Monday, December 3, 2012

Why "Threat Communication" is a Vital Part of Storm Warnings

For most of the last month, I've been writing about Hurricane Sandy, the National Weather Service's Sandy Assessment I co-chaired, and the aftermath of the cancellation of that assessment.

If you look through these many postings, you'll find a number of comments made, especially from meteorologists, about how good the forecasts were and how meteorologists (paraphrasing) "can't make people take precautions." You'll find similar comments left today with an article on this topic published in The Washington Post.
click to enlarge
Having worked with meteorologists every day of my professional life since 1971, we tend to be justifiably proud of what we do with good reason. That is why I wrote a book on that exact subject.

Over the years I have learned (based on actual data from industrial psychologists) most meteorologists have intrinsic motivations. That is usually a good thing: It explains why meteorologists work ridiculous hours in critical weather situations. You wouldn't want a "clock watcher" on duty when an F-5 tornado is bearing down on your town.

But, when intrinsic motivation is taken too far, a blind spot can exist: An excessive focus on meteorological minutiae (i.e., the temperature structure of Hurricane Sandy two miles above the ground) at the expense of clear threat communication with people outside the profession.

The New York Times had an extraordinary story today about why people in nursing homes were left in harm's way. Bryan Norcross published a wonderful piece about emergency communications on his blog. I'm going to use pieces of those to explain why threat communication is essential if we are to save lives and dollars. I'm going to juxtapose them for clarity. I urge you to read each piece in its entirety, both are well worth your time.

NYT:

Hurricane Sandy was swirling northward, four days before landfall, and at the Sea Crest Health Care Center, a nursing home overlooking the Coney Island Boardwalk in Brooklyn, workers were gathering medicines and other supplies as they prepared to evacuate.
Then the call came from health officials: Mayor Michael R. Bloomberg, acting on the advice of his aides and those of Gov. Andrew M. Cuomo, recommended that nursing homes and adult homes stay put. The 305 residents would ride out the storm.
The same advisory also took administrators by surprise at the Ocean Promenade nursing home, which faces the Atlantic Ocean in Queens. They canceled plans to move 105. 

Norcross: 

As I and others have discussed, the combined alerts to the public from the local National Weather Service office in Upton, NY and the National Hurricane Center in Miami were not helpful on that Friday, when critical information should have been dispensed to the public. Whether that contributed to the information void from local leaders, we can’t say. But it certainly limited the focus and attention of the public.

At 4:09 AM Friday morning, the Upton office issued their Hazardous Weather Outlook bulletin, which they do every day. In part it said,

THERE IS INCREASING CONFIDENCE THAT THE TRI-STATE AREA WILL FEEL THE IMPACTS OF A DANGEROUS COASTAL STORM LATE THIS WEEKEND INTO EARLY NEXT WEEK. THIS INCLUDES THE POTENTIAL FOR HEAVY RAINFALL AND RESULTANT SIGNIFICANT URBAN...SMALL STREAM...AND RIVER FLOODING...HIGH WINDS CAUSING WIDESPREAD DOWNING OF TREES AND POWER LINES...AND SIGNIFICANT SHORELINE IMPACTS FROM COASTAL FLOODING AND BEACH EROSION.

That sounds pretty tepid given the super, mega, historic storm that was forecast to hit, but at least it was something. Though it was hidden away in an arcane bulletin that few people in the public even know about. 

The Hazardous Weather Outlook continues:

PLEASE REFER TO THE NATIONAL HURRICANE CENTER FOR THE LATEST FORECASTS ON SANDY...AND MONITOR THE LATEST NATIONAL WEATHER SERVICE FORECASTS THROUGHOUT THE WEEK.

Well, first of all, this was issued Friday morning. What does “throughout the week” mean? Does it mean next week? No! It means right now. Second, they send you to the NHC advisory, which contains nothing about the threat of a storm surge in the northeast, but does say,

FOR STORM INFORMATION SPECIFIC TO YOUR AREA IN THE UNITED STATES...INCLUDING POSSIBLE INLAND WATCHES AND WARNINGS...PLEASE MONITOR PRODUCTS ISSUED BY YOUR LOCAL NATIONAL WEATHER SERVICE FORECAST OFFICE.

They send you back to the local NWS office. It’s a Chinese fire-drill of non-information… with a monster storm heading toward the coast! 


NYT:

While the National Hurricane Center in Miami had warned of “historic urban flooding” in New York City, local National Weather Service officials issued contradictory public advisories on Friday and early Saturday that said there would be only “moderate flooding.”
Inside the city’s emergency management center, the local weather officials reported that the storm surge would be similar to the one during Tropical Storm Irene — four to eight feet.
An eight-foot surge was an important marker: after Tropical Storm Irene, Dr. Shah and Dr. Farley had said in interviews that they believed many nursing homes and adult homes could not withstand a surge of that level.

