A new method to identify source areas of precipitation modification and their attribution: an illustration for the western Netherlands

Urban areas can influence local precipitation patterns and amounts. While many earlier studies focused on the magnitude of precipitation modification by cities, this study develops a method to identify the source areas of the precipitation modification. This is illustrated for the western part of the Netherlands for 2017-2023. Radar-derived precipitation data is used to analyze precipitation anomalies, which are traced back to their source areas by determining the precipitation’s motion for each rainfall event. Using machine learning techniques, the influence of hypothesized causes – urban heat, air pollution, surface roughness, and their interactions – on precipitation anomalies is evaluated. We find that urban and industrial areas generally enhance precipitation downwind, while large rural areas tend to weaken it. Nearly 30% of the study area frequently generates enhanced precipitation downwind, with an average increase of 16.1% of the total precipitation. The identified source locations of precipitation modification are consistent for different seasons. We also find that surface roughness and SO2 concentrations (used as a proxy for air pollution) show a positive correlation with precipitation anomalies. Air pollution is the most influential predictor of precipitation anomalies in this study, followed by the interaction between air pollution and urban heat, and then surface roughness. Compound effects are less important contributors than the individual variables. This methodology provides a novel perspective on precipitation modification, offering a foundation to refine assumptions about source locations and improve understanding of the mechanisms driving the effect.