Search

Precipitation events exhibit a spatiotemporal continuity since they move across both space and time, however, the joint spatiotemporal characteristics of precipitation events are often neglected, and the role of Atmospheric rivers (ARs) and tropical cyclones (TCs) in the events has not been systematically explored. To solve these problems, this study seeks to detect and characterize the three-dimensional (latitude + longitude + time) structures of spatiotemporally contiguous precipitation events, distinguish the heavy events associated with ARs or TCs, and systematically explore the different roles of these events. We do so by combining Forward-in-Time algorithm, an object-based physical event detecting algorithm, with TMPA precipitation estimates to detect spatiotemporally contiguous precipitation events over China, and an AR detection technique considering a combination of fixed lower limit and Integrated Water Vapor Transport percentile threshold is adopted to identify ARs. The heavy events are defined to analyze the differences of the events associated with TCs and ARs as a function of precipitation total, duration, and intensity. Results show that approximately half or more of precipitation events are shorter than 6-hour, and the precipitation total of each event is strongly correlated with maximum intensity and average intensity across China. The heavy events normally occur in southern, center, northern, and southwestern regions, where have approximately 10 to 40 heavy events at each grid. The fraction of heavy events associated with TCs generally increase for events with larger precipitation totals and for events with higher maximum intensity, until all of events are associated with TCs. The percentages of events associated with ARs as well normally increase as precipitation totals and maximum intensity increase up to around 150 mm and 30 mm/h, respectively. Our findings show that the events associated with TCs or ARs exhibit distinct different characteristics, and suggest that the different spatiotemporal dynamics should be sufficient considered in contiguous events analyses in different regions.