Table 2. RDM performance.

Model	Calibration	Project	Dataset	AUCTRAIN	AUCTEST	TSS	D
A	B. invadens	B. dorsalis	ENFA	0.888	0.762	0.559	0.51
B	B. invadens	B. dorsalis	Expert	0.881	0.683	0.480	0.51
C	B. dorsalis	B. invadens	ENFA	0.891	0.841	0.579	0.86
D	B. dorsalis	B. invadens	Expert	0.894	0.804	0.461	0.81
E	B. invadens	DPP	ENFA	0.888	0.787	0.568	0.76
F	B. invadens	DPP	Expert	0.881	0.731	0.494	0.68
G	DPP	B. invadens	ENFA	0.884	0.83	0.553	0.91
H	DPP	B. invadens	Expert	0.886	0.845	0.563	0.93

Models were assessed on their ability to predict the distribution of the other species with the AUC_TEST (area under the receiver operating characteristic curve) score (independent dataset not included in model construction). DPP = B. dorsalis + B. philippinensis + B. papayae. True Skill Statistic (TSS) values were calculated to evaluate model performance at the threshold of maximum training sensitivity plus specificity. Schoener’s D values are the overlap of the given model in the projected range where the reciprocal model was calibrated.