Ground Icing Formation: Experimental and Statistical Analyses of the Overflow Process

Abstract Icings are common hydrological phenomena in sub‐Arctic regions. These formations are supported by ground water seepage (ground icings), perennial springs (spring icings) and stream discharge (river icings). Although water is continuously supplied during the growth period, variations in icing topography cause the flooding areas to constantly shift, resulting in discontinuous growth at any specific location. In order to investigate the stochastic properties of this process, a field experiment was conducted under natural climatic conditions in which a series of overflow events was simulated. Autocorrelation analysis of both the time and space series shows that: (1) over time, there is no significant relationship between icing accumulation of current time and an earlier time step (20 minutes) for a single location; and (2) over the icing surface, the simultaneous accumulation can only occur over a very limited distance (60 cm). Using multiple regression analysis, it is further shown that the mean spreading length of icing layers during an overflow event is mainly controlled by five variables: discharge, water temperature, the product of mean air temperature and wind speed, air temperature, and slope of the icing surface prior to overflow. This analysis provides evidence that the maximum spreading length and the location of maximum icing deposition are determined by different sets of independent variables: the former being dependent on the first four of the above variables, and the latter being determined by time, water temperature, the product of air temperature and wind speed, and the slope of the icing surface. © 1997 John Wiley & Sons, Ltd.