National Flood Plan Committee To: "A - Division Chief and Staff", "B - Branch Chiefs and Offices", "DC - All District Chiefs", "CD - All Data Chiefs" From: "Robert M Hirsch, Chief Hydrologist, Reston, VA" Subject: National Flood Plan Committee Date: Thu, 24 Apr 1997 16:07:24 -0400 Sender: "L. Jane Rose, Secretary, Reston, VA" In Reply Refer To: Mail Stop 415 MEMORANDUM April 24, 1997 TO: Distribution FROM: Robert M. Hirsch (signed) Chief Hydrologist SUBJECT: National Flood Plan Committee This memorandum announces the formation of an ad hoc committee to develop a National Flood Plan for the Water Resources Division. The committee is an outgrowth of concerns expressed about our changing role related to flood emergency management and our ability to respond effectively to large floods. Issues of the 1996 flooding in the Pacific Northwest were identified in a retrospective analysis prepared by Derrill Cowing, Mike Nolan, Ken Wahl, and Tom Yorke. These issues are included in a memorandum, dated February 25, 1997, to the Chief Hydrologist from the Regional Hydrologist, Western Region, subject "Some Needed Planning for Large Floods." Both the retrospective analysis and the memorandum are attached. I have asked the following individuals to serve on the ad hoc committee for developing the National Flood Plan: Steve Blanchard, Deputy ACH Operations (sfblanch) Derrill Cowing, Idaho District Chief (dcowling) Tim Hale, Georgia District Chief (twhale) Ken Wahl, Regional Surface-Water Specialist (klwahl) Tom Yorke, Chief, Office of Surface Water (thyorke) Derrill Cowing has agreed to serve as Chairman of the committee. The committee will develop the National Flood Plan in consultation with the Senior Staff and other representatives of the Division's operational and research programs. The Western Region's memorandum will serve initially to define the issues that the committee should address. The committee will develop the plan in stages over the next year and submit parts of the plan to the Senior Staff for approval each quarter. The plan will include both short-term and long-term activities for improving our ability to respond effectively to major floods. It will provide a clear statement of our responsibilities for maintaining data service during floods and for conducting followup studies to document the magnitude and extent of flooding. It also will include alternatives for funding flood response efforts and options for enhancing the infrastructure of the national streamflow gaging station network. Part of the output of the committee should be a number of "how to" memorandums that are suitable for insertion into all District flood plans. These should include topics such as: how to work with FEMA (including copies of relevant agreements and examples of successful interactions in the recent past); how to communicate with the media (standard tables, press releases, etc. that can be expected to produce results); how to provide information to Headquarters on possible supplemental funding needs (a template); how to deal with possible conflicting statements from various agencies regarding actual discharges or recurrence intervals. The committee will use the attached documents as a starting point for developing a National Flood Plan, however, I am sure other issues will be considered by the committee. I encourage you to share this memorandum and its attachments with all employees in your respective offices and request their thoughts and concerns about our ability to respond to major floods. They should express their views to any of the committee members. 2 Attachments Distribution: A, B, DC, CD *************************************************************************** RETROSPECTIVE ANALYSIS OF FLOODING IN OREGON AND WASHINGTON, NOVEMEBER 1995 AND FEBRUARY, 1996 Prepared by: Derrill Cowing, K. Michael Nolan, Ken Wahl and Tom Yorke Introduction The winter of 1996 was extremely wet in Washington and northern Oregon. Rain and snowmelt combined to produce floods in November and February. February flooding was generally the most severe. Peaks of record were recorded on 7 streams in northern Oregon. Peak flows with probabilities less than 1.3 percent were common in western Washington. Figure 1 shows the 1996 winter hydrograph for the Sandy River near Bull Run, Oregon. The 1996 floods followed government furloughs in November and December, which meant that Field Office personnel were already struggling with a backlog of work when flooding occurred. Also, recent retirements had reduced the number of experienced personnel in both Washington and Oregon. Despite the fact that flooding occurred at a time when external circumstances made it particularly difficult for USGS personnel to respond, we felt the 1996 floods in the Northwest provided an opportunity to assess how well the USGS was able to respond to widespread large-magnitude floods. We felt it was particularly important to take advantage of this opportunity because the role of the USGS during significant floods seems to be changing rapidly. WRD regularly serves real-time or near real-time data on the Internet, which has increased WRD's visibility during floods. As we will discuss below, our retrospective analysis found that serving real-time data can profoundly affect USGS's role and the public's expectations of the USGS during regional flooding. The flood retrospective analysis was done to: 1. Understand how well the USGS presently responds to floods, 2. Understand how USGS's role during floods may be changing, and 3. Discover what the USGS could do to better respond to floods. The flood retrospective analysis was done by Derrill Cowing, Mike Nolan, Ken Wahl, and Tom Yorke during October 28 - 31, 1996. We spent approximately 2 days in each District, during which time we met with District managers and cooperators. In Tacoma, we met with the District Chief (Carl Goodwin), District Data Chief (Tom Zembrzuski), the Field Office Chiefs from Tacoma, Spokane and Pasco (Bill Wiggins, Ray Smith, and Greg Ruppert, respectively), District Data Relay Coordinator (Scott Knowles) and representatives from the U.S. Army Corps of Engineers (Wayne Wagner), Pierce County (Tim Ramsey) and King County (Tom Bean). In Portland, we met with the District Chief (Dennis Lynch), District Data Chief (Ed Hubbard), the acting Portland Field Office Chief (Rick Kittleson), and representatives from the U.S. Army Corps of Engineers (Dick Cassidy and Ron Mason) and the National Weather Service River Forecast Center (Chuck Orwig). The retrospective