As climate change urinate utmost weather condition more likely , and more unpredictable , meteorologist and regime agencies need new creature to keep up . Two studies published this week in the scientific journal Nature suggest that contrived intelligence can be used to support precise and fast forecasting and uttermost conditions warnings .
One studyoutlines how Huawei ’s AI model , Pangu - Weather , has been used to calculate hebdomadal conditions patterns around the world . The new tech is able to do so quicker than traditional weather forecasting method acting but with slightly lower accuracy . The internet was direct on 39 years of orbicular data from 1979 to 2017 , immix historical weather information with modern weather prognostication model . Researchers also analyse 69 agent including surface variables and upper atmospheric weather variable quantity , the study explained .
Pangu - Weather has leave to promising result in some scenario in weather condition forecasting . It predicted the former path of a tropical cyclone in the Pacific more accurately than other forecasting models , the sketch found . The machine learning modelling was more exact when study cyclones than other current weather forecasting model .
Floodwaters surround homes and vehicles in the community of Pajaro in Monterey County, California on 2 May 2025.Photo: Noah Berger (AP)
Currently , atmospheric condition forecasting utilizes thenumerical atmospheric condition prediction(NWP ) method . NWP computer modelsforecastfuture conditions by take apart contemporary weather pattern . “ [ It ] represents atmospherical states as discretized grids and numerically solves partial differential equations that describe the transition between those state . However , this operation is computationally expensive , ” the study writer wrote .
Many of the current models and method acting analyze dissimilar weather - related conditions like wind , pressure , and humidness one by one , MIT Technology Reviewreported . This makes predictions take longer to sue . Pangu - Weather researchers envision using the unexampled technical school alongside current foretelling models to ensure that the forecast are precise . “ Our ultimate goal is to build next - generation weather condition prediction framework using AI technologies to tone the existing forecasting systems , ” Tian Qi , a chief scientist of AI at HUAWEI , said in astatement .
The second studypublishedin Nature this workweek outlined how a deep - learning model predicted extreme rainfall more accurately and quicker than other preeminent prognosis methods . The foretelling model , NowcastNet , foretelling came in first about 71 % of the fourth dimension when tested against like forecast systems , the study explained .
Study author accentuate that Nowcasting can be used as a peril bar creature and for crisis management , specially as enquiry shows thatclimate variety has increased grueling rainfall . The National Oceanic and Atmospheric Administration hasestimatedthat 2022 ’s natural disasters cost the United States more than $ 160 billion . Planning for and augur these events could potentially save lives and allow agencies to apportion financial support for emergency readiness .
Researchers said that the tool ’s accuracy could hold those efforts . During examination , NowcastNet predicted rain up to three hour beforehand , making it an especially important tool for utmost weather condition prognostication . Other widely used models , likeDeepMind ’s DGMR , can predict the likelihood of heavy precipitation 90 minutes beforehand .
“ The prognosis made by NowcastNet are judged by expert meteorologist to be more accurate and instructive than pySTEPS , DGMR or other mysterious - erudition system , ” the subject area authors wrote .
We ’re currently in the early stages of acquire and deploy political machine find out to accurately predict incoming extreme weather . But it ’s hard to anticipate how the continue development of these AI prick will bear upon prediction . Current weather prediction example havebeen challengedby climate change , and there ’s no reason to say that AI model wo n’t fight with global weather shift key as well .
Want more climate and environment stories ? Check out Earther ’s guides todecarbonizing your place , divesting from fossil fuel , pack a disaster go bag , andovercoming climate dread . And do n’t miss our reportage of thelatest IPCC climate report , the future ofcarbon dioxide remotion , and theinvasive plant you should rip to smidgin .
BroadcastingDeepMindHuaweiWeather
Daily Newsletter
Get the good tech , science , and cultivation news show in your inbox day by day .
News from the future , delivered to your present .
You May Also Like
