The unpredictability of weather poses significant challenges for urban planners, city managers, and residents. Traditional weather forecasting methods, while advanced, often fall short in providing the hyper-local and highly accurate predictions necessary for modern cities. This gap can lead to inefficiencies in urban infrastructure management, emergency response, and everyday decision-making for citizens. Machine Learning Weather Forecasting (MLWF) solves these problems by leveraging advanced computational techniques to deliver precise and timely weather forecasts.
This solution employs sophisticated algorithms that process vast amounts of meteorological data from various sources, such as satellites, weather stations, and IoT devices. These weather models are designed to identify complex patterns and relationships within the data that traditional methods might miss. By continuously learning from new data, MLWF improve its accuracy over time. This adaptive capability enables it to provide forecasts with higher precision at much smaller spatial and temporal scales, making them invaluable for urban applications.
The core advantage of MLWF lies in its ability to integrate and analyse diverse datasets rapidly. For instance, it can process real-time data on temperature, humidity, wind speed, and atmospheric pressure, combining it with historical weather patterns and geographical information to predict localised weather conditions. This predictive power is particularly crucial for cities, where weather varies significantly across neighbourhoods. As a result, urban planners and citizens can better prepare for extreme weather events, optimise traffic management, and enhance public safety measures.
As urban areas continue to expand and climate change introduces more volatility into weather patterns, the need for precise and reliable weather forecasting becomes ever more critical. This solution can help cities become more resilient by providing early warnings for severe weather, thus reducing the risk of damage to infrastructure and loss of life. Moreover, these models support sustainable urban development by enabling more efficient use of resources, such as water and energy, based on accurate weather predictions.
Developers of the Gemini family of models, which are trained from the start to be multimodal across text, images, video, and audio.
Developing foundation models for robotics (Project GR00T) and vision-language models like VILA.
Operates proprietary radar satellites and uses generative AI ('Gale') for weather intelligence.
The European Centre for Medium-Range Weather Forecasts, traditionally a numerical modeling hub, now actively integrating ML.
China · Company
Developed Pangu-Weather, a 3D high-resolution AI model for global weather forecasting.
United States · Startup
Focuses on sub-seasonal to seasonal (S2S) forecasting using machine learning and ocean data.
United States · Startup
Provides AI-meteorology solutions to governments and defense using a proprietary neural network.
Switzerland · Startup
Developing a 'Large Physics Model' for weather prediction, aiming to provide high-resolution energy-focused forecasts.
Uses a constellation of nanosatellites to collect radio occultation data, fed into ML models for forecasting.
Provides climate risk analytics using cloud computing and AI to model extreme weather risks for asset planning.
Paper
A data-to-forecast machine learning system for global weatherNature Communications · Jul 19, 2025
FuXi Weather is a machine learning-based global forecasting system that assimilates multi-satellite data, generating reliable 10-day forecasts at 0.25° resolution using fewer observations than conventional NWP systems.
Paper
End-to-end data-driven weather predictionNature · Mar 20, 2025
Aardvark Weather demonstrates that a single machine learning model can replace the entire numerical weather prediction pipeline. It produces global gridded forecasts and local station forecasts that are skillful for up to ten days, competing with state-of-the-art systems.
Paper
End-to-end data-driven weather predictionNature · Mar 20, 2025
Aardvark Weather is an end-to-end data-driven weather prediction system that ingests observations to produce global gridded and local station forecasts, outperforming operational NWP baselines.
Paper
A regional high resolution AI weather model for the prediction of atmospheric rivers and extreme precipitationnpj Climate and Atmospheric Science · Dec 12, 2025
This study presents a stretched-grid AI weather model with 6-km horizontal resolution over the Western US. It effectively captures extreme precipitation events linked to atmospheric rivers, which coarser global models often underestimate.
Paper
A data-to-forecast machine learning system for global weatherNature Communications · Jul 19, 2025
FuXi Weather is a machine learning-based global forecasting system that assimilates multi-satellite data to generate reliable 10-day forecasts at 0.25° resolution. It outperforms ECMWF high-resolution forecasts beyond day one in observation-sparse regions.
Paper
FourCastNet 3: A geometric approach to probabilistic machine-learning weather forecasting at scalearXiv · Jul 1, 2025
FourCastNet 3 uses a geometric ML approach to probabilistic ensemble forecasting, surpassing leading conventional models and rivaling diffusion methods while being significantly faster.
Article
AI Might Be the Future for Weather Forecastinginterestingengineering.com
Weather forecasting has traditionally been a best guess, but could AI change all that?
Article
Accelerating Sustainability with AI: A Playbookblogs.microsoft.com
Today, Microsoft published a playbook for accelerating sustainability solutions with AI. You can read the foreword below and explore the piece in its entirety here.
