Geoscience Reference
In-Depth Information
Islands to Hokkaido. Although the Kuril Islands region produced many small basin-wide
tsunamis over the past ive years, all of these stations had failed by December 2008, and four
had failed in October 2008, or earlier. None were repaired until late June 2009, after weather
conditions had improved enough to reduce the risk of shipboard operations. As of May 2010,
three of these ive DART stations have been inoperative since September, 2009.
The optimization scheme used for planning the locations of the DART stations and testing
their ability to detect tsunamis basin-wide is based on an assumption of nearly 100 percent
performance (Spillane et al., 2008). There is a small amount of redundancy and overlap in the
DART network design in case of a single DART failure, but the consequences of multiple DART
failures have not been considered. Given the current geographic coverage, the DART network
is only useful for tsunami detection and forecasting if it is operational nearly 100 percent of
the time. In a practical sense, when one DART station is inoperative, its neighbors on either side
must be operational. If two neighboring DARTs become inoperative, then there must be an im-
mediate mitigating action. A minimum irst step in rectifying this situation is to establish more
explicit priorities for the DART stations in order to provide guidance for NDBC's maintenance
activities. Table 4.1 from Spillane et al. (2008) provides the coarsest priorities set for the initial
DART deployments, but the report does not provide justiications for the prioritizations, and
they are not speciic enough for the purpose of prioritization of maintenance schedules.
Figure 4.7 emphasizes that maintenance of inoperative gauges is slow and generally per-
formed on an annual cycle irrespective of the timing of outages. Even with many DART stations
inoperative in late 2008, NDBC's repair plan was to restore all nonoperational DARTs by the end
of July 2009. As a consequence of the pervasive outages of the DART stations, the TWCs can-
not depend on the DART network for tsunami forecasting. According to NDBC personnel, the
budget only allows for annual routine maintenance and no funds are available for “discrepancy
response” (that is, nonroutine maintenance for inoperative gauges) (National Data Buoy Center,
personal communication, 2009). The committee has assumed that summer time maintenance
cycles are, at least in large part, dictated by north Paciic weather. If this is the case, the main-
tenance of the high-priority DART buoys may not be practical or even possible. NDBC's bud-
get for maintaining the DART stations decreased the past few years, despite the mandate in
P.L. 109-424 for NOAA to “ensure that maintaining operational tsunami detection equipment is
the highest priority.” However, lack of maintenance funding explains only part of the present
problem with DART station failures. The number of DART II system failures is higher than
expected, with a current median time to failure of approximately one year when the design
lifetime was four years (Figure 4.8).
The task of building and deploying the DART buoys in two years, by Presidential directive,
has been challenging for NDBC. To meet the mid-2007 deadline, the DART II was rushed to pro-
duction and deployment without the customary level of testing required for a complex sys-
tem like the DART, with its relatively extreme operational environment. This rapid deployment
schedule required an active reliability improvement program, concurrent with initial operations
and funding, to sustain effective operations while reliability improvements were deined and
implemented. However, budget cuts slowed both maintenance and reliability improvement.
Furthermore, NDBC had no prior experience with sealoor instrumentation, acoustic modem
