A Comprehensive Evaluation of Surface Air Temperature Reanalyses over China against Urbanization Bias–Adjusted Observations

TL;DR: In this article, the authors compare the differences in surface air temperature (SAT) between observational that has been adjusted for urbanization bias and reanalysis data (NCEPV1, N CEPV2, ERA5, CFSR, MERRA, JRA55, 20CRV3 and CRA40) over mainland China during 1961-2015.

About: This article is published in Advances in Climate Change Research.The article was published on 2021-02-12 and is currently open access. It has received 13 citations till now.

Summary

Figures

• Figure 1 Distribution of the elevation distribution and 763 observation stations across mainland China.
• The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.
• This map has been provided by the authors.
• (The black dots represent signi cant deviation, the value of color bar means the annual mean SAT bias (units: °C)) Note:.

1961–2015 across China:(a) NCEPV1; (b) JRA55; (c) 20CRV3 (d) OBS. (The black dots represent

• Signi cant deviation, the value of color bar means the linear trend bias and linear trend (units: °C/10year)).
• This map has been provided by the authors.
• The biases of climate state and linear trends of annual mean temperature between the REAs and OBS for the period 1979–2015.
• The orange dots represent the altitude of stations between 500-1500m.

Figures

Table 5. Ranking and score of each of the reanalysis datasets in mainland China for the periods 742 1961–2015 and 1979–2015 743

Figure 1. Distribution of the elevation distribution and 763 observation stations across mainland 748 China. (The insert shows the number of stations at various elevations, the asl stands for elevation 749 (units: m), the red bar stands for the number of the elevation distribution) 750

Table 1. Atmospheric reanalysis datasets applied in this study 722

Figure 5. Annual mean SAT and annual mean SAT linear trends for the periods 1961–2015 and 768 1979–2015. (The red bar stand for climate state (units: °C), the orange bar stands for linear trends 769 (units: °C/10 years)) 770

Figure 4. Wavelet analysis based on REAs and OBS from 1961–2015: (a) NCEPV1; (b) JRA55; 763 (c) 20CRV3; (d) CRA40; (e) NCEPV2; (f) ERA5; (g) CFSR ; (h) MERRA ; (i) OBS. 764

Table 2. Relative bias between the REAs and OBS for climate state and linear trends for the 726 periods 1961–2015 and 1979–2015 727

Figure 9. The biases of climate state and linear trends of annual mean temperature between REAs 791 and OBS for the period 1961–2015. (a) NCEPV1; (b) JRA55; (c) 20CRV3. 792 (The asl stands for elevation (units: m); y-axis stand for the bias of climate state (units: °C); x-axis 793 stand for the bias of linear trends (units: °C/10 years); red bar stand for the number of stations with 794 different bias of linear trends; blue bar stand for the number of stations with different bias of climate 795 state; purple dots represent the altitude of stations above 3000m;blue dots represent the altitude of 796 stations between 1500-3000m; orange dots represent the altitude of stations between 500-1500m; 797 green dots represent the altitude of stations less than 500m) 798

Figure 8. Spatial distribution of linear trend bias between REA and OBS for the period 1979–2015 786 across China:(a) NCEPV1; (b) JRA55; (c) 20CRV3; (d) CRA40; (e) NCEPV2; (f) ERA5; (g) CFSR; 787 (h) MERRA.(The black dots represent significant deviation, the value of color bar means the linear 788 trend bias (units: °C/10year)) 789

Figure 10. The biases of climate state and linear trends of annual mean temperature between the 801 REAs and OBS for the period 1979–2015. (a) NCEPV1; (b) JRA55; (c) 20CRV3; (d) CRA40; (e) 802 NCEPV2; (f) ERA5; (g) CFSR; (h) MERRA. 803 (The asl stands for elevation (units: m); y-axis stand for the bias of climate state (units: °C); x-axis 804 stand for the bias of linear trends (units: °C/10 years); red bar stand for the number of stations with 805 different bias of linear trends; blue bar stand for the number of stations with different bias of climate 806 state; purple dots represent the altitude of stations above 3000m;blue dots represent the altitude of 807 stations between 1500-3000m; orange dots represent the altitude of stations between 500-1500m; 808 green dots represent the altitude of stations less than 500m) 809

Table 3. Standard deviation ratio (SDR) and correlation coefficient (R) of annual mean SAT across 731 China derived from REAs and OBS for the periods 1961–2015 and 1979–2015 (thresholds for the 732 95% confidence level are 0.26 and 0.32 respectively) 733

