Here's Why AI is Changing How We Predict The Weather

Technology has helped to improve the accuracy of weather forecasts for decades. Artificial intelligence-driven models are becoming more common, allowing forecasting using far less computation power than previous methods. But how accurate are AI's forecasts, and how is the industry adapting? Our weather reporter Mary Hui joins host Stephen Carroll to discuss.

See omnystudio.com/listener for privacy information.

Episode Transcript

Speaker 1 (00:02):
Bloomberg Audio Studios, podcasts, radio news.

Speaker 2 (00:08):
I'm Stephen Carroll and this is Here's Why, where we
take one news story and explain it in just a
few minutes with our experts here at Bloomberg.

Speaker 1 (00:21):
Severe weather is making its way east after deadly storms
ravage the South.

Speaker 2 (00:26):
It's a fact that we will see due to the
climate development, more events, more severe event The weather, as
always has a huge impact. Most shopping is still done
in person. I'm singlehandedly powering the gardening section of retail
sales data. Frankly, it's the center of so many conversations,
but also a key economic driver. A cold snap can

(00:49):
drive up energy demand. Sunny days can send shoppers out
to buy new clothes. Then there are extreme weather events
that can damage infrastructure, disrupt supply chains, and spark mass
of insurance claims. As weather gets more volatile in many
parts of the world, knowing what's on the way is
even more important. Technology has helped to improve the accuracy

(01:10):
of forecasts. So what does the explosion and artificial intelligence
mean for the science. Here's why AI is changing how
we predict the weather. Our weather reporter Mary Hoy joins
me now for more, Mary, before we get to the
effect of AI. Can you just bring us up to speed.
At how good have we gotten at being able to

(01:31):
predict the weather?

Speaker 1 (01:32):
We've gotten very, very good. We've made incremental but tremendous
progress in the past decades since the very first computer
based weather forecast was made in nineteen fifty and that's
all thanks to various technological and scientific advancements of scientists
have made around the world. That's allowed us to divide
the world into smaller grids and then also divide the

(01:55):
air above us into more and more layers, and all
of that means that we can look at the weather
in more detail, which then increases our ability to see
further into the future. So you know, a five day
forecast now is as accurate as a three day forecast
was in two thousand. That's about a day for every
decade of technological and in scientific progress. So that's a lot

(02:15):
of advancement. And as the weather gets more volatile and extreme,
will be depending on this weather forecasting capabilities.

Speaker 2 (02:24):
Well, I'm interested in that point. Actually, the increase in
volatile weather events seems to be something that we're observing
at pace. How does that affect weather forecasting? Does it
make it more difficult.

Speaker 1 (02:34):
So weather forecasting is just fundamentally tricky. The atmosphere is complicated,
it's chaotic, it's uncertain, So every forecast has an element
of uncertainty. And that's because of incomplete observations, so not
being able to see exactly everything that's going on in
the atmosphere, and also just approximations that weather models have
to make. And now extreme weather makes that slightly trickier

(02:58):
and even more challenging because extreme weather, I suppose by definition,
they just have more variation variability, and they've also been
relatively rare, so that means less data and fewer observations
over time, and so we just understand these events less
well and have done less research into them. And then
these conditions can change really quickly. A hurricane, for example,

(03:21):
given warmer waters can rapidly intensify, and that quick change
can catch forecasters off guard. So all of that is
making it slightly trickier, even though weather forecasting itself has
never been an easy exercise.

Speaker 2 (03:34):
Well, let's get to artificial intelligence. Then, how are forecasters
using AI? How can it help this process?

Speaker 1 (03:42):
There are lots of steps to creating a weather forecast,
and AI can be applied to some of those steps,
or maybe all of those steps. We can take the
first step gathering realms and reams of data, because to
forecast future weather, you need to know what the present
weather's like, and that requires getting a lot of observations,
whether that's temperatures or air pressure, humidity and the like.

(04:06):
So AI can help by helping us gather more data
from a quantified sense, but also more types of data.
Whereas current weather models typically are restricted to meteorological observations,
AI models can now allow us to expand it to
say maintenance logs for power grids, or even news articles,

(04:28):
just really gathering more and more information to inform models
with to then create perhaps hopefully more accurate forecasts.

Speaker 2 (04:37):
How widespread is the use of AI and forecasting currently.

Speaker 1 (04:42):
So a key intergovernmental weather forecasting agency called the European
Center for Medium Range Weather Forecasting or ECMWF. They've taken
their AI model into operations earlier this year and they're
running it side by side with their existing traditional physics
based the model, and they're soon to launch another iteration

(05:03):
of that AI model into operations too. So there's lots
going on in this world. So the ECMWF. They're also
running AI models from different providers, including China's Huawei the
tech giant, Google's graph Cast model, Microsoft's Aurora model, and
China's own Weather Bureau itself is also testing about a
dozen AI weather models, so a lot of experimentation going on.

(05:26):
There were probably going to see AI models and traditional
weather models working in tandem in the years to come,
rather than seeing AI can completely replace these traditional numerical
weather model forecasting methods.

Speaker 2 (05:41):
Yeah, I'm curious if there are risks to using AI
for weather forecasting. Is there things that technology might miss
if we see a greater implementation of it.

Speaker 1 (05:50):
The way a lot of these AI models are trained
right now is on historical climate data from decades and
decades observations in analyzes. But the thing is that the
past is never perfect predictor of the future, and so
if we're about to see in a warming climate, more
extreme weather that previously is not reflected in these historical

(06:15):
data sets, then there is a risk that AI weather
models will say underestimates maybe the intensity of a typhoon
or a tropical cyclone, even as it improves the forecasts
of a storm's track, for example. So there are some
risks there, and it's definitely an open and ongoing feel
of the research.

Speaker 2 (06:32):
So what's the next big development we should be watching
out for in this area?

Speaker 1 (06:37):
This is less one single big development but rather trend
and I'd be looking for kind of a shifting distribution
right of roles in this whole global weather enterprise. Are
public weather agencies which we've long depended on for weather forecasts,
are they going to play a slightly different role now
that private companies, tech firms and also small players increasingly

(07:00):
jumping into this world of creating and providing weather forecasts
at a more niche level. So it'll be interesting to
see how that division of labor between the public forecasters
and more private players will play out.

Speaker 2 (07:12):
Okay, Mary Hoy, our weather reporter. Thank you very much
for joining us. For more explanations like this from our
team of three thousand journalists and analysts around the world,
go to Bloomberg dot com slash explainers. I'm Stephen Carroll.
This is here's why. I'll be back next week with more.
Thanks for listening.