Measuring the impact of additional instrumentation on the skill of numerical weather prediction models at forecasting wind ramp events during the first Wind Forecast Improvement Project (WFIP)

Abstract

The first Wind Forecast Improvement Project (WFIP) was a DOE and NOAA‐funded 2‐year‐long observational, data assimilation, and modeling study with a 1‐year‐long field campaign aimed at demonstrating improvements in the accuracy of wind forecasts generated by the assimilation of additional observations for wind energy applications. In this paper, we present the results of applying a Ramp Tool and Metric (RT&M), developed during WFIP, to measure the skill of the 13‐km grid spacing National Oceanic and Atmospheric Administration/Earth System Research Laboratory (NOAA/ESRL) Rapid Refresh (RAP) model at forecasting wind ramp events. To measure the impact on model skill generated by the additional observations, controlled data‐denial RAP simulations were run for six separate 7 to 12‐day periods (for a total of 55 days) over different seasons. The RT&M identifies ramp events in the time series of observed and forecast power, matches in time each forecast ramp event with the most appropriate observed ramp event, and computes the skill score of the forecast model penalizing both timing and amplitude errors. Because no unique definition of a ramp event exists (in terms of a single threshold of change in power over a single time duration), the RT&M computes integrated skill over a range of power change (Δp) and time period (Δt) values. A statistically significant improvement of the ramp event forecast skill is found through the assimilation of the special WFIP data in two different study areas, and variations in model skill between up‐ramp versus down‐ramp events are found.

The Hydrometeorology Testbed–West Legacy Observing Network: Supporting Research to Applications for Atmospheric Rivers and Beyond

“…These observations support studies of model verification and validation, the impact of observations on model performance, and parameterization, which can ultimately lead to improvements in models and services. Some examples of current research in these areas include studies of the impact of instrumentation or data assimilation on model skill, as in the impact of observations on the skill of numerical weather prediction models at forecasting wind events and wind speeds [63,64]; model verification/validation, including parameterization and algorithms [65][66][67][68]; how model improvements impact skill of wind forecasts [41,69]; and evaluation of microphysical algorithms in the Advanced Research Weather and Research Forecast Model [70]. Observations have also been used in assessing model skill in forecasting ARs during CalWater-2015 [51], and in wind speeds for wind turbines [64].…”

“…Observations from the Washington and Oregon AROs have been used to study the effects of marine-layer-depth behavior on sea-breeze strength and inland penetration for wind-farm operations [41]; to improve the understanding of processes that influence moisture fluxes and precipitation in and through Oregon's Columbia River Gorge ( [13,26,73,74]; to understand mountain waves [75], understanding cold pool events [68,76] and model development for wind forecasting [66,69]. The first WFIP used 915-MHz radars in its suite of instrumentation (Table 2), and found that the additional observations had a positive impact on numerical weather prediction skill [63].…”

Utilizing physics-based input features within a machine learning model to predict wind speed forecasting error

“…), requiring bespoke forecasts for accurate predictions. Numerical weather prediction models (NWPs) fail at such complex sites due to a lack of appropriate parameterization schemes suitable for local conditions (Akish et al, 2019;Bianco et al, 2019;Olson et al, 2019;Stiperski et al, 2019;Bodini et al, 2020). Therefore, statistical models and computational learning systems (such as an artificial neural network or random forest) are likely better suited to provide accurate power forecasts.…”

Decreasing wind speed extrapolation error via domain-specific feature extraction and selection

“…This would likely be particularly difficult to accomplish by using numerical models. In complex terrain, these models' parameterization schemes must be tuned for each individual region in order to obtain optimal results (Stiperski et al, 2019;Olson et al, 2019;Akish et al, 2019), reducing the plug-and-play effectiveness of numerical modeling. Model results must also be scaled down to the desired finer resolutions, which can result D. Vassallo et al: Decreasing wind speed extrapolation error via domain-specific feature extraction and selection in representativity error (Dupré et al, 2020).…”