Aedes albopictus spatial models using Citizen Science and Non Citizen Science inputs

This deliverable compares spatial distribution models for the Aedes albopictus mosquito using data from Citizen Science (CS) and Non Citizen Science (NCS) surveillance datasets,. NCS datasets are often patchy and incomplete, and this research investigates the impact of incorporating CS data into spatial models to fill gaps. The analysis focused on Europe, where CS data is largely concentrated in the south, while NCS records more widespread in the north. Models based on CS< NCS, and a Combined dataset were generated using Boosted Regression Trees (BRT) and Random Forest (RF) techniques. Surprisingly, the Combined model's predicted distribution was more restricted in some areas than the NCS model alone, despite having more extensive training data. The Combined model showed marginally better accuracy metrics (e.g., higher Kappa and specificity scores). This suggests that integrating CS data, concentrated in warmer/drier regions, may effectively refine the statistical relationship defining suitable environmental conditions for the vector. The files are provides as an ARCMAP10.8 .mpk package. To access in other software change the extension to .zip and extract. open .mxd in folder /v108 if possible. Otherwise naviagte fodler to find .tif output or shapefile inputs. Filenames are designed to be self explanatory