Reducing errors of wind speed forecasts by an optimal combination of post‐processing methods

Sweeney, C.; Lynch, Peter; Nolan, P.

doi:10.1002/met.294

Cited by 59 publications

(47 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Wind speeds can show large spatial variability below typical NWP model grid cell sizes [e.g., Winstral et al , ]. Without using terrain parameters but also covering hilly areas Sweeney et al [] found reduced wind speed errors when applying adaptive statistical postprocessing methods on NWP output. This approach required training each day on data from previous days and was also only tested for stations around Ireland.…”

Section: Introductionmentioning

confidence: 99%

Parameterizing surface wind speed over complex topography

Helbig¹,

Mott²,

Herwijnen³

et al. 2017

JGR Atmospheres

View full text Add to dashboard Cite

Subgrid parameterizations are used in coarse‐scale meteorological and land surface models to account for the impact of unresolved topography on wind speed. While various parameterizations have been suggested, these were generally validated on a limited number of measurements in specific geographical areas. We used high‐resolution wind fields to investigate which terrain parameters most affect near‐surface wind speed over complex topography under neutral conditions. Wind fields were simulated using the Advanced Regional Prediction System (ARPS) on Gaussian random fields as model topographies to cover a wide range of terrain characteristics. We computed coarse‐scale wind speed, i.e., a spatial average over the large grid cell accounting for influence of unresolved topography, using a previously suggested subgrid parameterization for the sky view factor. We only require correlation length of subgrid topographic features and mean square slope in the coarse grid cell. Computed coarse‐scale wind speed compared well with domain‐averaged ARPS wind speed. To further statistically downscale coarse‐scale wind speed, we use local, fine‐scale topographic parameters, namely, the Laplacian of terrain elevations and mean square slope. Both parameters showed large correlations with fine‐scale ARPS wind speed. Comparing downscaled numerical weather prediction wind speed with measurements from a large number of stations throughout Switzerland resulted in overall improved correlations and distribution statistics. Since we used a large number of model topographies to derive the subgrid parameterization and the downscaling framework, both are not scale dependent nor bound to a specific geographic region. Both can readily be implemented since they are based on easy to derive terrain parameters.

show abstract

Section: Introductionmentioning

confidence: 99%

Parameterizing surface wind speed over complex topography

Helbig¹,

Mott²,

Herwijnen³

et al. 2017

JGR Atmospheres

View full text Add to dashboard Cite

show abstract

“…In contrast to the binning approach employed in Sweeney et al (), forecasted wind direction is here implemented in Eq. as Fourier expansion terms

\begin{matrix} alignleft \\ align-1 f (w_{d}) = \sum_{k = 1}^{N} a_{k} \cos (k w_{d}) + b_{k} \sin (k w_{d}), w_{d} \in (- π, π], & align-2 \end{matrix}

up to fifth order ( N = 5), to approximate wind speed dependence on a periodic function of forecasted wind direction w d .…”

Section: Methodsmentioning

confidence: 99%

Probing NWP model deficiencies by statistical postprocessing

Rosgaard

Nielsen²,

Nielsen³

et al. 2016

Quart J Royal Meteoro Soc

View full text Add to dashboard Cite

The objective in this article is twofold. On one hand, a Model Output Statistics (MOS) framework for improved wind speed forecast accuracy is described and evaluated. On the other hand, the approach explored identifies unintuitive explanatory value from a diagnostic variable in an operational numerical weather prediction (NWP) model generating global weather forecasts four times daily, with numerous users worldwide. The analysis is based on two years of hourly wind speed time series measured at three locations; offshore, in coastal and flat terrain, and inland in complex topography, respectively. Based on the statistical model candidates inferred from the data, the lifted index NWP model diagnostic is consistently found among the NWP model predictors of the best performing statistical models across sites.

show abstract

“…Wilks (2002) describes how the wind speed data could be transformed to a normal distribution by taking the square root of the distribution. Thorarinsdottir and Gneiting (2010) developed the EMOS approach for wind speed using a truncated normal distribution with a cut-off at zero to represent the data.…”

Section: Previous Studiesmentioning

confidence: 99%

High resolution forecasting for wind energy applications using Bayesian model averaging

Courtney

Lynch²,

Sweeney³

2013

Tellus A: Dynamic Meteorology and Oceanography

Self Cite

View full text Add to dashboard Cite

Two methods of post-processing the uncalibrated wind speed forecasts from the European Centre for Medium-Range Weather Forecasts (ECMWF) ensemble prediction system (EPS) are presented here. Both methods involve statistically post-processing the EPS or a downscaled version of it with Bayesian model averaging (BMA). The first method applies BMA directly to the EPS data. The second method involves clustering the EPS to eight representative members (RMs) and downscaling the data through two limited area models at two resolutions. Four weighted ensemble mean forecasts are produced and used as input to the BMA method. Both methods are tested against 13 meteorological stations around Ireland with 1 yr of forecast/observation data. Results show calibration and accuracy improvements using both methods, with the best results stemming from Method 2, which has comparatively low mean absolute error and continuous ranked probability scores

show abstract

Reducing errors of wind speed forecasts by an optimal combination of post‐processing methods

Cited by 59 publications

References 22 publications

Parameterizing surface wind speed over complex topography

Parameterizing surface wind speed over complex topography

Probing NWP model deficiencies by statistical postprocessing

High resolution forecasting for wind energy applications using Bayesian model averaging

Contact Info

Product

Resources

About