Interannual variability of diurnal temperature range in CMIP6 projections and the connection with large-scale circulation

Diurnal temperature range (DTR), as a core indicator of the Earth system, exhibits obvious temporal and spatial variations, which is not entirely consistent over global. The historical simulation capabilities of 19 Coupled Model Intercomparison Project 6 (CMIP6) models for DTR were firstly evaluated against CRU_TS v4.04 data. Future changes under three shared socioeconomic pathways (SSP1-2.6, SSP2-4.5, SSP5-8.5) in DTR were projected using a multi-model ensemble mean (MME), and its interannual variations were seasonally explored through empirical orthogonal function (EOF) analysis. The results indicated that CMIP6 models could reflect the decreasing trend of DTR during 1901-2014, with the global spatial correlation coefficient between models and observation ranging from 0.4 to 0.7. MME outperformed individual models in both spatial and temporal variations, indicating higher accuracy and reliability. The future changes of DTR exhibited significantly decrease across the northern hemisphere and increase in the South America, and change magnitude enlarged with time extension and emission intensity, especially by more than 0.4 degrees C under SSP5-8.5. The decreasing trend of global DTR was kept in SSP2-4.5 and SSP5-8.5, while SSP1-2.6 changed increasing trend during 2015-2100. DTR showed seasonal variations and was mainly influenced by colder months. The dominant modes of interannual DTR and their relationship with the 500 hPa geopotential height, the 200 hPa U wind, and the outgoing longwave radiation showed higher features in tropical regions. The highly positive correlation between the first mode of DTR and the Nino3.4 index in December/January/February (DJF) is 0.67, indicating significant influence of the El Nino-Southern Oscillation on DTR.