Spatial interpolation of climatic variables in a predominantly arid region with complex topography

The benefits of accurately interpolating spatial distribution patterns of precipitation and temperature are well recognized. However, precipitation and temperature patterns are difficult to understand in a region that has complex topography and poor meteorological information. In this study, geostatistical and deterministic interpolation methods are used to understand the best modeling approach for mapping precipitation and temperature in a data-scarce arid region of Pakistan, where elevation and climate vary widely within a short distance. Long-term climate data collected from 15 metrological stations distributed over the Balochistan province of Pakistan are used for this purpose. The performances of various deterministic and geostatistical methods are assessed by using root mean squared errors in interpolation. The results show a difference in accuracy among interpolation methods. Incorporation of elevation significantly improves the accuracy of the interpolation of climate variables. The study concluded that the most preferable models for reliable mapping of precipitation and temperature for such region are disjunctive and universal cokriging