Going back to the species distribution project, I am now considering not choosing my BIOCLIM layers based on correlation coefficients but rather running a PCA of random samples of points from each layer.

For the Alcippe peracensis extent, the first four principal components will be used:
PC1-->BIO4: Temperature seasonality
PC2-->altitude
PC3-->BIO12: Annual precipitation
PC4-->BIO19:Precipitation of the coldest quarter