A Comparison of Different Interpolation Techniques Used in Mapping Monthly Precipitation in Northern Turkey

Abstract

This study aimed to determine the most suitable method for modelling and mapping monthly precipitation using a geographical information system (GIS) in northern Turkey. Models were constructed using Inverse Distance Weighting (IDW), Spline, Simple Kriging, Cokriging and Linear Regression and were comparatively analyzed for effectiveness in interpolating precipitation data. Data was divided into 2 groups, with 55 stations used for analysis and 14 stations used for validation. Correlation coefficients for estimated and measured mean monthly precipitation varied between 0.69-0.98. Correlation coefficients for all months were found to be statistically significant (p<0.01). The method producing the best results varied by month, as follows: Spline: January-April, October; IDW: June, December (and annual results); Kriging: July-September, November; Linear regression: May. With the exception of linear regression, the differences between methods were statistically insignificant, indicating that IDW, Spline and Kriging (simple and cokriging) can all be used successfully in the preparation of monthly total precipitation maps in the study area.