Spatial interpolation of climate variables in Nebraska

Temperature and rainfall are important climatological parameters, and knowledge of their temporal and spatial patterns is useful for researchers working in many disciplines. In this study, spatial interpolation techniques were implemented in a Geographic Information System (GIS) to study the spatial variability of climate variables (maximum air temperature, minimum air temperature, and seasonal and annual rainfall) in Nebraska. Thirty years (1971-2000) of climate data (average monthly maximum and minimum temperatures and rainfall) from 215 National Weather Service Cooperative Observer Network (COOP) weather stations distributed throughout Nebraska and surrounding states were used in the analyses. Literature suggests that there is no single preferred method of interpolation, and the selection of interpolation method is usually based on the available data, desired level of accuracy, and available resources. We analyzed three different commonly used interpolation methods (inverse distance weighted, spline, and kriging) and evaluated their performance. Overall, the summary of all statistical parameters showed no significant difference between interpolation techniques in predicting the spatial variability in 30-year climate normals. Investigation of interpolation errors at individual weather stations agreed with summary statistics. Spatial variability, in this instance, is likely smoothed due to long-term averaging of the data (30 years), resulting in similar errors for all the interpolation techniques. Subjective assessment of maps for all climate variables showed that the kriging method produced smoother maps compared to spline and inverse distance weighted. Considering the degree to which accurate spatial interpolation could be accomplished with relative ease and less bias, the spline method proves the better option.