analysis was somewhat general in nature. We did not do a detailed review of data collected during the floods. Rather, we relied on discussions and tabulated summaries to understand flood-related activities. Summary of office activities during floods Washington District Spokane Field Office Personnel from the Spokane Field Office were able to effectively respond to the floods. The Spokane Office filled a traditional USGS role during flooding in eastern Washington. The observations discussed above were used to develop specific recommendations that fall under three broad categories--Be aware, Be prepared and Be decisive. It should be noted that cutting across all three categories is the recognition that WRD's role during floods is changing. Serving real-time data on the Internet is placing us in a position from which we must be ready to respond quickly to data needs of local, state and Federal emergency management agencies. We feel that to optimize our efforts during floods, the USGS should: 1) Be aware that: A) Serving real-time data on the Internet has the potential to greatly increase the role of WRD Districts in flood warning and flood forecasting. Making data available on the Internet tends to heighten the awareness of our data to a broad audience and raises the expectations of both cooperators and the public at large. Although this means we will be serving the needs of more people, it also means that WRD Districts serving real-time data on the Internet will feel increased pressures to keep stage data transmitting from many, or all, of the stations they operate. It may also mean that Districts will be asked, or expected, to provide 24-hour support to local, state and federal agencies during flood emergencies.n Washington. Data from the 40 DCPs operated by Spokane served as a backup to data disseminated by other agencies. Real-time data from eastern Washington were not served on the Internet. High-flow measurements were made at, or near, the peak at just about all gage sites. Only 2 indirect measurements were needed at discontinued sites. Pasco Field Office Data from sites operated by Pasco also were not served on the Internet during the floods. Pasco personnel, however, were forced to rely more heavily on indirect measurements to define rating curves. Significant peaks occurred on approximately 29 of the approximately 52 sites run by Pasco. Indirect measurements were needed at six sites. Two employees were brought in from Idaho to help with these indirect measurements. Tacoma Field Office Our discussions in the Washington District were dominated by how personnel from the Tacoma Field Office responded to the floods. Unlike Spokane and Pasco, data from most sites run by the Tacoma Field Office were served on the Internet. In addition, as per ongoing agreements, stage data and rating changes were supplied to 6 cooperators on a 24-hour basis. These two operating strategies profoundly affected the impact flooding had on operations at the Tacoma Office--District personnel found themselves under considerable pressure to maintain the stream of stage and discharge data to cooperators and to the public at large. Bill Wiggins felt that his office must serve the needs of the many diverse water agencies in the Tacoma-Seattle area who need flood- warning information. Floods in 1990 made the public believe in these systems and the USGS must, therefore, be actively involved. He realizes that the need may be becoming so large that many local agencies will eventually have to operate their own redundant telemetry system with advice, and perhaps stage- discharge relations, provided by the USGS. The Data Relay Section provided service for 110 gaging stations during the flood. Forty of those sites were designated National Weather Service flood- warning sites. By serving so much data on the Internet and by virtue of their agreement to provide 24-hour service to some Cooperators, the Tacoma Field Office was in the center of flood warning activities. District personnel were under constant pressure to maintain data transmissions from field sites and update the Washington District Home Page. The pressure to maintain the stream of stage data occasionally meant that making discharge measurements to update and improve ratings became a secondary priority. The need to provide 24-hour service seemed to physically drain personnel after 2 or 3 days of flooding. The Tacoma Field Office did four indirect measurements at discontinued sites after the floods and flagged high water marks at three other locations. Personnel were brought in from Alaska to help with these measurements. Cooperator comments In general, comments from the cooperators verified comments offered by District personnel. Corps of Engineers (Wayne Wagner) Although he was generally satisfied with service from the USGS, Wayne seemed to want more frequent updates of stage data and ratings. He would like to have update of ratings and shifts automated. In general, Corps projects run out of Seattle seem to be moving away from GOES in favor of VHF telemetry, which they install off of our sensors. This means that the Corps relies on their own CROHMS (Columbia River Operational Hydrometeorological System) network for stage data, but rely on USGS to maintain ratings. Stage data from GOES serves as backup for the Corps. Although Corps projects do not use GOES as their primary source of data, Corps "flood teams" do rely on our GOES data almost exclusively. The fact that USGS and Corps maintain two different data streams and two different data bases occasionally caused confusion and inconsistencies during the floods. The Corps update data on the Internet hourly, while USGS Internet updates were done only about every 6 to 8 hours. The computer computations needed to update graphics on our Home Pages took several hours to complete. In addition to confusion caused by different update frequencies, the two agencies may apply different rating shifts to the same stage data. The Corps did not have access to shifts being applied by the USGS as a result of measurements made during the flood. Wayne raised two issues that may make it even more difficult for the two agencies to cooperate in the future: 1. The Corps will be moving to an Oracle data base, and 2. The Corps has been directed to cut their operations by 30% over the next 5 years and to contract 35% of all engineering work, which may include stream gaging. Some conflict arose between the Corps and USGS over flood-frequency estimates at regulated sites. The Corps uses