Article
The rise of machine learning in weather forecastingecmwf.int
Machine learning (ML) has been one of the global topics of discussion this year. Everyone seems to be enjoying the exploration of generative artificial intelligence (AI), in the form of language and image models, to write their boring emails, do their homework or even fix their photos.
Article
GraphCast: AI model for faster and more accurate global weather forecastingdeepmind.google
Our state-of-the-art model delivers 10-day weather predictions at unprecedented accuracy in under one minute
Article
How AI models are transforming weather forecasting: a showcase of data-driven systemsecmwf.int
Developments in machine learning are continuing at breathtaking pace, both inside and outside of weather forecasting. To help assess machine learning weather forecasts from different sources, we now show a range of them in ECMWF’s charts catalogue.
Article
Deep-learning model speeds extreme weather predictionsphys.org
Climate change is one of the greatest challenges facing humanity today. To help address this, researchers from Lawrence Berkeley National Laboratory (Berkeley Lab), Caltech, and NVIDIA trained the Fourier Neural Operator (FNO) deep learning model—which learns complex physical systems accurately and efficiently—to emulate atmospheric dynamics and provide high-fidelity extreme weather predictions across the globe a full five days in advance.
Article
Machine Learning Methods in Weather and Climate Applications: A Surveymdpi.com
With the rapid development of artificial intelligence, machine learning is gradually becoming popular for predictions in all walks of life. In meteorology, it is gradually competing with traditional climate predictions dominated by physical models. This survey aims to consolidate the current understanding of Machine Learning (ML) applications in weather and climate prediction—a field of growing importance across multiple sectors, including agriculture and disaster management. Building upon an exhaustive review of more than 20 methods highlighted in existing literature, this survey pinpointed eight techniques that show particular promise for improving the accuracy of both short-term weather and medium-to-long-term climate forecasts. According to the survey, while ML demonstrates significant capabilities in short-term weather prediction, its application in medium-to-long-term climate forecasting remains limited, constrained by factors such as intricate climate variables and data limitations. Current literature tends to focus narrowly on either short-term weather or medium-to-long-term climate forecasting, often neglecting the relationship between the two, as well as general neglect of modeling structure and recent advances. By providing an integrated analysis of models spanning different time scales, this survey aims to bridge these gaps, thereby serving as a meaningful guide for future interdisciplinary research in this rapidly evolving field.
Article
Google says new AI models allow for ‘nearly instantaneous’ weather forecaststheverge.com
/ An increasingly important tool in a world shaped by climate change
Article
How Machine Learning Could Help to Improve Climate Forecastsscientificamerican.com
Mixing artificial intelligence with climate science helps researchers to identify previously unknown atmospheric processes and rank climate models
Article
Forecasting Climatic Trends Using Neural Networks: An Experimental Study Using Global Historical Datafrontiersin.org
Climate change is undoubtedly one of the biggest problems in the 21st century. Currently, however, most research efforts on climate forecasting are based on mechanistic, bottom-up approaches such as physics-based general circulation models and earth system models. In this study, we explore the performance of a phenomenological, top-down model constructed using a neural network and big data of global mean monthly temperature. By generating graphical images using the monthly temperature data of 30 years, the neural network system successfully predicts the rise and fall of temperatures for the next 10 years. Using LeNet for the convolutional neural network, the accuracy of the best global model is found to be 97.0%; we found that if more training images are used, a higher accuracy can be attained. We also found that the color scheme of the graphical images affects the performance of the model. Moreover, the prediction accuracy differs among climatic zones and temporal ranges. This study illustrated that the performance of the top-down approach is notably high in comparison to the conventional bottom-up approach for decadal-scale forecasting. We suggest using artificial intelligence-based forecasting methods along with conventional physics-based models because these two approaches can work together in a complementary manner.
Article
AI and machine learning are improving weather forecasts, but they won’t replace human expertstheconversation.com
A century ago, English mathematician Lewis Fry Richardson proposed a startling idea for that time: constructing a systematic process based on math for predicting the weather. In his 1922 book, “Weather Prediction By Numerical Process,” Richardson tried to write an equation that he could use to solve the dynamics of the atmosphere based on hand calculations.
Article
Artificial Intelligence Can Now Predict When Lightning Will Strikepopularmechanics.com
So, hopefully we can get out of the way before it's too late.
