Dust is an important source of atmospheric ice-nucleating particles (INPs) and plays a key role in ice cloud formation in mid- to high-latitudes and the upper troposphere. However, numerical studies on the impacts of severe dust weather on cloud and precipitation processes over China remain limited, and their accuracy is constrained by uncertainties in cloud microphysical and ice nucleation parameterization schemes.

In this study, a comprehensive identification method for dust-precipitation station occurrences and dust-precipitation events was developed based on surface meteorological observations, aerosol concentration data, and dust weather records from the China Meteorological Administration (CMA). Springtime dust-precipitation events during 2001-2021 were identified, and ERA5 reanalysis data were used to analyze the associated thermodynamic and dynamical characteristics of the atmosphere. In addition, an online aerosol-ice nucleation-cloud-precipitation coupling framework, including immersion freezing, deposition freezing, condensation freezing, and contact freezing, was introduced to explicitly represent the interactions between size-resolved dust aerosols and cloud microphysical processes. This framework was used to investigate the impacts of spring dust events on cloud and precipitation processes over China.
Results show a strong spatiotemporal co-occurrence between dust and precipitation events. A total of 293 dust weather events occurred during 2001-2021, of which approximately 77% were accompanied by dust-precipitation station occurrences. ERA5-based statistical analysis indicates that dust influences lead to a colder and drier lower and middle troposphere, enhanced mid-tropospheric subsidence, and conditions unfavorable for cloud formation. Specifically, relative humidity below 700 hPa decreases by 1%-3%, while temperature decreases by 0.2-0.7 K.

Simulations of a representative dust-precipitation case in 2018 show that dust significantly modifies the distribution of ice-nucleating particles. The original WDM6 scheme, which depends only on temperature, produces ice-nuclei concentrations of only 10²-10³ L⁻¹ at 4-7 km altitude, indicating a strong underestimation and unrealistic ice formation near the −40 °C layer. After introducing dust-related INP processes, ice-nuclei concentrations increase to ~10⁴ L⁻¹, and the peak heterogeneous ice production shifts to 4-7 km, consistent with observed vertical structures. Dust generally suppresses cloud development: ice-phase hydrometeors above 7 km decrease to less than 24%, mixed-phase hydrometeors at 4-7 km decrease by 10%-15%, and liquid-phase hydrometeors below 4 km decrease by ~5%, accompanied by a temperature increase of 0.16-0.52 K and a ~3% decrease in relative humidity. Precipitation biases are also reduced, with near-source overestimation decreasing by ~4.5 mm (RMSE reduced by 4.1 mm), while downstream underestimation increases by ~1.1 mm due to enhanced hydrometeor transport.