Establishment and assessment of hourly high‐resolution gridded air temperature data sets in Zhejiang, China

Abstract

A high temporal–spatial resolution (1 hr/1 km) gridded temperature data set with a long time‐series (2006–2015) is established for Zhejiang province, China, which has a complex terrain, including plains, hills, basins, mountains, islands and oceans. A comprehensive analysis method is used to combine regional surface station data with reanalysis data from the European Centre for Medium‐Range Weather Forecasts (ECMWF) such that the observations at the stations are reproduced, and the reanalysis data provide the spatial structure for the interpolation. The elevation correction method and inverse distance‐weighted interpolation method are selected for comparison. The quality of the data set is evaluated and analysed by comparison with the cross‐validation results, third‐party data validation and other gridded data sets. The results show that the comprehensive analysis method provides the best outcome, followed by the inverse distance‐weighted method, and the reanalysis data elevation correction method has the worst result. The temperature error obtained from the gridding methods increases with altitude, decreases with increasing station density and varies with different synoptic conditions. The obtained data set can describe the temperature feature of peaks and valleys more precisely, which produces lower errors and higher spatial correlation co‐efficients compared with observation than the other gridded data sets. Therefore, the established data set can meet the needs of fine meteorological businesses and has potential applications in many fields.