COMPARATIVE ANALYSIS OF DIFFERENT TECHNIQUES FOR SPATIAL INTERPOLATION OF RAINFALL DATASETS IN DUHOK GOVERNORATE
The spatial distribution of rainfall plays a key role in water management and hydrological modeling. This study employed rainfall datasets of different scales recorded at 22 meteorological stations distributed over Duhok governorate with a time span from 1998 to 2020. Four spatial interpolation methods were employed to transform the discrete values into a continuous spatial pattern over Duhok Governorate, Kurdistan Region, and Iraq. The interpolation methods evaluated were: Inverse Distance Weighted (IDW), Ordinary Kriging (OK), Universal Kriging (UK) and Spline. The reliability of the prediction techniques was examined via leave-one-out cross-validation besides giving ranks using TOPSIS algorithm. Several performance indicators were employed as criteria to evaluate the applied interpolation techniques. The results indicated that the IDW offered the best performance in most cases, while the spline method offered the poorest performance. In general, differences between these first three methods (IDW, OK and UK) are low but substantially different from that the spline method. Overall, the mean absolute percentage of error of estimation was 16% or less and the degree of prediction accuracy tended to increase with a decrease in time scale. Further, the interpolation methods provide similar spatial distributions of rainfall.
Spatial distribution, interpolation methods, rainfall estimation, comparative analysis