ACTIONS TAKEN FOLLOWING
THE RED RIVER OF THE NORTH FLOOD OF 1997

Following the devastating floods of 1997 on the Red River of the North (RRN), the National Weather Service (NWS) has initiated several actions to improve forecast services and take advantage of "lessons learned" from this event to improve NWS river forecast services for the area affected by this flood and for the Nation in general. Some of these actions were taken immediately and have been in place for some time; others are under way as of this writing (August 1998); still others are in planning phases and depend on budget resources in future years. Some of the actions are highly technical in nature; others involve improvements in coordination mechanisms, policies, procedures, etc. An assessment of NWS services along with a very technical study of the hydraulic characteristics of the 1997 flood in the area of East Grand Forks, Minnesota, and Grand Forks, North Dakota, has now been released in the report, "Red River of the North 1997 Floods Service Assessment and Hydraulic Analysis." This brief document reports the status of the major actions taken since the 1997 floods. Many of the actions were recommended in the preliminary findings of the survey team presented in July of 1997 and are within the final service assessment report.

Before the event was over, NWS had made changes to the rating curve used for the East Grand Forks forecast location, and NWS has continued to update this rating to be consistent with the latest information and analyses provided by the U.S. Geological Survey (which is the agency that collects stream flow information at the site). A widely attended meeting at the NWS North Central River Forecast Center on July 28-30, 1997, provided the forum for a post-flood technical review among Federal agencies (including NWS, USGS, the US Army Corps of Engineers, and the Federal Emergency Management Agency), the Manitoba Water Resource Branch, numerous state officials from North Dakota and Minnesota, and local officials from the Fargo and Grand Forks areas. This technical session addressed numerous coordination and information sharing issues raised by the service assessment. Following this technical meeting, all parties can confirm that they are aware of all of the relevant studies, reports, etc., that may be of value in a future event; and several procedural changes have been adopted in the handoff of information among these organizations (e.g., a revised method to pass forecasts to Canada). Additional planning activities under the aegis of the International Joint Commission (IJC) have continued to involve Canadian and U.S. officials in efforts to improve flood forecast capabilities for the RRN. And a flood planning meeting in Grand Forks on January 21, 1998, brought together local, state and Federal parties with interest in the 1998 snowmelt flood season in the RRN Basin (as well as the Souris basin).

The technical procedures used by the NWS to produce forecasts for the RRN have been looked at in great detail; and several efforts are either under way or planned to improve on these procedures. Those improvements that can be carried out within the existing resource base of the NWS are already under way. These include:

• using the recently completed hydraulic analysis to develop a dynamic routing operation for the East Grand Forks gage site,

• changing the NWS forecast software to provide a more explicit warning when a rating curve extension is in use,

• reviewing the established flood stage for every forecast point on the RRN along with the associated detailed information about flood forecast services and flood impacts,

• recalibrating the RRN Forecasting System using more complete historical data and models that are compatible with the latest NWS forecast methods,

• reviewing some of the unusual flow paths that water took during the RRN floods of 1997 to add an empirical estimate of these overland flows from one stream to another to the NWS modeling procedures where possible, and

• developing an enhanced system for analysis and use of snow information, and analyzing the existing flood outlook procedures to determine whether a useful estimate can be provided of the chances that the outlook flood crest will be exceeded.
  
  

A few of these technical improvements were available in rudimentary form for the 1998 snowmelt season on the RRN. Presuming that no unexpected difficulties are encountered, all but two will be fully in place for the 1999 RRN snowmelt season -- the transition of the RRN Forecast System to recalibrated models throughout will only be partially complete; and the development of an enhanced system for analysis and use of snow information will not be complete.

A far more sophisticated set of model improvements is planned based on the Advanced Hydrologic Prediction System (AHPS). This capability has been prototyped for the Des Moines River in 1997 and is included in the budget submitted for Federal fiscal year (FY) 1999 by the President. Based, in part, on the 1997 floods, the RRN has been identified as the highest priority implementation area for AHPS.

Presuming AHPS budget resources are eventually provided in the FY 1999 appropriation, NWS will:

• proceed to complete an operational forecast procedure for the RRN that includes advanced dynamic routing procedures to account for the complex hydraulics of the entire RRN (not just the East Grand Forks gage),

• develop a method to explicitly account for ponded meltwater in the flat terrain of the RRN,

• look carefully at physically-based methods to model unusual flow pathways at very high flood levels, and, most importantly,

• move the entire forecast system for the RRN into a longer time horizon forecasting methodology that explicitly accounts for forecasting uncertainty and allows for objective risk-based decision making.
  
  

As the first area (after the Des Moines prototype) scheduled for full-scale AHPS implementation, the RRN has National implications for the NWS plans for future advanced hydrologic services.

Some of the most important lessons from the RRN floods of 1997 have to do with the way that information is provided to, interpreted by, and used by NWS customers, not just the technical details of how the NWS produces forecasts. Here again, there are actions that have already been taken (e.g., inclusion of discharge forecasts in handoff of data from the NCRFC to Canada, additional explanatory information on outlooks, forecasts, and forecast uncertainty now included in NWS products) for the RRN as well as broader implications for the NWS as a whole. The NWS provided very early information on the risks of severe flooding in the RRN in 1997 and attempted to convey the uncertainty in its forecast products for this event. With the NWS planning changes in its hydrologic forecast products under the AHPS initiative, the lessons learned from the RRN floods of 1997 are of great interest.

The meeting in July of 1997 at NCRFC mentioned above settled many technical matters for the RRN, but the RRN flood event of 1997 also raised coordination issues of National significance:  The USGS and the NWS have agreed to review forecast sites for the whole Nation to determine where more complex variable rating models are needed. The NWS will also look into options for enhancing coordination with emergency management officials at state and local levels during similar events, consistent with NWS staff resources.

Finally, the RRN floods of 1997 have led to a review of certain NWS internal procedures:

• The NWS Central Region now has in place methods to order additional telephone services on an emergency basis, with other NWS regions expected to follow.

• An analysis of the feasibility of telephone services to route around a possible major telephone company failure is underway.

• Based on the extraordinary media interest in the RRN flood of 1997, NWS offices immediately affected by that flood have already been trained in ways to be more efficient in supporting the media; extending this training to National coverage is an NWS objective as is developing better and more explicit plans for escalating the NWS response to the media as media interest escalates in any future great flood.
  
  

Perhaps it is human nature to focus on the negative aspects of a devastating flood event, but it is well to remember that many of the "lessons learned" during the RRN floods of 1997 are things done well that should be repeated by other NWS offices in future floods -- things such as, treating an unusual event in an unusual way by releasing information early and by scheduling early and extra coordination meetings, taking full advantage of the capabilities of the Airborne Gamma Snow Survey Program to quantify the condition of the snow pack, and, overall, coming forward with the extra time and effort to be available to the media and to explain the severity of the threat early and often to local, state, and Federal officials. As a specific example, the NWS Central Region uses the flood preparedness activities led by the Eastern North Dakota office in 1997 as a guide to "best practices" for other offices to follow.

The National Weather Service is committed to providing the best possible forecast services for future floods on the Red River of the North and for the entire Nation within the limits of science, technology, and resources. Actions have been taken already to improve our services to the citizens of the Red River of the North based on the "lessons learned" from the devastating 1997 floods; other actions to improve services are underway or planned that will continue to improve the river forecast information that the NWS provides for the citizens of North Dakota, Minnesota, and the Nation.