Storm Hazard Assessment for San Pedro/Ambergris Caye, Belize Statistical Estimation of Hazard Frequency and Intensity Prepared For: Caribbean Development Bank Advanced technology and analysis solving problems in science, engineering, commerce, and public policy. R&D Division, 1204 E 49 th Street, Savannah, Georgia 31404, USA
Disclaimer and Cautionary Notes The user is strongly cautioned that natural hazards modeling and analysis are subject to many uncertainties. These uncertainties include, but are not limited to, incomplete or inaccurate data, changes to the natural and built environment, limited historical records, and limitations in the state of the art of modeling scientific understanding of some of the phenomena. Users should have an appropriate background in the field of application, or seek competent advice in such fields. Anyone making use of this data or the information contained within assumes all liability deriving from such use, and agrees to "hold harmless" any and all agencies or individuals associated with its creation. The publication of the material contained herein is not intended as a representation or warranty that this information is suitable for any general or particular use. Kinetic Analysis Corporation, the University of Central Florida, and any other agency or individual associated with the creation or presentation of this data assume no liability connected with your use of the data or the information it contains, and make no warranties, express or implied, as to its usability or accuracy. The user should coordinate the use of this or any other hazard information for the purpose of design and construction with the responsible local officials where such coordination and approval is required to ensure compliance with building codes and other legal requirements. TAOS TSAP Report for Belize Page 2 of 6
Introduction This document reports the results of the return period analysis for Ambergris Caye, with special emphasis on San Pedro. Details of the methodology used are available in the companion document, Methodology for Storm Hazard Mapping. The Arbiter of Storms (TAOS) hazard modeling system was used to create the storm hazard data sets for this study. Technical details of the TAOS model system are documented in various published papers noted in the bibliography, in particular Watson and Johnson (1999). The TAOS Statistical Analysis Package (TSAP) analysis results in a large volume of information for each analysis cell in the study area. This report describes the results of this return-period analysis wind hazards for one particular point on the coast of Ambergris Caye. Results of the analysis of wind, wave and surge hazard values for three periods (50-year Maximum Likely Event, 100-year Maximum Likely Event and 100- year 90% prediction limit event) are displayed on the maps produced as part of this study, and are available in digital form in geographic information system (GIS)- compatible shapefiles. Historical Storm Information The map in Figure one show past storms that passed near the study area. Intense storms are labeled with the name and year of the event. The historical track and intensity data was analyzed using the TSAP. TSAP simulates every storm in the historical data base (currently 1320 storms), computing the wind, wave, and storm surge at the site of interest. This data is analyzed to determine the return periods for both storm events and phenomena produced by these events. A critical feature of TSAP is the ability to compute prediction limits for the model results. A prediction limit allows the user to assess and incorporate uncertainty from both the relatively short historical record as well as modeling limitations. This will be discussed further in the context of the specific analysis below. For more information, see Johnson 1997 or Johnson and Watson, 1999. Results of the Statistical Analysis: A Wind Hazard Example Hurricane frequencies are not uniform through time, but vary along with regional and global climate patterns. The primary driver of variation in hurricane frequencies at Belize (as is typical in the Caribbean) is the El Niño/Southern Oscillation (ENSO). The ENSO is a large scale ocean/atmosphere phenomenon that changes upper level winds over the Atlantic. The ENSO is normally characterized in terms of the average temperature in the eastern Pacific. When temperatures are above normal, an El Niño is said to be in progress. When temperatures are below normal, a La Niña event is said to be in progress. The raw TSAP output for a selected point on the coast of Ambergris Caye (Location: 17.98-87.90) was extracted from the model results and is provided in Table One. TAOS TSAP Report for Belize Page 3 of 6
