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 2,487 m.
The habitat suitability model generated in Maxent showed areas suitable in climatic terms for this species in serrania del Baudo and few pixels sparsely in the Amazonia. These areas are not known to be occupied by the species and were excluded from the potential distribution map of this hummingbird. Some areas in the southern Caribbean and in extreme northeast Colombia likely suitable for this species were not predicted by our model, those areas were noted as such in our maps.
Distribution of specimens according to BioMap suggest a possible area (5,121 km2) of intergradation between subspecies saucerottei and australis in south central Cauca. Also a possible area (7,726 km2) of intergradation between subspecies saucerottei and warscewiczi in central Antioquia (Aburra Valley and north to the end of the Central Andes).
Assuming that the distribution of the species may have filled the complete climatic model generated, its distribution today in remnants of forest is about 101,562 km2, which corresponds to a loss of 70 % of its potential original distribution due to deforestation. Nonetheless, this species favours edges, secondary vegetation and plantations and therefore possibly deforestation has not negatively affected greatly its populations.
Some authors consider that variation in the southwestern Andes is clinal and that subspecies australis is well within variation of the nominate subspecies and therefore it does not deserve to be treated as a separate taxon (HBW Alive, 2017).
Regularized training gain is 1.359, training AUC is 0.920, unregularized training gain is 1.509.
Algorithm terminated after 2000 iterations (96 seconds).
The follow settings were used during the run:
925 presence records used for training.
10915 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: 0.050, categorical: 0.250, threshold: 1.000, hinge: 0.500
Feature types used: hinge product linear threshold quadratic
'Fixed Cumulative Value 5%', 'Equal Training Sensitivity and Specificity' and 'Equate Entropy of Thresholded and Original Distributions' thresholds and omission rates:
0.289-0.146-0.257-Fractional predicted area
0.025-0.146-0.038-Training omission rate