Assessment and Applicability Analysis of CLDAS Temperature Products in Gansu Province

Wang Yaping; Wang Suichan; Zhang Jin

doi:10.16765/j.cnki.1673-7148.2025.02.010

Abstract：Based on the hourly and daily temperature data,the quality of CLDAS air temperature products from December 1,2020,to November 30,2021 in Gansu Province and its applicability in weather and climate events are evaluated through statistical errors in different source testing data and multi time scales.The results show that the errors of all CLDAS temperature products are small and their correlation coefficients exceed 0.98.The quality of indicators in non-independent source test is better than in independent source test,and that of national stations is better than that of regional stations.The RMSE of CLDAS at independent source testing stations,national stations and regional stations are 1.502℃,0.908℃and 1.499℃,respectively.The sample size with negative deviation of CLDAS hourly tempera-ture,daily maximum temperature and daily minimum temperature is 50.49%,64.78%and 44.69%.The daily maximum temperatures are underestimated and the daily minimum temperatures overestimated.The RMSE of daily minimum temperature is higher than that of daily maximum temperature in winter,but the situation is on the contrary in summer.Larger errors tend to appear in the temperature rising processes in winter and spring,with the RMSE reaching 3.0℃or higher.There are similar diurnal variations of CLDAS temperature errors in all the four seasons,but the time getting to the lowest temperature is differ-ent.It is overestimated at sunrise but underestimated at sunset.The statistical error at night is higher than in the daytime in winter,but it is opposite in the summer time.All CLDAS products have smaller spatial differences.There are more stations with positive deviations for hourly and daily minimum temperatures but more stations with negative deviations for the daily maximum temperatures.The CLDAS temperature products can characterize the spatial and temporal variations of the mean temperature and the large scale of cold-wave process,and its difference from observation is in the range of±1.0℃.For high temperature and heat wave events,however,the coincidence rate is about 50%.In general,the quality of CLDAS temperature products is good,and it can be used in climate analysis and assessment,but the accuracy needs improving for the temperature-rise and scorching weather processes.