Long-term changes in the diurnal cycle of total cloud cover over the Tibetan Plateau

The diurnal cycle of total cloud cover (DCTCC) plays an important role in climate change over the Tibetan Plateau (TP). This study used the International Satellite Cloud Climate Project-H (ISCCP-H) observations, the fifth generation of reanalysis of the European Centre for Medium Range Weather Forecasts (ERA5), the secondgeneration Modern-ERA Retrospective analysis for Research and Applications (MERRA-2) reanalysis data, and 21 models from the Coupled Model Intercomparison Project Phase 6 (CMIP6) AMIP simulations from 1984 to 2014 to investigate the climatology and long-term variations in the DCTCC over the TP. The total cloud cover (TCC) over the TP exhibits sinusoidal-like unimodal diurnal cycles and is quantified with four metrics: the daily mean (mu), daily peak time (t(max)), daily coefficient of variation (c(v)), and full width at half-maximum of DCTCC (half width); the 30-year means of these four indices of TCC averaged over the TP were approximately 65.45%, 14:00 LST, 0.16 and 10 h, respectively. From 1984 to 2014, the TP-average mu of TCC decreased by -2.64%/decade, which was closely associated with the abrupt change in high cloud (clouds with cloud top pressure <440 hPa, defined by the ISCCP) cover (HCC) from 1998 to 1999. The t(max) was postponing by approximately 0.12 h/decade, and the DCTCC was narrowing and becoming steeper. These climatology observations and trends were dominated by the middle cloud (clouds with cloud top pressure between 440 and 680 hPa, defined by the ISCCP) cover and contributed to by the HCC. There were gross deviations in the climatological DCTCC from the reanalysis datasets and models, and they can hardly reproduce the long-term trends of the DCTCC. The climatology and long-term variations of the diurnal cycles of the 2 m temperature (T2m), 10 m wind speed (U10m), vertical integral of divergence of moisture flux (DMF) and relative humidity (RH) at 250 hPa and 500 hPa exhibited significant correlations with those of the DCTCC over the TP. Over most of the TP region, the secular variations in mu of TCC were negatively correlated with those of T2m, which were dominated by the correlations between TCC and T2m during the daytime. The secular variations in the nighttime-averaged HCC have a positive association with those in the T2m over the northern TP. The spatial pattern of correlations between mu of TCC and mu of U10m and DMF were mainly reflected in the HCC. The secular variations in mu of RH were positively correlated with those of the cloud cover. The lags in the t(max) of the TCC were negatively correlated with the variations in the tmax of T2m and positively associated with those of RH at 250 hPa over parts of the eastern TP. The long-term changes in the c(v) of the TCC were mainly positively correlated with those of T2m over most of the TP region and were partly correlated with the other factors. The decadal narrowing of the diurnal cycles of HCC was positively contributed to by the long-term trends in the half width of RH at 250 hPa over the central TP.