Article
Deep Learning-Based Weather Prediction: A Surveysciencedirect.com
Weather forecasting plays a fundamental role in the early warning of weather impacts on various aspects of human livelihood. For instance, weather forecasting provides decision making support for autonomous vehicles to reduce traffic accidents and congestions, which completely depend on the sensing and predicting of external environmental factors such as rainfall, air visibility and so on. Accurate and timely weather prediction has always been the goal of meteorological scientists. However, the conventional theory-driven numerical weather prediction (NWP) methods face many challenges, such as incomplete understanding of physical mechanisms, difficulties in obtaining useful knowledge from the deluge of observation data, and the requirement of powerful computing resources. With the successful application of data-driven deep learning method in various fields, such as computer vision, speech recognition, and time series prediction, it has been proven that deep learning method can effectively mine the temporal and spatial features from the spatio-temporal data. Meteorological data is a typical big geospatial data. Deep learning-based weather prediction (DLWP) is expected to be a strong supplement to the conventional method. At present, many researchers have tried to introduce data-driven deep learning into weather forecasting, and have achieved some preliminary results. In this paper, we survey the state-of-the-art studies of deep learning-based weather forecasting, in the aspects of the design of neural network (NN) architectures, spatial and temporal scales, as well as the datasets and benchmarks. Then we analyze the advantages and disadvantages of DLWP by comparing it with the conventional NWP, and summarize the potential future research topics of DLWP.
Article
The AI forecaster: Machine learning takes on weather predictionphys.org
According to a 2009 study, U.S. adults look at weather forecasts nearly 300 billion times a year. Reliable forecasts can predict hazardous weather―such as blizzards, hurricanes, and flash floods―as early as 9–10 days before the event. Estimates value these forecasts at $31.5 billion per year.
Article
Machine Learning−based Weather Support for the 2022 Winter Olympicslink.springer.com
Article
Sub-Seasonal Forecasting With a Large Ensemble of Deep-Learning Weather Prediction Modelsagupubs.onlinelibrary.wiley.com
We present an ensemble prediction system using a Deep Learning Weather Prediction (DLWP) model that recursively predicts six key atmospheric variables with six-hour time resolution. This computationally efficient model uses convolutional neural networks (CNNs) on a cubed sphere grid to produce global forecasts. The trained model requires just three minutes on a single GPU to produce a 320-member set of six-week forecasts at 1.4° resolution. Ensemble spread is primarily produced by randomizing the CNN training process to create a set of 32 DLWP models with slightly different learned weights. Although our DLWP model does not forecast precipitation, it does forecast total column water vapor and gives a reasonable 4.5-day deterministic forecast of Hurricane Irma. In addition to simulating mid-latitude weather systems, it spontaneously generates tropical cyclones in a one-year free-running simulation. Averaged globally and over a two-year test set, the ensemble mean RMSE retains skill relative to climatology beyond two-weeks, with anomaly correlation coefficients remaining above 0.6 through six days. Our primary application is to subseasonal-to-seasonal (S2S) forecasting at lead times from two to six weeks. Current forecast systems have low skill in predicting one- or 2-week-average weather patterns at S2S time scales. The continuous ranked probability score (CRPS) and the ranked probability skill score (RPSS) show that the DLWP ensemble is only modestly inferior in performance to the European Center for Medium Range Weather Forecasts (ECMWF) S2S ensemble over land at lead times of 4 and 5–6 weeks. At shorter lead times, the ECMWF ensemble performs better than DLWP.
Article
Artificial Intelligence—A Game Changer for Climate Change and the Environmentnews.climate.columbia.edu
As the planet continues to warm, climate change impacts are worsening. In 2016, there were 772 weather and disaster events, triple the number that occurred in 1980. Twenty percent of species currently face extinction, and that number could rise to 50 percent by 2100. And even if all countries keep their Paris climate pledges, by 2100, it’s likely that average global temperatures will be 3˚C higher than in pre-industrial times.
Article
Thunderstorm Predictions Using Artificial Neural Networksintechopen.com
Artificial neural network (ANN) model classifiers were developed to generate ≤15h predictions of thunderstorms within three 400-km2 domains. The feed-forward, multi-layer perceptron and single hidden layer network topology, scaled conjugate gradient learning algorithm, and the sigmoid (linear) transfer function in the hidden (output) layer were used. The optimal number of neurons in the hidden layer was determined iteratively based on training set performance. Three sets of nine ANN models were developed: two sets based on predictors chosen from feature selection (FS) techniques and one set with all 36 predictors. The predictors were based on output from a numerical weather prediction (NWP) model. This study amends an earlier study and involves the increase in available training data by two orders of magnitude. ANN model performance was compared to corresponding performances of operational forecasters and multi-linear regression (MLR) models. Results revealed improvement relative to ANN models from the previous study. Comparative results between the three sets of classifiers, NDFD, and MLR models for this study were mixed—the best performers were a function of prediction hour, domain, and FS technique. Boosting the fraction of total positive target data (lightning strikes) in the training set did not improve generalization.
