<b>Integrating Traditional Statistical Methods</b> <b>and Machine Learning for Enhanced</b> <b>Precipitation Forecasting</b>

Abstract

This study investigates the key meteorological factors influencing precipitation by integrating traditional statistical methods with advanced machine learning techniques. Five critical variables - sunshine, cloud cover, global radiation, barometric pressure and snow depth - were selected for analysis. The results show an inverse relationship between sunshine hours and precipitation, indicating that reduced sunshine correlates with an increased likelihood of precipitation. In addition, higher cloud cover significantly increases the probability of precipitation, while low air pressure is closely associated with greater precipitation activity. Although global radiation shows a weak positive correlation, its effect is overshadowed by other variables and snow depth has no significant effect overall. Despite the relatively low explanatory power of the model (R² = 0.148), this research highlights the complexity of precipitation dynamics. The results provide valuable insights for improving precipitation forecasting, particularly in urban environments such as London, and highlight the need for future studies to include additional variables such as topography and wind speed to improve prediction accuracy. By combining statistical and machine learning approaches, this study contributes to the ongoing discourse on effective precipitation forecasting in the face of evolving climate challenges.