The group has made research progress in the direction of multiphase precipitation energy inversion:

Precipitation-induced soil erosion shapes natural landscapes by promoting particle separation, soil aggregate decomposition, and sediment transport, and at the same time poses a serious challenge to ecosystems and human societies. The precipitation kinetic energy (PKE) of rainfall phases, as a key indicator for describing rainfall erosivity, intuitively reflects the ability of rainfall to decompose soil particles, i.e., the greater the PKE, the greater the ability to decompose soil particles, and the higher the potential risk of soil erosion. In many past studies, indirect methods have been used to characterize the microphysical properties of large-scale precipitation. PKE is calculated by reconstructing the precipitation particle size distribution (PSD) and establishing an empirical relationship between the kinetic energy per unit area per unit time (KE) and the rain intensity (I); however, a large number of studies have focused on rainfall at the expense of other precipitation phases, especially snowfall (which constitutes the second largest form of precipitation after rain), and the intensity of precipitation has been directly substituted into the KE-I relationship, neglecting the phase differences in the whole large-scale precipitation process. This ignores the phase differences in the whole large-scale precipitation process. To address this issue, Duanyang Ji, a graduate student of the team, published an article “Reshaping global rainfall erosivity rates: A study on precipitation phase correction from 2015 to 2022” in April 2025 in the Journal of Hydrology.

We innovatively adopts a precipitation phase delineation algorithm based on microphysical properties to solve the problem of significant overestimation in traditional precipitation kinetic energy calculations. By integrating high precision precipitation data from multiple sources, the team effectively breaks through the bottleneck of spatial resolution in phase-dependent PKE estimation, and realizes high precision PKE calculation at grid scale. On this basis, the team proposes for the first time a 0.1°×0.1° resolution PKE estimation model for the KE-I relational database, which comprehensively incorporates rain and snow multi-phase precipitation characteristics and significantly improves the model's universality and computational accuracy. The team further conducts a systematic analysis of global PKE and rainfall erosion rate at spatial and temporal scales from 2015 to 2022, which reveals the key elements that have been neglected in the previous studies, and provides a new theoretical framework and data support for the refined assessment of global rainfall erosion risk.

The results show that the accuracy of the KE-I fits for both liquid and solid precipitation is high, with an average pearson index of 0.8 in the validation area, and a significant overlap observed between the fitted curves of the raindrop spectrometer and the dual-frequency precipitation radar.

We use the global multi-phase KE-I database in conjunction with precipitation data from IMERG (2015-2022) to derive corrected global multi-year PKE data. The results emphasize the influence of different climate systems on regional and continental PKE patterns, highlighting the impact of the geographic extent and spatial distribution of precipitation hotspots on total and mean PKE values.

Meanwhile, the team calculated the global snowfall kinetic energy for the period 2015-2022 using both the traditional method and the improved PKE calculation method. Based on these calculations, the snowfall energy and its mean error ME were further evaluated, and it was found that the average annual snowfall energy decreases by a factor of ten, due to the misclassification of snow particles as rain particles during precipitation phase partitioning.

In addition, the team calculated the kinetic energy of snowfall and its relative error RE for each month of the year before and after the correction. despite variations in the error values, the monthly average RE remained relatively stable throughout the year, with an annual average RE of 11.3, suggesting a persistent overestimation of the energy of snowfall and flaws in the traditional calculation. Since snowfall is mainly concentrated in the Northern Hemisphere, the amount of snowfall and its associated energy decreases significantly as the global climate enters the warm season. Statistics show that the average snowfall kinetic energy is 12.6 MJ-ha-¹ during the warm season, while it reaches 102.6 MJ-ha-¹ during the cold season. Especially in the cold season, irregular snowflake patterns pose a challenge to accurate calculations, leading to a more pronounced problem of overestimation.

The study utilized the corrected PKE dataset and global rainfall events to calculate the global rainfall erosivity R-factor dataset for the years 2015-2022. Multi-year average annual R values usually correspond to the spatial distribution of annual precipitation. Errors generated by traditional calculation methods are more pronounced in the cool season than in the warm season. Seasonal alternation in each hemisphere leads to different patterns of error generation between continents.

In this study, a grid-based multiphase KE-I dataset was used to realize a fine estimation of PKE using IMERG data from 2015-2022, revealing the global spatial distribution and its seasonal variation patterns. Based on these findings, this study reconstructs the corrected global R-factor dataset to provide a reliable basis for future studies of rainfall erosion rates.