Geoscience Reference
In-Depth Information
numerous earthquake scenarios and under various DART failure scenarios, should continue to
help improve the network design (Spillane et al., 2008). The potential contributions of optimiza-
tion algorithms to the design process have not been exhausted.
A component of the periodic re-evaluations of the DART network needs to be the
re-evaluation of the prioritization of each group of DART stations, not just individual stations,
with detailed justiications for these determinations. In particular, the committee questions
the rationale for the very low priority of the group of ive DART stations deployed in the
Northwest Paciic ( Table 4.1) that provide coverage from the Dateline along the western
Aleutian Islands, and past the Kuril Islands to Hokkaido. The Kuril Islands in particular have
been the source of numerous tsunamis large enough to invoke tsunami watches and warn-
ings. At the very least, DART stations covering the Kuril Islands would have a high value for the
prevention of false alarms.
DART station prioritization could be reined by irst distinguishing prioritization criteria
based on the system's primary function in the detection process. A list of criteria might include:
•
detection of a large tsunami,
•
detection of a medium to small tsunami (to mitigate false alarms),
•
providing data for scaling forecast models during the occurrence of a large tsunami,
and
•
providing data for forecast model validation after the fact.
Depending on the order of importance of criteria such as these, quite different priori-
tizations of the DART stations might result. For instance, the value of the DART stations in
the Western Paciic south of 25
o
N (Figure 4.6) is primarily for scaling forecast models, since
the numerous island stations in the region (Figure 4.4) can adequately perform the tsunami
detection function. The value of the DART stations in the Northwest Paciic is primarily for the
detection of medium to small tsunamis, in order to conirm that a large tsunami has not been
generated and thus avoid the issuance of an unnecessary warning with its attendant costly
evacuation. Depending on the relative importance of the criteria in the list above, the North-
west Paciic DART stations may be more important than the Western Paciic DART stations,
contrary to the present prioritization represented in Table 4.1.
Conclusion:
NOAA is to be commended for having developed a prioritization scheme
for the distribution of the DART stations and for having rapidly deployed the DART array.
There are no serious gaps in the geographic coverage of the DART network as designed,
with regard to providing timely and accurate tsunami warnings and forecasts for at-risk
U.S. coasts and territories. However, the vulnerabilities of non-U.S. territories in the TWCs'
AORs were not a high priority in the network design, and the potential contributions of
optimization algorithms to the design process have not been exhausted.
Recommendation:
NOAA should regularly assess the numbers, locations, and
prioritizations of the DART stations, in light of constantly changing iscal realities,
