The effective thermal conductivity of snow (keff), which includes latent heat transfer due to vapor diffusion, was measured during three winters in Fairbanks, Alaska. In 1986-1987, keff of several layers of snow was monitored in detail as the snow metamorphosed into depth hoar. Measurements were made using a needle probe with an estimated accuracy of +8%; keff was found to decrease and then increase as the snow passed from new snow through several distinct stages of depth hoar. For depth hoar, keff ranged from 0.026 to 0.105, with an average value of 0.063 W m -1 K -1 . This is one half to one fourth the value suggested by most studies for snow of similar density. For depth hoar of a given type, keff can be represented as a linear function of temperature between 0 ø and -20øC but requires a nonlinear function for the range from 0 ø to -196øC. At -196øC the thermal conductivity of depth hoar approached that of still air, suggesting that conduction through the ice skeleton of the snow was limited and that the increase in keff at temperatures near 0øC is the result of the strong temperature dependence of water vapor density. This conclusion is consistent with the nature of the ice bonds in depth hoar, which are thin and relatively few in number. INTRODUCTIONThermal conductivity is a critical parameter that must be known in order to make calculations of heat transfer through the snow cover. Such calculations are important in modeling the energy exchange in snow-covered areas, modeling metamorphism of snow grains, and in determining the dominant processes that move heat and mass through the snow cover.Unfortunately, published values of the thermal conductivity of depth hoar are rare. Yet subarctic snow, consisting primarily of depth hoar, can be found in a circumpolar band that covers thousands of square kilometers [Pruitt, 1970[Pruitt, , 1984Benson, 1982]. Because these subarctic regions are often underlain by permafrost, knowledge of the thermal conductivity of the snow cover is of great importance in determining the depth and rate of frost penetration into the ground. Also, subnivean animals rely on the high insulation value of depth hoar to protect them from low temperatures. Without the development of a thick depth hoar layer, the same snow thickness would provide insufficient insulation for their survival [Pruitt, 1984;Adams, 1981].Functional relationships between conductivity and snow density are widely used to estimate the conductivity of a particular snow cover [e.g., Mellor, 1977]. With the exception of the work of Izumi and Huzioka [1975] and Lange [1985], these relationships have been derived from measurement series that did not include depth hoar. Consequently, they are not applicable to the subarctic snow. Even without considering depth hoar, values derived from the relationships show substantial scatter at a given density (Figure 1). The scatter indicates that density is not the only snow property affecting thermal conductivity.Texture, which includes the size, shape, bonding, and spatial arrangement o...

No abstract

The purpose of this study was to determine if air convects in a natural snow cover. To detect convection, the temperature field in the subarctic snow cover in Fairbanks, Alaska, was measured hourly during three winters (1984)(1985)(1986)(1987) using an array of thermistors which were suspended on threads and allowed to be buried by snowfall. The results indicate that convection occurred sporadically in 1984-1985 and almost continuously in 1985-1986 and 1986-1987. The evidence was (1) simultaneous warming and cooling at different locations in a horizontal plane in the snow, and (2) horizontal temperature gradients of up to 16øC m -1 During the winter, warm and cold zones developed in the snow and remained relatively fixed in space. We interpret these zones to be the result of a diffuse plumelike convection pattern linked to spatial variations in the temperature of the snow-soil interface. Air flow was inferred to have been primarily horizontal near the base of the snow and vertical elsewhere, Calculated flow speeds were of the order of 0.2 mm s -l, with a maximum value of 2 mm s-1. The convective circulation was time-dependent, with perturbations such as high wind or rapid changes in air temperature triggering periods when horizontal temperature gradients were strongest, suggesting that these were also periods when the air flow was fastest. The coincidence of depth hoar crystals with horizontal c axes and the horizontal flow lines at the base of the snow suggests that convection may have affected crystal growth directions.

ABSTRACT:A micropenetrometer has been developed that produces snow grain bond ruptures at the microstructural level and provides a unique signal for different snow types . A micromechanical theory of penetration has been developed and used to recover microstructural and micromechanical parameters for different snow types from the penetration force-distance signal. These parameters are the microstructural element dimension, the mean grain size, the critical microstructural deflection at rupture and the microstructural coefficient of elastic restitution. Additional derived mechanical properties include the compression strength and elastic modulus of microstructural elements and continuum scale volumes of snow. Analysis of the forcedistance signal from a Monte Carlo simulation of micropenetration indicates that microstructural and micromechanical parameters may be recovered with an accuracy of better than 5% when spatial and force resolutions are high and the penetrometer tip area is of similar size to the structure dimension.

A new constant-speed penetrometer for field and laboratory measurements has been developed. The initially independent work of SFISAR and CRREL has been brought together, and a portable field device is now in an advanced stage of testing.The new penetrometer has high rigidity and a high-resolution large dynamic range force sensor. It uses a much smaller sensing head (5 mm ) than previous designs and has a constant.-speed drive. ''''ith th~s construction, the penetration rrsistance of ~'ery fine layers and th~ mf1uenc~ of the bond~ng strength between snow grains can be more accurately determmed than tS possible With the rammsonde or Pandalp. Artificial foam layers as thin as 2 mm and thin layers in snow have been detected by the penetrometer. Thin snow layers detecte? from penetration~resistance profiles have been correlated to finr layering as determmed from plane-sectIOn mlcrophotographs of samples taken adjacent to the profile. The instrument's measurements are highly repeatable and the lack of subjective decisions when operating the penetrometer makes the penetration resistance a quantitative measure of snow stratigraphy.