Chlorostilbon russatus

NOTES

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 below 250 m and above 3,301 m.

The habitat suitability model generated in Maxent showed a very few areas suitable in climatic terms for this species in the Southern Andes and the northern Eastern Andes. Those areas off from the area Sierra Nevada de Santa Marta/serrania del Perija and the northern end of the Eastern Andes are not known to be occupied by the species and were deleted from our final potential distribution map. Otherwise, our model failed two predict three areas that might be important to highlight. A first area around Manaure (La Guajira) and a second and third in the Eastern Andes in serrania de Los Yariguies (Santander) and further south in Lago Fuquene (Cundinamarca). The record from Fuquene is of particular interest since it is a study skin placed in the Field Museum of Natural History, and if correct extends the known range of the species some 215 km to the southwest. This needs further revision and 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 8,676 km2, which corresponds to a loss of 58 % of its potential original distribution due to deforestation.

BirdLife International (2017) has catalogued this hummingbird as of Least Concern (LC) since it is believed it does not approach to the thresholds to be considered Vulnerable (VU) under the range size, population trend or population size criterion. However, this evaluation has a series of pitfalls that must be considered. According to our maps the species has a extent of occurrence of 20,693 km2, which very close to the 20,000 km2 threshold for Vulnerable. Additionally, differently to what is believed in the evaluation of this species the forests part of its range in the eastern slope of Sierra Nevada de Santa Marta and serrania del perija are severely fragmented and forest part of its range continue being threatened by deforestation and fragmentation. On the other hand, the species is believed to be uncommon and although its population has not been quantified it is assumed it does not approach to the thresholds to be Vulnerable (VU), which is below 10,000 mature individuals (BirdLife International 2017). This is extremely uncertain, assuming a regular occupancy of the extent of occurrence estimated here and if the population is 10,000 individuals the average density will be around 0.5 Individuals/km2, giving that we know the species is uncommon maybe this estimation of density may be near the real average density of the species in nature. Until further studies are conducted to understand its dependance on large intact forest tracts and know better its general ecology, and given that its very small range makes it highly vulnerable in biological terms, as a precaution this hummingbird must be reclassified at least as Near Threatened (NT).

MODEL METADATA

Regularized training gain is 3.652, training AUC is 0.991, unregularized training gain is 3.836.

Algorithm converged after 780 iterations (24 seconds).

The follow settings were used during the run:

71 presence records used for training.

10071 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: 0.133, categorical: 0.250, threshold: 1.290, hinge: 0.500

Feature types used: hinge linear quadratic

responsecurves: true

jackknife: true

maximumiterations: 2000

'Equal Training Sensitivity and Specificity' and 'Equate Entropy of Thresholded and Original Distributions' thresholds and omission rates:

ETSS-EETOD-Description

5.806-8.940-Cumulative threshold

0.040-0.084-Logistic threshold

0.039-0.026-Fractional predicted area

0.042-0.042-Training omission rate