An internal review of the National Weather Service’s forecasts for the June 29, 2012 derecho finds it was unsuccessful in providing long-lead times for the thunderstorm complex that produced a trail of destruction from Indiana to the Mid-Atlantic coast. The self-assessment praised the short-term issuance of warnings, but stressed they should contain stronger language for conveying the urgency in such violent storm events.
The review is very upfront with the fact the storm was “not forecast well in advance.” As such, the public and key stakeholders (broadcast media and emergency management) had little time to prepare. The derecho downed untold numbers of trees and left several million customers without power along its path.
The review offers two main reasons forecasts fell short.
First, despite the explosive environment for thunderstorm development, a trigger for the storm appeared absent.
“There were no obvious disturbances embedded in the flow to help focus thunderstorm development,” the report says.
As such, the primary forecast models did not simulate the dangerous complex of storms even within hours of its initiation.
“For this event, the [general forecast] models did not produce much in the way of precipitation for the time the derecho was occurring,” the assessment says.
Second, so-called “convection-allowing models” - which specialize in predicting thunderstorm development - were either unavailable at longer lead times or offered mixed signals. One model (known as the HRRR) provided a decent forecast for the derecho, but forecasters had reservations about it due to “past erratic performance,” the report says.
The report recommends NWS beef up its computing capacities to improve model performance.
“Fully supported, ensemble, high-resolution model data extending further out in time are needed to improve the lead time in forecasting these extreme wind storms,” the review says.
Despite knocking its longer-range forecasts, the assessment generally praises the accuracy and timeliness of NWS watches and warnings once the storms were within several hours of striking.
“[T]he offices generally did an excellent job issuing warnings,” the assessment says. “Overall, [warning] lead times were greater than 30 minutes, especially on the eastern end [of the region impacted by storms].”
But the reports notes the unusual intensity of the storms, relative to the typical thunderstorm, could’ve been better communicated in some cases.
“A common theme that emerged from talking to emergency managers, media, and the public was that although they received the warnings, they were surprised by the intensity of the winds,” the report says.
For the part of the NWS office serving Washington, D.C. in Sterling, Va., the assessment highlights its use of compelling language as a best practice in conveying the storm threat. It calls special attention to its special weather statement issued at 9:35 p.m. warning “extremely dangerous thunderstorms” would affect the area. This statement also mentioned the storm’s history of producing damaging winds, and concluded by stating “this is a particularly dangerous situation.”
The report makes the key recommendation that this kind of urgent language be tested and potentially integrated into severe thunderstorm warnings “especially during unusual events.”
A few other observations on the assessment:
* It smartly recommends NWS “should expand the use of social media for reaching the public and receiving obserrvations and damage reports during high-impact events.”
* Due to the heat-related fatalities following the derecho, it adds this important recommendation: “NWS should develop a statement to include in the impacts portion of heat products emphasizing the dangers of staying in homes without air conditioning during prolonged heat waves.”
* I would’ve liked to have seen some more input (i.e. direct quotes) from outside stakeholders in the assessment. There is limited input included from the media and emergency management officials compared to some other NWS assessments I’ve read.
* It’s interesting to see the assessment included three non-Federal participants on the review team, yet the report does not state it complies with the Federal Advisory Committee Act, which provides legal guidelines for such participation. The legal process required to comply with FACA was used by NWS - controversially - as justification to exclude non-Federal participation in the Hurricane Sandy review, now underway.
Update, 4:40 p.m. - on the issue of non-Federal participation (above), a spokesperson for the National Weather Service says: “The derecho service assessment team completed its work before we received new legal guidance on FACA compliance. Therefore, composition of the derecho service assessment team pre-dates this guidance.”
(Ian Livingston contributed to this post)