Getting Better At Predicting The Weather
ARUN RATH, HOST:
It's ALL THINGS CONSIDERED from NPR West. I'm Arun Rath.
It has been an ugly winter. In the past two weeks, a pair of storms has made life miserable across the Eastern U.S. On Thursday, much of the south and northeast were buried in snow and ice. At least 26 people have died. Tens of thousands of flights have been canceled, rail service delayed, and roads in many cities are still impassable.
Meanwhile, more than a million customers lost power during the storm. Thousands of homes and businesses are still without electricity. From the Deep South to the northeast, forecasters are working double time.
(SOUNDBITE OF WEATHER REPORTS)
UNIDENTIFIED MAN #1: The Weather Channel has just named its 17th winter storm of the season...
UNIDENTIFIED WOMAN #1: ...from the heaviest snowfall this winter.
UNIDENTIFIED MAN #2: It's going to be a cold and windy afternoon.
UNIDENTIFIED WOMAN #2: More than 75,000 domestic flights canceled and...
UNIDENTIFIED MAN #3: ...and that is where the forecast gets very tricky. So here...
RATH: So who takes the blame when cities are unprepared for weather disasters, weather forecasters or politicians?
Thomas Bogdan heads up the University Corporation for Atmospheric Research. It's a group that oversees national weather studies, and he says that forecasting can definitely get a lot better.
THOMAS BOGDAN: Forecasts have improved a lot over the years, but the demands on those forecasts have probably increased even faster than they have. You know, as society gets increasingly complex and interconnected, and we operate closer and closer to the margins of our systems, we're increasingly vulnerable to winter storms such as the one that hit Atlanta last month.
So first of all, it's trying to improve our models to deal with the timing and location, which are critical to what's going on. And then it's also completing that forecast, finding a way to take the information we have about the physical system and translate it into information that the average person can use to make good decisions about whether they go out, when they come home from work, if they pick up the kids from day care and things like that.
RATH: What are some of the things that can be done on the ground to make weather forecasting more accurate?
BOGDAN: Certainly on the ground, one of the key things is to bring more data into the forecast. What's exciting now is the potential of crowdsourcing. You think about smart automobiles. They know the temperature outside, they know whether they've been using their antilock brakes, they know whether the high beams are on or whether they've used the windshield wipers.
The Department of Transportation has funded a winter study in the states of Michigan, Minnesota and Nevada where they are deploying small sensors on plows. Even your cellphone, if you think about it, knows some critical things about the weather around you, and so each and every one of us with a smartphone can potentially be a small weather station. And that knowledge, that very dense knowledge is what we need to nudge the forecasts and keep them accurate.
RATH: You've also pointed out that we're good at predicting two things: That would be immediate weather, say, that there's a big storm coming in tomorrow; and long-term trends like an ongoing drought. But you've noted we're terrible at predicting weather over the medium term, like months or a year. Why is that?
BOGDAN: Indeed. The intermediate forecast is an area where scientists are very excited about trying to make some progress. We're talking about timescales from maybe one month to one decade as this intermediate area that I like to refer to as wimate or cleather, where you put weather and climate together and make a word that's unpronounceable.
And this area, it's fascinating that the ocean is the planet's memory on these timescales. And so the atmospheric sciences community and the oceanic community are really marshaling their forces to try to put together global earth system models that really include the ocean on the same par as the atmosphere.
Of course, we will never be making predictions that it's going to rain at 2:30 on February 17, 2018. That's not the way it's going to work. It's going to be probabilistic, but the economic impacts of getting this right are huge. If you think about companies, water resource managers, people in agriculture - there are so many decisions that our society has to make. And right now, without really good information on those timescales, they are encountering a lot of risk.
RATH: What are some of the ways that meteorologists can get better at delivering their message to the public?
BOGDAN: I think part of it is working with public officials that actually have to make decisions and understand what their criteria are. What is the threshold at which they have to begin considering taking actions that are out of the ordinary to respond to a situation? And once those thresholds are understood by the weather community, then you don't give information before you're at that threshold because it's not information that's actionable.
RATH: Now, in terms of the decisions that affect people's lives, obviously, the meteorologist, the National Weather Service, don't decide when to close schools and that sort of thing. So on the side of local governments, are there things that they could be doing that could - to be doing a better job?
BOGDAN: Well, I think communication is the key here. Atlanta's a really interesting situation to observe. They had the same type of event come back basically one week later after the first event, and the reaction was totally different from what they did there. And it was because relationships had formed between the officials and the weather people that provide that information. That trust was there, and lessons had been learned from the first time through.
RATH: That's Thomas Bogdan, president of the University Corporation for Atmospheric Research, on a rough couple of weeks of winter weather.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.