It’s winter in South America and the forecast in Santiago, Chile called for snow. It didn’t snow. Snow is a rarity in Santiago, Chile. They get snow, not often. It’s a Mediterranean climate — just what we have here in SoCal.
Which brings us to the message I got Monday afternoon from a reporter at Las Últimas Noticias, a national newspaper in Chile.
Thank you for answering.
Last week, all the forecast said it will be a snowfall this Monday morning. Even the schools closed. Most of people took the prediction as something extremely catastrophic. Well, the predictions failed. The TV weather men and women had to give some explanations. There are a lot of complains and jokes on Twitter about this. So here are some questions:
-How difficult is to predict a snowfall?
-What factors do you have to consider make it?
-Is it a matter of technology? The snowfall forecasts in US don’t fail too much?
-Have you made mistakes in situations like this? What do you do next? Is it good or not to try a kind of apologize?
-In Chile, at least, the weather in TV channels is a kind of terrorific issue. A heavy rain prediction, for example, is almost the end of the world, even with a terrorific music in the background. What do you think about that?
-Do the weather sell in TV?
Las Últimas Noticias
I’m not sure how he found me, but of course I answered.
Of all the predictions meteorologist make, snow is arguably the most difficult. It is a multistage forecast. It’s not just how much moisture is in the atmosphere, but the temperature at different levels in the atmosphere. A small change in temperature is the deciding factor between rain and snow, or even sleet (ice pellets) or freezing rain (rain falls as liquid but freeze immediately on contact with trees, power lines, roadways, etc.).
We used to do this all ourselves. Computers have made life much easier because computer models follow the drops as they fall and figure out their state at every stage. Humans can’t do this alone — much too complex.
The technology has gotten much better over time, but snow forecasts are much more complex than most others. Cloud temperatures decided how fluffy snow is. Snow can range from 6 or 7:1 (7 cm snow from 1 cm liquid) to 30:1 or higher! So, how much snow will fall is very difficult to forecast and almost never exactly right.
I was on TV for 30 years in Connecticut, about 100 km northeast of NYC. If I got a snow forecast wrong people knew and were VERY upset. After all, I claim I can predict the future. People would come up to me on the street or at the market and let me know. No one wants to disappoint the people who trust you — certainly not me.
The number of wrong forecasts is definitely a fraction of what it was 30 years ago. No one is happier about that than me!
I have apologized on-the-air. Absolutely. I take responsibility. The forecast is mine, not the computer’s. I ask you to believe me.
Weather is the single most cited reason for people to watch local TV news. People accuse us of hyping storms, or making them scarier than they are. No. No. No. No. There is no upside to being wrong.
One more thing — Chile doesn’t make the forecast easier by sitting alongside the Pacific. Though we now have satellite observations, much of the Pacific has no surface weather observations. This leads to poor initialization of the computer models and garbage in/garbage out.
Ari – let me know if there’s anything else you need.
Glad to help,
A Connecticut Facebook friend in Santiago promises to keep an eye out for the story.