The MaxEnt software package is one of the most popular tools for species-distribution modeling. Despite its popularity, researchers usually underestimate the influence of the initial parameters and spatial bias in the occurrence data. The choice of the feature types and regularization multiplier notably affects the spatial “compactness” or “smoothness” of the model. A nonrandom distribution of occurrence points in geographical space requires data correction. The determination of the ecological factors that influence the range formation require the minimization of multicollinearity in the predictor dataset.
5 February 2010 in Biological Journal of the Linnean Society
For a long time, studies of the distribution of living beings in a geographical space were performed only with empirical methods. A change in the view of a species distribution as a projection of a Hutchinsonian ecological niche led to the formation of the discipline of ecological modeling of the species distribution, which switched faunistics/floristics from data accumulation to a full-fledged scientific industry with experiment planning and result verification. The various methods of species-distribution modeling make it possible to analyze the patterns of the geographical distributional of organisms in the presence of methodological challenges: nonrandomness of the occurrence data, inhomogeneity of the collection efforts, landscape heterogeneity in different scales, etc. The results of species-distribution modeling represent spatially continuous data of habitat suitability and are valuable not only for studies of the habitats themselves but also for a number of disciplines that involve species distributions.
Twenty-five chipmunk species occur in the world, of which only the Siberian chipmunk, Tamias sibiricus, inhabits Asia. To investigate mitochondrial cytochrome b sequence variations and population structure of the Siberian chipmunk in northeastern Asia, we examined mitochondrial cytochrome b sequences (1140 bp) from 3 countries. Analyses of 41 individuals from South Korea and 33 individuals from Russia and northeast China resulted in 37 haplotypes and 27 haplotypes, respectively. There were no shared haplotypes between South Korea and Russia--northeast China. Phylogenetic trees and network analysis showed 2 major maternal lineages for haplotypes, referred to as the S and R lineages. Haplotype grouping in each cluster was nearly coincident with its geographic affinity. In particular, 3 distinct groups were found that mostly clustered in the northern, central and southern parts of South Korea. Nucleotide diversity of the S lineage was twice that of lineage R. The divergence between S and R lineages was estimated to be 2.98-0.98 Myr. During the ice age, there may have been at least 2 refuges in South Korea and Russia--northeast China. The sequence variation between the S and R lineages was 11.3% (K2P), which is indicative of specific recognition in rodents. These results suggest that T. sibiricus from South Korea could be considered a separate species. However, additional information, such as details of distribution, nuclear genes data or morphology, is required to strengthen this hypothesis.