the entire time-series and adjusts the natural flows that occurred before regulation based on the rule curves of the reservoir. The USGS uses only post-regulation time series. We suggest that agreements be worked out in advance of flooding as to who will provide recurrence-interval information for specific stations. Although our discussion focused on flooding, Wayne pointed out that everyone is willing to accept the fact that flood-related data will be prone to inaccuracies. Rating shifts and the rush to serve real-time data will probably always cause problems. The need for highly accurate data are actually higher during periods of low and normal flow. Errors in estimating low and normal flows can prove very costly, given the rising cost of water. Pierce County (Tim Ramsey) Pierce County maintains approximately 90 river miles of levees and has implemented a flood-warning system that relies on USGS data telemetered using GOES. Pierce County gets most of their stage and discharge data directly from ADAPS, although they seem to also use data from the Internet because it is so easy to access those data. A gage at a key flood warning inflow point was destroyed during the flood, causing significant concerns. The county suggested we review locations of critical gages and see what needs to be done to make them as invulnerable as possible to large flood flows. Site selection for streamgages becomes increasingly important as we are drawn more and more into flood forecasting and flood warning. Because they rely so heavily on USGS data during floods, Pierce County personnel felt they needed 24-hour access to USGS personnel during floods, similar to the service WRD provides (for an added charge) to 6 other cooperators. King County (Tom Bean) King County was pleased with USGS's response during the flood. The County River Management Section issues their own flood warnings relying on data they receive from the Corps and USGS. Although King County pays the USGS to deliver data to them from ADAPS, County personnel have found themselves relying more and more on data from the Corps because the Corps updates their home page hourly. The scrutiny WRD received during the flood was highlighted by transcripts from phone logs kept by the County. These logs showed that WRD responded quickly following one call for help but did not respond to a later call for help at another site. King County has a four-phase flood warning program that is keyed to discharge at one of USGS's gages. This is a nice example of how Cooperators rely on USGS data. ADAPS data are automatically sent to a printer at County offices. Occasionally, rapid rises in stage have been missed because no one was monitoring the printer. Tom would like to cooperate in developing a system that would sound an alarm when stages reach certain levels. In general, Tom did not think the USGS system of serving data on the WWW was set up for flood warning because updates were too infrequent. In addition, although they are satisfied with data fed directly to their Emergency Operations Center, there is no capability within the present data dissemination system to sound physical alarms when stages reach preset thresholds. The County would like to have data available on the WWW for longer than seven days so post-flood analyses could be done. FEMA Although we did not meet with representatives from FEMA we did discuss the District's involvement with FEMA. Tom Zembrzuski was able to enter into OFA funding agreements with FEMA to flag highwater marks and to do indirect measurements at several discontinued sites. Tom spent much time finding out what FEMA needed and/or could fund. Oregon District In Oregon, only the Portland Field Office was involved with flood work. Flooding was not significant in southern Oregon, which is served by the Medford Field Office. Personnel from Medford were called in to help after the flood was well along. Portland Field Office The Portland Field Office serves real-time data on the Internet only for selected sites. This, coupled with the fact that the District does not have agreements for 24-hour service with any cooperator, meant that the Portland Field Office was much less involved in flood warning and flood forecasting than Tacoma. The Corps and National Weather Service primarily used stage data telemetered by their own equipment installed at USGS gaging stations to issue flood warnings and flood forecasts. The reliance on telemetry of other Agencies was so great that the USGS did not even activate the GOES alert channel--GOES data were updated only every 4 hours. Much attention was focused on downstream reaches of the Willamette and Columbia rivers, which are much less flashy than the smaller streams around Seattle and Tacoma. This accounts, in part, for why Portland was able to get by with only 4-hour updates. Despite the fact that the Portland Field Office was not thrust into flood- warning/flood-forecasting activities on a 24-hour basis, flooding did profoundly affect the office. After the flood the office was faced with the need to do indirect discharge measurements at 29 sites. Field work associated with the flood and indirect measurements imposed such a large work load on the office that 1995 surface-water records had not been finalized at the time of our visit and, although most field work had been completed, no indirect measurements have been computed. Personnel were not brought in from other Districts to help with work during, or after, flooding. Two people from the Medford office were brought up to help with field work. Dennis was initially concerned about requesting support personnel from other Districts. Dennis offered the observation that several years of below-normal flows may have led District managers to believe the office was adequately staffed. Corps of Engineers, District Office (Ron Mason an d Dick Cassidy) The Oregon District seems to enjoy a good working relationship with the cooperators we met. We had frank discussions with both the Corps and National Weather Service. The Portland Corps District works closely with the State Emergency Management Agency and FEMA during floods and feel that they are forced into the role of a flood clearing house. For this reason they were pleased to have USGS data presented on the Internet. This cut down on the number of phone calls to the Corps. Because data are transmitted by both the USGS and Corps, inconsistencies in the data released to the public sometimes occurred, although the Corps tried hard not to release information that was