Skillful forecasting of offshore winds from satellite scatterometer constellations

2026-07-29

AI Summary

This study addresses the challenge of high-precision nowcasting of wind speed and direction over maritime regions at minute- to hour-scale lead times by introducing WindCastNet—the first nowcasting framework leveraging multi-national satellite scatterometer constellations. The method innovatively fuses asynchronous and non-uniform microwave scatterometer observations from European, Chinese, and Indian missions through a continuous-time representation and an observation masking mechanism. A novel partially convolutional LSTM architecture is designed to encode spatial masks and inter-observation temporal gaps, enabling forecasts at arbitrary lead times. Evaluated over the North Sea, WindCastNet reduces root-mean-square error by 23% and 7% for 1-hour and 2-hour wind field forecasts, respectively, compared to the HARMONIE MEPS model, and outperforms a persistence baseline by 9–15% in the first three hours.

Abstract

Accurate intraday forecasts of offshore wind are becoming increasingly important for power system operation and the integration of growing shares of offshore wind energy. Operational forecasts rely predominantly on numerical weather prediction (NWP), which is not optimized for lead times of minutes to hours, where initial-condition accuracy dominates forecast skill. Although satellite scatterometer observations are routinely assimilated into NWP, they have not previously been used directly for forecasting. Here we present WindCastNet, the first satellite-based nowcasting framework for offshore wind speed and direction, introducing a new paradigm for intraday forecasting that learns from spatiotemporally irregular satellite observations. WindCastNet predicts offshore wind fields from observations acquired by satellite scatterometer constellations. WindCastNet employs a partial convolutional long short-term memory network that exploits microwave radar observations from the European, Chinese, and Indian scatterometers despite their irregular spatial coverage, asynchronous sampling, and variable revisit times. Spatial observation masks and inter-observation intervals are encoded, while a continuous temporal representation enables forecasts at arbitrary lead times. Evaluated over the North Sea, WindCastNet reduces the root-mean-square error by 23% and 7% relative to the HARMONIE MEPS model at lead times of 1 and 2 h, respectively, and outperforms persistence by 9-15% during the first three forecast hours. Forecast skill decreases under strong-wind conditions and spatially non-uniform flow. These results demonstrate that satellite scatterometer constellations can provide an independent and competitive source of short-term offshore wind forecasts, opening new opportunities for renewable energy forecasting but also broader marine weather applications, including tropical cyclone nowcasting.

Problem

• offshore wind forecasting
• nowcasting
• satellite scatterometer
• intraday prediction
• wind energy integration

Innovation

• satellite scatterometer
• nowcasting
• offshore wind forecasting
• partial ConvLSTM
• spatiotemporally irregular data