||
|
|
||||||||||
The Spatially Explicit Exposure Module (SEEM): An Expanded Landscape For Wildlife Exposure Assessment Tools Charles A. Menzie, PhD
Summary: Ecological risk assessors continue to increase the realism and management value of risk analyses through the development of new analytical tools. While acknowledging that wildlife exposure areas often extend to an area that is larger than the site boundary and exclude areas lacking sufficient habitat, assessors are often left applying the assumption that a representative wildlife species’ exposure occurs entirely within the site boundary. Ecological risk assessors may draw from a toolbox of exposure assessment methods that vary from simple area use factors to the most complex wildlife population models in order to capture more realistic wildlife exposures and variability. At Menzie-Cura we are developing a spatially explicit exposure module (SEEM) for the U.S. Army Risk Assessment Modeling System (ARAMS) that balances analytical power with ease of use. SEEM is a powerful screening tool that provides assessors with the opportunity to explore wildlife exposure from the perspective of all of individuals in a population. Individuals in the population move across the landscape guided by a set of foraging rules. While the foraging behaviors are species-specific, the model also incorporates habitat characteristics that underlie individual movement. Specifically, SEEM uses Markov Chain Monte Carlo to create a habitat quality weighted, though randomized, movement pattern for each individual. The result is that SEEM captures and compiles exposure for all individuals in a local population. Instead of assuming that a representative individual is exposed to an entire site, SEEM incorporates habitat quality avoiding a situation in which risk is identified based on an exposure to chemicals in a portion of the site that does not contain habitat. SEEM is constructed to be flexible and may be adapted to answer a number of different questions. Screening templates will also be developed to allow users to exposure patterns resulting from the convergence of specific site features, wildlife species characteristics, habitat quality and chemical distribution. SEEM will become a module within ARAMS and in this capacity it will draw upon the large databases of species – specific information, toxicity reference values and fate and transport models within ARAMS.
CA Menze, WT Wickwire, D Burmistrov and BK Hope. An Expanded Landscape for Wildlife Exposure Assessment Tools: The Spatially Explicit Exposure Module (SEEM )
Biography: Dr. Menzie is founder and President of Menzie-Cura & Associates, Inc., a firm internationally recognized for its work in human health and ecological risk assessment. He has over 20 years of professional experience in the field of ecological and human health risk assessment, risk communication, and risk management. He has been involved in the development and review of risk assessment guidance, for the United States, Canada, several states, Gas Research Institute, and the Electric Power Research Institute. Dr. Menzie is a member of the National Research Council Committee on Bioavailability of Chemicals in Soils and Sediments. He is a member of the Executive Board for SETAC and has been involved in the Pellston Workshops related to sediments. Dr. Menzie has facilitated many peer reviews, workshops, and meetings related to risk assessment, ecotoxicology, soils, and sediments. His work has focused on applying weight-of-evidence approaches to the evaluation and management of contaminated soils and sediments. He is a past president for the local SRA Chapter and a past councilor for SRA National. Using Landscape Analysis to Estimate Exposure—Estimating PCB Concentrations in Floodplain Soils Using Habitat Constrained Spatial WeightingJohn P. Lortie
Summary: The USEPA is conducting an ecological risk assessment on the Housatonic River in Pittsfield, Massachusetts in accordance with a Consent Decree with General Electric Co. As part of the site characterization and conceptual model development, landscape analysis was used to predict the occurrence of species in a particular area and spatial weighting was used to estimate levels of PCB contamination in floodplain soils per area. Landscape analysis is the process of using known information on the physical, chemical, and biological features of a site to predict the occurrence of similar features where data is lacking. As a first step in landscape analysis, aerial photographs were used to map plant communities or habitats in the study area. Ground-truthing was then performed to correct any interpretation errors and to collect additional data on site micro-topography that could influence contaminant and species distributions. Digital maps of the boundaries of nineteen plant communities were created, and a Geographic Information System (GIS) data base was constructed using plant communities, man-made features, hydrology, topography, and contaminant concentrations (e.g., total PCBs, PCB congeners, dioxins/furans). Approximately 5,000 floodplain soil and sediment (inundated floodplain soils) samples had been collected along transects and in other areas as part of biological sampling programs (e.g., invertebrates, small mammals, amphibians). Total PCB concentrations were greatest in floodplain areas that were most frequently flooded and in areas close to the main channel. Two methods were investigated for use in creating an interpolated grid of total PCB contamination throughout the study area: Thiessen polygons and inverse distance weighting (IDW). IDW was chosen as the preferred method because it allows the user more flexibility over the mathematical form of the weighting function and the size of the neighborhood. IDW was developed using ESRI ArcView 3.2 and Spatial Analyst and selected extensions from the FIELDS tools developed by USEPA Region 5. Neighborhood size was constrained by habitats based on fate and transport hypotheses about PCB transport during flood events, and geomorphological observations of river channel meandering. Several iterations of IDW were performed; with a 1-m grid; with a 1-m grid constrained by habitats; with a 3-m grid constrained by habitats; and finally using a cross-validation process to minimize root mean square error within each polygon. The final product is a grid of the entire study area populated by total PCB concentrations, which can be used for exposure assessment (e.g., 95% UCL), fate and transport modeling, and for visually depicting the spatial extent of varying-levels of contamination. JP Lortie, RA McGrath, J Cassels, S Svirsky. Using Landscape Analysis to Estimate Exposure – estimating PCB concentrations in floodplain soils using habitat constrained spatial weighting.
Biography: Mr. Lortie is co-founder and President of Woodlot Alternatives, a firm specializing in applied field biology. He is a certified Professional Wetland Scientist, a Certified Wildlife Biologist, and an accomplished botanist. He has directed a number of large-scale ecological characterization studies, and in the last 8 years has been working with risk assessors to identify suitable receptors, develop site-specific exposure assessments, and characterize risk based on prey items and foraging habits. As a Professional Wetland Scientist, Mr. Lortie specializes in river and wetland restoration following the removal of contaminated sediments. |
||||||||||
|
|