inconsistent with data served by the USGS or NWS. The Corps feels that data must be sent over the GOES Alert channel during subsequent floods. More frequent updates are needed even though our data often serve only as backup. Unlike the Seattle District, the Portland District of the Corps would like to move away from VHF and phones and telemeter most data via GOES. Dick and Ron, pointed out that no matter how fast data are updated, someone will always want it faster. They would, none-the-less, like to have updates to GOES data more frequently than every 4 hours. Both the Corps and NWS ran into problems during the flood because USGS ratings had not been extended very far above recent high-flow measurements. As a result, Corps and NWS hydrologists were forced to extend ratings with little, or no, familiarity with the sites. River Forecast Center, National Weather Service (Chuck Orwig). The River Forecast Center in Portland relies on the Sacramento River Model for forecasting. Data are gathered from NOAA (Wallops Island), Corps, BPA and USGS. Forecasters are concerned about the reduction in USGS funds which has caused some gage sites to be discontinued. Over the years 15-20 gages at flood warning sites have been discontinued. The NWS feels they have been forced to continue to operate some discontinued gages because they are at key flood warning locations. The NWS only collects stage data but computes discharges at such sites, despite the fact that ratings are not maintained. Chuck felt that the NWS forecasts could be improved if USGS was better able to keep up with rating changes. He gave the example of the Cowlitz River, where the channel had changed several feet, but data available on CROHMS had not been updated. NWS responsibility does not include tidally-affected flooding. More data are needed in these areas as well as better models. This might be a place where USGS modeling expertise could help NWS produce better stage forecasts. Outflows from Corps and USBR reservoirs are based upon spillway and turbine ratings that clearly have problems. Chuck would like USGS to work with the Corps and NWS to verify as many such ratings as possible. Combined suggestions from Oregon cooperators Chuck Orwig sat in on our discussions with the Corps. Here is a summary of their combined suggestions: 1. Activate the GOES Alert channel during floods. 2. Collect precipitation data at all USGS gages. This would greatly aid river forecasts. 3. Collect additional data in tidally affected areas and consider how the USGS could help NWS predict flooding in such areas. 4. Find ways to increase funding for gaging stations. Do not reduce the number of stations and find ways to reactivate stations that have been discontinued. These are generally at sites that were previously agreed to be important flood-warning locations. 5. All Federal Agencies should participate in a post-flood forum where they can present how they cooperated to provide flood warning and forecasting. This forum should also be used to demonstrate what is needed to allow Federal Agencies to do a better job. 6. Nationally, USGS should cooperate with NWS as they begin to develop their next generation of flood-forecasting models. 7. Evaluate the location of all critical gages used for flood warning. Relocate those gages if they will not withstand a 100-year flood. 8. USGS should extend all ratings to the peak of record, or to twice the 1-percent flood. 9. USGS should consider making personnel available on a 24-hour basis during life-threatening floods. Summary The observations discussed above were used to develop specific recommendations that fall under three broad categories--Be aware, Be prepared and Be decisive. It should be noted that cutting across all three categories is the recognition that WRD's role during floods is changing. Serving real-time data on the Internet is placing us in a position from which we must be ready to respond quickly to data needs of local, state and Federal emergency management agencies. We feel that to optimize our efforts during floods, the USGS should: 1) Be aware that: A) Serving real-time data on the Internet has the potential to greatly increase the role of WRD Districts in flood warning and flood forecasting. Making data available on the Internet tends to heighten the awareness of our data to a broad audience and raises the expectations of both cooperators and the public at large. Although this means we will be serving the needs of more people, it also means that WRD Districts serving real-time data on the Internet will feel increased pressures to keep stage data transmitting from many, or all, of the stations they operate. It may also mean that Districts will be asked, or expected, to provide 24-hour support to local, state and federal agencies during flood emergencies. Districts must also be aware that data conflicts will likely arise as other agencies serve WRD data, or their own data, on the Internet. Coordination with all such agencies is needed before flooding occurs. This leads to the suggestion that, B) We must continually update and improve ties with local, state and federal agencies responsible for emergency response. It is too late to develop such relationships when floods occur. WRD must establish ways to quickly and reliably transmit stage data and rating changes to other agencies during floods. Increased flood warning and flood forecasting activities provides WRD the opportunity to work closely with local, state and federal agencies. We should take advantage of this and demonstrate how our products help other agencies and discuss what we and the other agencies need to do a better job. Jointly sponsoring post-flood forums would help demonstrate these points. FEMA needs during and following floods are rather specific. As demonstrated by the Washington District, the USGS is in a position to help FEMA, but we must understand their needs. Districts need to develop a continuing dialogue with appropriate technical people on regional FEMA offices regarding interagency communication and support before, during and after floods. C) WRD's software that serves real-time data is not adequate for flood-warning purposes. Updates to WWW pages take far too long and thresholds cannot be set to adequately warn users that stages have exceeded flood levels. D) Cooperators who rely on USGS data for flood warning and flood forecasting feel continued cuts in data collection funds are hurting their flood warning and flood forecasting capabilities. The NWS representative we talked with seemed to feel that WRD has an agreement to support the NWS despite the fact that WRD does not receive direct funding for such gages. 