Abstract Detail

Ecology

Randle, Christopher [1], Bianchi, Laura [2], Brenek, Austin [3], Castillo, Jesus [1], Galle, Nicholas [1], Hankins, Kayla [1], Nobleza, Kenneth [1], Reger, Nicholas [1], Williams, Justin [1].

Predictive niche modeling of pests and pathogens of major US crop commodities.

The United States has the third highest agricultural output of any country, and exports far more food products than any other country. Global food security therefore depends heavily on agricultural practices in the United States. The Cooperative Agricultural Pest Survey (USDA-CAPS) has identified more than 30 priority pests of corn, cotton, grapes, small grains, and soybeans, most of them with broad distributions outside of the US. Ecological niche modeling (ENM) is a computational technique for predicting most suitable habitats for a species in a region. Niche models work by quantifying current distributions of species and deriving response functions that maximize correlation between species occurrence and environmental data, such as temperature, precipitation, soil type, or forest cover. These optimized functions can then be used to project the most probable sites in which those species may persist onto a different region. This may be a useful tool in predicting suitable habitat for future invasions of croplands. However, some of the assumptions underlying ENM are likely to be violated by invasive pests, namely the assumption of ecological equilibrium. In this study, we incorporate three tools, the General Additive Models, Maximum Entropy Models, and Boosted Regression Tree Models, to make predictions about habitat suitability of these pests in the US. Ensemble predictions consisting of AUC (Area-Under-the-receiver-operator-characteristic-Curve) weighted averages were analyzed to 1) develop a ranking of pests with the highest probability of finding suitable habitat in US croplands, 2) identify US regions most susceptible to new agricultural pests, and 3) to identify regions of the world from which pests and pathogens are most likely to enter the US.

1 - Sam Houston State University, Department of Biological Sciences, 2000 Avenue I, Huntsville, TX, 77340, United States
2 - Sam Houston State University, Department of Biological Sciences, LSB 105, Huntsville, TX, 77340, United States
3 - Sam Houston State University, Department of Biological Sciences, Huntsville, TX, 77340, United States
4 - Sam Houston State University, Department of Biological Sciences, 2000 Avenue I, Huntsville, TX, 77340, United States

Keywords:
Species Distribution Modeling
Invasive
Pest
agriculture.

Presentation Type: Oral Paper
Number: EC10006
Abstract ID:170
Candidate for Awards:None