Climatology of snowfall and related meteorological variables with application to roof snow load specifications

Abstract

Representative statistical models of snowfall and related meteorological variables, important in snow hydrology research, are also required for many engineering design problems. Design roof snow loads, if based on the balance between deposition of snow by individual snowfalls and subsequent depletion by wind action and various thermodynamic mechanisms, in addition to aerodynamic characteristics of particular roof shapes, require the development of statistical models of various meteorological variables which influence the process.Mathematical models suitably describing the statistics of individual snowfall amounts, wind speed and directions, and air temperature are presented in this paper. These results are based on the analysis of daily meteorological data obtained at some 28 Canadian stations with typical record lengths of about 30 years. The suitability of the developed probability distribution of daily snowfall magnitudes is examined through comparisons of both derived parent and extreme value statistics with available full scale observations. The joint statistics of these meteorological variables are also examined and a methodology for approximately accounting for their interdependence is presented.The relevance of the overall winter climatology of an area to roof snow load formation is discussed. Results of a Monte Carlo simulation, which show the dependence of the snow load for a particular roof shape on the climatology of snowfall and related meteorological variables, are presented.