Tuesday, March 20, 2012

Nowcasting Gone Wrong (Redux)

Last July, we wrote about "A Case of Nowcasting Gone Terribly Wrong," in which we wrote about Jonathan Yuhas's efforts to provide a nowcast, essentially a very short-term forecast, for how the weather might or might not affect a Twins game and a U2 concert. We wanted to believe his incorrect forecast was a one-time thing, a combination of bad luck and poor judgement. Now, however, we're not so sure.

At 4:03 p.m. on Monday, he tweeted this:

We looked at the radar at the time of the tweet, and this is what it showed (the radar crosshair is pegged at Lakeville):
At the time, it struck us as odd that the tiny, isolated cell (in yellow, about a county to a county and a half just east of due south of Lakeville) could be expected to hold together and land on Lakeville by 5 p.m.

Moments later, we expressed our thoughts to @JonathanYuhas:
Jonathan did not respond to our inquiry.

Fast forward to 5 p.m., and this was the Lakeville-centered radar depiction:

As the radar shows, there were no showers or thunderstorms anywhere remotely close to Lakeville. The t-storm near Owatonna, small as it was, had completely dissipated.

We understand the value of short-term forecasting and warning people about precipitation, but it's our feeling that the science is not yet to the point where such precise forecasts can be made. We think weather consumers are better served by the more general precipitation "nowcast" tweets that many other local meteorologists were tweeting at about the same time.

13 comments:

  1. It would have been interesting to see what the software 'RainAware' (discussed on this site in the past) would have shown for Lakeville during this time!

    ReplyDelete
    Replies
    1. @DysonGuy Indeed, it would have been interesting. So far, I'm pleased with RainAware. It seems to function as an informed observer of the radar and a good match for humans.

      Delete
  2. Not that I really care what local TV weather actors do, but are these Yuhas tweets done under the auspices of his employer? Or is it something he does for fun? If it's part of his job, then yes, I think he's wrongly lead people astray.

    ReplyDelete
    Replies
    1. I think you have to see them as part of his employment. That's why he has the platform he does.

      Delete
  3. Not only was he wrong, but is something like that really worth tweeting? It would be one thing if it was already a nasty storm with history of producing hail or gusty winds, but it is a yellow speck that probably amounted to 30 seconds of rain for those it hit.

    ReplyDelete
  4. Makes no sense at all. He must be trying to prove his worth or something.

    ReplyDelete
  5. To the TMF site readers I think many of you may have seen what I posted on March 19th at 10:02am..

    randyinchamplinMar 19, 2012 10:02 AM

    Most all of western MN is now in a slight risk area, the biggest threat is wind and hail. However looking at the short range modeling I'm mildly surprised they didn't include eastern MN as well. Best time frame seems to be around 3pm far sw MN to around 8pm eastern MN.

    Since I'm not on Twitter or Facebook (for privacy reason's, after all we have all heard about employer's asking for facebook passwords), I don't know what other mets were tweeting. IMO Jonathan saw the same set up I was seeing, (timing just a little off,can't trust models with timing).

    He saw the threat and tweeted what he thought was a prudent thing to do, trying to raise the overall threat to his follower's. At 625 pm a tornado touched down in Elysian MN.

    As Mr Brickmann said in a recent interview that was posted at this site, one of the biggest challenges he has since he moved here is forecasting for a entire state, after all the local markets like Duluth and Rochester etc only forecast for their viewing area, the Metro stations are far more reaching.

    So my question to those of you that follow the mets and their tweet's, did they adequately alert the area of the severe potential?

    I saw larger Convective Available Potential Energy (cape) numbers were forecast to pour into the metro from the se, suggesting the winds would turn to the se (which they did) while at the same time the wind just west of the metro would be out of the WSW, allowing for the chance of twisting in lower atmosphere, thus enhancing the treat of severe weather.

    I don't want to put words in Jonathan's mouth, but if he was the only met out there that was relaying the T-storm threat, than I say job well done!!!

    Call me nuts but I think he did a good job,( even though that cell near Owatonna died, it was a precursor as to what was to come.)

    ReplyDelete
  6. It was a precursor as to what was to come - to another area, not the one he was forecasting.

    ReplyDelete
  7. I think he was just jumping the gun on the weather that is to come. Soonn we will have the big storms where established cells are more forecastable.

    ReplyDelete
  8. Plymouth Weather LoverMarch 22, 2012 at 10:18 PM

    Love this continued warm air and rain. Grass is just so green. Snow on the long range models?

    ReplyDelete
  9. That's too bad that such a forecast can't still be made... I'd be glad to know what the weather will be like on some certain "big" days...

    ReplyDelete
  10. This comment has been removed by a blog administrator.

    ReplyDelete