To elaborate our model we removed uncertain localities such as those localities with no coordinates and all records with coordinates laying down in sites with estimated elevations above 1,500 m. Also one record in EBIRD from Quindio, which may represent corrupted data in the database was deleted before modelling.
The habitat suitability model generated in Maxent showed some areas suitable in climatic terms for this species west of the Andes including areas in the Pacific, the slopes of the mid Cauca Valley and the head of the Magdalena Valley. Those areas are not known to be occupied by this species and were deleted from our final potential distribution map. It is important to highlight that an extensive area in the northern Amazonia was not predicted as suitable by our model and this was noted in our maps.
Assuming that the distribution of the species may have filled the complete climatic model generated plus the areas not predicted but likely suitable, its distribution today in remnants of forest is about 439,870 km2, which corresponds to a loss of 9 % of its potential original distribution due to deforestation.
Regularized training gain is 1.215, training AUC is 0.917, unregularized training gain is 1.672.
Algorithm converged after 280 iterations (1 seconds).
The follow settings were used during the run:
14 presence records used for training.
10014 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: 0.629, categorical: 0.357, threshold: 1.860, hinge: 0.500
Feature types used: linear quadratic
'Equal Training Sensitivity and Specificity' and 'Equate Entropy of Thresholded and Original Distributions' thresholds and omission rates:
0.143-0.297-Fractional predicted area
0.143-0.071-Training omission rate