Friday, November 2, 2012

The National Weather Service Should Not Investigate Itself

The National Weather Service has put together a "service assessment" team to review its performance in Hurricane Sandy. I recommend they stop right now and not waste our money.

Instead, the NWS should ask the National Research Council or some other independent organization to do an independent assessment of its performance along with that of Mayor Bloomberg and other emergency officials. We need a comprehensive overview of the strengths and weaknesses of the warning system and response.

The NWS service assessment (SA) program is badly broken. The Joplin tornado SA was a complete joke. It never mentions its own mis-location and misreporting of the tornado's location and direction of movement. The SA report issued in the wake of the April 27, 2011, tornadoes in the South omitted several crucial issues.

I'm a huge fan of the NWS (as anyone who has read Warnings: The True Story of How Science Tamed the Weather can attest), but I also believe we have to be scientifically honest in order to continue to improve as an applied science. The NWS is hardly disinterested in its own performance. A report on Sandy, at minimum, should examine:

  • Its decision not to issue a hurricane warning as a hurricane was approaching the coast was, in my opinion, disastrous. That is just my opinion. We need objective facts. 
  • To what extent did this decision influence Mayor Bloomberg and others to delay evacuations and/or make the evacuations less extensive than they should have been?
  • Did this decision contribute to the drownings on Staten Island and elsewhere when the hurricane's storm surge overtook the island and the coastline?
  • What was Sandy's real intensity? At least three independent meteorologists have come to the conclusion it was category 2 before landfall. I believe they may be right. The NWS has a long history of "adjusting" hurricane intensities after the fact (Hurricane Andrew, that many of us believed was a Cat 5 at the time, wasn't named a Cat 5 by the NWS until the tenth anniversary of the storm). This does not inspire confidence we are getting the accurate story now.
  • We have made relatively little progress in forecasting hurricane intensity. Is it time for a different approach to that problem?
So, NWS, stop your plans to investigate yourself. The magnitude of the issues raised by Sandy are too great to the Nation. It is time for an independent investigation.  

4 comments:

  1. Great post, Mike, and highlights a number of issues about the Hurricane Sandy forecast of interest. You make a good point that self-assessments aren't a substitute for an independent outside look. In a case like this, there could be usefulness from both. More at http://www.livingontherealworld.org/?p=762

    ReplyDelete
  2. When will Accuweather invite someone from the NWS or a competitor (Weather Channel, maybe) to investigate the repeated erroneous seasonal predictions produced by Accuweather? Last winter's forecast by Accuweather was nearly 100% wrong. Their forecasts are usually sensational and extreme to garner attention, but are usually wrong.

    ReplyDelete
  3. @ RealWX: Your comment went into the spam filter, thus the delay in it appearing. I approved it, even though it is ridiculous because it is on topic and there is no obscenity or profanity.

    1. AW takes no tax dollars.
    2. If you don't pay your taxes on April 15 (which is what supports the NWS), you go to jail. No one is forced to use our forecasts.
    3. If you don't like ours, feel free to use some other forecast you believe is superior.

    ReplyDelete