Soil moisture responses associated with significantropical cyclone rainfall events

Abstract

Several historic rainfall and flooding events associated with Atlantic Basin tropical cyclones have occurred in recent years within the conterminous United States: Hurricane Joaquin (2015) in early October over South Carolina; Hurricane Harvey (2017) in late August over southeastern Texas; Hurricane Florence (2018) in September over North Carolina; and Tropical Storm Imelda (2019) in September, again over southeastern Texas. A common attribute of these events includes a dramatic transition from dry soils to exceptional flooding in a very short time. We use an observations-driven land surface model to measure the response of modeled soil moisture to these tropical cyclone rainfall events and quantify the soil moisture anomalies relative to a daily, county-based model climatology spanning 1981 to 2013. Modeled soil moisture evolution is highlighted, including a comparison of the total column (0-2 m) soil moisture percentiles (derived from analysis values) to the 1981-2013 climatological database. The South Carolina event associated with Hurricane Joaquin resulted in a sudden transition from severe drought to significant flooding in the span of a few days, due to locally 700+ mm of rainfall. The prolonged heavy rainfall associated with Hurricane Harvey resulted in record soil moisture values well in excess of the tail of the climatological distribution. The soil moisture west of the Houston, Texas, metropolitan area was anomalously dry prior to Harvey, but quickly transitioned to near saturation in the top 1 m, while east of the Houston area antecedent soil moisture values were more moist prior to the local 1200+ mm of rainfall and catastrophic flooding in the Beaumont/Port Arthur area. Hurricane Florence led to widespread 500-700+ mm of rainfall in North Carolina, and another dramatic transition from anomalously dry conditions to record wetness. Once again, with Tropical Storm Imelda, portions of southeastern Texas experienced extreme rainfall amounts up to 1000+ mm, resulting in another sharp transition from drought conditions to extreme flooding in <3 days. An experimental forecast soil moisture percentile is presented for the Imelda event, showing the potential to increase situational awareness for upcoming flooding episodes, along with a discussion of how an ensemble-based approach could be explored to address forecast model error and uncertainty.