Improved prediction of Canada lynx distribution through regional model transferability and data efficiency

These datasets were used to generate an ensemble species distribution model using the R package Bioclim2. The file "Model_WA30VegOnly_TrainData_15Dec2020" contains the used and background locations used to make the best-performing model in the paper referenced below. It contains n=5026 actual lynx GPS locations, with the covariate values used in the analysis appended to each location. The coordinates of the points have been removed due to the sensitivity of the species. The background data were used in equal sample size to the use data, but subsampled repeatedly from this full dataset for use as multiple instances of psuedo-absences. The file "Moel_WA30VegOnly_TestData_15Dec2020" gives an equal number of used and background locations for use as model validation data. The data in this file were not used for model construction, but were collected on the same GPS collars, so are strongly spatially correlated with the model construction data.<br>Further information and details on the datasets can be found in "Olson et al., 2021, Improved Prediction of Canada lynx distribution through regional model transferability and data efficiency, Ecology and Evolution".<br>