Spatial interpolation of air temperature in Himalayas

Abstract

The daily air temperature measured at few stations in Sutlej basin need to be interpolated spatially across the basin to account for spatial variability in the proposed snowmelt runoff model. GIDS algorithm combines the multiple regression and inverse-distance-squared approaches in which the inputs includes location and elevation details of the stations with reference to inverse squared distance between each station. The model was created in GIS environment and executed daily for April and May by varying the coefficients generated using MATLAB software with 30m resolution. The model contains two modules, with Kaza and Racksham as check stations. Statistical analysis result shows, according to Kaza station the percentage of error is +/− 5%. The standard deviation for both the months is 0.60 and 0.44. In case of Racksham, the percentage of error is within +/− 20%. The standard deviation for both the months is 3.4 and 1.12. Based on analysis, GIDS model was governed by the range of inputs like spatial distance, elevations and temperatures.The computed temperatures were compared with those derived using lapse rate method (LRM). On comparison GIDS produced better results than LRM. Thus the study concludes that GIDS can be used for interpolation with certain boundary conditions.