Norcross:  

Meanwhile, more boilerplate gobbledygook was distributed to the public. Here’s the forecast for New York City for Monday from the NWS Upton office that was issued that Friday morning:

.MONDAY...RAIN LIKELY. RAIN MAY BE HEAVY AT TIMES. WINDY WITH HIGHS IN THE UPPER 50S. CHANCE OF RAIN 70 PERCENT. 
.MONDAY NIGHT...RAIN LIKELY. RAIN MAY BE HEAVY AT TIMES. WINDY WITH LOWS IN THE UPPER 40S. CHANCE OF RAIN 70 PERCENT.

Does that sound like the storm of the century is coming? It sounds like there’s a 70 percent chance of a nasty day, not that the ocean was going to rise up and swallow the coastline. I could clog your computer with more examples of how the communications system failed, but you get the point. 

The bottom line is, the National Weather Service needs an entirely new way of thinking about mega events, so they aren’t trapped by a boilerplate, one-size-fits-all set of rules.


For the record, here is what this blog was saying that Friday morning. I invite you to compare it to the above. In particular, two paragraphs (verbatim):

The storm surge will be destructive.  Because wind speed and direction are so important, I can't offer a map of predicted storm surges yet. But, there will be areas where the combination of full moon, waves of more than 20 feet, and storm surge cause great destruction in coastal areas. 

Power will fail in a large geographic area. There is a fair amount of inconsistent information this morning pertaining to peak gusts, i.e., will they be 60 mph or 80 mph? But, since power often fails with a 60 mph gust, it may not matter.

It is not my intention to boast: I have made a career out of forecasting extreme weather and specializing in the communication of these threats. My intention is to make two vital points:

The National Weather Service cannot be allowed to do the sole assessment of its performance in Sandy. As blogger Nate Johnson puts it, the NWS investigating its own forecasts is like "a parent investigating its child."

Any investigation needs to include social scientists familiar with weather communications (we had two on the aborted Sandy Assessment team) and messaging in critical situations.

In spite of Hurricanes Isaac and Sandy this year, the United States is in a record "hurricane drought." In the near future, it is inevitable hurricanes will return to "normal" (higher) levels. As I see it, we have two choices:

1.  We can learn the lessons of Sandy and improve the way we warn of hurricanes starting in 2013, saving lives and billions, or,

2.  We can continue on our present, and often confusing, path.

If you agree that #1 is the better course of action, I urge you to let Congress and the Obama Administration know your preference. 

1 comment:

  1. I'm posting this comment on behalf of reader Jim Caruso of Philadelphia:


    "On the core issue of the NHC not issuing hurricane or TS warnings for the Northeast during Hurricane Sandy, I call your attention to the following from Section 4.1.2 of the Hurricane Irene (2011) Service Assessment:


    '…when Irene transitioned to an extra-tropical storm over WFO Caribou’s County Warning Area (CWA), NHC ceased its support. Because of this action, WFO Caribou could no longer issue a Hurricane Local Statement (HLS), thus the WFO had to cancel the tropical storm warning and issue an inland high wind warning instead. While significant wind impacts were still associated with the extra-tropical period of Irene, the inland high wind warnings did not convey the same level of risk as a tropical storm warning.

    Finding 3: The change in terminology from tropical storm warning to inland high wind warning is inconsistent in terms of expected impacts and can be misinterpreted as a lower-level risk from wind impacts.

    Recommendation 3 (Strategic): The NWS product suite for wind risk from hurricanes and tropical storms should consistently convey potential impacts when a storm transitions from tropical to extra-tropical. Policies should be revised to ensure that any transition is transparent to partners.'

    Looks like a lesson that was not learned, doesn't it?

    I could give them the benefit of the doubt and assume that the Irene situation was different because it involved CANCELLATION of a TS warning, and that may be why they avoided issuing a hurricane or TS warning for Sandy in the northeast to begin with, i.e. to avoid the confusion of a later cancellation. But it could also be a failure to follow up on a recommendation..."

    Thanks, Jim.

    I believe they didn't follow their own recommendation from the Irene Assessment. By Saturday morning, it was obvious a hurricane was threatening the coast. Hurricane watch/warning should have been issued as stated on this blog at the time.

    See: http://meteorologicalmusings.blogspot.com/2012/10/unfortunate-decision-by-national.html et. al.



    ReplyDelete