C34A-01
Estimation of Winter Leaf Area Index and Sky View Fraction for Snow Modelling in Boreal Coniferous Forests
Leaf area index (LAI) and sky view fraction have an important role in controlling snow interception capacity, throughfall fraction, and canopy radiation transfer in snow energy balance models developed for forest conditions. In most models winter LAI and sky view fraction are provided by the user a priori as model parameters. Because stems, branches and needles are capable of intercepting water, snow and radiation, LAI needs to be defined in terms of the total plant area. Plant area indices (PAI) or winter 'effective' LAI values have been introduced in the snow models. Canopy parameters are typically adopted from literature, where the relationship between the winter LAI and sky view fraction is described with different mathematical functions. Winter LAI and sky view fraction can be measured using optical methods. Optical measurements include radiation transmission observations by plant canopy analysers or analysis of hemispheric photographs. Measurements of tree biomass provide another option for the determination of LAI. Computation of LAI from the biomass is a tempting option, because an estimate of stand biomass can easily be derived from operational forest inventory data. The first objective of this study was to estimate winter LAI and sky view fraction using different methods. Optical and biomass-based approximations of winter LAI and sky view fraction were available from coniferous forests in Scandinavia with different stand density and site latitude. The biomass-based estimate of LAI was found to be comparable with the values derived from the optical measurements in most sites. Heterogeneity of tree species and site fertility, as well as edge effects between different forest compartments caused differences in the LAI estimates in some sites. The second goal was to apply a snow energy balance model (SNOWPACK) to detect, how the differences in the estimated values of the winter LAI and sky view fraction are reflected in simulated snow processes. An increase in LAI and a decrease in the sky view fraction changed the snow surface energy balance by decreasing short-wave radiation input and increasing long wave radiation input. Changes in the studied canopy parameters had a direct impact on snow accumulation through altered throughfall fraction, and an indirect and less visible impact on snowmelt through the changed surface energy balance.
C34A-02 INVITED
Sensing dynamics of snow – vegetation interaction
Snow accumulation and melt governs the hydrology in many parts of the world. Vegetation can further alter these processes. While sub-disciplines in hydrology, ecology and meteorology have been researching the influence of vegetation cover and vegetation structure on (i) water storage as snow and (ii) water release during snowmelt for a long time, experimental techniques have been weak at addressing two important aspects of these two processes: their spatial variability and temporal dynamics. We have developed new sensors and low-cost sensor packages to observe the dynamics of snow processes in the vegetation cover at an adequate spatial scale and a number of sample locations. Complemented by a better description of the vegetation structure derived from LiDAR and other remote sensing methods, the collected data allows exploring and analyzing new aspects of the role of vegetation in the dynamics of the storage and release process. In this contribution we introduce experimental design and deployed sensors, present first results on the impact of vegetation cover and structure on snow processes at individual sites and outline the potential to address space-time dynamics over larger areas.
C34A-03 INVITED
The Effect of Forest Structure on Snow Hydrology in the Canadian Rocky Mountains
Mountain evergreen forests exert two primary effects on the formation and ablation of the seasonal snowcover. The first is the interception of snowfall in the forest canopy, resulting in sublimation, melt or unloading from the canopy. Snow interception can capture a very large percentage of the seasonal snowfall and in cold conditions this snow can be held for many days in the canopy. In the canopy snow is exposed to high rates of turbulent transfer and so can sublimate rapidly where unsaturated conditions prevail. Snow also metamorphoses in the canopy and some intercepted snow is unloaded to the surface, depending on the mechanics of slippage between branch and intercepted snow and the failure of intercepted snow clumps. The result of the snow interception, sublimation and unloading processes is that snow accumulation usually declines with increasing forest cover. This effect is most pronounced in cold, dry and windy conditions. The second primary effect of mountain forests is that snow on the ground is subjected to modified radiation regimes due to the extinction of shortwave radiation and emission of longwave radiation by the forest canopy. Some extinguished shortwave energy is also emitted as longwave energy. The result is to sometimes reduce and sometimes increase the incoming radiation available to the snowpack compared to open areas, the direction of change depending on cloudiness, solar angle, canopy structure, snow albedo, degree of slope and aspect. In cold continental mountain environments, net radiation available for snowmelt is generally reduced by increasing forest cover on level sites and south facing slopes but is enhanced by increasing forest cover on north facing slopes. The dramatic implications of the combined effects on snow hydrology of changing snow interception and radiation regimes as mountain forest cover is reduced from insect infestation, logging and burning are demonstrated and discussed.
