Comparison of spatial interpolation methods for estimating the rainfall in Lumbini Province, Nepal

Abstract

For strategic management and long-term use, it is important to have a better understanding of the spatial distribution of rainfall data. The current study assesses the geographic distribution of rainfall over Nepal’s Lumbini Province, 39 data sites were used for the study. The spatial distribution of rainfall was obtained by interpolating the observed data using the various geostatistical interpolation algorithms offered by ArcGIS. The Ordinary Kriging (OK), Simple Kriging (SK), Inverse Distance Weighting (IDW), and Radial Basis Function (RBF), spatial interpolation methods were investigated. Based on the cross-validation results and the statistical parameters Correlation (R) and Root Mean Square Error (RMSE), the performances of different interpolation algorithms were assessed with R and RMSE values of 0.61 and 26.35 respectively. The Ordinary Kriging seems to have performed well for the region. The Ordinary Kriging interpolation method was used to map the spatial distribution of rainfall over the low and high lands of Lumbini Province. The results showed that the average annual rainfall during 22 years period ranged between (1058 to 2555 mm). At the Meantime, province’s southwest region has lower rainfall of 1058 to 1431 mm, whilst high rainfall in the range of 2046 to 2555 mm occurs in the province’s northeast region and appears to be gradually decreasing towards the lower altitude. Additionally, seasonal precipitation range of pre-monsoon, monsoon, winter monsoon and post monsoon are observed. Furthermore, pre-monsoon, monsoon and winter monsoon follow the highly similar pattern of overall precipitation pattern throughout the years. However, distribution of post monsoon rainfall pattern is discreate over the territory.