92nd American Meteorological Society Annual Meeting (January 22-26, 2012)

Wednesday, 25 January 2012
A Comparison of WRF Simulated Low-Level Shear to Observations in the 10-80 m Layer
Hall E (New Orleans Convention Center )
William A. Gallus Jr., Iowa State University, Ames, IA; and A. Robertson, R. Showers, and A. J. Deppe

Accurate forecasts of winds are needed to optimize wind power generation, and as wind turbines increase in height with longer blades, more attention may need to be paid to lower tropospheric wind shear. Simulations of 32 different cases covering all four seasons were performed with the Weather Research and Forecasting (WRF) model and compared to wind observations from an 80 m tall meteorological tower at the Pomeroy wind farm in northwestern Iowa. A detailed analysis of diurnal and monthly trends was performed on the observed shear in several layers within the lower troposphere, including 10-50 m, 50-80 m, and 10-80 m. Wind shear obtained from six different WRF runs for each case, each using a different Planetary Boundary Layer (PBL) scheme, was then compared to observations in these three layers. The comparison showed that despite some systematic errors in wind speed, forecasts of wind shear had relatively small errors at most times when most PBL schemes were used. As would be expected due to the decoupling of the boundary layer at night and the frequent occurrence of nocturnal low-level jets in this region, shear was strongest at night and weakest during the daytime, and the various model runs depicted these diurnal trends fairly well. However, the Yonsei State University (YSU) PBL scheme behaved noticeably differently, having a much weaker diurnal signal. In addition, the YSU scheme was much more likely to depict weaker shear than observed, usually at night, while the other five schemes almost always overestimated shear. Peak observed shear in the 10-80 m layer in most months averaged no more than 5 m/s of speed difference, while some model runs showed values around 6 m/s. At some times, the YSU underestimates were smaller than the overestimates of other schemes, but at other times, the YSU scheme had worse absolute errors. In addition to the comparison of WRF simulations with observations, wind shear in the WRF runs was also computed in the 80-120 m layer as a tool to estimate how much additional power might be generated if standard hub heights increased from 80 to 120 m.

Supplementary URL: