Note: This paper was orginally given at the 1995 Conference of the American Meteorology Society.

Engaging Planners and Investors in the Assessment of Storm Risk in Jamaica

Jan C. Vermeiren, Keith Ford, Charles C. Watson

1) Department of Regional Development and Environment, Organization of American States (OAS) ; 2) Former Coordinator, OAS/Caribbean Disaster Mitigation Project; 3) GIS Consultant, OAS/Caribbean Disaster Mitigation Project;

1. Introduction

The General Secretariat of the Organization of American States (OAS), under an agreement with the Office of Foreign Disaster Assistance of the U.S. Agency for International Development, is executing a five year Caribbean Disaster Mitigation Project (CDMP). Principal objective of the OAS/CDMP is to establish sustainable public/private sector mechanisms which measurably lessen loss of life, reduce the potential for physical and economic damage, and shorten the disaster recovery.

One component of this project is the assessment of potential hazards generated by tropical storms, in terms of storm surge, coastal flooding and extreme wind.

For this purpose, the OAS/CDMP has developed a computer based numerical model that produces estimates of maximum sustained wind vectors at surface, and still water surge height and wave height at the coastline, for any coastal area in the Caribbean basin. The model relies on a generic data base structure, and uses USGS digital data for deep ocean bathymetry, DMA digital data for land boundaries and rough topography, satellite imagery for foreshore bathymetry and land cover, and the National Hurricane Center data base for storm data. Model runs can be made for any historical storm, or for probable maximum events associated with different return periods. Model outputs can be exported to common GIS formats.

2. Storm Hazard and Land Use Planning in the Caribbean

In several of the Caribbean countries, migration from rural areas towards centers of economic opportunity starting in the 1970's far outstripped the development planning and regulatory capacity of their governments. Migrants settled on marginal lands, along coastal strips, in floodplains, and on steeply sloped lands. Simultaneously, the growth of tourism in many of the islands attracted substantial investments in tourism and other economic infrastructure to the coastal areas exposed to significant storm risk.

Hurricanes David and Allen prompted a growing recognition of the vulnerability of development to tropical storms, and led to various initiatives in identifying and mapping, usually at large (national) scales, hazard prone areas. Demonstrated use of this information remains limited.

Some limited coastal hazard information is generated by engineering firms when designing major coastal civil works, using the traditional "single transect" method to estimate storm surge and wave set-up. Such estimates may be sufficient for design purposes, but are of little value to other applications due to simplifying assumptions and site specificity.

In the mid 1980's, the Federal Emergency Management Agency and the Department of Marine Sciences of the University of Puerto Rico undertook a Flood Insurance study for Puerto Rico and the US Virgin Islands using the SLOSH model. The University of Puerto Rico continues to develop further applications and to support the preparation of hazard maps in Puerto Rico.

More recently in 1992, the National Weather Service of the Bahamas and the National Hurricane Center collaborated in the preparation of a SLOSH model for the northern section of the Bahamas, including New Providence. Attempts to use the information for land use planning are constrained by the low resolution of the SLOSH data, relying on a telescoping grid with single cell sizes averaging 10 square miles.

3. Storm Hazard Assessment in Jamaica

Two hurricanes hit Jamaica during the 1980's. Hurricane Allen (August 1980) passed approximately 30 miles north of the island, causing coastal damage through storm surge and wave action. The Geological Survey Division inspected the north coast and analyzed the extent of the damage, estimating surge and wave height from debris lines [1]. The report was widely distributed but was not in a format that could be easily translated into land use planning guidelines.

Following hurricane Gilbert, the Caribbean Environment Programme of UNEP supported a study of the economic impact of the hurricane on coastal and marine resources.[2] The report details the storm's effects on coastal ecosystems and economic activities such as fishing and oyster culture. No direct coastal land use recommendations were made.

In 1986, the Government of Jamaica, with financing from the United Nations System embarked on a Floodplain Mapping Project preparing large scale flood hazard maps for all major rivers. Information generated by this project is now being fed systematically into the physical planning and development control process.

The need for improved understanding of storm hazard and risk in the Caribbean has been driven home by the recent upsurge in insurance costs following the underwriting losses suffered by insurance and reinsurance companies in the aftermath of Gilbert, Hugo and Andrew. Accurate hazard maps will allow the insurance industry to improve risk management and underwriting practices. This fact greatly increases the demand for hazard maps and therefore the likelihood of their systematic production.

4. National Institutional Aspects of Storm Hazard Assessment

Upon embarking on an assessment of storm hazards in Jamaica, the OAS/CDMP invited the public and private sector agencies with interests in the activity and its applications to form a technical working group. The group includes development planning and natural resource agencies, the University of the West Indies, and the Jamaica Association of General Insurance Companies (JAGIC).

A high resolution coastal digital elevation model produced by the Geological Survey Division, one of the cooperating agencies, was incorporated in the storm assessment model to enable production of large scale storm surge and coastal flooding maps for the island. Staffs of participating agencies were trained in its operation.

The project's original design, in which one of the cooperating agencies was to serve as a platform for installation and applications development has not been achieved as of this date. Principal reasons for this are: (a) agencies which are mainly sectoral have difficulty in accepting a responsibility with multi-sectoral implications; (b) a tendency among professionals in government and university agencies not to share information freely; and (c), slow progress in achieving institutional commitments and support from policy makers.

