Fig. 3: Predictions of suitable habitat for Cochemiea halei . The map is of predictions of habitat suitability, on a scale of zero (transparent) to 1 (dark green). The model predictions derive from a BRT method, using WorldClim V. 2.0 data at 30 arcsec resolution. 44 presences and 207 pseudoabsences were used. The following variables were used: annual temperature range, the mean temperature of the warmest quarter (July-September), precipitation of the warmest quarter (July-September), precipitation of the coldest quarter (December-February), and soil type. The model fitted 4950 trees, with a 10-fold cross-validated AUC of .88. The parameters used for the boosted regression tree analysis were a tree complexity of 2, a learning rate of .0007, bag fraction of .7 and a step size of 25.