Heavy precipitation events extending over many days can promote extensive flooding. We identify both short- (5-day) and long-term (30-day), seasonal heavy precipitation events over the Missouri River basin using an object-oriented analysis method, Tempest Extremes (TE). TE identifies and tracks 3-dimensional space-time objects, which we analyze here as events based on precipitation. Further, we examine whether 30-day precipitation events are an outcome of clustering 5-day precipitation events, or steady precipitation.

We analyze precipitation simulated by the regional climate model RegCM4, which was driven by two GCM simulations. Simulations used 25-km grid spacing. We use PRISM observational data to evaluate the quality of the simulations. Analysis time periods use contemporary (1981 - 2000) and RCP 8.5 scenario (2041 - 2060) time ranges. Daily precipitation is accumulated over both 5- and 30-day time periods and input into TE to identify sequences of seasonal precipitation events. We use accumulated precipitation thresholds of 50- (5-day) and 150-mm (30-day) to identify events with substantial precipitation. TE provides frequency, area and duration of the events, where duration is the number of days in consecutive, overlapping wet periods given by 5- and 30-day windows that shift each day.

In our analysis, contemporary simulations tend to have event durations and areas similar to observations, though frequencies differ somewhat from observations. Scenario results indicate a future with increasing frequency of our heavy precipitation events, though durations and areas covered by individual events have little change. Further analysis shows the degree to which long-term heavy precipitation is an outcome of accumulated short-term events.