Search for a command to run...
Abstract 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 potential distributions of invading vectors to target surveillance and control efforts and identify at risk populations. Species Distribution Models (SDMs) are widely 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. Here, we use long-term mosquito 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 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 if 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. Author Summary Mosquitoes are greatly impacted by their surrounding environment. Environmental changes such as those driven by climate change or changes in land use (i.e. urbanization, etc.) can have profound impacts on mosquito ranges. Species distribution models (SDMs), which use the occurrence data of a species in combination with environmental data (e.g., landscape characteristics, temperature, etc.) to estimate suitable habitats for the species, are helpful tools to understand how species ranges can change with the environment through time and space. There are many different species distribution models to choose from – each with different methods for estimating suitable habitat – making it important to compare the performance of different models. Furthermore, because invasive species often violate assumptions of SDMs, it is important to rigorously explore how different methods perform when predicting invasive species ranges in newly invaded areas. In this study, we tested the predictive accuracy of different species distribution models, and we found it varied across modeling and validation method. We also tested whether models could use early-stage invasion data to predict the late-stage invasion distributions of Ae. albopictus. We found that some modeling methods performed well while others needed more data to improve accuracy of late-stage invasion range predictions. In summary, when modeling species ranges using SDMs, it is important to use multiple methods and compare results because methods will likely disagree, and further sampling may be required to determine which model is most accurate.