Summary

The objective of this study was to determine current and future development condition frequency functions to be used in flood damage analysis in the CRS watershed. To calculate the increase in EAD, the analyst needed the current and future development flow-frequency functions. Using an available calibrated rainfall-runoff model of current development in the watershed, precipitation events of known frequency were used in HEC-HMS to compute a flow-frequency function. Eight precipitation frequency events were used to ensure adequate resolution of this function. The flow-frequency function developed was then compared to functions developed using alternative procedures. A best-estimate current development condition frequency function was then adopted. Then, the rainfall-runoff model was altered to reflect the watershed with future development. For both development conditions, several methods, as shown in the Alternative Methods for Deriving Flow-Frequency Functions and discussed in the text, can be used to compute flow-frequency functions.

Key Takeaways

Key Steps in HEC-HMS

  1. Start from a calibrated current-condition basin model, calibrated where possible to observed events of magnitude similar to the frequencies of interest, and keep parameters consistent across events, varying only what the event warrants (for example, initial loss).
  2. Build the events across the range of annual exceedance probabilities from a current precipitation-frequency source (NOAA Atlas 14 through the Precipitation Frequency Data Server, or local depth-duration-frequency data), using either the Frequency Storm method (a balanced, nested storm) or the Hypothetical Storm method (temporal pattern supplied from historical storms or NOAA Atlas 14 patterns), and apply a depth-area reduction with the storm area set to the drainage area at the evaluation location.
  3. Compute the events. A Frequency Analysis compute assembles the flow-frequency curve at the selected element in one compute. Use enough events to resolve the curve (this case study uses eight, spanning 0.50 to 0.002 AEP).
  4. Compare the computed curve with independent estimates (a gage analysis in HEC-SSP under Bulletin 17C, or regional regression equations from StreamStats) before adopting a best-estimate curve.
  5. Re-run the adopted approach with basin models modified for future or with-project conditions, and carry the curve and its uncertainty description into the risk assessment (EM 1110-2-1619 (USACE, 2025), Chapter 3).

Common Pitfalls

  • Assuming AEP neutrality without checking it. Taking the AEP of the design rainfall as the AEP of the computed peak is standard practice, but it is an assumption. Antecedent conditions, temporal pattern, and storm area all affect the peak, and the depths within a real storm rarely share one exceedance probability. Compare the computed curve with independent estimates before adopting it.
  • Using point rainfall without a depth-area reduction. Precipitation-frequency products such as NOAA Atlas 14 report point depths. Apply an area reduction except for very small study areas or depths already reduced, set the storm area to the drainage area at the evaluation location, and do not apply the TP-40/TP-49 curves to study areas larger than about 400 square miles.
  • Calibrating only to small events. Where possible, calibrate to observed events of magnitude similar to the frequencies of interest. Parameters fitted to frequent events may not represent response during rarer ones.
  • Fitting a curve to a short gage record. ER 1110-2-1450 (USACE, 1994) recommends avoiding statistical analysis of short samples. The ten-year record in this case study is shown for illustration only.
  • Mixing flood populations. Annual peaks produced by different mechanisms, such as snowmelt and thunderstorm runoff, should be identified and analyzed separately (EM 1110-2-1619 (USACE, 2025), paragraph 3-3).
  • Assuming stationarity. Flow-frequency relationships assume the statistical characteristics of the record hold into the future. ECB 2026-1 (USACE, 2026) requires an assessment of long-term hydrometeorological conditions.
  • Using superseded precipitation-frequency data. Check whether NOAA Atlas 14 covers the study area before falling back to NOAA Atlas 2 or the older Technical Papers, and watch for NOAA Atlas 15.