A spatiotemporal investigation of multi-hazard associated with tropical cyclones

As a tropical cyclone (TC) approaches land and makes landfall, both coastal and inland communities can be impacted by a range of hazards, including storm surges, intense winds, and heavy rainfall. Characterizing the multi-hazard events will enhance disaster mitigation efforts by identifying likely locations and combinations of hazards. This work aims to identify the multi-hazard events and their association with TCs originating from the North Atlantic and East Pacific over the Contiguous United States (CONUS) from 2006 to 2022. The hazard data are obtained from the National Climatic Data Center’s Storm Data. This database provides the location, time, and damage of hydro-meteorological hazards based on daily report entries since 2006. It also provides text descriptions of the synoptic weather systems associated with hazards. First, TC-associated hazards are identified based on text descriptions explicitly mentioning their connection with TCs. Next, a Geographic Information System is used to map and aggregate all TC-associated hazards at the county level. Subsequently, temporal characteristics of these hazards within each county are analyzed to distinguish between single, co-occurring, or consecutive hazards. We have found that during the year 2022, a total of 643 hazards were linked to ten TCs across the CONUS. The most frequent hazard types included tropical storm-force winds, floods, and heavy rainfall. A total of 250 counties were affected, with an average of 2.57 hazards per county and a maximum of 30 hazards recorded in a single county. In total, 121 co-occurring multi-hazard events were identified, 80 occurring within a single day and 41 spanning over multiple days. A further aspect of this research involves the use of spatial techniques and natural language processing to investigate the interactions between hazards and their association with TCs, including multivariate hazards, triggering, and change conditions.