The Scaled Wind Farm Technology (SWiFT) facility in
Lubbock, Texas, is hosted at the Texas Tech University (TTU) National Wind Institute
[Hirth and Schroeder, 2014] and operated by Sandia National Laboratories. This site was previously chosen as
an International Energy Agency Task 31 (“Wakebench”) benchmark for wind-turbine-wake evolution and dynamics. For MMC
research, this site is an attractive validation dataset because it has flat terrain with uniform
land cover. Essential validation data include a heavily instrumented 200-m meteorological mast with
10 measurements heights offering high-frequency wind measurements from 0.9 to 200 m above ground
level (AGL). These are complemented by a radar wind profiler with radio acoustic sounding system
(RASS) for measurements of wind and temperature profiles above 200 m, as well as data from the West
Texas Mesonet. Details of the site characterization are provided in Kelley and Ennis [2016].
Fig. 2 The SWiFT facility with adjacent TTU atmospheric facilities.
A classic diurnal cycle was identified between 8-9 November 2013 based on criteria outlined in
Haupt et al. [2017]. Conditions were generally clear and considered quiescent. This diurnal
cycle featured a morning transition, daytime convective boundary layer, evening transition, and
nocturnal low-level jet (LLJ). The convective, neutral, and stable periods were separately assessed
for simulations with MYNN and YSU PBL schemes. Despite having a warm bias of up to ~5 K,
WRF captures the correct trends in wind speed, direction, and virtual potential temperature
[Haupt et al., 2017].
A separate study into the effect of the terra incognita on modeling was conducted by
Rai et al. [2019]. This study considered 3 cloud free days at SWiFT: 13 July 2013, 22 September
2013, and 6 June 2014. Proper Orthogonal Decomposition analysis clearly indicates that the
microscale contains energetic modes that originated from the mesoscale flow. Depending on horizontal
grid spacing and turbulence modeling choices, flow from the terra incognita may downscale into
unrealistic flow in the microscale.
Relevance to Wind Energy
Nonstationary conditions result in time-varying hub-height wind speed and direction, wind shear
and veer, and turbulence intensity.
Accurate downscaling of energy from the microscale is important for predicting realistic
turbulent flow features in the wind-farm operating environment.
MMC Techniques Demonstrated
Ensemble mesoscale modeling and assessing best performers + model sensitivity
Online (WRF/WRF-LES) and offline (WRF/SOWFA) coupling between NWP models and microscale LES
Internal coupling with two methods: the profile assimilation and mesoscale budget components
approaches
The WRF simulation setup contains 3 nested domains with 27 km, 9 km, and 3 km grid spacing, centered
at the SWiFT site (Fig. 3). Simulations were started at 00:00 UTC on 8
November 2013. Initial and boundary conditions were set by the final analysis data from the Global
Forecast System. Turbulence modeling was provided by the MYNN Level 2.5 PBL scheme. Other physics
parameterization are detailed in the WRF namelist.
Fig. 3 WRF domain configuration for SWiFT mesoscale simulation.
200-m tower [Ennis, 2014]: The tower data analysis
includes data standardization and sonic tilt correction, and calculates turbulence quantities of
interest. Additional calculations
include virtual potential temperature, stability (bulk Richardson number and Obukhov stability
parameter), and the surface heat flux. A more indepth analysis of the TKE
shows anomalous observations around 10:00 UTC on 9 November 2013.
The mean quantities are combined with the radar dataset (described next) to form the input data
for the Allaerts et al. [2022] study. The turbulence quantities are used to validate the LES
predicted turbulence in Allaerts et al. [2020], Allaerts et al. [2022], Draxl et al. [2021].
Radar wind profiler with RASS [Ennis, 2014]: The radar data analysis shows
two wind scanning modes, short range (up to 2 km AGL) and long range (up to 6 km) and
corresponding temperature profiles up to a maximum of 800 m AGL. The same notebook also performs
the data reconstruction to create the mesoscale forcing dataset for the Allaerts et al. [2022]
study.
Results from the development and validation of the profile assimilation technique
[Allaerts et al., 2020], which couples the WRF mesoscale NWP model to SOWFA LES, are postprocessed
in the
studies/SWiFT/profile_assimilation_wrf/produce_figures.ipynb
notebook. This study demonstrated that simple data assimilation techniques (i.e., direct profile
assimilation) can lead to nonphysical shear and turbulence production, due to the algorithm’s
inability to cope with inaccuracies in the mesoscale data. Applying mesoscale forcing with
vertical smoothing (i.e., indirect profile assimilation) improves predictions of turbulence
statistics (Fig. 4).
