WMOD00562 - Laserfiche WebLink

Home Browse Search

computerized search of the scientific literature revealed that the large majority of articles dealing with , snow density" are concerned with relatively long time scales such as days, weeks, or even months. For xample, both hydrologic and avalanche forecasting are concerned with long-term snow density changes. nly a limited number of references were found that addressed density of freshly-fallen snow, and these r ferences generally used daily observations. For the purpose of the SAA, we are concemedwith snow ensity on time scales from 1 hour to the duration of a single storm, usually no more than a day or two. e most practical approach for real-time SAA estimation of SD appears to be to divide the radar's hourly WE estimate for any range bin by the median density for that locale. We recommended this approach r last winter's real-time tests in Cleveland and Minneapolis. he SAA provides SWE estimates for 1-h, 3-h, and storm total periods, all updated each volume scan. is approach is appropriate for SWE, which, after reaching the ground, does not change significantly vel' a storm's duration unless snow melt occurs. It will be shown that snow density changes over f latively short time periods. The density of falling snow is affected by changes in cloud microphysics, hich typically produce varying ice crystal types and sizes and varying degrees of riming and aggregation ver the course of a storm. now density can change after the snow reaches the ground because of compaction over time caused by number of factors. Snow compaction is apparently primarily related to temperature and vapor variations nd wind effects as discussed by Grant and Rhea (1974). They cite evidence from LaChapelle (1962) i dicating the weight of the snow itself has a major influence on compaction. However, LaChapelle r ported a 28-inch mountain snowfall with a resulting low density of 0.05 g cm-3, indicating surprising s ctural strength. So the crystal structure of the snow partially controls its rate of settling (compaction) fter it reaches the surface. However, attempting to predict snow density changes with time after the snow falls is beyond the scope of the SAA development. Furthermore, the SAA does not have the omplexity to attempt to deal with snow melt on the surface. now density will vary markedly in time and space, and the degree of windiness will certainly affect not nly snow density but SD and SWE directly as snow is scoured and drifted. The attempt here will be to stimate the depth of snow fallen over the past hour to past day in locations protected from the wind. The AA cannot be expected to accurately predict SWE, much less SD, in locations with significant snow r distribution caused by the wind. Such changes are caused by small and medium scale variations in t pography, plant cover including forests, man-made structures, and other factors. These factors are the ame ones that make accurate and representative snowfall measurement so difficult. ven if the SAA accurately estimated hourly SD in a protected location, it should be recognized that ompaction could occur over the course of an entire storm period. Moreover, the rate of compaction may e quite variable among a population of storms. Therefore, even accurate hourly SD predictions could suIt in an unknown degree of SD overprediction over a storm's duration. .2 A verage Values for Snow Density he "ten to one rule" is often cited, meaning that, on average, the SD is ten times the SWE (i.e., snow ensity is 0.1). Potter (1965) notes that this rule has long been used as it is described in an 1878 'Instructions to Observers" in Canada. Potter states that its use in Canada was based on a long series of xperiments in Toronto. However, Potter compared its use with 1963-64 measurements of both SWE d SD across Canada. He shows a map of Canada with the ratio of measured SWE (presumably daily) t that estimated from the "ten to one rule." Most of Canada had ratios smaller than 0.9, corresponding 20