Heavy precipitation can have major impacts on infrastructure, and different regions have different definitions of heavy precipitation. This therefore necessitates accurate background knowledge and forecasting of the weather regimes that can cause heavy precipitation for a given region, as well as common regimes that precede and follow them. Investigation of these regimes uses self-organizing maps (SOMs) of Multi-Region Multi-Sensor (MRMS) observed precipitation data. Six SOMs are trained on MRMS 1-h and 24-h quantitative precipitation estimate (QPE) data for six regions within CONUS to parse out storm types affecting each region. QPE data are filtered by average recurrence interval data to prevent pattern smoothing in the SOM output nodes due to relatively low-precipitation cases decreasing node-average precipitation, thus providing more focus on cases with heavier precipitation.