Fig. 4 Time history of turbulent kinetic energy at 80 m AGL for direct/indirect profile assimilation
and budget components coupling, in comparison with the TTU meteorological tower; vertical dashed
black lines indicate sunrise and sunset
Results from the evaluation of coupling the WRF mesoscale NWP model to SOWFA through mesoscale
budget components [Draxl et al., 2021] are postprocessed in the
studies/SWiFT/budget_components_coupling/plot_*
notebooks. This work shows that mesoscale models can have difficulties predicting profiles of
shear and veer. While LES can improve shear and veer predictions, the wind speed and direction are
not adjusted. When forcing the LES with the mesoscale budget, spatiotemporal averaging of the
forcing terms is not necessary.
D. J. N. Allaerts, Eliot Quon, Caroline Draxl, and Matthew Churchfield. Development of a Time–Height Profile Assimilation Technique for Large-Eddy Simulation. Boundary-Layer Meteorology, 2020. doi:10.1007/s10546-020-00538-5.
Dries Allaerts, Eliot Quon, and Matt Churchfield. Using observational mean-flow data to drive large-eddy simulations of a diurnal cycle at the SWiFT site. Wind Energy: Progress in Wind Energy Research from the NAWEA/WindTech 2022 Conference, 2022.
Caroline Draxl, Dries Allaerts, Eliot Quon, and Matt Churchfield. Coupling Mesoscale Budget Components to Large-Eddy Simulations for Wind-Energy Applications. Boundary-Layer Meteorology, 179(1):73–98, April 2021. doi:10.1007/s10546-020-00584-z.
Haupt2019
Sue Ellen Haupt, Branko Kosovic, William Shaw, Larry K. Berg, Matthew Churchfield, Joel Cline, Caroline Draxl, Brandon Ennis, Eunmo Koo, Rao Kotamarthi, Laura Mazzaro, Jeffrey Mirocha, Patrick Moriarty, Domingo Muñoz-Esparza, Eliot Quon, Raj K. Rai, Michael Robinson, and Gokhan Sever. On Bridging a Modeling Scale Gap: Mesoscale to Microscale Coupling for Wind Energy. Bulletin of the American Meteorological Society, August 2019. doi:10.1175/BAMS-D-18-0033.1.
Sue Ellen Haupt, Rao Kotamarthi, Yan Feng, Jeffrey D. Mirocha, Eunmo Koo, Rod Linn, Branko Kosovic, Barbara Brown, Amanda Anderson, Matthew J. Churchfield, Caroline Draxl, Eliot Quon, William Shaw, Larry Berg, Raj Rai, and Brandon L. Ennis. Second Year Report of the Atmosphere to Electrons Mesoscale to Microscale Coupling Project: Nonstationary Modeling Techniques and Assessment. Technical Report PNNL-26267, Pacific Northwest National Laboratory, February 2017.
Raj K. Rai, Larry K. Berg, Branko Kosović, Sue Ellen Haupt, Jeffrey D. Mirocha, Brandon L. Ennis, and Caroline Draxl. Evaluation of the Impact of Horizontal Grid Spacing in Terra Incognita on Coupled Mesoscale–Microscale Simulations Using the WRF Framework. Monthly Weather Review, 147(3):1007–1027, March 2019. doi:10.1175/MWR-D-18-0282.1.
Brandon Ennis. Radar - TTU Radar - Raw Data. Maintained by A2e Data Archive and Portal for U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, 2014. doi:10.21947/1329730.
Brandon Ennis. Tower - TTU (SWiFT) Tower, All levels - Raw Data. Maintained by A2e Data Archive and Portal for U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, 2014. doi:10.21947/1329252.
B Hirth and J Schroeder. A summary of the National Wind Institute Meteorological Measurement facilities at the Texas Tech University's Reese Technology Center Field Site. Technical Report, Texas Tech University, Lubbock, Texas, 2014.
Christopher Lee Kelley and Brandon Lee Ennis. Swift Site Atmospheric Characterization. Technical Report SAND–2016-0216, Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States), January 2016.
