Evaluation of reanalyzed surface air temperature over the western Tibetan Plateau

Very few evaluation studies have focused on the western Tibetan Plateau (TP), despite considerable scientific interests from climatologists and glaciologists globally. This region exhibits anomalous glacial stability under climate warming, but the lack of long-term and high-density observational data poses a major research challenge. Evaluating and selecting reanalysis datasets with higher accuracy over the western TP is thereby vital for advancing climate research. Using ground-based 2-m air temperature (T_2m) observations from multiple neighboring countries across the western TP, we establish an integrated observational network to evaluate the performance of widely used reanalysis datasets, including NCEP1, NCEP2, ERA-Interim, JRA55, MERRA2, CRA-Land, CRA, and ERA5.

The results indicate that three reanalysis datasets (CRA-Land, CRA, and ERA5) generally outperform the others in capturing T_2m climatology and year-to-year variability across the western TP as a whole. However, over the Karakoram, the core glacial zone in the western TP, the above three datasets demonstrate the lowest skills in simulating T_2m climatology, although they still perform well in capturing year-to-year variability. For CRA-Land and CRA, the error sources are unclear, although their terrain elevation differences respectively explain ∼67.4 % and ∼69.1 % of their climatological T_2m errors. For ERA5, the terrain elevation differences (explaining ∼58.4 %) and collective radiative processes are the two primary error sources. Moreover, higher surface albedo in ERA5 is identified as the key contributor causing radiative errors. This study identifies the optimization of surface albedo in the Karakoram and its neighboring regions as a critical pathway to enhance ERA5 data quality in the future.