Thursday, December 9, 2010

Snow forecast for Saturday, December 11

As the first flakes began to fall, here is the final pre-storm forecast as of 11:00 p.m. on Friday, December 10. Summary: While there is clear consensus that a big snowstorm is on the way, there is considerable variance among Minneapolis weathercasters as to how much snow will fall. The Star Tribune/Paul Douglas predicts the most with a 15 to 20 inch forecast while KARE11 (11) is on the low range with 7 to 14 inches. Other weather outlets are in the general 12 to 16 range. The final forecasts appear below. Upon the conclusion of this snow event, we will grade the forecasters. Enjoy the snow!

WCCO (4): 12 to 16 inches

KSTP (5): 12 to 18 inches

KMSP Fox9 (9): 10-14 inches

KARE11 (11): 7 to 14 inches

Star Tribune/Paul Douglas: 15 to 20 inches.

National Weather Service: 13 to 16 inches

Forecast as of 10 a.m. on Friday, December 10

WCCO (4): 8 to 12 inches

KSTP (5): 7 to 11 inches

KMSP Fox9 (9): 8 to 12 inches with headline "Probable Blizzard in Twin Cities"

KARE11 (11): 5 to 10 inches

Star Tribune/Paul Douglas: 12 to 16 inches. Tweets: Biggest since Halloween '91? Models hinting at 1-2 feet , 2-5 ft. drifts poss.

Forecast as of 10 p.m. on Thursday, December 9

Channel 4: 5 to 11 inches

Channel 5: 9 to 12 inches

Channel 9: 6 to 10 inches. Highest amounts in southern suburbs.

Channel 11: 6 to 10 inches

Paul Douglas/Star Tribune: 10 to 16 inches. Highest amounts for nw suburbs.

Forecast as of 11 p.m. on Wednesday, December 8


Channel 4: An inch or two

Channel 5: A couple of inches possible on Saturday.

Channel 9: old forecast on website

Channel 11: Light snow. No mention of accumulation

Paul Douglas/Star Tribune: 3 to 6 inches with a few 8" amounts close to home

22 comments:

  1. Thanks for starting this! It'll be great to see how close to the mark each meteorologist is to reality. :)

    ReplyDelete
  2. Thank you. Appreciate the feedback.

    ReplyDelete
  3. You should also include the NWS.

    As far as why they are different - there are many different weather models to use, and each model is typically run four times per day. You can see how you can quickly get many different "answers" from these models. Additionally, weather models aren't perfect and the snow/rain amounts they output are usually the most incorrect field they predict. Which is why we have professionally trained people to not only interpret them but to analyze past and current real data (surface observations, radar/satellite data, weather balloons, etc.).

    ReplyDelete
  4. I like the fact that Paul Douglas lays out all the facts and shows that true weather forecasting, opposed to the news weather 'hype', is often unpredictable and can change drastically over a short period of time, especially with snow. Just 2 tenth's of an inch of total precipitation can skew snow totals by inches. That, said, he was more right than wrong on the storm last week...actually underforecasted the total. Nice idea for the blog though. I also love the idea of percentage forecasting.

    ReplyDelete
  5. Thanks for your interest. Keep checking back!

    ReplyDelete
  6. I always thought the forecasters used the same models. If so, how does Douglas have 12-16 and KARE have 5-10? Do they add some subjective forecasting onto the models?

    ReplyDelete
  7. Paul Douglas has the luxury of writing as long of a blog post as he wants to explain how he arrived at his totals. An on-air meteorologist has three minutes and an audience with an average understanding of 3-5th grade (demographic data clearly show the average online user has a higher education level than the average TV viewer).

    As for KARE11 including a 7" total, it is completely plausible. With storm systems like this, you get mesoscale banding - bands of snowfall with higher/more intense snowfall rates. Snowfall rates under these bands can reach 2-3"/hr, but the bands can be as small as 50 miles wide. A slight shift in these bands can easily mean the difference between 8" and 14" of total snowfall amounts. That is a lot of uncertainty to deal with (forecasting a feature like that, the exact location, intensity, and duration, with accuracy is beyond the current level of weather models).

    ReplyDelete
  8. Oh, this should be fun! My co-workers and I always scan each website and each forecaster to see the various predictions and try to guess who'll be right. Glad to see all the data in one location!

    ReplyDelete
  9. Actually, there are several models that are used (GFS, NAM, ECMWF, RUC, etc) and there is a lot of subjective forecasting that does go on with the forecast models. The best models can still be influenced by bad data or a even a bad simulation.

    ReplyDelete
  10. Odd that you would not include the NWS. I do not know how all the information is accessed, but I do not doubt that the NWS provides the bulk of the data.

    ReplyDelete
  11. There are others that have the luxury of trying to discredit Paul Douglas, the so called weather terrorist. The bias is so obvious, because Douglas blogs in support of AGW. In all of the polls, how about recognizing those who consistently oppose certain forecasters solely because of one issue. This would weed out those who want to skew the results. Of course, who knows what biases the owner of this blog has.

    ReplyDelete
  12. The goal of this blog is to provide facts regarding short-term weather forecasts. Discussions about global warming will have to find another outlet.

    ReplyDelete
  13. You can access the information from weather.gov. For the Twin Cities office it is weather.gov/minneapolis. There is more than a ton of data on the NWS website. Easiest way to access the forecast is to go to weather.gov, click on the US map of your area. Then at the local office, click on the map again of where you want a forecast for and you'll get a seven day forecast.

    ReplyDelete
  14. More than that even. GFS, NAM, SREF, RUC, ECMWF (EU run), GEM (Canadian run), and a lot of local NWS offices are now running higher resolution models based on one of those larger models (typically the GFS or NAM). AND each of these models are typically generated four times per day. So you could just copy "the model", but which one of the 30+ models/model iterations is going to be the "right" one? Forecasting is far more complicated than people realize.

    ReplyDelete
  15. It should be noted that the Twin Cities area is very large, about 40 miles in diameter which is over 1,200 mi^2. In the public's eye, how can you provide a very narrow range and yet be "right"? This blog doesn't address that. What are you considering to be the "right" answer? The spread of observed snowfall totals across the Twin Cities? What is measured at the NWS office in Chanhassen? Measured at MSP airport? Downtown MSP? My house?

    ReplyDelete
  16. P is correct. Someone in the metro will only get 8 inches. Someone will get 20. Whats the correct answer?

    ReplyDelete
  17. By no means was I attempting to introduce global warming at this site. In fact, I was describing biases against certain forecasters because of that issue. I would love this site to just stick with the forecasting facts.

    ReplyDelete
  18. What are the criteria for the grade? As previous posters have mentioned, the forecast is for a large area, as I write this the lower tier counties if the seven county Twin City metropolitan area are under a NWS Blizzard Warning, and seem to be well on the way to receiving the higher snow amounts. It looks as if the north and northwest side of town might get a little less.

    ReplyDelete
  19. What are the criteria for the grade? As previous posters have mentioned, the forecast is for a large area, as I write this the lower tier counties if the seven county Twin City metropolitan area are under a NWS Blizzard Warning, and seem to be well on the way to receiving the higher snow amounts. It looks as if the north and northwest side of town might get a little less.

    I suspect that you will be able to grade this anyway you want! Good Luck.

    ReplyDelete
  20. I'll explain the rationale behind my grading when the report comes out. :-)

    ReplyDelete