Reduction of uncertainties in surface heat flux over the Tibetan Plateau from ERA-Interim to ERA5

Serving as the successor to the ERA-Interim dataset, a new global reanalysis (i.e., ERA5) has been released by the European Centre for Medium-Range Weather Forecasts. However, evaluation of the performance of ERA5 in relation to the Tibetan Plateau (TP) remains lacking. Using in situ eddy covariance observations from eight stations located on the TP, this study evaluated the uncertainties of both ERA5 and ERA-Interim. Results showed that ERA5 effectively corrects the overestimation of both sensible heat flux (H) and latent heat flux (LE) over the TP, which is severe in ERA-Interim. In most seasons, the bias errors (BEs) of H and LE are generally smaller in ERA5 than in ERA-Interim, although the BEs of wintertime H and LE in spring and autumn are not decreased significantly from ERA-Interim to ERA5. The spatiotemporal relationships with the Global Land Data Assimilation System (GLDAS) and the Global Land Evaporation Amsterdam Model (GLEAM) reanalysis data were investigated. Analysis of correlation coefficients and root mean squared errors revealed large inconsistency in H between ERA5/ERA-Interim and GLDAS, but good agreement in LE between ERA5/ERA-Interim and GLEAM. Overall, the level of agreement of ERA5 with GLEAM was better than that of ERA-Interim. Anomaly correlation coefficients (ACCs) between the two ERA datasets (ERA5 and ERA-Interim) and the two reference datasets (GLDAS and GLEAM) were also investigated. The ACC results further confirmed that ERA5 has better agreement than ERA-Interim with GLEAM in terms of LE over the TP.