Figure One: Tropical Cyclone tracks in the vicinity of San Pedro, Belize Table One shows the expected events for 10, 25, 50, and 100-year events at 50%, 75%, 90%, and 95% prediction limits, for various phases of the ENSO cycles. The 50% prediction limit (which is equivalent to what is reported in a traditional extreme event analysis), is the value which would be high in half of a large number of periods and low in the other half. In other words, if we divided 1000 years into ten 100-year periods, we would expect 5 of them to exceed our 50% prediction limit and 5 to fall below it. For our 90% limit, we would expect only one of our 100-year periods to exceed our forecast value. Knowledge of prediction limits allows the user to decide how much protection to build in to their design, as well as to assess the uncertainty in the forecast. As noted earlier, the ENSO cycle has a significant impact on frequencies in the Caribbean. In Table One, ALL values are the compute return periods using all 134 years in the historical record. NOR years are Normal years, ELN years are years with warm pacific sea surface temperatures ( El Niño in progress), LAN (La Niña) are cold pacific SST years. TAOS TSAP Report for Belize Page 4 of 6
A review of the 100-year return period analysis in the last section of Table One clearly shows the effect that the ENSO cycle has on these return-period estimates. This analysis of both El Niño and La Niña years results in lower expected values during these periods than for all 134 years taken together, while normal years are 15 knots higher for the 50% prediction limit than the case where all 134 years were analyzed together. Based on this analysis, we recommend using the Normal cycle return period analysis results, which are higher than the other results, although it must be noted that even using these values, during La Niña years the probabilities of higher intensity storms are significant. While using the ALL years analysis is in keeping with traditional returnperiod studies, we feel this would result in an under design as the El Niño years of suppressed activity may artificially lower the design criteria. TSAP Tropical Cyclone Return Period Analysis for Ambergris Caye Location: 17.98N, 87.90W All Years 10. yrs ALL 134 32.2 34.7 36.7 38.1 25. yrs ALL 134 49.4 53.5 57.7 61.4 50. yrs ALL 134 63.2 69.7 78.0 83.8 100. yrs ALL 134 77.7 86.1 97.4 106.5 Normal Years 10. yrs NOR 87 34.7 37.8 40.6 42.4 25. yrs NOR 87 52.5 58.0 63.8 68.0 50. yrs NOR 87 66.9 74.7 84.3 92.6 100. yrs NOR 87 81.3 91.9 106.2 117.4 El Nino Years 10. yrs ELN 26 10.8 12.6 14.4 15.4 25. yrs ELN 26 15.6 18.3 21.4 23.6 50. yrs ELN 26 19.5 23.1 27.4 30.7 100. yrs ELN 26 23.0 27.7 33.5 37.8 La Nina Years 10. yrs LAN 21 48.9 57.3 65.7 73.4 25. yrs LAN 21 70.0 84.9 105.7 127.5 50. yrs LAN 21 88.5 108.2 140.3 174.3 100. yrs LAN 21 105.5 131.6 178.6 224.6 100 year return period analysis 100. yrs ALL 134 77.7 86.1 97.4 106.5 100. yrs NOR 87 81.3 91.9 106.2 117.4 100. yrs ELN 26 23.0 27.7 33.5 37.8 100. yrs LAN 21 105.5 131.6 178.6 224.6 Table One: Raw TSAP report for wind speeds (in knots) at selected location TAOS TSAP Report for Belize Page 5 of 6
References Johnson, M. E., 1997: Caribbean Storm Surge Return Periods: Final Report. Organization of American States, Caribbean Disaster Mitigation Project Workshop, Kingston, Jamaica, October 31, 1997. Johnson, M. E. and C. Watson, Jr., 1999: Hurricane Return Period Estimation, 10th Symposium on Global Change Studies, Dallas, TX, 478-479. Watson, C., Jr., 1995: The Arbiter Of Storms: a high resolution, GIS based storm hazard model, National Weather Digest, 20, 2-9. Watson, C., Jr., 2002a: Using integrated multi-hazard numerical models in coastal storm hazard planning, Solutions for Coastal Disasters 02 Conference Proceedings, American Society of Civil Engineers, Reston, VA. Watson, C., Jr., 2002b: Implications of climate change for modeling coastal hazards, Solutions for Coastal Disaster 02 Conference Proceedings, American Society of Civil Engineers, Reston, VA. TAOS TSAP Report for Belize Page 6 of 6