Latest Headlines

Forecasting the weather isn't easy

March 16, 2016 4:55 A.M.
(B.C.Govt photo)

Today's Feature Comment has been posted to the story "Weather reports weren't always this wrong."

Hi, I'm a meteorologist, living in Kamloops. I don't work for Environment Canada, but I do do forecasts throughout British Columbia, including Kamloops. You're claiming that the forecasts are 'wrong' far more than they're 'right'. I'm not sure how you're determining this. Are you keeping track of the forecast daily, including the probability of precipitation, and comparing the forecasts to the actual weather? How do you determine if a forecast is 'wrong'?

If the probability of precipitation is 30%, and it rains, was the forecast wrong? If the high temperature was forecast to be -2, and it's 0, was the forecast wrong? If the forecast was for 2 cm and it was 3 cm...or 1 cm, was the forecast wrong? If you're not actually looking at this quantitatively, then you're subject, as you are, to confirmation bias, one of the most frequent ways humans fool themselves.

You've decided that 'we don't have local forecasters therefore the forecasts are more wrong than they're right'. Then you only remember the situations where the forecasters were wrong. "Welp, they blew it again", and forget about the multitude of times they got it correct. Especially if one of those times we get it wrong is a major event. https://en.wikipedia.org/wiki/Confirmation_bias The fact is that weather forecasting isn't easy, ESPECIALLY when it comes to forecasting snowfall, and it's amazing we get it right as much as we do.

In fact, according to Nate Silver, in his essay "The Weatherman is not a Moron", weather forecasting is the one area in which humans make predictions about the future that is not only good, but improving. In fact, the short-term forecasts for the interior improved dramatically in my opinion, once the Silver Star radar became operational.

I know that my confidence in being correct sure improved. And local knowledge isn't a panacea. While it can help apply the forecasts locally (ie. I often am discussing snow only being in Aberdeen and Mt. Dufferin) if the overall forecast is incorrect, the local forecast will also be incorrect.

I've been on the delivery of some spectacular busts while doing weather forecasts for Kamloops, but also have done many, many, many good ones. But my forecasts don't get transmitted to the public.

As for that snowstorm this past winter, I was also surprised by how intense it was. Although I wasn't actually on duty for that storm (it was one of my days off), I was monitoring and was caught off guard by how much snow we were getting in such a short amount of time. Why didn't EC send out a warning earlier? Because the criteria for a snowfall warning for our area is 15 cm or more.

This is a rather large hurdle to breach. But once they _did_ realize that it was going to be breached, they sent out the warning. So yeah, that wasn't EC's finest hour, but the local meteorologist _also_ got it wrong. You just didn't get to see it.

DEAN GILBERT

Stewart Duncan says:
March 16, 2016 09:00am

While I rarely consult forecasts, I can't help wonder if Dean Gilbert et al forecast the heavy blanket of snow that fell over Lower Sahali at 8:30 a.m.

Reply

Emails will not be published

Leave a Reply

Emails will not be published