Investigation of Storm Temporal Patterns using Gridded Meteorological Datasets for Hydrologic Modeling

Rainfall-runoff models require spatial and temporal patterns of precipitation as input. The duration and intensity of this rainfall directly influences the runoff peak and volume generated. In dam safety applications, the spatial and temporal characteristics of extreme storms are typically derived from hydrometeorological reports, storm databases, and frequency analyses available from multiple sources, including, but not limited to, the National Oceanic and Atmospheric Administration, Army Corps of Engineers, and Bureau of Reclamation. The usual format depicting the spatial and temporal patterns is depth-area-duration tables. These tables are typically limited to discrete, 6-hour increments; storm spatial patterns are often limited to storm total isohyetal maps. In reality, we recognize that storms: vary spatially with respect to time; rarely fit into discrete time periods; and have accumulation rates and event durations that are modulated by multiple factors (e.g., geographical region, time of year, and storm type).

In this study, we present a new, applied methodology to explore and examine the variations in storm durations for select locations in the western United States. Comparisons between point precipitation data, in situ-based precipitation estimates, and reanalysis products (e.g., the Climate Forecast System – Reanalysis) are performed to evaluate the utility of each product in defining the appropriate duration(s) for flood risk assessment applications. Automated methodologies for defining the duration of each precipitation event are explored using criteria from reanalysis datasets to define the synoptic period.

We present the hydroclimatology of precipitation events developed using the new methodology for estimation of storm duration, a comparison of key significant precipitation events across datasets, and a rationale for the use of varying-duration rainfall temporal distributions in hydrologic studies.