Examination of potential biases in air temperature caused by poor station locations

Questions have been raised about whether poor siting practices that have existed in recent years are causing a bias in the U.S. temperature record. This potential bias was examined using homogeneity adjusted mean temperature data from five stations in eastern Colorado, two with good current siting and three with poor current siting. No siting induced bias was found in the homogeneity adjusted data. One of the two stations with good current siting had no homogeneity adjustments while the other only had small adjustments to account for minor changes in observing time, so their raw and homogeneity adjusted time series are very similar. By contrast, the stations with poor siting experienced changes in time of observation, instrumentation and location resulting in their homogeneity adjusted time series being very different from their unadjusted time series. Therefore, these results indicate that homogeneity adjustments of time series from the stations with poor current siting work well at making their data very similar to data from stations that have excellent siting and are clearly representative of the region as a whole.