Skip to main content
U.S. flag

An official website of the United States government

Delaware Flood Frequency

In Delaware, the current flood-frequency characteristics for sites monitored by streamflow gages, and regression equations for estimating flood magnitudes at unmonitored sites, are based on data collected through 2004 (Ries and Dillow, 2006). Thirteen additional years of peak-flow data and new techniques for flood-frequency analyses have become available since the 2006 report was completed.


1/2 Knowledge of the magnitude and frequency of floods and droughts is needed for the effective and safe design of bridges, culverts, and other infrastructure; for flood-plain planning, management, and for flood-inundation mapping, The drought/low flow analysis will provide an improved understanding of the spatiotemporal response of streamflow to the drought of record and of the timing of conditions prior to drought will improve our ability to predict current and future vulnerability under increased water demands in the basin. .

In the United States, flood-frequency characteristics are typically computed based on fitting a log-Pearson Type III (LPIII) distribution to a time series of annual-peak streamflow values based on observations from 10 or more years at a streamflow gage. Mean, variance, and skewness of the log-transformed annual-peak streamflow time series are used as parameters of the LPIII distribution to estimate flood quantiles. Those quantiles describe the annual exceedance probabilities (AEPs) associated with peak flows of various magnitudes. For example, the 0.01-quantile corresponds to the 1-percent AEP (also referred to as the flood with a 100-year recurrence interval).

Recent improvements in computational methods for flood-frequency analysis include improved ability to incorporate information about uncertainty in the peak-flow values (Cohn and others, 1997, 2001), the ability to incorporate periods of missing record by using non-exceedance thresholds (Cohn and others, 2001), and improvements in the ability to detect and include low outliers (Cohn and others, 2013; Griffis and others, 2004). In addition, improved techniques have been developed for estimating the skew (lack of symmetry of the peak-flow distribution) of each sites peak-flow data set by weighting skews determined from a sites peak-flow data with better estimates of regional generalized-skew values (Veilleux, 2009; England and others, 2017).

In using historically-observed peak flows to determine the peak design flow, the standard methods assume that the peak design flow will occur in the future with the same probability and magnitude as determined from the historical analysis.

With changing climate, land cover, and agricultural and land drainage practices occurring across the United States, the assumption of stationarity in the observed peak-flow record may not be valid. In fact, 18 of the 116 streamflow gages whose data were used in the preceding study (Ries and Dillow, 2006) had increasing trends in annual peak flows. In cases where such trends exist at sites with long periods of record, it may be best to make at-site AEP estimates using only more recent data rather than entire periods of record, or otherwise treat the pertinent peak-flow time series to account for the trends. The proposed study takes advantage of the additional years of peak-flow data, improved analysis techniques, and improved Geographic Information System (GIS) data layers for determining explanatory variables for regression equations. Newer techniques and recent data will provide more accurate at-site AEP estimates, which will result in improved regional regression equations for AEP estimates at unmonitored sites in Delaware.



  • Objective 1: Calculate magnitude of peak flows with 50-, 20-, 10-, 4-, 2-, 1-, 0.5-, and 0.2-percent annual-exceedance probabilities (AEP) at non-regulated streamflow monitoring sites in and near Delaware with a minimum of 10 years of peak-discharge data, using data collected through the 2017 water year and the most current statistical methods;
  • Objective 2: Update all GIS and hydrologic databases based on current LiDAR coverage and most recent DEMs to update the StreamStats application in Delaware with best-available basin and land-use characteristics
  • Objective 3: Produce updated regional regression equations for estimating the 50-, 20-, 10-, 4-, 2-, 1-, 0.5-, and 0.2- percent AEP peak flows at unmonitored sites in Delaware;
  • Objective 4: Document and make available the methods, and results of this work in a USGS report, and in StreamStats, a USGS Web-based application for calculating streamflow statistics for any nontidal stream location in Delaware.
  • Low Flow Objective: A journal article describing the spatiotemporal variability in low flows in the Delaware River Basin. A second peer reviewed report describing the enhanced prediction equations, and an associated Drought Mapper tool available to the public. Low flow regression equations will be published with peak flow equations in an SIR and incorporated into StreamStats.