China races to dethrone Europe’s weather data with homegrown AI dataset

China is ramping up efforts to wean itself off European weather data as it seeks to become a leader in the global field of meteorology and climate science. Officials are promoting a homegrown data set designed for the era of artificial intelligence, part of Beijing’s comprehensive effort to secure data and achieve technological independence.

Europe’s ERA5 dataset is the global standard for climate and weather data, providing measurements of variables such as rainfall, temperature, and wind spread over more than 80 years. It has also become the foundation of AI-powered weather forecasting, with many cutting-edge models, including those developed in China by companies like Huawei, trained on its data.

However, depending on a foreign system, projects may conflict with China’s national strategy. For Chinese policymakers, meteorological data is increasingly seen as strategic infrastructure rather than just scientific input. That view has become more urgent as AI models require increasingly larger and more detailed datasets.

“Weather forecasting is national security,” said Andreas Prein, a professor at ETH Zurich and an expert in weather and climate modeling. “If you solely depend on external data sources, you make yourself vulnerable.”

China advances in-house data in bid to boost AI forecasting

ERA5 is created through climate reanalysis, a sophisticated method that combines millions of global observations into a consistent picture of past weather. Created by the European Centre for Medium-Range Weather Forecasts, the dataset is continually updated and heavily relied upon by governments, insurers, and researchers globally.

Public agencies utilize ERA5 to mitigate risks from floods, droughts, and wildfires, as well as for insurers to model catastrophe exposure. The European Union calculates that the data set is delivering savings worth hundreds of millions of dollars every year, underscoring its central role in public policymaking and private markets.

Now, China sees that dominance as a strategic weakness. Months earlier, the National Data Administration stated that it is Beijing’s goal for China to shift from overdependence on European and American reanalysis products. At the same time, the China Meteorological Administration also released a new global dataset (CMA-RA V1) over this period in June. 5, which it launched for a global download.

The agency had said earlier that some Chinese AI weather models were now being trained on the new data. Officials also described the release as part of a broader effort to disseminate “high-value data” — information that can be used to support innovation — while also enhancing national capabilities.

Global researchers test China’s reanalysis as alternatives expand

Already, researchers and companies are starting to examine how the Chinese dataset works in practice. Hui Su, a professor at the Hong Kong University of Science and Technology, is using the data in her startup, Stellerus, to train AI models at regional scales and evaluate traditional numerical forecasts.

One big plus, Su said, is the higher spatial and temporal resolution of the dataset. The smaller grid, compared to the ERA5, enables the system to produce higher-resolution data, which can be useful for both AI model training and regional investigation. This particular aspect is very useful for local prediction studies.

Beyond the scope of forecasting, more readily available Chinese reanalysis data could open up global markets for weather risk hedging. Far more developed in the US and Europe, where rich data supports financial products linked to climate and weather extremes, such markets are still developing elsewhere.

David Whitehead, head of weather risk management at Vaisala, a Finnish company, stated that greater access to Chinese weather data would likely attract significant interest from many companies in international markets. He added that the company has already begun analyzing how the Chinese dataset could underpin financial hedging products.

China’s meteorological authority says its reanalysis is more accurate for wind speeds at 100 meters above ground level, which are of key importance to the world’s largest wind power market. However, experts caution that in the near term, the dataset is still unlikely to displace ERA5.

Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.