Developing Ecological Forecasting Models

[Staff]

Developing Ecological Forecasting Models
Research Task: 8327CMC.1.0
Task Manager: Catherine Jarnevich

Developing Ecological Forecasting Models
excerpt from above:

High-performance modeling and use of space-based observations are essential elements of the Nation’s emerging assault on invasive plants, animals, and disease organisms. As a part of this effort, and in cooperation with Colorado State University and others, FORT scientists developed the Global Organism Detection and Monitoring system (www.niiss.org). This Web-based data management system features online spatial modeling tools to assist DOI land managers and others with management decisions and activities related to invasive species. FORT also recently created an “Advanced Invasive Species Modeling Room,” where scientists are exploring the latest in predictive modeling techniques, including which techniques work better with different datasets, taxa, and spatial extents and resolutions. Predictive models developed at FORT using the modeling room are being used to create on-demand, regional and national-scale assessments of invasion patterns, vulnerable habitats, potential distributions of specific invaders, and how all of these may be affected by changing climate. For example, researchers developed a model using current climate conditions that identifies suitable habitat for the invasive kudzu vine in the southeastern United States. They then applied the model to future climate conditions, creating predictions for where populations might decrease, be stable, or increase in the future. Both GODM and the Advanced Modeling Room provide the means for delivering advanced decision-support capabilities that can be used in a wide range of management applications. These systems have successfully been tested in wildlife refuges, national parks, and in research areas of other USGS scientists. Species tested have included weeds like dalmation toadflax and leafy spurge, tamarisk, the invasive diatom Didymosphenia geminata, Burmese pythons, the problematic rodent nutria, and the avian disease Salmonella.

[Staff]

Modelling spread of foot-and-mouth disease in wild white-tailed deer and feral pig populations using a geographic-automata model and animal distribution
Michael P. Ward, Shawn W. Laffan, and Linda D. Highfield

ScienceDirect - Preventive Veterinary Medicine : Modelling spread of foot-and-mouth disease in wild white-tailed deer and feral pig populations using a geographic-automata model and animal distributions
excerpt from above:

Abstract

We investigated how the size and distribution of wild deer and feral pigs – species that might act as potential foot-and-mouth disease (FMD) virus maintenance hosts – might affect the persistence and spread of FMD. We used a susceptible-latent-infected-recovered geographic-automata model and spatially referenced data from southern Texas, USA. Within this study area, 100 locations were randomly selected and FMD virus spread was simulated (50 simulations each) at each location. As expected, the predicted sizes (km2) of the wild deer outbreaks were highly correlated (rSP > 0.95) with the number of deer at incursion locations, the total number of deer within 2 km of incursion locations, and the minimum and maximum deer herd size within 2 km of incursion locations. However, the predicted sizes of the feral pig outbreaks were only moderately correlated (rSP 0.63–0.67) with the total, maximum and variance of the number of feral pigs within 2 km of incursion locations. Lack of continuity within the feral pig herd distribution across the landscape makes predicting disease spread more difficult than for deer, a more homogenously distributed species. When assessing the potential of wild and feral animal species at a locality to act as maintenance hosts of FMD virus, estimates of the population size and distribution might serve as a useful indicator of potential outbreaks in some circumstances.