ISFS participates in Ecological Society of America 2004 Annual Meeting

The Invasive Species Forecasting System project will be presenting a paper & accompanying poster at this year's 89th Annual ESA Meeting held in Portland, Oregon August 1-6.

Schedule

Our USGS partners from the Fort Collins Science Center and the National Institute of Invasive Species Science will also be presenting at the conference.

Schedule

The Ecological Society of America (ESA) is a nonpartisan, nonprofit organization of scientists founded in 1915 to:

• promote ecological science by improving communication among ecologists;
• raise the public's level of awareness of the importance of ecological science;
• increase the resources available for the conduct of ecological science; and
• ensure the appropriate use of ecological science in environmental decision making by enhancing communication between the ecological community and policy-makers.

Paper Abstract: "Data synergy and spatial modeling of leafy spurge."

Crosier, Catherine^{1, 2}, Stohlgren, Tom^{1, 2}, ¹ Natural Resource Ecology Laboratory, Fort Collins, CO, USA² Fort Collins Science Center, Fort Collins, CO, USA

Potential distribution maps for many non-native species are unavailable, making early detection a difficult task for land managers. We collected over 30 different datasets from across the state of Colorado and combined them into a spatial database to use as input to generate large-extent, high-resolution models. We extracted presence/absence locations for Euphorbia esula (leafy spurge) to produce a map of probability of occurrence using a generalized linear model (GLM) with stepwise variable selection and kriging of spatial autocorrelated residuals. The evaluation dataset was a statewide survey of county weed managers for acreage estimates of E. esula per quarter quad. The GLM explained 62% of the deviance. However, a Mann-Whitney test to approximate the area under a Receiver Operating Characteristic (ROC) curve indicated poor discrimination between presence/ absence sites. Before discrediting the model completely, it should be noted that the survey was based on county managers incomplete knowledge of current distributions, not potential ones, and was conducted at a lower resolution. Modeling techniques such as these used in concert with synergy of datasets have promise as aids to land managers to detect, map, and control the spread of weeds such as E. esula across large spatial extents. This is an iterative process with the results of the model driving further field efforts which are in turn used in the model.

Paper Abstract: "Predictable patterns of plant, bird, and fish invasions in the United States: The huge role of local determinism."

Crosier, Catherine^{1, 2}, Stohlgren, Tom^{1, 2}, ¹ Natural Resource Ecology Laboratory, Fort Collins, CO, USA² Stohlgren, Thomas¹, Barnett, David ², Flather, Curtis ³, Fuller, Pam¹, Peterjohn, Bruce¹, Kartesz, John⁴, Master, Larry⁵, ¹ U.S. Geological Survey, Fort Collins² Natural Resource Ecology Laboratory, Fort Collins, CO³ USDA Forest Service Rocky Mountain Research Station, Fort Collins, CO⁴ Biota of North America Program, Chapel Hill, NC⁵ NatureServe, Boston, MA

Following global patterns of biodiversity, we show that the richness and density of species of native and non-indigenous vascular plants, birds, and fishes predictably decreases in northern latitudes and higher elevations following declines in PET, mean temperatures, and precipitation or productivity. We used county-level data on plants (n = 2,018 counties) and birds (n = 3,077 counties), and watershed (hux-8) data on fishes (n = 874 watersheds) to show that native and non-indigenous plant species richness were strongly positively correlated (r = 0.78, P < 0.0001), as were native and non-indigenous bird species (r = 0.31, P < 0.001). Multiple regression models showed that the density (species/km²) of native plant, bird, and fish species could be strongly predicted (R ² = 0.77, 0.71, and 0.80 respectively) at county and watershed levels. Similarly, non-indigenous plant and bird species densities were strongly predictable (R² = 0.83 and 0.92 respectively), but non-indigenous fish species density was less predictable (R² = 0.35). We conclude that while humans facilitate the initial establishment of non-indigenous species, the spread and subsequent distributions of non-indigenous species may be controlled largely by environmental factors.

Paper Abstract: "Use of temporal indicators for invasive plant species in the NASA/USGS invasive species forecasting system"

J.T. Morisette, J.A. Pedelty, J.A. Smith, T.J. Stohlgren, C. Crosier, M.A. Kalkhan, R. Reich and J.L. Schnase.

