An evaluation of methods for modelling distribution of Patagonian insects

Abstract

Various studies have shown that model performance may vary depending on the species being modelled, the study area, or the number of sampled localities, and suggest that it is necessary to assess which model is better for a particular situation. Thus, in this study we evaluate the performance of different techniques for modelling the distribution of Patagonian insects. We applied eight of the most widely used modelling methods (artificial neural networks, BIOCLIM, classification and regression trees, DOMAIN, generalized additive models, GARP, generalized linear models, and Maxent) to the distribution of ten Patagonian insect species. We compared model performance with five accuracy measures. To overcome the problem of not having reliable absence data with which to evaluate model performance, we used randomly selected pseudo-absences located outside of the polygon area defined by taxonomic experts. Our analyses show significant differences among modelling methods depending on the chosen accuracy measure. Maxent performed the best according to four out of the five accuracy measures, although its accuracy did not differ significantly from that obtained with artificial neural networks. When assessed on per species basis, Maxent was also one of the strongest performing methods, particularly for species sampled from a relatively low number of localities. Overall, our study identified four groups of modelling techniques based on model performance. The top-performing group is composed of Maxent and artificial neural networks, followed closely by the DOMAIN technique. The third group includes GARP, GAM, GLM, and CART, and the fourth best performer is the BIOCLIM technique. Although these results may allow obtaining better distributional predictions for reserve selection, it is necessary to be cautious in their use due to the provisional nature of these simulations.

Modeling Potential Distribution of Oligoryzomys longicaudatus, the Andes Virus (Genus: Hantavirus) Reservoir, in Argentina

“…The extensive time period may hide temporal dynamical changes that might be detected by finer resolution climate information but generating a new nationwide data set would, itself, be subject to further concerns of its accuracy. These climate data were chosen, in part, because of the consistent methods in the construction (Hijmans et al 2005) and because they have been applied successfully in earlier species distribution modeling studies in Argentina (Jayat et al 2009;Tognelli et al 2009;Martin 2010;Torres and Jayat 2010). The potential distribution maps for O. longicaudatus predicted the highest occurrence probabilities along the Andes range, from 32°S and narrowing southwards.…”

Development and Comparison of Species Distribution Models for Forest Inventories

“…Results showed that new methods, such as maximum entropy (MaxEnt), have greater predictive power than other methods, such as logistic regression (both adjusted generalized linear models, GLM, and adjusted generalized additive models, GAM). Subsequent studies have also obtained better predictive capacity for MaxEnt than for logistic regression [7][8][9][10][11].…”

Consequences of land use change on bird distribution at Sakaerat Environmental Research Station

“…edu/~schapire/maxent), the maximum-entropy approach for species habitat/distribution modeling Dudik 2008) to estimate the probability occurrence of Black-crested Bulbul because it requires only presence data and environmental information. The Maxent is considered to be better than other methods using presence-only data (Elith et al 2006, Tognelli et al 2009) and performs best when few presence records are available (Wisz et al 2008).…”

“…We selected the Maxent model to predict potential occurrence for Black-crested Bulbul because it requires only presence data and habitat factors and its performance is considered to be better than other methods using presence-only data (Elith et al 2006, Tognelli et al 2009). This performance can be observed by the evaluation of predictive accuracy using the independent occurrence data (testing data) of 20% of the total presence observations.…”