Abstract:With an increasing share of the electricity grid relying on wind to provide generating capacity and energy, there is an expanding global need for historically accurate high-resolution wind data. Conventional downscaling methods for generating these data have a high computational burden and require extensive tuning for historical accuracy. In this work, we present a novel deep learning-based spatiotemporal downscaling method, using generative adversarial networks (GANs), for generating historically accurate high-resolution wind resource data from the European Centre for Medium-Range Weather Forecasting Reanalysis version 5 data (ERA5). We achieve results comparable in historical accuracy and spatiotemporal variability to conventional downscaling by training a GAN model with ERA5 low-resolution input and high-resolution targets from the Wind Integration National Dataset, while reducing computational costs over dynamical downscaling by two orders of magnitude. Spatiotemporal cross-validation shows low error and high correlations with observations and excellent agreement with holdout data across distributions of physical metrics. We apply this approach to downscale 30-km hourly ERA5 data to 2-km 5-minute wind data for January 2000 through December 2023 at multiple hub heights over Eastern Europe. Uncertainty is estimated over the period with observational data by additionally downscaling the members of the European Centre for Medium-Range Weather Forecasting Ensemble of Data Assimilations. Comparisons against observational data from the Meteorological Assimilation Data Ingest System and multiple wind farms show comparable performance to the CONUS validation. This 24-year data record is the first member of the super resolution for renewable energy resource data with wind from reanalysis data dataset (Sup3rWind).

What problem does this paper attempt to address?

The main objective of this paper is to develop a novel spatiotemporal downscaling method based on Generative Adversarial Networks (GANs) to generate high-resolution historical wind energy resource data from low-resolution ERA5 reanalysis data from the European Centre for Medium-Range Weather Forecasts. Specifically, the paper addresses the following key issues: 1. **Improving Wind Power Data Accuracy**: As the proportion of wind energy in power systems increases, the demand for wind energy resource datasets with high spatial and temporal resolution is also growing. Such historically accurate meteorological data are crucial for ensuring resource adequacy, reliable system operation, and efficient power market functioning. 2. **Reducing Computational Costs**: Traditional downscaling methods, especially those based on numerical weather prediction, can provide high-accuracy data but are computationally intensive and require extensive tuning to ensure historical accuracy. The proposed method can reduce computational costs by 2 orders of magnitude compared to traditional dynamic downscaling methods. 3. **Expanding Application Scope**: The method in the paper has been validated not only in the continental United States (CONUS) but also in parts of Ukraine, Moldova, and Romania to support the post-conflict reconstruction and decentralization of power grids in these regions. The effectiveness and generalizability of the method have been demonstrated through comparisons with observational data. 4. **Uncertainty Estimation**: The paper also explores how to estimate uncertainty for the observation periods of 2012-2015 and 2019-2023 by downscaling the ensemble data assimilation (EDA) members from the European Centre for Medium-Range Weather Forecasts. In summary, the main contribution of this paper is the proposal of an efficient GAN-based spatiotemporal downscaling method that significantly reduces computational costs while maintaining historical accuracy and exhibits strong generalizability and scalability.

Super Resolution for Renewable Energy Resource Data With Wind From Reanalysis Data (Sup3rWind) and Application to Ukraine

Wind Power Assessment based on Super-Resolution and Downscaling -- A Comparison of Deep Learning Methods

Super-resolution perception for wind power forecasting by enhancing historical data

A 3D super-resolution of wind fields via physics-informed pixel-wise self-attention generative adversarial network

Algorithmic Hallucinations of Near-Surface Winds: Statistical Downscaling with Generative Adversarial Networks to Convection-Permitting Scales

Improving trajectory calculations using deep learning inspired single image superresolution

Downscaling and Wind Resource Assessment of Climatic Wind Speed Data Based on Deep Learning: A Case Study of the Tengger Desert Wind Farm

Statistical treatment of convolutional neural network super-resolution of inland surface wind for subgrid-scale variability quantification

A novel wind model downscaling with statistical regression and forecast for the cleaner energy

An Unsupervised Learning Approach for Predicting Wind Farm Power and Downstream Wakes Using Weather Patterns

Interpolation-Free Deep Learning for Meteorological Downscaling on Unaligned Grids Across Multiple Domains with Application to Wind Power

High-resolution meteorology with climate change impacts from global climate model data using generative machine learning

Wind speed super-resolution and validation: from ERA5 to CERRA via diffusion models

Data-Driven Multi-Resolution Probabilistic Energy and Reserve Bidding of Wind Power

Enhancing operational wind downscaling capabilities over Canada: Application of a Conditional Wasserstein GAN methodology

The impact of the spatial resolution of wind data on multi-decadal wind power forecasts in Germany

Enhancing wind field resolution in complex terrain through a knowledge-driven machine learning approach

Forecasting Spatio-Temporal Renewable Scenarios: a Deep Generative Approach

Spatio-temporal estimation of wind speed and wind power using machine learning: predictions, uncertainty and technical potential

AIRU-WRF: A Physics-Guided Spatio-Temporal Wind Forecasting Model and its Application to the U.S. Mid Atlantic Offshore Wind Energy Areas