Abstract: Using historical observation data from meteorological stations near a hydropower station, with the XGBoost machine learning algorithm, a computational model for relative humidity was developed, analyzing the impact of flow aeration on relative humidity. The results indicate that: (1)Flow aeration from the hydropower station leads to an increase in relative humidity downstream, with the magnitude of influence diminishing rapidly as the spatial distance increases. (2) Differences in sky conditions affect the extent of the aeration-induced spray, showing distinct diurnal variation characteristics. (3) The discharge volume has a relatively minor impact on the spray range, suggesting that the energy dissipation achieved by the submerged hydraulic jump effectively mitigates flow aeration and subsequent spray generation.