Recinto Universitario de Mayagüez Programa de Alerta y Mitigación Contra Maremotos de Puerto Rico FEMA (Agencia Federal para el Manejo de Emergencias)
 
THE PUERTO RICO TSUNAMI WARNING AND MITIGATION PROGRAM
(October 2000 to March 2003)
[ Prev ]   Page - 1 - 2 - 3 - 4 - 5 - 6 - 7 - 8 - 9 - 10 - 11 - 12 - 13 - 14 - 15   [ Next ]
It is important to let map users know the limitations involved in the preparation of these maps:
 
  • The topographic data (from the USGS DEM model for Puerto Rico) is old (in some cases by as much as 40 years). In some locations (a good example is the Punta Miquillo/Punta Picua area in Rio Grande) the changes to the topography have been enormous. The same holds in the Isabela area, where coastal dunes, 10 or more meters high, appear in the USGS DEM data, but those dunes are no longer there, having being mined decades ago. This fact, together with the computational grid cell size of approximately 90 meters, has to be considered in assessing the reliability of the maps. Hopefully, in the near future we may have available high-resolution LIDAR data taken in December of 2000 along a 1 km wide coastal strip. The use of this LIDAR data hinges on the processing of it to obtain “bare earth elevations”, and in the determination of an accurate geoid surface model for Puerto Rico. 
  • As mentioned above, the SHOALS data contains, at some locations, elevations above MSL that correspond to the vegetation cover at the site. The same for nearshore infrastructure like for example, the city of Aguadilla. These erroneous elevations will present a solid barrier to the inland penetration of the sea. Although if the infrastructure is well built it will really act as a barrier to the inland flood. The same holds for any sand dunes present in the data, whether the dune is still there or not. It will act as a solid barrier since no erosion, or scouring, is contemplated in the model.
  • The high-resolution computational grid used in the tsunami simulations, the grid where the tsunami runup phase was computed – Grid C, had a constant Manning friction coefficient, n, all over it. The value used, 0.025, corresponds to water flow over slime-covered cobbles, gravel, and grass (see http://wwwrcamnl.wr.usgs.gov/sws/fieldmethods/Indirects/nvalues/ for a good visual description of Manning’s n values). This is the default value of the model and it is the same value used in the numerical simulation of the 1918 Puerto Rico tsunami (Mercado and McCann, 1998), where the comparison of model results and runup observations (see report by Reid and Taber, 1919, on the Internet site) was acceptable. It is obvious that the value of ‘n’ will vary significantly all over the island, and especially in developed areas. The value adopted in this study will be valid only up to a certain limited distance from the MSL shoreline. For this reason, it is felt that the extensive inland flooding observed in some coastal regions of the island represents quite conservative results. For this reason, and in order to somewhat compensate for this, the inland limit of the flood was taken at the 1 meter elevation (above Mean Sea Level) contour of the inland flood. That is, in low lying areas, where the inland flood peaks at some distance inland and then decreases, the digitization of the inland limit of flooding was done along the 1 m flood elevation contour. In future studies, when more up-to-date, higher resolution, and more accurate topography becomes available, one could run the model with variable ‘n’ values based on land-use information available in GIS format. 
  • As explained above, the flood results shown do not discriminate between earthquake scenarios of different return periods. As will be described in the final report about the seismology of the region upon which the tsunami scenarios were based, different earthquake-prone regions around the island have different expected earthquake return periods. The practical approach adopted here, which is typical in the field of hazard mitigation, is that since the uncertainty in return periods is so large then we better be prepared for the worst case scenario. We can leave for another study refinements of this adopted approach. Since the original flood data for each earthquake has been stored, if necessary we could prepare flood maps based on the region of the earthquake epicenter. 
  • In the calculation of the initial sea surface deformation in the earthquake source region, computed using the Masinha and Smylie model, we had the option of using either a mean (over the fault length) slip magnitude or the maximum magnitude (occurring somewhere along the fault length), both output by the methodology used by Dr. Huerfano. In this study we adopted the maximum value as applicable all along the fault length. This is the first time this type of maps are being prepared for Puerto Rico and we wanted to err on the conservative side, leaving for a later re-evaluation the use of smaller values of the slip. As expected, this will also add to the conservatism of the results. [Note: There are nowadays more sophisticated, and possibly reliable, models to estimate the initial sea surface condition than the Masinha and Smylie model. But they require seismic data not available to us.]
  • For the same reason, the fault depth below the bottom of the ocean was assumed to be as small as the algorithm used to compute the initial sea surface deformation allowed. The algorithm tends to become unstable if fault depths below the ocean bottom are assumed less than approximately 200 meters. But this threshold varies, and so we did computations with different values and adopted the smaller one allowed.
  • Finally, it should be stated that there is the possibility that a tsunami may propagate upriver and, in some cases, overflow its banks. This is something that has been observed in the past at many locations. This possibility has not been addressed in this study. People living near river banks in the proximity to the coastline should be careful about the use of these maps since they will not show this possibility.
  • There are other assumptions involved in the methodology used to obtain the fault parameters, but these are made in the study by Dr. Huerfano.
These maps have been distributed to local emergency response agencies in both paper and CD versions. All of this information is also available in GIS Shape files format. The results show that western and northern Puerto Rico will be the most affected zones, with the hazard being smaller in the eastern and southern zones. 


[ Prev ]   Page - 1 - 2 - 3 - 4 - 5 - 6 - 7 - 8 - 9 - 10 - 11 - 12 - 13 - 14 - 15   [ Next ]