Measurement of snowfall by radar.

Two aspects of one major snowstorm were analysed using 3-cm radar observations taken at a number of heights: (i) the correlation of radar returns taken at low heights with snowfall measurements at the ground, (ii) the variation of radar returns with height, and with time. For 20 grid points within 42-miles range, the ratio of amounts measured by radar at 5000 ft to amounts measured at the ground had a log normal distribution, the standard deviation corresponding to ratio 1.3 and its reciprocal. It was found that useful measurements of snowfall amount could be derived from radar observations out to 100-miles range, and that the variation of snowfall rate at a point could be determined from observations taken earlier along the trajectory of the snow falling toward that point. Radar data from greater heights showed that more than 95% of the snow that fell within 100-miles range had been generated below 20,000 ft, and that the growth of snow with distance fallen varied in space and time. Thus in regions north of the radar hall the growth was at heights between 10000 ft and 5000 ft; in regions south of the radar half the growth was between 15000 and 10000 feet