The habitat suitability model generated in Maxent showed areas that might be possibly occupied by this species in Sierra Nevada de Santa Marta and in serrania de San Lucas. However, we do not have knowledge of any records from those areas and believe they must be excluded at the moment from the potential distribution map of the species.
There are two old records from Medellin and two more records from Samana for which the specimens have been identified as race intercedens. The specimens from Medellin, one of them dated from 1879, might not have been collected in the cordillera Central. However, the specimens from Samana, which are dated from 1951, might be correctly identified. Thus, the extension to which race intercedens is present in the cordillera Central it is unknown. On the other hand, there are no specimens collected from the extreme south Andean region and it is not clear which subspecies might be present in that area.
The subspecies discrepans is known from just two records in the tropical belt of Tolima and Meta, the estimated extent of the areas where this subspecies might be present today is extremely narrow (≈ 5,377 km2, with forest just 2,215 km2). It is completely unknown if it might be present in other areas of the cordillera Oriental or if it might be already extinct.
Assuming that the distribution of the species may have filled the complete climatic model generated, its distribution today in remnants of forest is about 86,058 km2, which corresponds to a loss of 63 % of its potential original distribution due to deforestation.
Regularized training gain is 1.476, training AUC is 0.937, unregularized training gain is 1.764.
Algorithm converged after 720 iterations (19 seconds).
The follow settings were used during the run:
44 presence records used for training.
10044 points used to determine the Maxent distribution (background points and presence points).
Environmental layers used (all continuous): bio10co bio11co bio12co bio13co bio14co bio15co bio16co bio17co bio18co bio19co bio1co bio2co bio3co bio4co bio5co bio6co bio7co bio8co bio9co
Regularization values: linear/quadratic/product: 0.210, categorical: 0.250, threshold: 1.560, hinge: 0.500
Feature types used: linear quadratic hinge
'Equal Training Sensitivity and Specificity' and 'Equate Entropy of Thresholded and Original Distributions' thresholds and omission rates:
0.15-0.229-Fractional predicted area
0.159-0.045-Training omission rate