Field validation of multiple species distribution models shows variation in performance for predicting Aedes...

Climate and land use changes have resulted in range expansion of many species. In this shifting disease landscape, it is important to leverage tools that can predict the distributions of invading vectors to target surveillance and control efforts and identify at-risk populations. Species distribution models (SDMs) are used to predict ranges of invasive species; however, invasive species often violate assumptions of equilibrium and niche conservatism. Moreover, these studies are rarely validated using independent data.

We use long-term surveillance data for Aedes albopictus, a highly invasive mosquito capable of transmitting several arboviruses, at its range edge to evaluate a variety of SDMs (MaxEnt, GAM, Random Forest, Boosted Regression Tree) in predicting the Ae. albopictus range. We identify key environmental drivers of distributions and areas where models tended to disagree in predicting occurrence. At sites where models disagree, we sampled for Ae. albopictus to generate an independent dataset for field validation of models in addition to the common practice of cross-validation. Finally, we determine whether models based on early invasion data can predict later stage invasion ranges.

We found that landscape and climatic variables are important drivers of population distributions. SDM methods varied in predictive accuracy between models and across validation methods (i.e. cross vs. field validation). GAM and MaxEnt best predicted later-stage invasion distributions, requiring fewer years of training data.

Our work shows that SDMs can be useful tools to predict the ranges of invasive species and highlights the importance of comparing predictions of invasive species' range.