资源环境科技发展态势分析平台

Diurnal cycle of surface air temperature T is an important metric indicating the feedback of land-atmospheric interaction to global warming, whereas the ability of current reanalyses to reproduce its variation had not been assessed adequately. Here, we evaluate the daily maximum temperature T-max, daily minimum temperature T-min, and diurnal temperature range (DTR) in five reanalyses based on observations collected at 2253 weather stations over China. Our results show that the reanalyses reproduce T-min very well; however, except for Modern-Era Retrospective Analysis for Research and Applications version 2 (MERRA-2), they substantially underestimate T-max and DTR by 1.21 degrees-6.84 degrees C over China during the period of 1980-2014. MERRA-2 overestimates T-max and DTR by 0.35 degrees and 0.81 degrees C, which are closest with observation. The reanalyses are skillful in reproducing the interannual variability of T-max and T-min but relatively poor for DTR. All reanalyses underestimate the warming trend of T-min by 0.13 degrees-0.17 degrees C (10 yr) 21 throughout China during 1980-2014, and underestimate the warming trend of T-max by 0.24 degrees-0.40 degrees C (10 yr)(-1) in northwestern China while overestimating this quantity by 0.18 degrees-0.33 degrees C (10 yr)(-1) in southeastern China. These trend biases in T-max and T-min introduce a positive trend bias in DTR of 0.01 degrees-0.26 degrees C (10 yr)(-1) within China, especially in the north China plain and southeastern China. In the five reanalyses, owing to the sensitivity discrepancies and trend biases, the surface solar radiation R-s and precipitation frequency (PF) are notable deviation sources of the diurnal cycle of air temperature, which explain 31.0%-38.7% (31.9%-37.8%) and 9.8%-22.2% (7.4%-15.3%) of the trend bias in T-max (DTR) over China, respectively.