The habitat suitability model generated in Maxent showed areas that might be possibly occupied by this species in serrania del Baudo. However, we do not have knowledge of any records from that area and believe it must be excluded at the moment from the potential distribution map of the species. Similarly, the model point to some areas in the far east, which are very unlikely to be part of the distribution of this species. Otherwise, we included some areas in the Sierra Nevada de Santa Marta predicted by the model because we know of several observations made in the area of San Lorenzo (EBIRD), which we did not use in elaboration of the model.
The distribution of this species is extremely unknown in Colombia. There is one record from San Antonio in Valle del Cauca, dated 1911, identified as race kleei. This subspecies is distributed in the Eastern Andes of Ecuador and east Peru to Bolivia and Brasil (HBW Alive). In Colombia there are no known records in the southern Andean region. However, it is assumed to exist since there are records ≈ 10 km south from the border into Ecuador. At present time it is not clear if the specimen from San Antonio has been correctly identified or may represent an extension in the distribution of this subspecies. On the other hand, it might be that the populations in the Western Andes represent an undescribed subspecies.
The subspecies larensis allegedly occupies areas in the Sierra Nevada the Santa Marta (HBW Alive). Nevertheless, there are no collected specimens from that area and this needs confirmation.
Assuming that the distribution of the species may have filled the complete climatic model generated, its distribution today in remnants of forest is about 93,586 km2, which corresponds to a loss of 63 % of its potential original distribution due to deforestation.
Regularized training gain is 0.963, training AUC is 0.940, unregularized training gain is 1.629.
Algorithm converged after 480 iterations (13 seconds).
The follow settings were used during the run:
17 presence records used for training.
10016 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: 0.500, categorical: 0.250, threshold: 1.830, 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:
0.13-0.382-Fractional predicted area
0.118-0-Training omission rate