Comparison of Alternative Strategies for Invasive Species Distribution Modeling

Abstract

Species distribution models (SDMs) can provide useful information for managing biological invasions,such as identification of priority areas for early detection or for determining containment boundaries.However, prediction of invasive species using SDMs can be challenging because they typically violatethe core assumption of being at equilibrium with their environment, which may lead to poorly guidedmanagement resulting from high levels of omission. Our goal was to provide a suite of potential decisionstrategies (DSs) that were not reliant on the equilibrium assumption but rather could be chosento better match the management application, which in this case was to ensure containment throughadequate surveillance. We used presence-only data and expert knowledge for model calibration andpresence/absence data to evaluate the potential distribution of an introduced mesquite (Leguminoseae:Prosopis) invasion located in the Pilbara Region of northwest Western Australia. Five different DSs withvarying levels of conservatism/risk were derived from a multi-criteria evaluation model using orderedweighted averaging. The performance of DSs over all possible thresholds was examined using receiveroperating characteristic (ROC) analysis. DSs not on the convex hull of the ROC curves were discarded. Twothreshold determination methods (TDMs) were compared on the two remaining DSs, one that assumedequilibrium (by maximizing overall prediction success) and another that assumed the invasion was ongoing(using a 95% threshold for true positives). The most conservative DS fitted the validation data mostclosely but could only predict 75% of the presence data. A more risk-taking DS could predict 95% of thepresence data, which identified 8.5 times more area for surveillance, and better highlighted known populationsthat are still rapidly invading. ThisDSandTDMcoupling was considered to be the most appropriatefor our management application. Our results show that predictive niche modeling was highly sensitiveto risk levels, but that these can be tailored to match specified management objectives. The methodsimplemented can be readily adapted to other invasive species or for conservation purposes.