Abstract
At three Canadian test locations during the cold seasons of 2006 and 2007/08, snowfall measurements are derived from changes in the total depth of snow on the ground using multiple Campbell Scientific, Inc., SR50 ultrasonic ranging sensors over very short (minute–hour) time scales. Data analysis reveals that, because of the interplay of numerous essential factors that influence snow cover levels, the measurements exhibit a strong dependence on the time interval between consecutive measurements used to generate the snowfall value. This finding brings into question the reasonable accuracy of snowfall measurements that are derived from the snow cover surface using automated methods over very short time scales. In this study, two mathematical methods are developed to assist in quantifying the magnitude of the snowfall measurement error. From time-series analysis, the suggested characteristics of the snowdrift signal in the snow depth time series is shown by using measurements taken by FlowCapt Snowdrift acoustic sensors. Furthermore, the use of three collocated SR50s shows that repeated snow depth measurements represent three pairwise essentially different time series. These results question the reasonable accuracy of snowfall measurements derived using only a single ultrasonic ranging sensor, especially in cases in which the snow cover is redistributed by the wind and in which snow depth spatial variability is prominent.
Corresponding author address: Alexandre P. Fischer, Observing Systems and Engineering, Weather and Environmental Monitoring, Environment Canada, 4905 Dufferin St., Toronto, ON M3H 5T4, Canada. Email: alexandre.fischer@ec.gc.ca
