The rapid urbanization in China and the long‐range transport dust (LRTD) from arid and semi‐arid areas has resulted in an increase of PM₁₀ concentration. In this study, an interpretable deep learning model [deep forest (DF)] with FY‐4A top‐of‐the‐atmosphere reflectance (TOAR) data were used to obtain the hourly PM₁₀ in China. The optimal hourly average R² of 10‐fold cross validation can achieve 0.85 (13:00 Beijing time); The R² (RMSE, μg/m³) of the daily, monthly, and annual averages were 0.82 (24.16), 0.97 (6.53), and 0.99 (2.30), respectively. Using TOAR data, the DF model performed better than other machine learning models. The feature importance of the TOAR‐PM₁₀ model showed that TOAR and meteorological elements both contributed significantly to the model. In spring, the PM₁₀ in northern China was greater than that in southern China, which may be related to the LRTD. Excluding the dust weather periods, the areas with high PM₁₀ values in China were mainly in cities and their suburbs, where were correlated with human activities. During a dust weather process, LRTD increased PM₁₀ in northern China by 80.4%. During a mixture haze and dust weather process, the PM₁₀ increased by 130.2% in northern China, of which LRTD led to an increase of 73.7%. The sources (from the Taklimakan Desert in China) and transmission paths of these two LRTD processes were similar. The contribution of LRTD to PM₁₀ was related to dust intensity and meteorological conditions. The results showed that LRTD and local pollution to PM₁₀ was both important in haze periods.

Sea surface temperature (SST) could significantly affect the dynamic and thermodynamic conditions of the atmospheric circulation and consequently the cloud variations. Here we use several different satellite records to extract the spatial‐time modes of total cloud cover (TCC) by employing a pairwise rotation of Empirical Orthogonal Function analysis. The results show that the first two principal oscillation modes of TCC are closely associated with the Central Pacific El Niño Southern Oscillation (CP ENSO) and Eastern Pacific (EP) ENSO during the 1980s‐2000s, while the ENSO‐like mode of TCC can provide an evident contribution to the TCC change during the 2000s‐2010s. In CP El Niño, the cloud vertical structure decomposed from CloudSat observations shows an increase of cloud occurrence frequency near the equator and around 40°, and a decrease around 10° in both hemispheres, suggesting a symmetric tightening of Hadley cell (HC). In addition, cloud occurrence frequency increases around 180°, which is accompanied by an eastward shift of Walker circulation (WC). In EP El Niño, TCC increases (decreases) over the Equatorial Eastern Pacific (Western Pacific warm pool), and decreases asymmetrically over the subtropical Pacific Ocean, indicating a weakening of WC and an asymmetric tightening of HC, respectively. The different response of circulation and clouds to CP and EP El Niño highlights the nonlinearity of El Niño SST forcing. We also construct a trend mode of TCC to investigate cloud long‐term responses to SST warming by transferring the linear trends of the rest modes to a specific mode. The principal components (PCs) of TCC trend modes are strongly correlated with global‐mean SST (GSST) with correlation coefficients of about 0.60 during the 1980s‐2000s and 0.45 during the 2000s‐2010s, suggesting a continued influence of global SST warming on TCC. The global TCC change is mainly influenced by the combined effect of Atlantic Multi‐decadal Oscillation (AMO), Pacific Decadal Oscillation (PDO) and Indian Ocean Dipole (IOD). The variation about the trend mode of TCC is closely associated with PDO and IOD.