Spatio-temporal modes of global land surface air temperature variability investigated with ERA5-Land

As global warming intensifies, changes in land surface air temperature (LSAT) exert significant influence on both natural systems and human society. This study investigated the spatio-temporal heterogeneity of global LSAT variability using the ERA5-Land reanalysis dataset from 1950 to 2022. The results indicate a significant global warming trend of 0.035 ± 0.003 °C/year during 1980-2022, with pronounced hemispheric asymmetry. The warming rate in the Northern Hemisphere (0.038 ± 0.003 °C/year) far surpassed that of the Southern Hemisphere (0.018 ± 0.003 °C/year). Seasonal analysis revealed that peak warming occurred during the Sep-Oct-Nov period, corresponding to boreal autumn and austral spring. Spatially, 99.23% of the global land area experienced warming, with 46.07% exceeding the global mean rate, primarily in northern mid-upper latitudes. Arctic amplification was evident, with warming exceeding twice the global average. Empirical orthogonal function (EOF) analysis revealed that the first three modes explained 35.95%, 14.82%, and 6.03% of the variance, respectively. The first mode (EOF1) exhibits a globally coherent warming pattern associated with the long-term trend. The second mode (EOF2) displays a distribution characteristic of "North Eurasia + /Rest - ," strongly correlated with the North Atlantic Oscillation/Arctic Oscillation (NAO/AO). The third mode (EOF3) is linked to tropical Pacific sea surface temperature (SST) anomalies and correlates significantly with the El Niño-Southern Oscillation (ENSO).