Accuracy assessment of spatial interpolations methods using ArcGIS

¹ Climatology Department, School of Meteorology, Climatology, and Geophysics (STMKG), Banten, Indonesia
² Taiwan International Graduate Program (TIGP), Earth System Science Program, Academia Sinica, Taipei, Taiwan
³ Graduate Institute of Applied Geology, National Central University (NCU), Taoyuan City, Taiwan
⁴ Education and Culture Board of Medan City Government (PEMKO MEDAN), North Sumatra, Indonesia

^* Corresponding author: yahya.darmawan@bmkg.go.id

Abstract

The Inverse Distance Weight (IDW), Spline, and Ordinary Kriging (OK) methods are widely used for interpolating spatial data. Many GIS software options are available, particularly for spatial interpolation. This study aims to assess the accuracy of interpolation approaches based on Climate Hazards Group Infrared Precipitation with Station (CHIRPS) data in ArcGIS 10.8. The research used monthly precipitation data extracted from 147 rain gauge station locations in North Sumatra Province. Additionally, eight stations from BMKG's network were excluded from the interpolation process and instead used as validation data. Validation was conducted by calculating error statistics between the predicted and actual data at these excluded points. The predicted data was generated by interpolating values from the 147 rain gauge stations, excluding the eight points corresponding to the validation data locations. The interpolation was done using ArcGIS 10.8 with the IDW, Spline, and Kriging toolbox. A comparison was performed by assessing error parameters and Pearson correlation coefficients. According to the accuracy assessment, both IDW and OK show comparable performance; both are better than spline interpolation. In addition, the box plot chart demonstrated that the spline technique had a tendency to generate outlier values along the perimeters of the study area.