Toward Improved Off-Shore Wind Predictions by Combining Observations with Models through State Estimation – An Analysis of Marine Boundary Layer Parameterizations

Thursday, 18 December 2014: 9:45 AM
Branko Kosovic1, Luca Delle Monache1, Joshua Hacker1, Jared A Lee1, Francois C Vandenberghe1, Yonghui Wu1, Andrew Clifton2, Sam Hawkins3, Jesper Nissen3 and Dorita Rostkier-Edelstein4, (1)National Center for Atmospheric Research, Boulder, CO, United States, (2)National Renewable Energy Lab, Golden, CO, United States, (3)Vattenfall, Solna, Sweden, (4)Israel Institute for Biological Research, Dept. of Applied Mathematics, Environment Division, Ness Ziona, Israel
In recent years, significant advances have been achieved in model representation of atmospheric boundary layers (ABL). However, fundamental understanding of the processes governing the evolution of the Marine Boundary Layer (MBL) is still incomplete. We address this problem by combining available atmosphere and ocean observations with advanced coupled atmosphere-wave models, via state estimation (SE) methodologies. The goal is to improve wind prediction for off-shore wind energy applications through advances in understanding and parameterization of underlying physical processes, with an emphasis on the coupling between the atmosphere and the ocean via momentum and heat fluxes. We systematically investigate the errors in the treatment of the surface layer of the MBL in the Weather Research and Forecasting (WRF) model and identify structural model inadequacies associated with the MBL parameterization. For this purpose we are using both the single-column model (SCM) and three-dimensional (3D) versions of the WRF model, observations of MBL structure provided by offshore observational platform FINO1, and probabilistic SE. We have also developed an atmosphere-wave coupled modeling system by interfacing WRF with a wave model (WaveWatch III – WWIII). This modeling system is used for evaluating errors in the representation of wave-induced forcing on the energy balance at the interface between atmosphere and ocean. Probabilistic SE is based on the Data Assimilation Research Testbed (DART). DART is the framework for obtaining spatial and temporal statistics of wind-error evolution (and hence the surface-layer fluxes), along with objective tuning of model parameters.

We explore one of the potential sources of MBL model errors associated with roughness length parameterized using Charnock’s relation. Charnock’s roughness length parameterization assumes wind-driven waves are in equilibrium. However, it has been shown that swells propagating at different speeds and angles with respect to the wind direction can have a significant effect on winds in the lower boundary layer. Results over a domain centered at the FINO 1 tower covering northern Europe for the months of October, November, and December of 2006 will be presented.