Comparison of Different Kriging Methods to Estimate the Tree Density. A Case Study: West of Karkheh in Southwest of Iran
Abstract
Kriging the geostatistical methods used to create continuous maps of spatially auto correlated characteristic, the ability to estimate values at unsampled locations with regard to their spatial correlation values. The attributes and location (spatial coordinates) of trees is important because these factors lead to spatial heterogeneity. Therefore, this various modeling techniques have been applied to discovery of spatial heterogeneity in environmental and forest studies. In this study, geostatistical methods of Simple Kriging and Cokriging to interpolate the tree density in the forest of Beyt - Qasham in 4350 ha in the margin of Karkheh river of Khuzestan province were comparison and evaluated. Field sampling was performed based on a systematic random sampling a circular plot with intensity of each 100 hectare in the 1000 m2 area, tree density information and geographic coordinates of each plot center were recorded. The performance assessment of the above mentioned was carried out using cross-validation techniques and criteria for the comparison between their ME, RMSE, ASE, MSE, RMSSE were used. The results showed that the tree density have been high spatial correlation in this region. The cross validation results also showed that among the used methods, the Cokriging method is superior accuracy with minimum of ME, RMSE, ASE, MSE and RMSSE.
