Friday, October 18, 2013

Bit More On Outlooks From NWS CPC


I looked a bit more at CPC's web page this morning to find information about how these outlooks are made. There is a discussion of the forecast tools used (7 of them, including a 40-member ensemble forecast model) at:       http://www.cpc.ncep.noaa.gov/products/predictions/90day/tools.html

However, I didn't find anything about how the outputs from these tools are distilled into the final outlook products. Perhaps I just didn't look in the right spots? The following is from the discussion of the skill of these forecasts at the link above (which is similar to what I wrote yesterday):

FOR ALL MODELS PRECIPITATION FORECASTS ARE GENERALLY LESS SKILLFUL THAN 
TEMPERATURE -- WITH MARGINAL SKILL FOR ALL TOOLS EVEN IN THEIR BEST SEASONS AND 
LOCATIONS UNDER NORMAL CIRCUMSTANCES.  HOWEVER WHEN STRONG EL NINO OR LA NINA 
CONDITIONS ARE PRESENT - PRECIPITATION SKILL CAN BE AS HIGH AS TEMPERATURE SKILL 
FOR COOL SEASON FORECASTS FOR A NUMBER OF AREAS OF THE U.S. - INCLUDING THE 
SOUTHERN THIRD - THE NORTHERN ROCKIES - THE HIGH PLAINS AND THE OHIO VALLEY.  
STRONG LA NINA CONDITIONS IMPLY THE POSSIBILITY OF MODERATE PRECIPITATION SKILL 
FOR SOME PARTS OF THE WARM SEASON AS WELL. 

I've shown two more outlooks here that illustrate another head scratcher. At the top is the current 8-14 day temperature outlook valid through October 31st. Below is the current outlook for November temperature. To me these indicate almost a binary transition to a different pattern at the end of October, since I don't see any evidence of a transition between the two maps.


4 comments:

  1. 'Twas I that asked the question. Other than being interested in Meteorology, I have a number of citrus trees. I need to have an accurate as possible forecast to know whether we are in for cold weather and whether I need to take cold precautions (which take a while per tree to install) for my trees. You can see a few of them in the East Webcam view on my weather site noted below. Any suggestions as to who to use?

    Your penpal,

    Peter
    http://tucsonweather.homeserver.com

    ReplyDelete
  2. Peter,

    It's not clear what kind of forecast lead time you need to take action - is it a day, a week, longer?

    Bob

    ReplyDelete
  3. Hi Bob: Sorry, I guess I didn't specify a time frame. I would like a relatively longer term (perhaps 30days) forecast, if that is even possible. This would be for planning purposes. For periods of 7 days or less, I also use the NWS short range forecast at http://www.wrh.noaa.gov/forecast/wxtables/index.php?lat=32.18743&lon=-110.756495&table=custom&duration=7&interval=1. I have found this forecast to be exceptionally accurate (temps to within a degree or two, esp. in winter) and gives we detail as to when I should begin to prepare for freezing conditions. IMHO, it is as accurate as the long range one seems to be inaccurate. Perhaps it is only a coincidence.... Also, I have prepared a spreadsheet and graph showing the October highs and lows back to 2010 (when my weather station went online). I don't know if any of my data would be useful but you are welcome to use it. I don't know how to attach a file here or I would have done it as an example.

    Regards,

    Peter

    ReplyDelete
    Replies
    1. Hi Peter - the forecast link you sent is to the grid point forecasts produced here at the NWS Forecast Office - they are also available under "Tables" on the local web site. These are about as good as you'll find with such fine spatial resolution. They can go pretty badly awry, especially when actual weather events are occurring. I also think they tend to be a bit too cold at high elevations and too warm at low elevations - for example they are running 8 to 10 degrees too warm for the grid point that I'm located in.

      Some European weather services have fairly specific longer term forecasts available - driven by the ECMWF - but mostly for US cities. See
      http://www.yr.no/place/United_States/Colorado/University_of_Colorado_at_Boulder/long.html
      for an interesting example from Norway.

      Bob

      Delete