Table 1. Evaluation of species distribution models using interpolated precipitation or cloud product for the montane woodcreeper (Lepidocolaptes lacrymiger) and king protea (Protea cynaroides).
species | model | nPresence | nTrials | AUC | cor | DIC | MPSRF | MoransI | GearyC |
---|---|---|---|---|---|---|---|---|---|
Lepidocolaptes lacrymiger | Cloud | 215 | 4318 | 0.87 | 0.35 | 1861.15 | 1.02 | 0.92 | 0.08 |
Lepidocolaptes lacrymiger | Precipitation | 215 | 4318 | 0.68 | 0.10 | 2176.33 | 1.00 | 0.97 | 0.03 |
Protea cynaroides | Cloud | 3012 | 18878 | 0.87 | 0.52 | 21258.07 | 1.01 | 0.82 | 0.17 |
Protea cynaroides | Precipitation | 3012 | 18878 | 0.82 | 0.40 | 23137.50 | 1.00 | 0.90 | 0.08 |
nPresence is the total number of observed presences. nTrials is the total number of trials (species survey/checklists). AUC is the area under the receiver operating curve, a measure of model performance. Cor is the correlation between the observed data and the modeled probability of presence. DIC is the deviance information criterion. MPSRF is Gelman and Rubin's multivariate potential scale reduction factor measuring model convergence. MoransI and GearyC are metrics of spatial autocorrelation. Data available at http://doi.org/10.6084/m9.figshare.1531955.