1986 Volume 21 Issue 4 Pages 278-288
Optimum design and quantative evaluation of parameters of dispersion model are proposed by experimental regression analysis with orthogonal array table and Signal-Noise ratio (SN ratio). The advantages of this method are as follows.
1) The same effect as is obtained by calculations of all combinations of parameters can be obtained by much less calculations.
2) The effectiveness of each parameter can be evaluated quantatively by Analasis of Variance (ANOVA).
3) The precision of prediction by dispersion model can be evaluated quantatively by SN ratio.
The proposed method is applied to the dispersion experiment in Garfield area by using the prediction system of pollutant dispersion in complex terrain proposed by the authors and is found to be applicable.
