The New Year's Eve Flood on Oahu, Hawaii: December 31, 1987 - January 1, 1988 (1991)

The reconnaissance team members offer the following recommendations as a result of their postdisaster study of the December 31, 1987, Oahu, Hawaii flood.

The performance of weather forecasters was hampered by the level of technology available, resulting in a lack of adequate predisaster warnings. The unavailability of adequate radar information, the mediocre performance of the raingauge network, and the limits of satellite imagery conspired to leave forecasters without sufficient data to anticipate the flood threat.

• The automatic telemetered raingauge network on Oahu needs immediate improvement. Improvements should include increased-capacity raingauges to preclude overflow, increased raingauge density, and consideration of the implementation of high-rate sampling, such as the ALERT system (LaMarche, 1985).

• The National Weather Service (NWS) should devise guidelines to assure higher-frequency monitoring of existing telemetered raingauges.

• The Phoenix experience that documented a local agency's experience with a real-time hydrologic telemetry and warning system (LaMarche, 1985) demonstrates that raingauges alone are not sufficient. The state of Hawaii should provide some radar assistance to the NWS until the scheduled mid-1990s installation of the Next Generation Radar (NEXRAD). This assistance may include but should not be restricted to

• Providing direct access by the NWS to imagery from Hickam Air Force Base and/or University of Hawaii radars.

• Acquiring a series of inexpensive radars similar to the University of Hawaii's system to provide a network of coverage for Oahu.

• Additional research should be encouraged and funded by the federal and state governments on the subject of intense nonthunderstorm orographic rains.

The city and county of Honolulu and the state of Hawaii urgently need to incorporate the consideration of debris loads carried by streams into the planning and design of flood control works in Oahu.

The following actions are recommended:

• Studies should be initiated immediately to determine the volume of debris produced from storms occurring in the urbanized watersheds.

• Data from these studies should be used to delineate areas of different debris production throughout Oahu.

• Data from these studies should be used to develop a series of debris production curves for each of the debris production areas.

• Based on these studies, an envelope curve for maximum debris flows per storm as a function of drainage area should be developed.

• A physical-process-based computer simulation model for implementing a management strategy to mitigate potential debris flow damage needs to be developed. The modeling approach could be used to estimate channel erosion, debris deposition, bedload stabilization, and the effectiveness of drainage systems under different hydrologic inputs and operation alternatives.

• Based on these studies, design criteria for adequately controlling debris loads in the planning and design of flood control works should be developed.

The development of a realistic plan or strategy for natural disaster damage mitigation has been one of the most intractable problems in management science. Perhaps the most promising approach to this complex problem is computer simulation modeling. Studies by Simons and Li (1978, 1982) indicate that physical process computer modeling provides a reliable methodology for analyzing the dynamic process of the extreme event, as well as for developing solution strategies in damage-mitigation plans.

The problem of destructive debris flows occurring during major storms remains one of the most difficult issues in flood hazard mitigation. Counties in Southern California have grappled with this problem since the turn of the century (see Los Angeles County Flood Control District, 1971, 1987; Riverside County Flood Control and Water Conservation District, 1978). The solution remains elusive and the problem plagues Southern California to this day (McPhee, 1988a, 1988b).

Certainly, one of the major problems in the Oahu disaster was the lack of adequate lead time for the emergency management agencies to prepare themselves and for the general public to protect itself from the flood episodes. The recommendations listed in the Meteorology section of this chapter will do much to improve

the NWS's ability to monitor and provide better severe-weather information to responsible emergency management agencies.

However, a warning system implies more than identification and quantification of the developing hazard. It implies the transference of relevant information to the appropriate agencies and the general public. While emergency management agencies and response organizations were eventually contacted, and some representatives were brought into the emergency operations center (EOC), a delay did occur in the ability to respond to requests for assistance. This delay prompts the following recommendation:

• Emergency response exercises, modeled after this holiday disaster, should be enacted to identify strategies to bring knowledgeable agency representatives into the EOC and to mobilize public sector resources as rapidly as possible.

A second feature of the warning system was the transmission of the flood warning message to endangered residents. Certainly, the broadcast media quickly began to broadcast information about flooding in progress. Those who were watching television or listening to the radio on New Year's Eve would have heard about what was going on and, possibly, about the NWS's flood warning. However, for warning messages to motivate adaptive responses, they must include information on the types of actions people should take to protect themselves and must specifically identify the areas that are at risk (Nigg, 1987). These following recommendations are made:

• Attention should be given to the content of weather advisory messages and to the most effective way to transmit them to the general public.

• Although there were two island-wide warning systems available (an air-raid warning system and a system used for tsunami warnings), neither was used in this instance even though unusual circumstances may have warranted their implementation. Consideration should be given to enhancing the ability of these systems to alert residents to monitor their radios or televisions for emergency broadcasts.

With regard to the identification of areas that experienced flooding in the Oahu disaster, two recommendations are made:

• The flood hazard mapping efforts for the city and county of Honolulu should be expanded to assess currently unevaluated areas (Zone D).

• A reassessment of Zone C areas, especially in the Kailua area, should be conducted to determine whether the hazard designation is appropriate.

For a variety of reasons involving the complexities of land ownership and the timing and status of their home mortgages, many homeowners were exempt from

the mandatory flood insurance requirement and simply chose not to buy flood insurance. Others were not required to buy such insurance since they lived in areas currently unevaluated for flood potential. As a result, the number of homeowners who purchased flood insurance was low. For this reason, two additional recommendations are made:

• Although a number of loan and grant programs were made available after the disaster, an evaluation should be undertaken to determine what monetary resources homeowners and renters used to repair damage and replace possessions, especially those not covered by the federal flood insurance program.

• An assessment should be undertaken to determine whether adequate resources were available to working-class homeowners and to those in homeland areas to assist them in recovering from the disaster.