Advanced meteorological modelling across scales

Marc Imberger*

*Corresponding author for this work

Research output: Book/ReportPh.D. thesisResearch

90 Downloads (Pure)

Abstract

Wind farms are affected by a variety of atmospheric flow phenomena across spatial and temporal scales ranging from the macroscale and mesoscale to the microscale. This PhD
project focuses on simulating a few of these phenomena using the atmospheric component of the Model For Prediction Across Scales (MPAS) to investigate its capabilities for wind energy applications. MPAS is a global model that utilizes a subset of the physics schemes from the limited-area Weather Research and Forecasting (WRF) model. It has an unstructured mesh discretization (Spherical Centroidal Voronoi Tessellations) that enables quasi-uniform and variable-resolution simulations with smooth transition zones, allowing for regional refinement of the model simulations. In this thesis, the strengths and challenges of a global modeling approach with regional refinement, using MPAS, are examined and compared with measurements and traditional modeling, using WRF. Three different application areas are explored. The first area focuses on difficulties related with simulating a mid-latitude cyclone at medium resolution, using both WRF and MPAS. The WRF simulations have challenges due to the lateral boundaries, but with different setups a reasonable forecast can be produced. MPAS is able to improve storm intensity compared to WRF, but deviations in the model forecast increased time shifts between simulated and observed local minimum sea level pressure. The second application focuses on an episode of open cellular convection over the North Sea, allowing the investigation of MPAS’ capabilities for simulating mesoscale wind variability. Given sufficiently high horizontal resolution and an adequate model initialization time, MPAS represents the increased spatial and temporal wind variability quite well, compared to point measurements and satellite derived observations. The third application area explores MPAS to study the impact of wind farms on the atmosphere. For this purpose, the ’Explicit Wake Parametrisation’ scheme, previously implemented in WRF, is implemented in MPAS. After verifying the implementation in MPAS, a short-term case study is simulated, for two wind farms in the southern North Sea. During the implementation, issues with model divergence due to pressure field dependent random number generation were found and fixed, but the additional presence and impact of the propagation of small scale numerical noise requires further investigations. Across all the studies, it was found that the lack of lateral boundary conditions in MPAS, which help constrain the model simulations in WRF, allows for more significant model deviations from the real atmospheric state in case studies. This requires further exploration of data assimilation techniques to prevent the simulation drift.
Original languageEnglish
Place of PublicationRoskilde, Denmark
PublisherDTU Wind Energy
Number of pages180
DOIs
Publication statusPublished - 2020
SeriesDTU Wind Energy PhD
Number0104(EN)

Fingerprint

Dive into the research topics of 'Advanced meteorological modelling across scales'. Together they form a unique fingerprint.

Cite this