The IPCC AR6 employs five global land temperature benchmark datasets: NOAA/NCEI's GHCN (1850–), NASA/GISS's GISTEMP (1880–), Berkeley Earth's BEST (1850–), UEA/CRU's CRUTEM (1850–), and Sun Yat-sen University's C-LSAT (1850–). While these datasets exhibit minimal discrepancy in describing long-term trends in global mean land temperature, significant differences exist regarding observed changes in global mean land Diurnal Temperature Range (DTR)—whether derived from these five datasets or the extreme climate dataset HadEX3 developed by the WMO/CLIVAR/JCOMM Expert Team on Climate Change Detection and Indices (ETCCDI) (Fig. 1). Consequently, considerable uncertainty remains regarding the long-term trends and decadal variability of global land DTR. Current research suggests a trend reversal in global DTR beginning in the late 20th century, implying significant impacts on global climate and ecosystems. However, this reversal has largely been inferred through emergent constraints relying on climate models and observations, lacking robust direct evidence from benchmark observational data.

We identified that DTR exhibits stronger spatial heterogeneity compared to mean temperature; the station density in existing global datasets is insufficient to overcome biases arising from this heterogeneity, leading to vast discrepancies in reported DTR changes. To address this, our team collated observational data series from approximately 4,000 additional land stations. We applied advanced techniques to re-homogenize all latest data, resolving widespread non-homogeneity issues in station series. Simultaneously, we reconstructed and corrected climatological normals, upgrading C-LSAT2.0 to C-LSAT2.1 and specifically developing the C-LDTR dataset.

Analysis based on the latest C-LDTR dataset reveals that while global land DTR has predominantly shown a decreasing trend since the 20th century, it primarily increased during the first half of the century. Since 1979, global and hemispheric mean DTR has fundamentally reversed to an upward trend, particularly in the recent period (with increasing trends in 3/4 of global regions). DTR characteristics are closely linked to regional climate change, land-use patterns, and environmental factors such as solar radiation and aerosols. Notably, recent "global brightening" (driven by shortwave radiation) and greenhouse gas accumulation have reversed the earlier declining DTR trend. These findings provide solid observational evidence for long-term global DTR changes and supplement current understanding of the mechanisms driving global and regional DTR variability.

Xu, Q., Wei, S., Li, Z., Li, Q.*, 2025, A New Evaluation of Observed Changes in Diurnal Temperature Range, Geophysical Research Letters, 52, e2024GL113406. DOI:10.1029/2024GL113406.