85th AMS Annual Meeting

Tuesday, 11 January 2005
Lidar validation of numerical simulations of a Lake Michigan land breeze
Gijs De Boer, University of Wisconsin, Madison, WI; and E. W. Eloranta and G. J. Tripoli
As a part of the Lake-Induced Convection Experiments (Lake-ICE; Kristovich, 2000), the University of Wisconsin Volume Imaging Lidar (UW-VIL) was stationed at Sheboygan, Wisconsin, during the winter of 1997-1998. On December 21, 1997, an intriguing shallow land breeze circulation was observed over Lake Michigan with the VIL.

To test the merits of nested small-scale simulation of large-scale induced events such as the land breeze, numerical simulations were run utilizing the University of Wisconsin Non-Hydrostatic Modeling System (UW-NMS; Tripoli, 1992). Grid nesting capabilities of the model were used to achieve 32-meter horizontal resolution in the inner domain, while the outer domain featured 60-kilometer resolution. This combination allowed both large and small scales to be simulated simultaneously, and allowed for interaction between the evolution of the large-scale and the form of the small-scale land breeze.

Validation of these simulations is difficult utilizing traditional point measurement methods because of the high temporal and spatial resolutions needed to resolve the simulated structures. To address this issue, the VIL data was utilized in order to test the accuracy of the numerical simulations. Parameters such as the horizontal extent of the land breeze, the vertical extent of the land breeze and land breeze head, wind speed and direction, and timing of the simulation were compared between the simulation and the lidar data through the utilization of several analysis methods. This comparison shows good agreement between the simulation and the observed structure, and reveals some weaknesses of the nesting technique when used to simulate very small-scale events. Additionally, comparison between the simulation and traditional point measurements shows the importance of validation of the simulation output using lidar data.

Supplementary URL: