Heterogeneous Boundary Layers through the Diurnal Cycle: Evaluation of the WRF Wind Farm Parameterization using Scanning Lidar Observations and Wind Turbine Power Measurements during a Range of Stability Conditions

Abstract:

As wind energy deployment increases, questions arise regarding impacts on local climates and how these impacts evolve with the diurnal cycle of the boundary layer. Satellite observations suggest nocturnal increases of surface temperatures, and measurements of turbine wakes document stronger and more persistent reductions of wind speed and increases in turbulence downwind of turbines during stable conditions. Validations of mesoscale parameterizations of these effects have been constrained to idealized conditions defined by neutrally-stratified conditions and/or limited wind directions and wind speeds, or by comparison to idealized large-eddy simulations.

Synthesis of conventional meteorological measurements and unconventional measurements can offer unique insights for validating models over a large heterogeneous domain. The CWEX-13 field experiment provides an extensive dataset for such validation at spatial scales on the order of 10 km in a range of atmospheric stability and wind conditions. CWEX-13 took place within a 300 MW wind farm in central Iowa during summer 2013 and featured strong diurnal cycles. The wind turbines are sited irregularly, creating a heterogenous “canopy”. Three profiling lidars, numerous surface flux stations, and a scanning lidar sampled wakes from multiple turbines. Further, the wind farm owner/operator has provided access to turbine power production and wind speed measurement data for model validation, providing ~ 200 measurements of proxies that integrate the wind profile over the rotor disk, from 40 m to 120 m above the surface.

Building on previous work that identified optimal physics options, grid configurations, and boundary condition data sets by comparison to lidar wind profile measurements, we execute simulations with the WRF Wind Farm Parameterization for a ten-day period featuring moderate winds and strong diurnal cycles. We evaluate simulations with different modeling choices (e.g., vertical resolution, approaches to representing boundary-layer turbulence) by comparison to both the conventional meteorological measurements as well as the unconventional measurements from wind turbines. We find that model performance depends on these modeling choices. Recommendations for generalization of these results to other locations will be presented.