Transferability of a Three-Dimensional Air Quality Model between Two Different Sites in Complex Terrain

Rolf Lange Lawrence Livermore National Laboratory, University of California, Livermore, California

Search for other papers by Rolf Lange in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

The three-dimensional, diagnostic, particle-in-cell transport and diffusion model MATHEW/ADPIC is used to test its transferability from one site in complex terrain to another with different characteristics, under stable nighttime drainage flow conditions. The two sites were subject to extensive drainage flow tracer experiments under the multilaboratory Atmospheric Studies in Complex Terrain (ASCOT) program: the first being a valley in the Geysers geothermal region of northern California, and the second a canyon in western Colorado. The domain in each case is approximately 10 × 10 km. The 1980 Geysers model evaluation is only quoted. The 1984 Brush Creek model evaluation is described in detail.

Results from comparing computed with measured concentrations from a variety of tracer releases indicate that 52% of the 4531 samples from five experiments in Brush Creek and 50% of the 831 samples from four experiments in the Geysers agreed within a factor of 5. When an angular 10° uncertainty, consistent with anemometer reliability limits in complex terrain, was allowed to be applied to the model results, model performance improved such that 78% of samples compared within a factor of 5 for Brush Creek and 77% for the Geysers. Looking at the range of other factors of concentration ratios, results indicate that the model is satisfactorily transferable without tuning it to a specific site.

Abstract

The three-dimensional, diagnostic, particle-in-cell transport and diffusion model MATHEW/ADPIC is used to test its transferability from one site in complex terrain to another with different characteristics, under stable nighttime drainage flow conditions. The two sites were subject to extensive drainage flow tracer experiments under the multilaboratory Atmospheric Studies in Complex Terrain (ASCOT) program: the first being a valley in the Geysers geothermal region of northern California, and the second a canyon in western Colorado. The domain in each case is approximately 10 × 10 km. The 1980 Geysers model evaluation is only quoted. The 1984 Brush Creek model evaluation is described in detail.

Results from comparing computed with measured concentrations from a variety of tracer releases indicate that 52% of the 4531 samples from five experiments in Brush Creek and 50% of the 831 samples from four experiments in the Geysers agreed within a factor of 5. When an angular 10° uncertainty, consistent with anemometer reliability limits in complex terrain, was allowed to be applied to the model results, model performance improved such that 78% of samples compared within a factor of 5 for Brush Creek and 77% for the Geysers. Looking at the range of other factors of concentration ratios, results indicate that the model is satisfactorily transferable without tuning it to a specific site.

Save