An improved climate division database for the Conterminous United States

The U.S. Climate Division database is widely employed in climate research and monitoring activities in the United States. Unfortunately, trends derived from the database can be significantly biased by the different methods used to compute the averages for the 344 divisions over the period 1895-present. Consequently, the purpose of this paper is to describe a new approach for computing climate division averages and to then compare the new divisional database with its predecessor. In brief, the new approach involves creating 5 km gridded fields for each year and month, then averaging the gridded values for each division. The gridded fields are developed using an approach known as “climatologically aided interpolation,” which entails the combination of gridded base-period normals (produced with thin plate smoothing splines) and gridded anomalies for each year and month (produced via inverse distance weighting interpolation). The new database exhibits more stable and accurate long-term trends that are consistent with other databases, with the largest improvements occurring in the first third of the 20th century.