2) Be prepared by: A) Developing some formal mechanism to support Districts during floods. District Chiefs must know that they will be supported both in terms of money and personnel. We suggest that WRD consider the following support mechanisms: 1. An inter-District support agreement, where-by all Districts would agree to provide a limited number of employees to help neighboring Districts during and after floods. Districts would provide the support without compensation unless the receiving District obtained supplemental funds. 2. A working capital fund that could be built up by all Districts in a Region, or within the Division. This capital fund could be administered by the Region or Division. B) Evaluating the location of all gages used for flood warning. Try to insure gages that are critically needed will withstand the 1% flood. The loss of such gages is devastating to flood warning efforts of our cooperators. This issue becomes even more acute as funding reduces flood-warning networks. C) Extending all ratings used for flood warning and flood forecasting to the peak of record or to twice the 1-percent flood. These ratings must be extended prior to floods so cooperators are not forced to make adhoc extensions during floods. Although WRD has traditionally cautioned against extending ratings more than twice the measured discharge, ratings can be extended farther than that with some confidence by optimizing scale offsets and utilizing hydraulic techniques such as slope-conveyance and step-backwater. Serving real-time data has made this an especially important issue. Cooperators will have critical needs for discharge data even though stages may have exceeded ratings. Also, serving real-time data seems to have reduced time traditionally spent on making high-flow measurements. There is more emphasis on keeping gages transmitting than on making discharge measurements. D) Making sure offices are adequately staffed and trained to handle floods. Managers should not let prolonged flood-free periods lull them into a false sense of security. An annual flood exercise might be useful for identifying shortcomings. E) Cross training personnel. Be sure personnel normally involved in investigations receive training that will allow them to assist in collecting data during floods. It may take too long to bring trained personnel in from other Districts in time to help during early phases of flooding. Having personnel within the District who have been trained in data collection activities would help the overall response to flooding. F) Making sure adequate flood-measurement equipment exists in all offices. All offices should have adequate supplies of measuring cranes, Columbus weights, current meters, etc. Also, the adequacy of cableways and bridges as high- flow measuring locations should be assessed. 3) Be decisive by: Acting quickly to authorize overtime and arranging for personnel to come in from outside the District, or from other offices of the same District, to support data personnel. Outside personnel should be used both during and after floods. During the floods, they can help with measurements, station maintenance, etc. After the flood, they can help with indirect measurements, gage reconstruction etc. This external help will allow normal District operations to resume as quickly as possible after floods. Without such help, flood effects can linger in District operations for years. Developing a funding support mechanism, as mentioned above, would greatly facilitate decisive action. Decisive action will help maximize the amount of data collected during the flood. Every effort possible should be made to define ratings using current meter measurements and reduce reliance on indirect measurements. Indirect measurements are much more costly and time consuming and are associated with much more uncertainty than even fair or poor measurements made during high stages. We should not get in the habit of relying on indirect measurements. ************************************************************************* Water Resources Division Western Region 345 Middlefield Road, Mail Stop 470 Menlo Park, California 94025 MEMORANDUM February 25, 1997 To: Chief Hydrologist From: T. John Conomos Regional Hydrologist Subject: PROGRAMS AND PLANS: Some Needed Planning for Large Floods When you visited Menlo Park on January 14, we discussed the need for more formal documentation of the Division's expectations regarding district flood responses during very large floods. You asked that we prepare a memorandum to you reiterating our ideas. This memorandum serves that purpose. We recommend that the Water Resources Division prepare a document that clearly defines the authorities, expectations, responsibilities, and communication channels during very large floods. We refer to this document as a national flood plan, although we do not presuppose that it will take the same format or content as the district flood plans. In the past, the district flood plans were perhaps sufficient for the purpose. Four major factors have come into play over the past few years, however, that make a national flood plan essential: 1. Funding problems related to more restrictive auditing of the districts' use of funds. 2. The flood-management value of our near-real-time data. 3. Our increased role in flood-hazard response, mostly brought about by our near-real-time data and the distribution of those data on the WEB. 4. Changing roles of Federal agencies, such as the increased take-charge role of Federal Emergency Management Agency (FEMA), the decreased flood-documentation funding for Corps of Engineers (COE), and the ambiguous role of the National Resources Conservation Service (NRCS). In addition to these four major factors, there are at least five other reasons for us to develop a national flood plan: 5. The appropriate level of USGS field response to very large floods is not clear in the minds of many district chiefs, nor is it in the minds of several FEMA managers. 6. USGS documentation of very large floods is usually quite slow, often uncoordinated with Other Federal Agencies, and variable in content. 7. The effort to allow the USGS more credit in the media needs more structure, guidelines, and enunciation of authorities; 8. COE and USGS are increasingly in dispute over calculated flood frequencies, and the disagreement is causing great frustration to FEMA and local managers; 9. Standardized, WEB-based reporting forms would greatly aid all parties, and also would allow WRD managers to focus their thoughts and actions on strategic actions during the floods. More detail on all these issues follows: Issue 1: Funding.