Geoscience Reference
In-Depth Information
engineering, operational, logistical, and political constraints. The process that began at this
workshop was augmented by an optimization analysis, which was subsequently completed
at the NOAA Center for Tsunami Research (NCTR) at PMEL. To the extent that the constraints
on siting can be quantiied and the beneits expressed in functional form, array design can
be approached as a problem in optimization. This avenue was explored using a tool called
NOMAD (Nonlinear Optimization for Mixed vAriables and Derivatives; Audet and Dennis, 2006).
Although the scheme was tested for only relatively simple cases, the methodology shows
promise as an example of a scientiically robust process for siting and prioritizing stations in an
operational sensor network.
The methodology and inal rationale for the siting of DART stations are the subjects of a
NOAA technical memorandum (Spillane et al., 2008). The inal siting decisions were based on
the workshop recommendations, as well as site recommendation reports produced at NCTR
in consultation with the TWCs, with input from the USGS, NDBC, and other interested parties.
The technical memorandum provides a starting point for continued reinement of the siting
decisions and extension of the DART array, if necessary, while also providing information to aid
efforts by the international community to extend the network coverage.
The net result of the deliberations on the siting of the DART stations is the current array
displayed in Figure 4.6. The prioritization of groups of these sites is presented in Table 4.1
(Spillane et al., 2008). Some of the more important issues involved in site selection are
described in Box 4.1.
The committee does not ind any serious gaps in the geographic coverage of the DART
network as designed, with regard to providing timely and accurate warnings and forecasts of
far-ield tsunamis on U.S. coasts.
It can certainly be argued that denser coverage of open-ocean
sensors would provide important redundancy capacity (in light of current reliability problems
discussed below) and would provide more opportunities to improve the accuracy of model-
generated wave forecasts. From a more global perspective, gaps in coastal sea level station
coverage (as revealed in the Caribbean region, for instance; see previous section), which expose
TABLE 4.1
Sub-Region Allocations and Priorities Within the Overall U.S. DART Array
Array Sub-Group
Instruments Assigned
Pre-Existing Sites
Priority
Alaska/Aleutians
6
3
1
Western Paciic
6
0
2
Puerto Rico/Caribbean
3
0
3
West Coast
5
2
4
Southwest Paciic
4
0
5
Central/South America
4
0
6
Atlantic
3
0
7
Gulf of Mexico
1
0
8
Northwest Paciic
5
0
9
Hawaii/Mid-Paciic
2
2
N/A
SOURCE: Spillane et al., 2008; NOAA.