Figure 3. The bias of monthly mean SAT between REAs and OBS from 1961–2015 across 758 mainland China: (a) NCEPV1; (b) JRA55; (c) 20CRV3; (d) CRA40; (e) NCEPV2; (f) ERA5; (g) 759 CFSR; (h) MERRA. (The value of color bar means the monthly mean SAT bias (units: °C)) 760

Figure 11. Taylor diagrams of annual mean SAT for 763 stations across mainland China derived by 812 the REAs and OBS for the period 1961–2015: (a) NCEPV1; (b) JRA55; (c) 20CRV3. 813 (The asl stands for elevation (units: m), the purple dots represent the altitude of stations above 814 3000m. The blue dots represent the altitude of stations between 1500-3000m. The orange dots 815 represent the altitude of stations between 500-1500m. The green dots represent the altitude of 816 stations less than 500m, threshold for the 95% confidence level is 0.26) 817

Figure 7. Spatial distribution of linear trend of OBS and linear trend bias between REA and OBS 780 for the period 1961–2015 across China:(a)NCEPV1; (b) JRA55; (c) 20CRV3 (d) OBS. 781 (The black dots represent significant deviation, the value of color bar means the linear trend bias 782 and linear trend (units: °C/10year)) 783

Figure 2. Time series of annual mean SAT across mainland China from 1961–2015 derived from 754 OBS and REAs: (a)SAT; (b) SAT anomalies; (c) SAT with linear trends removed. (units: °C) 755

Table 4. R and SDR values between the REAs and OBS at different elevations for the periods 737 1961–2015 and 1979–2015 738

Figure 12. Taylor diagrams of annual mean SAT for 763 stations across mainland China derived 820 by the REAs and OBS for the period 1979–2015:(a) NCEPV1; (b) JRA55; (c) 20CRV3; (d) CRA40; 821 (e) NCEPV2; (f) ERA5; (g) CFSR; (h) MERRA. 822 (The asl stands for elevation (units: m), the purple dots represent the altitude of stations above 823 3000m. The blue dots represent the altitude of stations between 1500-3000m. The orange dots 824 represent the altitude of stations between 500-1500m. The green dots represent the altitude of 825 stations less than 500m, threshold for the 95% confidence level is 0.32) 826

Figure 6. Spatial distribution of annual mean SAT bias of REAs from observations for the period 773 1979–2015 across mainland China:(a)NCEPV1; (b) JRA55; (c) 20CRV3; (d) CRA40; (e) NCEPV2; 774 (f) ERA5; (g) CFSR; (h) MERRA 775 (The black dots represent significant deviation, the value of color bar means the annual mean SAT 776 bias (units: °C)) 777

Figure 1. Distribution of the elevation distribution and 763 observation stations across mainland 748 China. (The insert shows the number of stations at various elevations, the asl stands for elevation 749 (units: m), the red bar stands for the number of the elevation distribution) 750

Frequently asked questions

Q1. How many reanalysis datasets have been developed?

The biases in 70 inter–annual variability, time series, and spatial distribution characteristics between reanalysis and 71 observational data have been analyzed. 

Q2. What is the overall performance of the REAs in the period 1979–2015?

The integrated score of 469 20CRV3 was 92.5 and 93.5 during the periods of 1979–2015 and 1961–2015 respectively, with the 470 better performance than other REAs in linear trend (1979–2015), dispersion (1961–2015), and 471 climate state (1961–2015). 

Q3. What are the disadvantages of the reanalysis?

Since these reanalyses have the disadvantages of incorporating the errors from 53 the numerical prediction models, the assimilation processes, and the observation systems 54 (Bengtsson et al., 2004a, 2004b; Zhao et al., 2010; Zhao et al., 2015), the reliability of the reanalysis 55 data needs to be proven on global and regional scales. 

Q4. How did the authors evaluate the applicability of the reanalysis data?

In this study, the authors evaluate the applicability of the reanalysis data by using the urbanization 125 bias–adjusted observational data from 1961–2015 across mainland China. 

Q5. What are the REA biases in correlation, standard deviation, climate state and linear trends?

527 (4) The REA biases in correlation, standard deviation, climate state and linear trends generally 528 increase with increasing elevation of stations. 

Q6. What is the average of the data used in the reanalysis?

The time series, correlation, standard deviation, 177 climate state, and linear trends of the reanalysis data are evaluated against the observations. 

Q7. What are the reasons for the lower SATs of CRA40, NCEPV1?

The 512 cold biases of CRA40, NCEPV1 and JRA55 are mainly caused by the lower SAT in summer, 513 whereas those of NCEPV2, CFSR, and ERA5 are mainly caused by the lower winter mean 514 temperature.