SURVEY OF STATISTICAL MODELS FOR OXIDANT AIR QUALITY PREDICTION

The statistical-empirical approach to oxidant modeling is reviewed with particular reference to oxidant prediction. Such modeling is currently proceeding in 2 directions. The first, short term oxidant forecasting over the range of hours to days, has the goal of episode control by short-term emission control as well as the issuance of health and perhaps agricultural warnings. The major tools as Box-Jenkins time series analysis, multiple regression, and aspects of pattern recognition. The second direction is long term prediction over the scale of years of the expected changes in oxidant levels due to changes in total reactive hydrocarbons (RHC) and nitrogen oxides (NOX) emissions, either aggregated for an air basin or more rarely spatially resolved. The goal of such prediction is the assessment of probable effects on oxidant air quality of various emission control strategies, regional planning, transportation planning, and environmental impact reports. The major categories of models are rollback, prediction from HC, prediction from NOX, and spatial resolution. The advantages of using statistical-empirical models include (a) their close relation to the actual air measurement data from which they are derived, allowing the possibility of correct prediction even in the absence of understanding of the underlying phenomena, and (b) their simplicity and low cost of development and use. Disadvantages and dangers in the use of this modeling techniques and other techniques (smog chamber modeling, mechanistic modeling) are noted.

Media Info

  • Media Type: Print
  • Features: Figures; References;
  • Pagination: pp 46-62
  • Monograph Title: Assessing transportation-related air quality impacts
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00139651
  • Record Type: Publication
  • Files: TRIS, TRB, ATRI
  • Created Date: Sep 21 1976 12:00AM