Penelope perspicax

NOTES

To elaborate a first model we removed uncertain localities such as 'Bogota skins' and all records with coordinates laying down in sites with estimated elevations out of the range within 800-2,200 m.

The habitat suitability model generated in Maxent showed a few areas that are suitable in climatic terms for this species in the east slope of the Central Andes and the west slope of the Eastern Andes. These areas are known not to be occupied historically by the species and were excluded from the potential distribution map of this guan. Interestingly, in this first exercise we deleted a couple of records from DatAves allegedly from Otun-Quimbaya Sanctuary, which lay down at approximately 2,600 m and which most likely come instead from Ucumari Regional Park or received coordinates from this park; both areas are well known to be important strongholds for the population of this species. Consequently, the model failed to predict areas in both reserve areas.

We repeated the modelling exercise including these localities to elaborate a new model. As result we obtained much better predictions in this particular zone, southwest of Los Nevados National Park. Its is possible that this caused certain degree of overprediction at higher elevations than those from where this species has been usually recorded. Nevertheless, In recent decades apparently this species has been observed at similar elevations (Renjifo et al. 2002), suggesting that this mainly subtropical species has been pushed to use forest remnants at higher elevations than usual.

Assuming that the distribution of the species may have filled the complete climatic model generated (first model), its potential distribution today in remnants of forest is about 13,566 km2, which corresponds to a loss of 73 % of its potential original distribution due to deforestation. If we restrict this model to the most likely historical distribution, the species can be in about 10,391 km2 of forested areas, being lost from 75 % of what originally it may have occupied. Interestingly, these figures just changed slightly when we repeated the modelling exercise (second model), in that case its potential distribution today in remnants of forest may be about 14,764 km2 (10,907 km2 if restricted), which corresponds to a loss of 72 % (74 % if restricted) of its potential original distribution due to deforestation.

Renjifo et al. (2002) estimated this species have lost nearly 95 % of its original habitat. A closer look to our estimation of forested areas showed some inaccuracies for example in the zone between Manizales and Chinchina (Caldas) in the western slope of the Central Andes; where many areas of shade grown coffee were predicted as suitable and forested. De facto, these areas observed from a few kilometres may resemble forests, although they are not known to be used by this guan. This suggests that our estimation of the climatically suitable and forested habitat must be over the actual value. Nevertheless, the inspection of the potential areas left in the Western Andes shows several tracts of forest that indicate potential habitat must be definitely over the value estimated by Renjifo et al. (2002). Thus, it is urgent for this species to define its presence in other localities, particularly in the Western Andes.

Both Renjifo et al. (2002) and BirdLife International (2015) consider the loss of suitable habitat as the most important reason causing the inferred reduction in the population of this guan, while portray hunting pressure at a second level. Although loss of habitat must have been an important factor, differently to those authors we believe that the lack of records from important tracts of forested areas (particularly in the Western Andes) support the idea that, during at least 200 years or more of colonisation of the Cauca valley, this species has been subject to a extremely heavy hunting pressure, which has driven it selectively and locally extinct at many sites part of its historical distribution. Being the biggest bird of game in its range (Renjifo et al. 2002), this species must have been a highly appreciated trophy by colonisers and hunters.

MODEL METADATA

First model

Regularized training gain is 2.650, training AUC is 0.987, unregularized training gain is 3.423.

Algorithm converged after 480 iterations (14 seconds).

The follow settings were used during the run:

18 presence records used for training.

10018 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: 0.481, categorical: 0.250, threshold: 1.820, hinge: 0.500

Feature types used: linear quadratic hinge

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

17.439-13.363-Cumulative threshold

0.187-0.139-Logistic threshold

0.056-0.071-Fractional predicted area

0.056-0.056-Training omission rate

Second model

Regularized training gain is 2.587, training AUC is 0.987, unregularized training gain is 3.345.

Algorithm converged after 620 iterations (17 seconds).

The follow settings were used during the run:

19 presence records used for training.

10019 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: 0.462, categorical: 0.250, threshold: 1.810, hinge: 0.500

Feature types used: linear quadratic hinge

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

20.053-13.463-Cumulative threshold

0.219-0.145-Logistic threshold

0.053-0.075-Fractional predicted area

0.053-0.053-Training omission rate