By AI Trends Staff
Traditional models of weather forecasting are based on statistical measures based on data collected from deep space satellites, such as NOAA’s Deep Space Climate Observatory, weather balloons, radar systems, and sometimes from IoT-based sensors. Today, AI is finding a role in weather forecasting with machine learning being employed to process more complex data in less time, with the hope of improving accuracy.
For example, the Numerical Weather Prediction (NWP) site from NOAA offers a range of data sets for use by researchers, from temperature and precipitation data to wave heights, according to a recent account in Analytics Insight. The site offers vast data sets relayed from weather satellites, relay stations, and radiosondes to help deliver short-term weather forecasts or long-term climate predictions.
Besides machine learning, other AI techniques for weather predictions include Artificial Neural Networks, Ensemble Neural Networks, Backpropagation Networks, Radial Basis Function Networks, General Regression Neural Networks, Genetic Algorithms, Multilayer Perceptrons and fuzzy clustering.
IBM’s Weather Company Seeks to Transform Weather Forecasting
Weather prediction is big business. With a history of using computers to improve weather forecasting, IBM acquired The Weather Company and all its properties in 2016, including weather.com. IBM plans to use the Weather Companies extensive weather data with IBM Watson’s advanced cognitive capabilities and its Cloud platform to transform weather forecasting.
IBM late last year announced IBM GRAF, the Global High-Resolution Atmospheric Forecasting System, according to an IBM press release, to predict conditions up to 12 hours in advance with a detail and frequency previously unavailable.
Current global weather models cover 10-15 square kilometers (6.2-9.3 miles) and are updated every 6-12 hours. By contrast, IBM GRAF forecasts down to 3 kilometers (1.9 miles) and is updated hourly, IBM stated.
The Weather Company collaborated with the National Center for Atmospheric Research (NCAR) to create IBM GRAF, based on NCAR’s next-generation open-source global model, the Model for Prediction Across Scales (MPAS). That model is said to use state-of-the-art science to forecast the atmosphere down to thunderstorm level on a global scale.
The new IBM GRAF system runs on an IBM POWER9-based supercomputer optimized for both CPUs and GPUs (graphics processing units). The Weather Company and IBM, together with NCAR, the University of Wyoming’s Department of Electrical and Computer Engineering, and others applied OpenACC directives to MPAS, to take advantage of NVIDIA V100 Tensor Core GPUs on an IBM Power Systems AC922 server.
IBM states that this is the world’s first global weather model to run operationally on a GPU-based high-performance computing architecture. The hope is to put customers in a position to make better-informed, weather-related decisions.
Google Prediction Tool Tries to Predict Rain Six Hours Ahead
Google has developed a weather forecast tool making use of AI techniques to make accurate rainfall predictions six hours ahead of when the rain falls. The tool is based on the U-Net convolutional neural network (CNN) (developed for biomedical image segmentation at the Computer Science Department of the University of Freiburg, Germany), a sequence of layers of mathematical operations arranged in an encoding phase. It takes input images from satellites and transforms them into output images in a series of steps producing higher resolution.
In his team’s research paper, ML for Precipitation Nowcasting from Radar Images, Google Research’s Senior Software Engineer, Jason Hickey states, “If it takes 6 hours to compute a forecast, that allows only 3-4 runs per day and results in forecasts based on 6+ hour old data, which limits our knowledge of what is happening right now.” This UNET proposed tool of Google was reported to outperform alternatives.
In its conclusions, the researchers stated, “An open question remains as to whether pure Machine Learning data-driven approaches can outperform the traditional numerical methods, or perhaps ultimately, the best predictions will need to come from a combination of both approaches.”
Climate Corporation, a subsidiary of Bayer (formerly a division of Monsanto, which was acquired by Bayer in 2018), is using satellite imagery and hyper-local weather data with machine learning. The company’s FieldView digital farming platforms aims to provide farmers with advanced connectivity and easy access to machine-generated agronomic data.
The company recently reached an agreement with CLAAS, the German manufacturer of agricultural machinery, to give customers of CLAAS Telematics access to machine-generated weather data within FieldView.
“Farmers have been collecting data from their farm equipment for decades. The same is true for weather data, soil data, crop performance data, the list goes on and on,” stated Mike Stern, CEO of The Climate Corporation in a press release. “These data sets become even more valuable to our customers when they can be combined with the advanced AI tools we are developing to help drive profitability and reduce risk on their farms.”
Renewable Energy Industry Invests in AI Weather Startup
Weather prediction is proving valuable in the renewable energy business as well. The recent acquisition of AI and machine learning startup Climate Connect of India by clear energy firm ReNew Power, India’s largest renewable energy firm, is evidence of the trend.
ReNew Power plans to operate Climate Connect as an independent subsidiary that continues to develop software and its business. “The first wave of growth in the renewable energy industry came through the addition of physical assets on the ground,” stated Sumant Sinha, current chairman and managing director of ReNew Power, as reported in an account from CNBC. “The next wave will come through the development of digital products that help optimize power flow from generators to distribution companies to customers.”
Climate Connect CEO and Co-founder Nitin Tanwar stated, “We believe that the company’s acquisition by ReNew Power will help us create long-term value for our existing distribution utility and IPP customers and provide us the much-needed scale for the next leg of our journey.”
Read the source articles from Analytics Insight, an IBM press release on GRAF, a Google research paper, ML for Precipitation Nowcasting from Radar Images, a press release from Climate Corp./Bayer, and a report from CNBC.