Depth hoar

The Taiga snowpack covers cold forested regions in North America and Eurasia, as represented in detail in Sturm et al. It is typically 50 cm thick (Figure 1) and covers the ground from October-November to April. In mid-winter, it is composed of a thick basal layer formed of centimetric depth hoar crystals, that has a very low mechanical strength, and a density near 0.2 g.cm It is topped by a layer of faceted crystals 1 to 2 mm in size, that eventually transform into depth hoar."" Layers of decomposing crystals and of fresh snow are observed after snow falls. All of these snow layers have a low density, typically < 0.2 g.cm as shown in Figure 1. 

The Alpine and maritime snowpacks have many similarities. A specific figure is not necessary here and can be found in Sturm et al, " Alpine snowpacks form in regions slightly colder than the maritime snowpack and signs of melting are less frequent. The temperature gradient in late fall can be sufficient to form a depth hoar layer 5 to 20 cm... 

The impact of the temperature gradient on metamorphism explains many of the features of Figure 1. The typical HGM-type metamorphism of the taiga snowpack eventually transforms most of the snowpack into depth hoar, " while the QIM-type metamorphism of the maritime and Alpine snowpacks forms, in the absence of melting, layers of small rounded grains 0.2 to 0.4 mm in diameter. However, considering the effects of other climate variables such as wind speed is necessary to explain features such as the presence of windpacks formed of small rounded grains in the tundra snowpack. 

In dense windpacks, crystals do not appear to have the space required to grow to large sizes and this has been invoked to explain why depth hoar does not grow in dense snow." However, we found depth hoar of density up to 0.4 g.cm on the tundra, so that space is probably not the only factor. As detailed below, windpacks have a high heat conductivity that hinders the establishment of a high temperature gradient across them and this may also explain why little transformation is observed in such snow layers. 

In the absence of wind, snow density increases because of loading by subsequent layers and of destruction of small structures by sublimation, leading to the collapse of crystals higher up. Without wind and melting and under QTM conditions, snow density is around 0.35 g.cm at 1 m depth. Under HGM conditions, compaction is compensated by upward vapor fluxes,limiting the density of basal depth hoar layers to about 0.2 g.cm. ... 

Most snow SSA values were obtained by measuring the adsorption isotherm of methane on snow at 77 For dry snow, values range from around 1500 cm .g for fresh dendritic snow to about 100 cm. g for depth hoar. Melting severely decreases SSA and values of 18 cm. g" have been measured for melt-freeze crusts." ... 

Figure 3 Measured snow SSA profiles in central Alaska in two different snowpacks a natural snowpack on the ground, where a strong temperature gradient led to depth hoar formation (HGM snowpack) and the same snowpack on Tables, under which the air circulation prevented the establishment of a significant temperature gradient (QIM snowpack). (a) comparison of the QIM snowpack of 16 February with the HGM snowpack sampled 8 days before and 10 days after (b) comparison of both snowpacks sampled just 4 days apart. In both cases the SSA is much higher in the QIM snowpack. Typical snowpacks heights were 50 cm for HGM and 40 cm for QIM.

Climate change will modify kr values in complex ways. Warming will limit depth hoar formation, increasing kj More frequent melting events in temperate climates will form ice layers with high kr values. In contrast, the growth of shrubs on the tundra will limit the effect of wind, transforming windpacks into depth hoar of much lower values. 

Besides temperature (Figure 5a), the cold and warm scenarios differ by the structure of the snowpack. In both cases, the snow water equivalent have the same temporal evolution 2 cm at the end of October, 11 cm at the end of January and 15.7 cm in late April. Stratigraphies and heat conductivities are very different. In the cold scenario, depth hoar layers of low densities (0.21 to 0.26 g.cm O alternate with denser windpacks (0.38 to 0.48). Transient layers of fresh snow and of faceted crystals are also present. values range from 0.06 W.m K for aged depth hoar to 0.46 W.m K for dense windpacks. In the warm scenario, two melt-freeze layers (densities 0.40 to 0.55) alternate with hard windpacks (0.34 to 0.41) while layers of fresh snow are sometimes included in the mean monthly stratigraphies, kr values are 0.45 and 0.63 W.m K for the melt-freeze layers and range from 0.36 to 0.48 W.m K for dense windpacks. Recent snow has values around 0.2 W.m lC Overall, the warm snowpack has a greater heat conductivity than the cold one. 

In the natural snowpack on the ground, evolving under HGM conditions, depth hoar of density 0.20 g.cm formed rapidly and the permeability in the lower half of the snowpack increased to beyond 500x10 ° m in late March. In contrast, on the tables under QIM conditions, fine-grained snow of density 0.28 g.cm formed and the permeability decreased to values between 30 and 70x10 m in early March. 

If the southern tundra turns into shrub tundra or taiga the transformation of the tundra windpack into faceted crystals or depth hoar of lower SSA, longer e-folding depth and higher permeability will lead to enhanced release of adsorbed and dissolved species, greater photochemical activity (modulated by tree shading) and more efficient release to the atmosphere, so that emissions may in this case increase. 

Crystal shapes of each layer were observed and classified into three types the initial crystal type (lightly compacted snow (Class 2dc)), solid type depth hoar (Class 3mx, 4a) and skeleton type depth hoar (Class 5cp), Snow crystal photomicrographs of each layer were taken and then the projected area of the crystals were obtained. 

The water vapor flux seems to affect significantly the growth rate of depth hoar snow. It seemed that water vapor transportation also caused a density change in snow. The initial and final density of each layer is illustrated in Figure 6. Before and after the experiment, the lost mass was only 0.5 %. The density of layer I (warmest) decreased. 

The dimensionless snow-pore space partition coefficient Ksa is notably lower in the sub-Arctic snow pack largely because of the influence of temperature on ATja (see Equation 18.2). The bulk of the snow packin sub-Arctic regions consists of depth hoar that exhibits both low density and SSA, which further contributes to this discrepancy. The differences in partitioning of PCB-28 within the sub-Arctic and Arctic snow pack cause significantly different concentration gradients between snow cover and lower atmosphere and thus, different exchange behavior. While PCB-28 is released from the sub-Arctic snow pack, the comparably low pore space concentration in the Arctic snow leads to a net deposition of this chemical (Table 18.1). 

---