Detecting occurrence of blowing snow events with decision tree model

Blowing snow processes are crucial in shaping the strongly heterogeneous spatiotemporal distribution of snow and in regulating subsequent snowpack evolution in mountainous terrain. Although empirical formulae and constant threshold wind speeds have been widely used to estimate the occurrence of blowing snow in regions with sparse observations, these methods struggle to accurately capture the high local variability of blowing snow. This study investigated the potential capability of the decision tree model to detect blowing snow occurrence based on routine meteorological observations (mean wind speed, maximum wind speed, air temperature and relative humidity) and snow measurements. Results show that the maximum wind speed contributes the most to the classification accuracy, and the inclusion of more predictor variables improves the overall accuracy. Besides, the overall accuracy of blowing snow occurrence detected by the decision tree model is comparable or even better than traditional methods, indicating it is a promising approach requiring limited meteorological variables and having the potential to scale to multiple stations across different regions, such as the Tibetan Plateau.