--Now that discretionary funds are so tight in Headquarters and essentially gone from district offices, owing to recent Inspector General rulings, the means of funding a major flood response are not as clear as they need to be. District chiefs face a no-win situation because there are no guidelines. If district chiefs respond conservatively, they run the risk of being questioned about their weak flood-response efforts, and if they respond aggressively, they could be placed in an end-of-the-year deficit and then be questioned about the "excessive" expenses they incurred. The expected response, therefore, needs to be better defined in a national flood plan. In addition, at least some of these reserve funds and any supplemental appropriations need to be carryover or multiyear funds. It is nearly impossible to fix flood damage and do all the necessary scientific response in the same year. It is also essential that we have the ability to retain reserves, for example, in order to respond to September floods without the risk of losing the funds on September 30. The issue of reserve flood funds at the district level and at Headquarters level also needs resolution. The latest Department ruling on "above normal expenses" was that we are in the hydrology business, and we know that floods happen, so USGS should be holding reserve funds. The question of the level of funding which should be held needs to be more clearly defined, probably in consultation with the Department and OMB, if it is to benefit us during subsequent supplemental appropriations requests. Issue 2: Increased Demand for Near-Real-Time Data.--It is clear from the February 1996 floods of the Pacific Northwest, and the January 1997 floods of Nevada and California that the USGS has a new role to play in flood response. Historically, measurement of flood flows was our principal flood response, with leisurely (by today's standards) post-flood documentation of flood magnitude in flood reports. Now, our near-real-time data are in such demand that serving this demand is interfering with our ability to measure the flood flows. Serving this demand entails (1) maintaining the lines of communication from the gage to the end user, not just during floods but at all times so that we are ready for floods, and (2) having redundancy built into our gages so that the streamflow data dependably comes through. Redundancy includes multiple floats/transducers, multiple lines of communication (DCP, land line, cellular, radio), and multiple downlinks and computer systems to accept and process the data. The choices regarding serving near-real-time data during major floods are principally three: a. go back to the "good old days" approach and refuse to serve the near-real-time data needs; or b. serve the near-real-time data needs and neglect the flood-flow measurements; or c. serve both needs, at considerable increase in cost and manpower requirements. In our opinion, option c is the only one that will keep us viable, but the problems of funding this option have not been resolved. Because options b or c place USGS squarely in the role of assisting in the protection of life and property, we cannot implement these options in a less than total manner. The serving of real-time-data during a major flood must work efficiently and reliably. The USGS liability and embarrassment potential for inadequate planning and execution are too great to ignore; but at the same time, these concerns should not prevent us from fulfilling this needed service to the Nation. If the Division is going to transition to options b or c, the national flood plan needs to address the issues of new expectations, interdistrict manpower coordination, and new sources of funding. Issue 3: USGS Hazard Response Role.--Our near-real-time data have squarely thrust us into the public eye as the agency possessing the data that are needed for flood-hazard response. Increasingly, we are invited "to the conference table" by the National Weather Service (NWS), FEMA, and the water management agencies during a flood crisis. At these meetings some district chiefs are being criticized by NWS, and to a lesser degree by FEMA, over unmet expectations. Some offices of NWS have repeatedly and publicly stated that USGS is not fulfilling its Memorandum of Agreement (MOU) with NWS. The NWS needs more flood-flow data, and feels it is our obligation to provide it. Their point of view may be far removed from ours, but it is not good for either agency for these issues to be aired publicly. The districts should not be in the position of trying to defend to every River Forecast Center (RFC) manager why our Collection of Basic Records (CBR) funds are so tight. That explanation needs to be made to NWS at the national level, and the request made to NWS that the RFC managers be apprised of the CBR-funding problems that USGS has faced. We need to more clearly articulate when and why gaging stations are discontinued, so our customers realize how greatly our network is controlled by the funding actions of Congress and other agencies. It may be necessary for USGS to amend the MOU with NWS to reflect current reality, rather than a goal we can no longer reach. Other issues relative to our increasing role in flood-hazard response are: NWS feels that a lack of telemetered data is hindering their flood- forecasting efforts. Should Congress be asked to fund a plan for updating all forecast stations if NWS would support it? This plan could easily find support from many other agencies--FEMA, COE, NRCS, BOR, BPA, TVA, BLM, USFS, FWS. The NWS should strongly endorse this effort because their LARC system (land-line telephone) is not reliable during very large floods. It works adequately on the small to intermediate floods, but telephone infrastructure often fails during the large floods. We must work with multiple agencies to fashion a plan that works for Federal and local interests. Perhaps we can partner with State Engineers, who would most likely support the notion that more money should be spent on floods, which might actually lead to less money spent later by FEMA to clean up the aftermath. The more near-real-time data we serve, the more coordination we need with other Federal, State and local agencies. This consequence is not negative. It