Decreased dust particles amplify the cloud cooling effect by regulating cloud ice formation over the Tibetan Plateau

Ice-nucleating particles (INPs) can initiate cloud ice formation, influencing cloud radiative effects (CRE) and climate. However, the knowledge of INP sources, concentrations, and their impact on CRE over the Tibetan Plateau (TP) - a highly climate-sensitive region - remains largely hypothetical. Here, we integrated data from multi-source satellite observations and snowpack samples collected from five glaciers to demonstrate that dust particles constitute primary INP sources over the TP. The springtime dust influxes lead to seasonally elevated ice concentrations in mixed-phase clouds. Furthermore, the decadal reduction in dustiness from 2007 to 2019 results in decreased springtime dust INPs, thereby amplifying the cooling effect of clouds over the TP, with a 1.98±0.39 W m-2 reduction in surface net CRE corresponding to a 0.01 decrease in dust optical depth. Our findings elucidate new pathways of climate feedback from an atmospheric INP perspective, especially highlighting the crucial ro..., , , 1. Dataset_1 showed the ice-nucleating particle (INP) concentrations per volume of sample-melt water (N_(INP_water)) and the INP concentrations per volume of air (N_(INP_air)). These data were used in Figs. 3, 4, S7, and S8.
2. Dataset_2 showed the element concentrations (Ca, Mg, Al, K, Mn, Fe, and Sr) of the LHG and AL samples. These data were used in the Fig. 4.
3. Dataset_3 showed the ratio of water-soluble calcium ion to elemental calcium (Ca2+/Ca) for samples collected in LHG and AL. These data were used in the Fig. 4.
4. Dataset_4 showed the hydrogen and oxygen isotopes (Delta D and Delta 18O) of collected snowpack samples. These data were used in the Fig. S8.
5. Dataset_5 with 3 sheets showed the statistical analysis results of backward trajectories coupled with land cover types. These data were used in the Fig. S16.
6. Dataset_6 with 2 sheets showed the cloud ice number concentration over the TP from June 2006 to July 2019. These data were used in the Fig. 2.
7. Dataset_7 showed...