Assessment of ECMWF reanalysis data in complex terrain: Can the CERA‐20C and ERA‐Interim data sets replicate the variation in surface air temperatures over Sichuan, China?

Surface air temperatures (SATs) derived from the European Centre for Medium‐Range Weather Forecasts (ECMWF) ERA‐Interim and CERA‐20C reanalysis data sets are compared with data from 43 observation stations in Sichuan for 1979–2010. The results show (a) the temperatures from the ERA‐Interim and CERA‐...

Full description

Saved in:
Bibliographic Details
Published in:International journal of climatology 2019-12, Vol.39 (15), p.5619-5634
Main Authors: Luo, Haolin, Ge, Fei, Yang, Kangquan, Zhu, Shoupeng, Peng, Ting, Cai, Wenyue, Liu, Xiaoran, Tang, Weiwei
Format: Article
Language:English
Subjects:
Air temperature
Calibration
Climate change
Coefficients
Correlation coefficient
Correlation coefficients
Data
Datasets
ECMWF reanalysis
Elevation
Environmental assessment
High altitude
Lapse rate
observational data analysis
Regional analysis
Regional climates
Regions
Sichuan
Stations
surface air temperature
Surface-air temperature relationships
Topography
Topography (geology)
Weather forecasting
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Surface air temperatures (SATs) derived from the European Centre for Medium‐Range Weather Forecasts (ECMWF) ERA‐Interim and CERA‐20C reanalysis data sets are compared with data from 43 observation stations in Sichuan for 1979–2010. The results show (a) the temperatures from the ERA‐Interim and CERA‐20C data sets are strongly correlated with those from the observation stations, although significant cold biases are seen on both annual and seasonal timescales. (b) The biases in SATs are predominately influenced by the differences between the actual topography and the topography used in the reanalysis models. Larger differences in temperature are observed in the plateau and mountainous regions of Sichuan. We confirmed larger SAT biases at high altitudes by categorizing the elevation into four bands, each with a spacing of 1,000 m. (c) We reduced the biases resulting from elevation by using an elevation correction method with internal lapse rates derived from different reanalysis pressure levels. The annual mean bias was reduced from −2.86 to −0.75°C for the ERA‐Interim data set and from −5.27 to −2.21°C for the CERA‐20C data set. After calibration, the correlation coefficients between the difference in SAT (observed minus reanalysis data) and the difference in elevation (station elevation minus model elevation) decreased from −0.97 and −0.91 to −0.29 and −0.30 for the ERA‐Interim and CERA‐20C data sets, respectively. These significant differences should not be ignored in the application of reanalysis data sets to climate research. The evaluation and calibration of reanalysis data sets are essential before making assessments of regional climate change, especially over regions with complex topography. Temperatures from CERA‐20C and ERA‐Interim correlate quite well with those from observed stations in Sichuan during 1979–2010, but significant cold biases between reanalysis data sets and observations are captured. The cold biases of the surface air temperature derived from ECMWF reanalysis could be mainly attributed to the elevation differences between stations and reanalysis model. It is demonstrated that the elevation correction using model internal lapse rates is an effective method to diminish the influence of altitude differences on temperature biases and calibrate the surface air temperature errors of the reanalysis.
ISSN:0899-8418
1097-0088
DOI:10.1002/joc.6175