The NAS TWS Report

Summary of Tsunami Waning Center Operational Changes Recommended by the Report
Tsunami Warning and Preparation:  An Assessment of the U.S. Tsunami Program and the Nation’s Preparedness Efforts

This report by the National Research Council of the National Academy of Sciences (NAS) assesses improvements made to the nation’s tsunami warning system (TWS) as a result of two laws enacted by the U.S. Congress in 2005. Thus it evaluates the current state of the system, which includes the National Weather Service's (NWS’s) two Tsunami Warning Centers (TWCs), but it also seeks to identify ways in which to further improve the whole system, including other NOAA and NWS offices. This report would be valuable to any NWS administrator who was motivated to learn how the TWCs actually function, rather than try to understand our operations through the lens of administering more typical NWS forecast offices (see the recent service assessment and the workforce study for examples). The report’s authors took great pains to describe both the science of tsunami warning as well as the current reality of putting the science into practice, warts-and-all. They criticize NWS and NOAA management for current shortcomings, but praise TWC staff for their efforts to meet mission objectives despite this lack of external support and poor management (p. 93, p. 177). 

The report’s observations and recommendations are not new. TWC staff members have raised all of them within the organization before. What is new is the collection of all of these concerns into a single document, their public airing, and the possibility that something may actually be done about them. The report contains errors, especially errors of omission, but corrections to these errors would not significantly change the report’s conclusions and recommendations, which cover the range from heavily technical subjects (such as the design of DART buoys) to the day-to-day operations of the TWCs, which, if implemented, could result in significant changes to TWC operations. My summary, presented below in outline form, tries to identify all of the TWC operational changes that would result if the NWS or NOAA implemented all of the report’s recommendations. My personal observations are in italics, and I have highlighted those topics that I believe are especially expensive, complicated, or otherwise contentious. Of course, these highlighted topics would also represent the most significant changes to TWC operations.

“NOAA/NWS should undertake a comprehensive, enterprise-wide, long-range planning effort for the TWCs.”

This “effort” (p. 154) should consider:
  • Better integration of TWC warning functions with emergency managers by using old and new messaging technologies, including social media (p. 94, p. 169)

  • Long-range technology planning, especially information technology (IT) (pp. 170-171)
    • greater IT resources, including continued, substantial funding
    • compliance with professional IT standards

  • The location(s) of the TWCs (pp. 176-177)
    • near or co-located with academic and/or government research or operations centers (i.e., universities, NEIC, or NCEP centers)
    • combining the two centers into a single facility (a view shared by the WSSPC's report)
      • removes operational differences between the two TWCs
      • relieves 24/7 shift-worker staffing shortages by combining existing staff
    • proximity to TWC customers
    • protection from natural hazards

  • The culture of the TWCs
    • Should the TWCs be administered by NWS regions or NCEP? (p. 160, pp. 174-175) In other words, administered more like Weather Forecast Offices (WFOs, in NWS regions) or more like the National Hurricane Center (NHC, under NCEP)? The report does not recommend NCEP outright, but repeatedly holds up NCEP centers as examples of “highly reliable organizations” while it criticizes NWS regional administration of the TWCs.

    • NOAA/NWS should provide for recruiting, training, development, mentoring, and professional exchanges of TWC scientists (e.g., scientific meetings, workshops, sabbaticals, and grant-funded research) (pp. 172-173). Currently official support of such activities is uncommon--I'm one of the lucky few. More often our scientists will attend meetings like AGU on their own time and their own dime.

    • NOAA/NWS should incentivize high employee performance to promote adherence to standards and continuous improvements to the tsunami warning system (p. 171, p.178). 

The tsunami program needs additional reviews

  • NWS should establish “committee of experts in the social science of warning messaging” to review message format, style, content, and delivery method of messages (p. 92).

  • NOAA/NWS should use external science review boards (i.e., not a NWS service assessment) for post-event analysis (p. 94).

  • NOAA should coordinate post-event field analysis of areas impacted by tsunamis with other agecies (p. 100). The report does not say TWC staff should be included in these surveys, but TWC staff have participated in such surveys in the past and I believe that they should be included in the future.
  • TWC scientists should help select DART buoy locations due to their experience with the system (p. 136).

  • A committee including non-NOAA/NWS members (i.e., from the USGS) should conduct ongoing IT reviews  (p. 170).

Other TWC recommendations

  • Add staff to the TWCs 
    • “public relations officer” (p. 93) The TWCs do not currently have dedicated outreach personnel.
    • IT professionals (p. 170), since the TWC scientists developing software are not software engineers (p. 166, p. 168). Currently only one of the TWCs has a full-time IT specialist.

  • Standardize procedures between the centers (p. 167), including transmission of a single message from both centers (p. 92). This recommendation also implies consistent styles of 24/7 shift coverage, so one of the TWCs would have to change theirs...unless they're merged, of course.

  • The TWCs should use an ensemble of multiple forecast models similar to the NHC (p. 141). Since this report does not describe in any detail the models developed within the TWCs, I am concerned that someone may try to leverage this report to prevent the TWCs from developing their forecast models any further and require them to rely solely on external developers such as the Pacific Marine Environmental Laboratory (PMEL) or the Environmental Modelling Center (EMC). At least one TWC scientist would resign if he could not continue such work.