C34A-04 INVITED
Experimental Increases in Snow Depth Alters the Seasonality, Structure and Function of Ecosystems in Alaska and Greenland
Establishing how changes in winter snow depth can alter the seasonality and magnitude of important feedback processes, linkages between trophic levels and connectivity between terrestrial and aquatic ecosystems is central to understanding tundra ecosystems of the future. For the past 15 years we have carried out an extensive set of Low and High Arctic snow depth manipulations, including a new set of experiments as part of IPY. We will discuss our most important findings, including a new framework for a collaborative Pan Arctic Snow Experiment Network (PASEN). Our studies indicate that: 1) modest increases in snow depth (2-3X) stimulate shrub expansion, while large increases (5-6X) in snow depth lead to permafrost thawing and the release of old carbon; 2) higher rates of winter N mineralization under deeper snow leads to higher leaf N in many plant species during the subsequent summer, a change that is associated with greater forage quality and higher rates of ecosystem carbon fixation. It is worth noting that deeper snow shifts the timing of maximum leaf N by up to 2 weeks, a change that may disrupt plant-herbivore relations; 3) deeper snow increases the concentration of DOC in soils which can lead to the transfer of carbon from terrestrial to aquatic systems, altering the biogeochemistry of streams and rivers. We believe that a continued emphasis on Arctic winter biogeochemical processes is needed throughout the North and is central to understanding the future Arctic system. This can be accomplished by establishing a PASEN, with a common set of monitoring and process-level studies.
C34A-05 INVITED
Using NIR Photography to Document Snow Stratigraphy Quickly: Lessons from Three Field Campaigns
We began using near-infrared (NIR) photography as a quick way to document snow stratigraphy in 2006 as part a snow validation campaign in Barrow, Alaska. Seventeen snow pits (30 to 80 cm deep) were photographed using a Sony DSC-P200 Cybershot 7.2-megapixel digital camera supported on a mini-tripod and equipped with a NIR filter (850 nm). Standard layer measurements of thickness, density, grain size, hardness, and grain type were also made in each pit. During the 2007 SnowSTAR traverse across Alaska and Canada, 43 snow pits were photographed and measured in the same fashion. During the CLPX-Alaska campaign of 2008, three trenches each about 10 m long were photographed in their entirety as well as documented in a traditional manner, this time using a Fuji S9100 9 mega-pixel digital camera with an 850 nm filter. For the trenches, the camera was supported on a sliding rail system. NIR photographs were processed using Image-J software and a simple algorithm that enhanced contrast based on grain size. Our goal is to develop a method of documenting stratigraphy that is faster than recording the results in a field book. For a 50 cm deep pit prior methods of recording stratigraphy would have required about 30 minutes. We succeeded in reducing the average time to acquire a pit photograph to less than 15 minutes. However, pit face preparation time increased by about 15 minutes because of the need to produce a smooth, divot-free snow surface. Required protocols to compute grain size from the photos added a further 20 minutes or more if used, so frequently these were omitted. While at present there is no real net reduction in the time to record stratigraphy using NIR photography vs. older methods, the result is superior in all ways to our best previous efforts to "map" the stratigraphy though hand-recorded data. A combination of older traditional methods and NIR photography is strongly recommended as the best method to document the snow stratigraphy.
