Time Zone Dependence of Diurnal Cycle Errors in Surface Temperature Analyses

Abstract

Surface temperatures from both the NCEP analysis and ECMWF Re-Analysis (ERA-Interim) in January 2008 over the Africa?Eurasian region were compared with surface station measurements to study analysis errors in the diurnal cycle, with data sampled at 3-h time intervals. The results show the dominance of the diurnal cycle in surface temperature analyses and the significance of the time zone dependence of diurnal cycle errors in global analyses. While a distinct diurnal cycle of surface temperature is evidenced in all three datasets, with a nearly constant thermal response of about 2 h in surface station data, the thermal response in the two global analyses varies with longitude. A smaller error in analysis is found in time zones where the observed maximum temperatures (Tmax) occurred close to a global analysis time (e.g., 0000, 0600, 1200, or 1800 UTC). The thermal response errors in surface temperature analyses increase linearly with the difference between the observed diurnal peak time and the analysis time closest to the peak time. Large similarities between the results obtained using both NCEP and ERA-Interim analyses give confidence in the significance of these results. The results are potentially useful for forecast verification, surface data assimilation, and climate research.