Hydrometeorological Forecast Performance and Sensitivities for the 27 May 2018 Ellicott City Flood

For the second time in just 22 months, a “1000-year” torrential rainfall event brought catastrophic flood damage and one fatality to central Maryland on 27 May 2018, especially Historic Ellicott City. Lasting flood damage from road washout, land erosion, and localized landslides continues to present challenges as the region recovers from the second such extreme precipitation event in less than two years.

Hydrometeorological forecast challenges for this event will be evaluated, with a particular focus on short-range forecasts from the NOAA National Water Model (NWM), the quantitative precipitation forecasts (QPFs) from the High-Resolution Rapid Refresh (HRRR) model, and the corresponding streamflow and flood inundation response from the NWM.

Forecasts are evaluated on various time and space scales and from multiple forecast perspectives, along with discussion of possible sensitivities to urban development and land-use, upstream reservoir water management processes, limitations due to sparse observations, and potential NWM operational forecast benefits. We have encouraged local, regional, and national community involvement on the event evaluation, and we will include these viewpoints in the presentation. Our results to date highlight that “good” meteorological forecasts do not necessarily guarantee accurate streamflow forecasts. That is, small QPF displacement errors from HRRR model forecasts that would be considered quite skillful at many meteorological scales of interest still translate into notably different streamflow response forecasts in the NWM given the very small size of the study basin. Thus, the possibility of additional forecast tools - and perhaps more specifically, forecast communication and dissemination tools and strategies - could be considered to better anticipate impacts from these events and their inevitable forecast uncertainties in the future.