Evaluation of methods for preparing wind roses and 3-way joint frequency distribution data summaries

摘要

One class of input data necessary to run long-term Gaussian air quality (health risk assessment) models is 3-way joint frequency distribution data for wind speed, wind direction, and atmospheric stability. Currently, 3-way joint frequency distributions are generated from the hourly meteorological data by using the STAR (STability ARray) utility program. The STAR program uses a wind sector allocation routine to convert 36-category wind direction data to 16-category data. This routine distorts the frequency distribution of wind directions and thereby the magnitude of the directional average wind speed as well. When such distorted wind data are used in long-term air quality modeling calculations, ambient air quality impacts and health risks are over- or underestimated, depending on wind directions. This note was prepared (1) to discuss the nature and extent of the distortion in wind direction pattern and air quality impact estimates that results from using the STAR utility program, and (2) to evaluate alternative methods for eliminating such distortion and its consequences.