The NASA Office of Earth Science and the US Geological Survey are working together to develop a National Invasive Species Forecasting System (ISFS) for the early detection, remediation, management, and control of invasive species on Department of Interior and adjacent lands. The system provides a framework for using USGS's early detection and monitoring protocols and predictive models to process MODIS, ETM+, ASTER and commercial remote sensing data, and create on-demand, regional-scale assessments of invasive species patterns and vulnerable habitats. Recent work on the ISFS project has shown the importance of remotely-sensed time-series data in geostatistical models for mapping the distribution of invasive species. The study used field surveys of species richness, one 30m spatial resolution Landsat 7 Enhanced Thematic Mapper plus (ETM+) image, and a three year time-series of 250m spatial resolution Moderate Resolution Imaging Spectrometer (MODIS) imagery over three sites. Summary values from the MODIS time series were added as explanatory variables in the ISFS framework. In a model to predict species richness, results show the MODIS summary values are at least as statistically significant as ETM+-derived predictors. This indicates the importance of including temporal information for large area mapping of invasive species, even at the more coarse 250m spatial resolution.

Poster Abstract: "Predicting the Extent of Tamarisk Habitat on a Quarter Quadrangle Scale"

Davern, T.R.^{1, 2}, R.M. Reich³, T.J. Stohlgren²

Tamarisk (Tamarix spp.) is an invasive shrub that is native to Eurasia. It was introduced into the United States in the 1800s. By the 1920s tamarisk was becoming a problem (Brotherson and Field, 1987, Di Tomaso, 1998) . Since then it has covered almost a million hectares (~2.5 million acres) of the American West, (Pearce and Smith, 2003) occurring mainly in waterways and riparian habitats. We identified potential tamarisk habitat using spatial statistical modeling. Field work identified the presence/absence and percent cover of tamarisk on four different 7.5’ topographical maps. Two of these quarter quadrangles were in western Colorado and two were in southern California. We created a model predicting the extent of tamarisk habitat within the selected areas. To test the models we compared the Colorado models to the California models to see if the same variables were significant to the habitat of tamarisk in both locations. We also looked at presence/absence models to determine the area delineated as tamarisk . Tamarisk habitat is found in riparian areas as expected, but tamarisk has not reached its full potential. Habitat information can be used by land managers for early detection, rapid response, and monitoring purposes. Infestations could then be controlled before they expand into large areas.

Poster Abstract: "The National Invasive Species Forecasting System: A geostatistical modeling infrastructure for predicting biological and ecological invasion"

J.W. Closs^{1, 2}, N.F. Most^{1, 3}, D.J. Kendig^{1, 2}, M.A. Kalkhan⁴, J.T. Morisette¹, J.A. Pedelty¹, T.J. Stohlgren⁴ and J.L. Schnase¹.
¹NASA Goddard Space Flight Center, Earth and Space Data Computing Division, Greenbelt, MDUSA.
²Science Systems and Applications, Inc., Lanham, MDUSA.
³Intelview, Greenbelt, MDUSA.
⁴Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, COUSA

The ability to model small scale variability in landscape characteristics requires the generation of full coverage maps depicting characteristics measured in the field. While many spatial datasets describing land characteristics have proven effective for macroscale ecological monitoring, these relatively coarse scale data fall short in providing the precision required by more refined ecosystem resource models. Geostatistical modeling provides a means of developing spatial models that can be used to correlate coarse scale geographical data with microscale field measurements of biotic variables. To address this need, the NASA Office of Earth Science and the US Geological Survey are working together to develop the National Invasive Species Forecasting System for the early detection, remediation, management, and control of invasive species on Department of Interior and adjacent lands. The ISFS is a Web-based information management and modeling environment tailored to the needs of the invasive species community. The system provides a framework for using USGS's early detection and monitoring protocols and predictive models to process MODIS, ETM+, ASTER and commercial remote sensing data, and create on-demand, regional-scale assessments of invasive species patterns and vulnerable habitats. When fully implemented, it will provide a mechanism for managing the data and modeling activities that underpin invasive species research, management, and policy decision-making. NASA is working directly with the USGS National Institute of Invasive Species in Ft. Collins, CO. The Institute is part of the Fort Collins Science Center, which is in turn part of the USGS Biological Resources Division (BRD). The major public interface to the data and capabilities of the Invasive Species Forecasting System is through the National Biological Information Infrastructure (NBII) program. Initial test sites include the Cerro Grande wildfire site in Los Alamos, New Mexico, Rocky Mountain National Park, and the Grand Staircase Escalante National Monument on the Colorado plateau.

Data synergy and spatial modeling of leafy spurge

+ download presentation

+ view Stohlgren bio
+ view Crosier bio





Use of Temporal Indicators for Invasive Plant Species in the NASA/USGS Invasive Species Forecasting System

+ download presentation (Morisette)
+ download presentation (Schnase)

+ view Morisette bio
+ view Schnase bio



Predicting the Extent of Tamarisk Habitat on a Quarter Quadrangle Scale

+ download poster



The National Invasive Species Forecasting System: A geostatistical modeling infrastructure for predicting biological and ecological invasion

+ download poster