Geoscience Reference
In-Depth Information
bitmap
FIGURE 4.9
Comparisons of the February 27, 2010, Chilean tsunami recorded at two U.S. sea level gauges
with forecasts obtained from high-resolution model runs. The forecast models were run in near-real time
before the tsunami reached the locations shown. The model data for Santa Barbara exhibited a 9-minute
early arrival (0.8-1 percent error accumulated during the propagation simulation) that has been removed
for the purposes of this comparison. SOURCE: http://nctr.pmel.noaa.gov/chile20100227/; Center for Tsu-
nami Research, NOAA.
Hurricane Center (NHC) or more generally the weather service, where ensemble modeling is
commonplace.
Conclusion:
Metrics are needed to objectively measure each model performance. In
addition, a process is needed by which multiple model outputs can be used to develop a
single solution (e.g., ensemble model approach in the NWS and NHC).
Recommendation:
The TWCs and the NOAA Center for Tsunami Research at PMEL should
continue to work together to bring the SIFT tsunami forecast methodologies into full
operational use. The utility of the methodologies could be improved by ensuring that TWC
staffs undergo a continuous education and training program as the forecast products are
introduced, upgraded, and enhanced.
INSTRUMENTAL DETECTION OF NEAR-FIELD TSUNAMIS
Near-ield tsunamis present a daunting challenge for emergency managers. Even if the
near-shore populace is well informed about the potential for a tsunami when the ground
shakes, and even if local managers receive information from forecasters of an impending
