• temperature estimates;
  • spatial interpolation;
  • quality control;
  • quality assurance;
  • lapse rate


The inverse distance weighting (IDW) and spatial regression test (SRT) methods provide data estimates for a station of interest based on the measurements at neighbouring stations. This paper evaluates the performance of the two approaches across the USA in estimating maximum and minimum daily temperature where the estimates are compared to actual measured data. The performance of these approaches was assessed using the coefficient of efficiency, explained variance, root mean square error, systematic and non-systematic errors. The t-test and variance test were also used to compare the performances of the two methods. In addition, two other versions of the IDW were tested. The first IDW modification was intended to determine the importance of adding a lapse rate correction to the surrounding stations. The second IDW modification used the intermediate estimates from the SRT method and therefore, by comparison to SRT, showed the relative importance of using SRT weights. The spatial regression approach was found to be superior to all versions of the IDW method especially in the coastal and mountainous regions. The spatial regression approach successfully resolves the systematic differences caused by temperature lapse rate with elevation, which is not accounted for in the inverse distance weighting method. Both the SRT and the IDW methods are found to perform relatively poorly when the weather station density is low. Copyright © 2007 Royal Meteorological Society