Field NotesWhat we're learning about AI and the environment

Wind forecasting for India's grid with Google DeepMind weather models

⚠ Not yet claimed by the project owner or their organization. This information may not be completely accurate. Is this your project? Claim it.

Open Climate Fix and Google DeepMind tested DeepMind's WeatherNext weather models inside Open Climate Fix's wind forecast for India. The forecast is used by Indian state grid operators and by the power producer Adani Green, and the partners are now moving it into live operation.

Why it mattersIndia is adding renewable power fast, and producers pay penalties when their output misses the forecast. Better wind forecasts cut those penalties and make it easier for the grid to rely on clean power.

What they learned

Worked

AI weather models improved wind forecasts by up to 8% in testing

From public sources: Open Climate Fix reports up to an 8% gain in forecast accuracy for Indian state operators when tested on historical wind data, and an average 13% cut in imbalance penalties for Adani Green. The partners say these test results still need to be confirmed at scale, which is the purpose of the live rollout.

The problem

Wind output is hard to predict, and in India forecast misses carry regulatory penalties, which are especially costly at very large sites.

What the AI does

AI weather models predict the weather, and a second model turns those predictions into expected wind-power output for each site.

Sources

This entry was summarized in our own words from these public sources.

Claim this entry Suggest an edit Ask for a correction or removal

Contact the team

We forward your message. We never show anyone's email address.