The researchers developed a technique called Self-Organizing Map Downscaling (SOMD) to predict how much rain will fall in a month based on historical weather data. They used this method to make detailed predictions at individual weather stations, which can help us understand how global climate change affects local weather patterns.