Continued efforts are directed at overcoming these obstacles. Meetings with the political directorate and top level technical personnel are planned to discuss institutional commitments and interagency cooperation. Cross-cutting policy issues, e.g. the impact on the ministries of tourism and construction of more stringent enforcement of coastal development standards by the planning department, and recommended policy changes, will be the main focus of these meetings.

The project is exploring different institutional arrangements around specific applications of the storm hazard assessment model. The project collaborated directly with the Greater Montego Bay Development company, a mainly private group, in the preparation of 1:10,000 coastal flooding maps for 50 and 100 year return periods. And the Meteorological Office, the Office of Disaster Preparedness, and the Jamaica Broadcasting Service recently agreed to collaborate in the installation of the model to project wind and possible surge and coastal flooding impacts using the NHC tropical weather forecasts.

5. Incorporating Hurricane Loss Reduction in Development Decisionmaking in Jamaica

The local Office of Disaster Preparedness and Emergency Management (ODPEM) has been one of the leading proponents for the incorporation of vulnerability reduction measures in development planning at all levels. In 1982, the ODPEM developed a Natural Hazards Management Plan for the country. This plan, which was completed in 1984, recommended a combination of structural and non- structural mitigation measures to reduce the island's vulnerability to disasters and also recommended the development of a comprehensive hazard mapping project. Recommendations from this plan were incorporated in the Kingston Metropolitan Regional Plan which was completed in 1986, and also led to the Flood Plain Mapping Project mentioned earlier.

The ODPEM is again an important actor in disseminating the information generated in the storm hazard mapping exercise of the OAS/CDMP. There is evidence that the information is being used by both public and private sector agencies/companies including private engineering companies. The ODPEM regularly receives requests for information generated by the project, from both the public and the private sectors. Additionally, some insurance companies are contracting services to help determine their exposure to hazards, including their probable maximum loss from extreme events.

A substantial part of the project's resources are directed at training and technology transfer. Staffs of the collaborating agencies are trained in the use and applications of the storm hazard assessment model. Workshops are held for planners and representatives from the insurance and development financing sectors in the use of hazard mapping. The investment in training is aimed at ensuring that the activity can be sustained beyond the project's timeframe.

6. Reliability of the Storm Hazard Model

The storm hazard assessment model used by the OAS/CDMP in this project is new, therefore placing demands on the project to demonstrate the reliability of its outputs, and to gain acceptance in the professional community. The TAOS model (The Arbiter of Storms), was originally developed by Charles Watson, who used it to analyze tropical storm risk to Hilton Head, South Carolina. The model is described in detail in a forthcoming issue of the NWA Digest of the National Weather Association. [3]

The ultimate test of any computer model is how well it compares with reality. A new model must not only prove itself in comparison to the real world, but also must be shown to outperform existing techniques in order to gain acceptance. In the case of storm hazard modeling, comparisons with reality are complex. Each of the four components of storm damage — wind, storm surge, rain induced damage (including inland flooding and landslides), and wave action — present particular challenges. Storm surges are not well documented in the Caribbean. Most islands have at best one or two functional tide gauges, the most accurate measurement of surge. In extreme events tide gauges are often damaged or their range exceeded.

In addition, the meteorological parameters for storms are rarely known with the precision, witness the ongoing debate over wind speeds in Hurricane Andrew, a relatively well-observed storm. Wind measurements are even spottier in the Caribbean than in the US. Wave data is also difficult to obtain, and often must be extrapolated from observed damage. This uncertainty in observational data makes verification difficult and comparisons between models tentative.

The TAOS model has been tested against reality and compared with storm surge models—SLOSH and a model developed by Meteo-France for the area around Guadeloupe and Martinique. TAOS has performed well when compared with tide gauge readings. For example, a relatively coarse run (4 km per cell) of the Atlantic regional model was made for Hugo ('89), Allen ('80), Camille ('69), Betsy ('65) and Carla ('61). Hourly comparisons were made against 15 tide gauges located near the tracks of these storms. The largest error encountered in the 208 readings in comparison with hourly readings was 23%, with an average error of 7%, including relatively small (<1 meter) surges. The published accuracy for SLOSH is "generally ± 20% for significant surges." [4]

7. Conclusion

A growing interest in storm hazard and risk assessment is emerging in the Caribbean. The traditional interest in this area of the planning and scientific communities is now being reinforced by insurance and development financing agencies, who are forced to improve their management of property risk. Stand-alone PC-based numerical models combined with Geographic Information Systems are offering affordable technological solutions in support of these efforts.

1. Effects of Hurricane Allen along the North Coast of Jamaica. Wilmot Simpson et al, Geological Survey Division, Jamaica, September 1980.

2. Assessment of the Economic Impact of Hurricane Gilbert on Coastal and Marine Resources in Jamaica. UNEP Regional Seas Reports and Studies No. 110, UNEP, 1989.

3. TAOS, The Arbiter of Storms: A High Resolution , GIS Based System for Storm Hazard Modeling. Charles C. Watson. NWA Digest, National Weather Association (in press).

4. NOAA Technical Report NWS-48.