Since the accumulation of snow due to wind action is a complex phenomenon, it is difficult to predict the probable accumulations in any particular set of circumstances except by the judicious use of small-scale model simulation techniques. These techniques have been used by Morrison Hershfield Limited on several hundred projects in Canada and the USA. The authors relate some of these experiences in this paper. They discuss both model test results and full-scale observations that throw light on the accuracy or shortcomings of the model simulations. Since a perfect simulation of the various facets of snow drifting is not possible with current techniques, it is of great importance in interpreting model results to draw on past full-scale experience. In some instances, the model tests are capable of giving valuable quantitative data. In other cases, they serve mainly as a diagnostic tool for identifying the locations, but not the exact size, of problematic drifts.