just requires considerable high-level attention paid to the issue. In the Columbia River basin, the Columbia River Water Management Group, a tight consortium of Federal agencies, serves as the ongoing link between the key water agencies during a flood. Each geographic region may need a different arrangement that is customized to its needs. The key to success is that water agencies in appropriate geographic regions need to meet frequently enough during tranquil periods to resolve all the impediments to communication they will need during emergencies. Some of this will require upgrading equipment or installing redundancy. These impediments need to be resolved on a case-by-case basis, as does the mechanism to fund the effort. If all the agencies work together and agree on the need for change, they can, region-by- region, help us convince Congress. This effort is well within the charge to USGS by OMB Memorandum 92-01. The Internet presently does not seem to be a universally adequate medium for serving near-real-time data during a flood crisis, nor are phone lines, telephones, or cell-phones. Our experience with the 1996 and 1997 floods indicates that a multimedia technology is needed to provide adequate service to the public-safety agencies. With the migration of WRD to Windows NT servers, the tools are available to serve data via the Internet as well as by automated fax technology via phone lines and agency teletype networks. To do so reliably will require a serious national investment in software and hardware that is compatible with systems used by regional emergency response agencies. This requirement is not a serious technical challenge--the technology is already in use by other agencies--but it is one that cannot be met by each district inventing its own solutions. Also, in some cases we do not provide the data that are most needed by the decision-makers. The Division needs to address this issue of how to best serve near-real-time data during a flood crisis, and provide mechanisms for doing it. This single effort could well be our best customer service (and outreach) effort in the immediate future. Issue 4: The Changing Roles of Other Federal Agencies.--In the period of immediate response and assistance to a "national" disaster, FEMA assumes the command-and-control role, theoretically in partnership with the State disaster agency. However, there seems to be no concurrence within FEMA as to the appropriate role of USGS in providing technical assistance. A national MOU is needed between the two agencies to enable USGS to quickly meet flood- documentation needs of FEMA and other agencies while the field data are still available. If FEMA has the expectation that USGS is willing and able to provide flood documentation, we will need to develop a National strategy to quickly detail help to the flooded region. These details must be considered a National priority. Currently, it is very difficult to receive this level of help. Our priorities must change if we plan on doing flood documentation promptly and correctly. The roles of the COE, BOR, and NRCS are also unclear relative to FEMA and USGS, and vary significantly from place to place. In some areas the COE presence in data collection and flood documentation is major. In other basins, we are finding that the COE has lost funding for flood documentation and is not providing the expected service. In basins where BOR has water/reservoir management purview, BOR's expectations of USGS in providing real-time data and immediate reestablishment of damaged data networks are similar to that of NWS and FEMA; that is, they believe that it is the USGS mission, to be done with USGS funds. High-level interagency MOUs defining the principal roles of each agency would greatly facilitate quick response of USGS to serve the public need for flood documentation and data-network recovery. We also must make it clear each year when we negotiate the operation of a cooperator- or an OFA-funded gage that the O&M costs do not cover catastrophic (50-year recurrence interval or greater?) floods or other catastrophic events. Currently, with almost no discretionary funds in districts or Headquarters, it is difficult, if not impossible, to replace or restore gages and other field equipment that are lost or damaged. Failure to make this point clear to all other agencies has resulted in unnecessarily unmet expectations. Issue 5: Appropriate USGS Field Response to Floods.--The USGS must reexamine how and why we acquire flood data. There are several issues. First, some of the old-guard USGS surface-water hydrologists, now in positions of authority in other agencies, feel that we are not doing everything we should to obtain peak flows. They cite what seems to be a proliferation of indirect measurements in lieu of actual measuring of peak flows. They also cite the lack of maps in the station folders showing preplanned alternative routes to gage stations, which they claim had been part of a standard contingency plan in the past. In contrast to these allegations, our district chiefs rightly say that in many cases the degree of flood inundation of highways was greater than ever experienced before; thousands of bridges were closed, and highways inundated; cableway accesses were under water; and debris in the floods was so extreme as to prohibit safe measurements. On the surface it would appear to be an argument for which there is no common ground attainable. We think otherwise. The debate leads to the question of what is a reasonable effort on our part today? In this era of tight funding, we must prioritize numerous aspects of the field effort and clearly enunciate those priorities to all interested parties well before a major flood. In the absence of such a clear statement of what USGS will do during major floods, expectations will be high and unattainable. One part of such a prioritized plan could be to designate all of the USGS/ NWS flood-forecast stations as priority-one sites for flood response. Conceivably, the national flood plan for these priority-one sites might specify that: * Cableways would be upgraded to assure flow-measurement access even during 500-year floods; * Data telemetry would be redundant; * Rating curves would be extended to the 500-year-flood level; and * There would be a plan for getting to the site to accomplish flood- flow measurements regardless of the road conditions. For these sites, a prearranged contingency plan to use helicopters to reach these critical sites might be reasonable. (The major issue in