The second meeting of the Hydrologic Frequency Analysis Work Group was convened in the North Penthouse Conference Room, Main Interior Building, Washington, DC on July 6, 2000. Martin Becker, acting co-chair, conducted the meeting. The attendees were:

The meeting consisted of discussions relative to the three priorities of the work group.

Frequently Asked Questions about Bulletin 17B

Bill Kirby and Jerry Coffey had previously provided the work group a compilation of all the Frequently Asked Questions (FAQs). Bill Kirby commented that all the questions were worthwhile but there was not consensus on the answers. Each person in attendance was asked to give his/her views on FAQs relative to Bulletin 17B.

Ken Bullard - Ken noted that the Bureau of Reclamation is more concerned about extreme events, such as the 500- and 1000-year events, that are beyond the limits of frequency analysis using observed flood data (i.e., Bulletin 17B guidelines). Ken indicated he does not routinely receive questions about Bulletin 17B guidelines. His FAQ is why use Bulletin 17B? He suggested that we not spend too much effort on the FAQs. He would like to see research pursued on new techniques.

Zhida Song-James - Zhida commented that questions from FEMA Study Contractors often relate to the use of skew and regulated watersheds. She noted that land-use change in recent years makes Bulletin 17B guidelines less applicable for frequency analysis for flood insurance studies. She suggested that we concentrate our efforts on issues (questions) for which we do not have consensus answers. She asked whether the procedure in Appendix 7 of Bulletin 17B was the way to go for extension of flood records.

Lesley Julian - Lesley asked the question whose uses Bulletin 17B other than FEMA. The response was that all Federal agencies use Bulletin 17B for their water-resources planning and design and that many State agencies, such as State Departments of Transportation, use Bulletin 17B in design of hydraulic structures.

Martin Becker - Martin indicated that all flood studies completed by Federal agencies have the potential for use by FEMA in the National Flood Insurance Program (NFIP) and that Bulletin 17B is the recommended approach for frequency analysis at gaging stations. Therefore, there is a strong link between FEMA and the use of Bulletin 17B. Martin also noted that the Bulletin 15 Work Group that first recommended the use of the log-Pearson Type III method was convened as part of House Document No. 465 (passed in 1966) entitled "A Unified National Program for Managing Flood Losses". Flood frequency guidelines and the NFIP are intrinsically linked.

Bill Kirby - Bill cautioned that we should not underestimate the effort to define answers to the hard questions for which we do not have consensus. He suggested that definitive answers to some of the questions may involve significant research.

Rick McCuen - Rick suggested we send out the FAQs to all the membership and ask them to prioritize the questions and identify needed research

Jerry Coffey - Jerry commented that a description of skew is not in recently published textbooks, it is difficult to estimate, and suggested that we consider using another approach that does not involve the estimation of skew. He also suggested that we could use less than annual data (i.e., partial-duration series) for frequency analysis.

Will Thomas - Will commented that Bulletin 17B does not provide clear guidance on all issues and, therefore, it is not surprising that there is no consensus on some FAQs. Those FAQs for which we do not have consensus could form a basis for future investigations and research by the work group.

Action item: By September 25, Bill Kirby and Jerry Coffey are to prioritize the FAQs, suggest answers and provide a basis for these answers, identify FAQs with no consensus answers, and suggest any needed research that may be applicable.

Ungaged Issue

Various persons offered their views on what the work group should do relative to developing guidance for estimating flood discharges for ungaged watersheds.

Martin Becker - Martin asked the question - Are there watershed sizes for which regional regression equations or rainfall-runoff models are more applicable? If so, this may provide some direction for our work group on the ungaged issue..

Rick McCuen - Rick believes that the paper entitled "Hydrologic Review Procedures for Flood Insurance Studies and Map Revisions" prepared by Will Thomas and Mike Grimm is not specific enough for addressing the ungaged issue. He noted that the example in the paper is one where flood discharges estimated by a rainfall-runoff model, regression equations and gaging station data are all within reasonable agreement. He asked what actions are required when these procedures are not in agreement and how often are flood discharges submitted for flood insurance studies outside of reasonable limits? Will Thomas and Zhida Song-James indicated hydrologic analyses for most studies are accepted without revisions but that approximately 10-20 percent of the time the flood discharges need revisions.

Jerry Coffey - Jerry felt that the most important objective was to achieve consistency in the estimation of flood discharges for ungaged watersheds. Jerry commented that we should concentrate on getting the correct model formulation and remove bias through calibration.

