Comparison of Geostatistical Methods to Determine the Best Bioclimatic Data Interpolation Method for Modelling Species Distribution in Central Iran

• Main Authors: R. Khosravi, M. R. Hemami, M. Malekian
• Format: Article
• Language: fas
• Published: Isfahan University of Technology 2014-09-01
• Series: Iranian Journal of Applied Ecology
• Subjects: Interpolation; Cokriging; Bioclimatic variables; Distribution modeling; Semivariogram; Central Iran.
• Online Access: http://ijae.iut.ac.ir/browse.php?a_code=A-10-1-58&slc_lang=en&sid=1

Climatic change can impose physiological constraints on species and can therefore affect species distribution. Bioclimatic predictors, including annual trends, regimes, thresholds and bio-limiting factors are the most important independent variables in species distribution models. Water and temperature are the most limiting factors in arid ecosystem in central Iran. Therefore, mapping of climatic factors in species distribution models seems necessary. In this study, we describe the extraction of 20 important bioclimatic variables from climatic data and compare different interpolation methods including inverse distance weighting, ordinary kriging, kriging with external trend, cokriging, and five radial basis functions. Normal climatic data (1950-2010) in 26 synoptic stations in central Iran were used to extract bioclimatic data. Spatial correlation, heterogeneity and trend in data were evaluated using three models of semivariogram (spherical, exponential and Gaussian) and the best model was selected using cross validation. The optimum model for bioclimatic variables was assessed based on the root mean square error and mean bias error. Exponential model was considered to be the best fit mathematical model to empirical semivariogram. IDW and cokriging were recognised as the best interpolating methods for average annual temperature and annual precipitation, respectively. Use of elevation as an auxiliary variable appeared to be necessary for optimizing interpolation methods of climatic and bioclimatic variables.