helicopter availability is not the supply, but advance coordination and MOUs with the emergency-response agencies to get airlifts for USGS.) A national MOU with the Air National Guard, Air Force Reserve, Navy, and Coast Guard would be very valuable. Without this agreement, local military/disaster commanders are too apt to say, "I'd like to help you, but I have no authorization. . . ." Regardless of a station's flood-response priority, if roads are impassable, districts mostly rely on indirect measurements after flooding has abated. In some cases the cost of a helicopter to get to stations and measure the peak flows would be cheaper, and the data would be better. There are certainly many situations where helicopter access would not be possible, but there are enough situations where it would be possible, that it should be considered, and planned as a contingency for those stations where the acquisition of peak flows is important and is possible by helicopter access. Again, a preflood agreement with the Air National Guard would probably be necessary. At some USGS flood-forecast stations, preflood contingency measures might even include creating and maintaining a cleared-landing area to assure helicopter access. Clearly a national flood plan must define what is expected of a district in terms of in-the-field flood-response needs. Expectations will differ among stations depending on the station's flood-response priority. Once this flood plan is published, there will be a major spinoff benefit from this particular element of the plan. That is, FEMA (and other OFAs) will be less apt to not fund some of our work because they think it is already our mission responsibility. If we can demonstrate a national flood plan which clearly defines our funded responsibilities, then anything beyond those responsibilities is more likely to be accepted as an above-mission request by FEMA. Flood-hazard response is one of our most important functions. In our view, we need a SWAT-team approach that is administered by regions and Headquarters. When a major flood occurs, capable investigators are flown in to help document what happened, repair gages, flag and run indirects, flag-flooded areas, calculate recurrence intervals, and handle the media. Each district should have two or three people designated as Flood-Crisis Assistance Team (FCAT) members, so as not to create confusion when a call for support and aid is made by region or Headquarters. Issue 6: Flood-Report Documentation of Floods.--Because the Division has no policy regarding what is expected in terms of flood-report documentation, some floods are better documented than others. A national flood plan could define the expectations, both as to when a flood report is necessary, and elements that should be included. Since the advent of empowered onsite Federal emergency coordinators during national disasters, MOUs that predefine the role of USGS in providing flood- documentation support to FEMA Disaster Field Offices are essential. Understandably, we must be willing and able to respond rapidly and well to any MOU that obligates us to document floods. Issue 7: Receiving adequate and accurate media coverage.--There is a greatly increased emphasis on the districts to assure that USGS gets proper credit in the media for our activities during major floods. This recent emphasis, added to the increased demands of near-real-time data and increased interagency coordination, leaves district management greatly overextended during major floods. Some consideration must be given to having the Public Information Officers (PIO) of the Division and bureau go onsite during a flood crisis (or other crisis) and write the news releases, have them reviewed by the resident scientists, and then transfer the releases to the media. Our current PIO infrastructure needs more support to reach our needs; that is, the level of sustained effort is presently ineffective and inadequate, even during nonemergency times. The one caution, however, is that the district chief should always have final approval on the accuracy and contents of the news releases. That is, the commander in the field should determine what is said publicly. Issue 8: Flood Frequency Disputes.--The USGS and COE use remarkably different methods to estimate flood frequencies for streams that are flow regulated. Not surprisingly, the results are different. These differences are confusing to other agencies and embarrassing to USGS and COE. An agreement on flood- frequency methodology that is negotiated at the national level among COE, FEMA, NRCS, NWS, and USGS is essential. Having these agencies reach consensus on methods and technology would be received positively by the Administration, Congress, and the public as a government-reinvention success story. The USGS should convene such a group; our authority to do such comes from OMB Memorandum 92-01. The public is confused as to why 100-year floods do not necessarily occur every 100 years. Because this term is so poorly understood we believe that the Washington District Fact Sheet on 100-year floods should be rewritten from a national perspective. Such a Fact Sheet would be very educational, particularly for the media, and it also should be used to illustrate that 100-year flood estimates could not be derived without data collected by WRD. Additionally, the USGS needs to reexamine the issue of how we analyze and report peak-flow data. Is our current policy regarding calculation of peak flow using only the preregulation record the best policy today? What do the water-management agencies, emergency-management agencies, and the public really need and want? Are we giving them the information they need, and in a form they can use? Issue 9: Standardization of Forms and Reporting.--Finally, a standard form would be very useful for supplying data in support of flood reporting and supplemental-appropriation requests. The current OSW template is a big step in the right direction, but it should be carried further with even more benefit. For example, a form could be created for districts to update information on flood effects (recurrence intervals obtained, gaging stations damaged, etc.). If such a form was created for the WEB, it would stop the very short turnaround-time requests from Headquarters, and would afford district personnel more time to respond to the unusual information requests. We are pleased to provide this memorandum based on our experiences, and offer to assist the Division in developing and implementing a national flood plan.