Wind Atlas for Germany Analysis

Wind Atlas for Germany and Remodeling Effect

• A study investigated the uncertainty of wind simulations for the wind energy industry during project planning and financial considerations.
• Measurements from 118 onshore and offshore German sites were analyzed using the mesoscale WRF model.
• A correction of the annual cycle and a remodeling approach were developed to minimize differences between simulations and observations.
• The remodeling methodology utilized linear regression analysis of simulated and observed wind speeds, accounting for sub-grid variations in orography and roughness.
• Averaging regression parameters from 26 sites resulted in a global parameter set applied to wind atlas data.
• The "raw" data (without optimization) exhibited up to 30% difference in annual mean wind speed compared to observations.
• The remodeling process reduced the bias to below 5% for most measurements.
• Comparing with NEWA, EMD-WRF Europe+ and anemos "raw" data, an overall bias of 0.6-0.8 m/s was found, but decreased to zero with remodeling and site-specific adaptation.

Model Setup and Wind Atlas Simulation

• The WRF (Weather Research and Forecasting Model) version 3.7.1 was employed to downscale ERA5 reanalysis data for Germany.
• A two-way nesting approach used a 30x30 km² horizontal resolution from ERA5, downscaled to 9x9 km² and 3x3 km² domains.
• 50 vertical levels, with 14 in the lower 300m, were prescribed, crucial for wind energy applications.
• Initial and boundary conditions were derived from ERA5 data and nudged into the WRF model hourly.
• Model output stored every 10 minutes from 1997 and is still continuously updated.
• Orography from SRTM (Shuttle Radar Topography Mission) and vegetation/roughness from CORINE data, both interpolated to model grid.
• ERA5 data included soil temperature, soil moisture, and snow cover.
• WRF physics parameterizations included YSU planetary boundary layer, Monin-Obukhov surface layer, Noah land surface model, RRTM longwave radiation, and Dudhia shortwave radiation.

Observational Data

• Data from more than 100m high meteorological towers and lidar measurements were used for data analysis.
• Data primarily focused on onshore and offshore research stations for wind farm planning and wind farm characteristics.
• The observational uncertainty was considered low due to these towers often being purpose-built for wind energy.
• Measurements taken every 10 minutes and aggregated to hourly values.
• Data sets included from 48 onshore met masts and 4 offshore masts, used for remodeling and verification.
• Independent wind data set provided by Ramboll for 66 locations using lidar and mast data.

Verification and Remodeling

• The study optimized a wind atlas for Germany through a remodeling process.
• The approach involved correcting the annual cycle of the raw wind speed data, adapting to height and roughness, and calculating site-specific time series.
• The remodeling involved a four-step process based on comparing simulated and observed data from 26 onshore met masts, separately accounting for offshore sites.
• Verification metrics evaluated the effect of the process by comparing "raw" data, data after remodeling, and site-specific data to observations.
• The remodeling process generally reduced bias and improved correlation with measurements at various heights (e.g., 60m, 80m, 100m, 140m).

Results and Correlation of Hourly Wind Speed

• A multiple linear regression model derived a general correction function to improve wind simulation accuracy from 26 onshore masts.
• The models were applied to raw data sets which were considered "semi-independent".
• Results indicated that bias was consistently less than 5% for the majority of data sites.
• Correlation between modeled and observed wind data increased with height.
• Correlation and bias calculations also included data from NEWA and EMD-WRF Europe+ data sets.

Conclusion and Outlook

• Reducing uncertainty in wind simulations is crucial, especially for wind energy investments, and the study demonstrated improved model accuracy.
• Improvements in the simulation data set allow more accurate estimation of electricity production by wind turbines.

This quiz explores a study on the uncertainty of wind simulations in the German wind energy sector. It delves into the methodologies used, including the WRF model and linear regression analysis to improve simulation accuracy. Participants will learn about the remodeling techniques that significantly reduced bias in annual wind speed measurements.