Correlation between predictions of Culex pipiens distribution models created using boosted regression trees (BRT) and maximum entropy (Maxent) algorithms. “Training region” consists of 158 points taken from within a similar geographical area to the samples used building the model, 79 background points and 78 independent occurrence points. “Expanded Region” measures model agreement throughout the entire modeled extent, and compares values at 158 random points taken with equal probability from the entire model extent. Left panels describe models built using human population density as a parameter (N = 9). Panels on the right describe models excluding human population density (N = 14). COR = Pearson correlation coefficient.