Lesley Julian - Lesley agreed with Jerry that having a consistent approach was important.

Ken Bullard - Ken believes there is a place for rainfall-runoff models in flood frequency analyses. He noted that the Bureau of Reclamation is concerned with estimating flood volumes and hydrographs. This effort requires time and data to make these estimates. He felt that rainfall-runoff models should be calibrated to the entire flood hydrograph.

Dave Wingerd - Dave commented that the U.S. Army Corps of Engineers (USACE) uses rainfall-runoff models in near real time for flood forecasting for operation of their flood-control projects. He noted that USACE has daily regulation models that utilize data from 8,800 streamgaging stations.

Bill Kirby - Bill commented that the Thomas-Grimm paper does not give a consistent, systematic approach for all cases. Like Rick McCuen, he felt there was a need for more guidance for a wider variety of cases.

Will Thomas - Will clarified that the objective of the Thomas-Grimm paper was to provide guidance to FEMA on a consistent approach for reviewing hydrologic analyses submitted for flood insurance studies and map revisions. The objective is to determine whether the flood discharges are reasonable (within some uncertainty limit) and acceptable for use within the NFIP. The review approach is primarily to compare flood discharges from rainfall-runoff models to regression estimates and gaging station data. He acknowledged that the Thomas-Grimm paper could be revised to be more applicable to the needs and objectives of the work group. In fact, he commented that the FEMA guidelines on this subject were still evolving and being improved.

Action item: By August 25, Will Thomas and Rick McCuen will produce a revised version of the Thomas-Grimm paper that provides more guidance on estimating flood discharges for ungaged streams. One approach will be to add additional examples that illustrate differences in flood estimates from different methods.

Regulated Flood Frequency Issue

Martin Becker began this discussion with the question - What should we do about regulated flood frequency analysis? Several persons offered their viewpoints.

Bill Kirby - Bill emphasized the range of conditions for regulated flood frequency analysis from the case where there are a few large reservoirs designed for flood control to the situation with several dozen small reservoirs in the upland watershed. He noted that different approaches are needed for these conditions

Ken Bullard - Ken noted that there are many factors difficult to predict relative to regulated flow analysis such as dam operators not following the rule curve. He also noted the Bureau of Reclamation reservoirs are not for flood control, rather for irrigation and water supply. These reservoirs may have incidental flood control storage

The work group defined the following types of regulation for floods under the assumption there is an existing reservoir(s) in the watershed.

1. Single reservoir,
2. Multi-reservoir system,
3. Distributed reservoirs throughout the watershed (similar to Natural Resources Conservation Service reservoirs),
4. Incidental flood storage similar to Bureau of Reclamation or Federal Energy Regulatory Commission reservoirs

The work group discussed the following issues related to regulated flood frequency analysis.

1. Available data and operating procedures (degree of regulation) for the reservoir(s),
2. Development of homogeneous unregulated flows or input data to the regulated analysis,
3. The transfer relation from unregulated to regulated flows,
4. Procedures for performing the regulated frequency analysis.

The work group generally agreed that there is no universal approach for regulated flood frequency analysis and the approach should be commensurate with the data available and the type of regulation. The objective will be to develop a flow chart or a sequence of steps for regulated frequency analysis under various scenarios.

Action item By October 10, each member of the work group is to develop a sequence of steps for regulated flood frequency analysis.

Election of Chair and Vice-Chair

The election of a Chair and Vice-Chair for the work group was postponed until the January 2001 meeting since there were only nine work group members in attendance and some members who attended last time were not present. It was felt that a broader representation of members should be present for the important task of electing a Chair and Vice-Chair. It was agreed that the acting Co-Chairs (Don Woodward and Martin Becker) continue in their present positions until permanent Co-Chairs are elected.

There are about 19 members on the email list but some of these persons are unlikely to attend meetings. In an effort to determine the active membership of the work group, Martin Becker will send an email to all email correspondents asking about their future involvement in the work group and whether they will attend meetings in person or by teleconference. If needed, the January 2001 meeting will be conducted at a site where there is a speaker telephone where members can call in and participate in the meeting. It is important to establish a list of active members of the work group so that we can determine if a quorum is available in case a vote is required on an issue. A permanent Chair and Vice-Chair will be elected at the January 2001 meeting.

Next Meeting

The next meeting of the work group will be in January 2001 on the same day as the Subcommittee on Hydrology meeting. The exact day and place will be announced a few months prior to the meeting.

Will Thomas
August 7, 2000
Revised May 14, 2001