Evaluation And Correction Of ERA-interim Air Temperature Data Over The Qilian Mountains Of China

Global climate change is the focus of the whole society.The Qilian Mountains(QLM)has important influence on ecological environment of western China due to its special geographical complex terrain and climate features.However,the conventional observation data are not enough for climate change researches because of limited meteorological stations.The research contents in this study are follows as:(1)Evaluating ERA-Interim analysis temperature data derived from the European Centre for Medium-Range Weather Forecast(ECMWF)in the period of 1979-2017 based on Bias,Correlation coefficient(r)and Root mean square error(RMSE)indices using observational meteorological data from 24 meteorological stations in the Qilian Mountains;(2)Correcting ERA-Interim data using temperature lapse rate method T_t=T_ref+(38)×?h,which is based on the ERA-Interim temperature data and geopotential heights at different pressure levels.Finally,Bias,r and RMSE are used to test the errors between corrected data and uncorrected data.The results are as follows:(1)In general,ERA-Interim captures the monthly,seasonal and annual variation very well.High daily correlations ranging from 0.956 to 0.996 indicate that ERA-Interim captures the observations at daily scale very well.However,an average root-mean-square error(RMSE)of 2.7?of all stations reveals that ERA-Interim should not be directly applied at individual sites.The biases are mainly attributed to the altitude differences between ERA-Interim grid points and stations.A significant positive trend(0.457?/decade)is found over the Qilian Mountains based on the observations in1979-2017.ERA-Interim captures the warming trend very well with an increase rate of0.384?/decade.The observations and ERA-Interim both show the largest positive trends in summer with the values of 0.552?/decade and 0.481?/decade,respectively.We conclude that in general ERA-Interim captures the observed 2 m air temperature trend very well and ERA-Interim is reliable for climate change investigation over the Qilian Mountains under the premise of cautious interpretation.(2)Generally,the five temperature lapse rates are around-6.0?/km,which is approximately according with general environmental temperature lapse rate.Observational temperature lapse rate and ERA-Interim temperature lapse rate are different at different months.Variation amplitude of temperature lapse rate in June and July are larger than other months.From seasonal scale,the absolute value of temperature lapse rate in summer and autumn are higher(>6?/km)than spring and winter(<6?/km).(3)The five correction methods can correct ERA-Interim data effectively via comparing the Bias between corrected data and uncorrected data.From seasonal Bias,the corrective results in spring and autumn are better than summer and winter.The mean absolute deviation of five methods in four seasons is 0.77?,0.71?,0.67?,0.68?and0.24?,respectively.The correlation coefficient(r)of five correction methods performed well,and the r values of most stations exceed 0.9.However,the r values between corrected and uncorrected data changed insignificant.The values of RMSE between corrected and uncorrected data indicate the best results.The corrected RMSE in 19stations are reduced when compared with uncorrected RMSE,which indicates that temperature lapse rate have an important effect.The mean RMSE of five methods in 24meteorological stations is 0.98?,0.94?,0.91?,0.92?and 0.70?,respectively.Finally,ERA-Interim and corrected temperature can capture the observational temperature data well from temperature increasing trend.In general,the lapse rate based on the observational data is best when compared with other four methods,but this method is independent with meteorological stations,so the lapse rates calculated form ERA-Interim can be functioned and used widely for these areas where the meteorological stations are rare.We can learn that Method?is the best one(based on the lapse rate(38)₇₀₀₈₅₀),which could be extended to other mountain regions.