Lafresnaya lafresnayi

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 1,338 m and above 3,877 m. Also one record in BioMap from San Miguel that was misplaced in Meta instead of Magdalena/La Guajira was deleted.

The habitat suitability model generated in Maxent showed very good fit not showing suitable areas out of the main Andean Ranges and the Sierra Nevada de Santa Marta.

Distribution of specimens according to BioMap suggest a zone of intergradation (25,942 km2) between subspecies saul and longirostris or saul and the nominate subspecies in the southern Andes from Valle del Cauca/Tolima south to Nariño/Putumayo. This area of the Central Andes needs collections to clarify better which subspecies are present.

Assuming that the distribution of the species may have filled the complete climatic model generated, its distribution today in remnants of forest is about 53,229 km2, which corresponds to a loss of 64 % of its potential original distribution due to deforestation.

Subspecies liriope that inhabits the unique Sierra Nevada de Santa Marta has a very small extent of occurrence, estimated in 4,648 km2. Despite the species is fairly common and uses border and secondary habitats, subspecies liriope must be considered of some importance for conservation if is intended its survival in Colombia.

MODEL METADATA

Regularized training gain is 1.989, training AUC is 0.960, unregularized training gain is 2.180.

Algorithm terminated after 2000 iterations (70 seconds).

The follow settings were used during the run:

293 presence records used for training.

10291 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

responsecurves: true

jackknife: true

maximumiterations: 2000

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

17.143-5.797-Cumulative threshold

0.337-0.221-Logistic threshold

0.097-0.136-Fractional predicted area

0.096-0.041-Training omission rate