Contradictory TWC-related observations

  • The USGS is planning to create a 24/7 operational backup to the National Earthquake Information Center (NEIC) (p. 109). If the NEIC wants its own 24/7 “hot spare” that suggests that the TWCs should remain as two centers, as opposed to being merged as sugested above. The TWCs have also previously served that function on an ad-hoc basis. This relationship could be formalized with a MOU instead of creating a NEIC backup within the USGS.

  • The report contains a brief suggestion that the TWCs should not do seismic data analysis at all (p. 109), but concludes that the TWCs should continue to work in parallel with NEIC (p. 110).

  • The claim in the NAS's press release that the TWCs cannot back each other up is contradicted by the report's statement on p. 160: “Despite these differences, however, the TWCs are able to perform federally-mandated back-up activities for each other and are able to issue each other’s messages.”

Omissions and Errors

  • The report states that the TWCs receive enough data from the existing global seismic network to detect all potentially tsunamigenic earthquakes (p. 103, p. 110). This conclusion ignores the configuration of the global network, which was installed without regard to tsunami hazard mitigation, and is therefore suboptimal for rapid detection and characterization of earthquakes in some parts of the world. In other parts of the world this network could easily be rendered suboptimal with the loss of a single station (due to communication trouble, vandalism, etc.), and thus the existing network also lacks redundancy for the purposes of tsunami warning. Please see my AGU presentation for a preliminary quantitative analysis.

  • The report frankly admits that the TWCs have “difficulties in 24/7 staffing” yet suggests only one remedy, to combine the staff of the two centers into a single center (p. 176). The report uses the example of WC/ATWC (p. 172) which has 9 scientists available for shift work. Two scientists are required to be on duty at all times, so in one week scientists need to cover 336 hours of shift work (2 x 24 x 7). Assuming a 40-hour work week, then the minimum staffing would be 8.4 scientists (336 ÷ 40), i.e., 9 scientists, which is the current staffing profile. Such minimal staffing, however, does not allow for any scientist to exit the duty rotation for official travel, training, illness, family emergencies, or vacation, despite the report's claim to the contrary (p. 172). As the report also recommends that TWC scientists interact with the scientific community at large by attending scientific conferences, workshops, etc. (pp. 173-173), the report should have included an alternative to merging the centers to the effect that more scientists will need to be added to the TWCs if the centers are not merged. For comparison, the NWS operates over a hundred WFOs that also have two scientists on shift at all times, yet each WFO typically has 13 staff members qualified for this type of work, or about 150% of minimum staffing, and they don't even have the research requirement of the TWCs (see NWS Directive 10-7, Section 3.3b). This lack of such an obvious “Plan B” suggests that the report's authors did not seriously consider alternatives to merging the centers, an expensive and complex proposition that NOAA/NWS may not implement.

  • The report notes that the Mwp method of magnitude determination is insufficient for earthquakes with magnitudes greater than 8.0 (p. 109). However, the report neglects to mention that because the magnitude threshold the TWCs use to issue a tsunami warning is 7.6 this method is sufficient for operational use. Furthermore, though the report rightly encourages adoption of “W-phase” magnitudes, it neglects to mention that this magnitude method takes too long (more than half an hour) to be useful for the rapid magnitude determinations required for initial tsunami warning.

  • The report claims that the TWCs cannot detect landslide-generated tsunamis (p. 156).  Though the report describes the coastal sea-level gauge network (pp. 113-122) and the DART sensor network (pp. 112-137), it ignores the land-based “runup detectors” employed by PTWC that indicate when sea level has exceeded its maximum tidal range, and that these instruments could detect a landslide-generated tsunami in Hawaii.

  • The report argues that operating two TWCs without identical software do not represent redundant systems (p. 174). However, just as identical systems would work the same, they would fail the same, so heterogeneous systems should be less likely to succumb to the same mode of failure.

  • The report claims, without evidence or attribution, that its authors “believe” that during the 2010 tsunami from Chile that “PTWC chose to forecast wave heights using a model developed in-house” (p. 251). This “belief” is false. PTWC tested its new forecast model, but did not use it operationally.

So what does this all mean?

At the AGU 2010 Fall Meeting I spoke with one of the authors of the report and I complained to him that its text repeatedly leads up to a Big Recommendation then wimps out, such as the suggestion for future studies and what they should consider, rather than giving outright recommendations. This individual told me that the authors were in fact prohibited from making such Big Recommendations, as they were not allowed to set government policy.  If that's the case, then I believe that the report's implied vision for TWC operations is this: that the two TWCs should be combined into a single NCEP-operated National Center very much like the NHC with additional staffing, significantly upgraded IT support, multiple forecast models, and support for the staff's professional growth. Furthermore, I believe the authors envision the combined TWC to be co-located with other operational and research institutions. The obvious candidates to host a combined TWC are 1) Boulder, CO, home to several NOAA operations including the National Geophysical Data Center (including historic tsunami data), Eath Systems Research Laboratory (including tsunami research), and NCEP's Space Weather Prediction Center, as well as the nearby NEIC in Golden, CO;  2) Seattle, WA, home to NOAA's Center for Tsunami Research, part of PMEL and located with a WFO on a NOAA campus near the University of Washington; 3) Honolulu, HI, home to the University of Hawaii that is in turn home to the Central Pacific Hurricane Center; also, Pearl Harbor will be home to a new NOAA campus that is expected to house PTWC in 2013 and will have room for additional staff.

This summary refers to the prepublication version of this report released on September 17, 2010.