C34A-06
Practice of near-infrared photography of snowpits
Documentation and quantification of snow pits using near-infrared sensitive photography is a cheap and
efficient technique (Matzl and Schneebeli, 2006). However, the quantitative processing of images from
conventional digital cameras is not without pitfalls. The camera must be calibrated for intensity variation
caused by the optic, which must done under homogenous illumination. In the field, a simple way was found to
setup diffuse illumination, to prepare the pit, to position the calibration targets and to take the flat field
reference image. The processing of the raw image to determine the absolute reflectivity requires several
steps. First, the green channel of the raw image is extracted and interpolated. The green channel of most
digital CCD has the highest number of pixels. Because the red-green-blue filters on the chip filter near-
infrared red differently, a single color channel image is less noisy than a composite image. This raw image is
then normalized by the optical correction image, and subsequently corrected for illumination heterogeneity by
the field flat field image. This image can now be referenced to absolute reflectivity using the calibration
targets. The calibrated image is used to segment quantitatively for optical grain diameter and specific surface
area. A more qualitative interpretation of the snow stratigraphy, using image classification algorithms, is also
possible. The equipment developed for near-infrared photography is transportable in a backpack and is used
in alpine terrain. Images from different field campaigns in the Alps show the wide range of features, which
are not easily documented using traditional stratigraphy.
Matzl, M.; Schneebeli, M., 2006: Measuring specific surface area of snow by near-infrared photography. J.
Glaciol. 52, 179: 558-564
http://www.wsl.ch/personal_homepages/schneebm/nip
C34A-07 INVITED
Observing Microscale Variations in Snow Stratigraphy Using Near Infra-red Photography
The physical characteristics of snowpacks in arctic tundra can be highly spatially heterogeneous. Consequently, understanding the small scale (sub-metre) variability can provide valuable information for evaluating remotely sensed images that integrate snowpack variability over much greater spatial scales. Here we outline a new method, developing on work by Matzl and Schneebeli (2006), to quickly obtain in-situ observations of centimetre resolution snowpack stratigraphy along 10 m trenches. Mie theory suggests that the reflectance of snow in the near infra-red (NIR) spectrum is largely controlled by grain size. This reflectance is not unduly influenced by impurities in the snow or variations in density, so long as densities are less than 650 kg/m3. Consequently, stratigraphic boundaries where grain size differs can be defined by the reflectance intensity in the NIR. NIR digital photography, therefore, provides the potential for rapid capture and quantitative analysis of spatial variations in stratigraphy along a trench face. NIR imagery is presented of snow trenches at Imnaviat Creek and Toolik Lake, Alaska, in February 2008 as part of the 2nd NASA Cold Land Processes Experiment. Issues involved in acquiring suitable images, stitching a series of adjacent images, picking stratigraphic layers, evaluating against manual stratigraphic observations and quantifying the uncertainties involved in each stage of the process are discussed.
C34A-08 INVITED
Three-Dimensional Reconstruction of Snow from Serial Surface Sections
The thermal, mechanical, and optical properties of snow depend directly on the configuration of its microstructure. Bulk properties, such as density, are related to the microstructure; but we cannot estimate the thermal conductivity, elastic modulus, or optical grain size without knowing the size, shape, and distribution of the structural elements. Scientists have used a variety of techniques to examine the microstructure of snow. We can determine the crystallographic orientation from thin sections and estimate density and specific surface area from surface sections. Although these techniques are very useful they still require statistical techniques to estimate the desired parameters. Recent advances in computing power, laboratory technique, and image processing allow us to reconstruct the three-dimensional microstructure of a snow sample from a series of surface sections. The resulting model not only allows us to view the snow structure in situ (without desegregation), but provides a level of detail seen only with expensive methods such as electron microscopy and X-ray micro-tomography. We can use these models to measure bulk quantities directly or as inputs for numerical models that calculate the thermal, mechanical, and optical parameters.