Interpolating Surface Air Temperature for Use in a Semi-distributed Snowmelt Runoff Models

Surface air temperature is an important meteorological input variable for snowmelt runoff models, as well as for models of other hydrologic processes. Methods for incorporating this variable into models range in complexity from simply calculating average values from available weather station measurements and then applying a constant lapse rate, to more spatially-explicit methods, such as ordinary kriging. In complex terrain, such as that found throughout much of Idaho, it is necessary for the method of spatial interpolation to specifically account for orographic effects. Apart from most other past studies that compare spatial interpolation, this research examines techniques at a relatively fine spatial scale (basin size of ~1600 km2) and at a daily temporal scale. This research compares the performance of various temperature spatial interpolation techniques, including the lapse rate method, ordinary kriging of elevationally-detrended data, and the use of climate interpolation models (PRISM).