An equation for predicting l10 levels in urban streets has been derived from an analysis of measurements made at 190 sites in Edinburgh and elsewhere. The validity of the equation has been confirmed by an analysis of 134 measurements made in Sheffield and Rotherham. This equation included traffic variables which are at present difficult to predict, and an alternative equation that uses easily predictable variables has been derived from the Sheffield and Rotherham data. The standard errors of the two prediction methods were almost 3 db(a). It has been shown that this error is not significantly affected by the instrumentation deployed, provided that it reaches industrial grade standards, or by the sampling period of the measured levels, provided this is not less than 15 minutes. Over the range of data studied the standard error was not significantly affected by non-linearities or discontinuities in the 22 numerical variables considered in the regression analysis nor did it vary systematically with any of 11 descriptive variables that were examined. There was some indication, however, that the regression analysis technique was not detecting significant interactions between variables. When certain of the variables dismissed as non-significant were aggregated in a level of service term and the regression analysis repeated, a significant reduction in the prediction standard error was achieved.(a) /TRRL/

  • Availability:
  • Corporate Authors:

    Academic Press Incorporated

    Berkeley Square House, Berkeley Square
    London W1,   England 
  • Authors:
    • Gilbert, D
  • Publication Date: 1977-3-22


  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 00159794
  • Record Type: Publication
  • Source Agency: Transport and Road Research Laboratory (TRRL)
  • Report/Paper Numbers: Analytic
  • Files: ITRD, TRIS
  • Created Date: Apr 12 1978 12:00AM