Comparison between the Results of a Snow Metamorphism Model and Remote Sensing Derived Snow Parameters in the Alps

https://doi.org/10.1016/S0034-4257(98)00116-3Get rights and content

Abstract

The numerical snow metamorphism model CROCUS derives a complete description of the snow cover according to its geographical location (range), elevation, slope, and orientation. From the remote sensing data, snow parameters that were comparable to the model results were derived: the lower elevation of the snow cover, the surface grain size, and the surface temperature. The Landsat Thematic Mapper sensor was used because it has a good spatial resolution, a short wave infrared channel which is sensitive to grain size, and a thermal infrared channel. A first Landsat TM (Thematic Mapper) scene was acquired on 24 April 1992 and a second one on 11 December 1992. A DEM (digital elevation model) was used to obtain the local incidence angles and the elevation of each snow pixel. The pixels were then grouped according to the CROCUS classification (range, elevation, slope, orientation), and the mean snow characteristics for each class were compared with the CROCUS results. The lower limit of snow and the surface grain size derived from TM data compared favorably with the model results even if those parameters were not always easy to define. Larger differences were found for the temperature because it varies rapidly and is very sensitive to shadowing by surrounding mountains and also because its remote measurement is dependent on atmospheric conditions.

Introduction

For the purpose of climatological studies Hall 1988, Marshall and Oglesby 1994, avalanche forecasting Brun et al. 1989, Brun et al. 1992, and water resource management, it is necessary to measure and compute the main snow characteristics: area, depth, water equivalent, albedo, snow type, etc. The snow covered areas are often not easily accessible, and remote sensing is, therefore, an adequate tool for their study. To reach these goals, many experiments were carried out at different locations in the French Alps with simultaneous remote sensing and ground truth data (Fily et al., 1997).

The objective of this article is to compare the output of a snow metamorphism model (CROCUS) with the surface snow characteristics derived from remote sensing data (Landsat Thematic Mapper) obtained on a large scale, that is, many ranges of the Alps. The parameters that are compared are the lower limit of the snow cover elevation, its temperature, and the surface grain size. The Landsat Thematic Mapper is adequate because it has good spatial resolution, a thermal infrared channel, and a 1.65 μm channel, which is necessary to distinguish clouds and is very sensitive to the grain size.

In this article we describe the CROCUS snow metamorphism model, the remotely sensed data, and their processing. Then comparisons between model results and measurements are discussed.

Section snippets

The crocus snow metamorphism model

The snow cover evolution at a given location mainly depends on the prevailing meteorological conditions. They govern its energy and mass balance and therefore the metamorphism of each layer. They also govern the presence of liquid water inside the snowcover. A physically based numerical model, called CROCUS, has been developed to simulate all these phenomena Brun et al. 1989, Brun et al. 1992. It derives a complete description of the snow cover including temperature, density, liquid water

Remote sensing data

Two Landsat 5 TM (Thematic Mapper) images, quarter scenes, were acquired on 24 April and 11 December 1992 at 9:45 U.T. (Fig. 1). In April the solar incidence angle was 39.9° and its azimuth 134°; in December they were 69.1° and 168°, respectively. The main spectral characteristics of TM channels are given in Table 1. The spatial resolution is 30 m for Channels 1, 2, 3, 4, 5, and 7 and 120 m for Channel 6.

On 24 April 1992 the sky was very clear with only a few small clouds. On 11 December the

Lower limit of the snow cover

The lower limit of the snow cover is different for each range because the meteorological conditions are different. Slope and azimuth are also influential. On north-oriented slopes the snow should be present on the ground at a lower elevation than on south-oriented slopes because there is less solar irradiance. Comparisons were made between CROCUS- and TM-derived lower elevations for the 24 April 1992 image. This date was chosen because the lower elevation range is larger at the end of the

Surface snow grain size

The snow grain size has an important effect on the snow albedo and is therefore an important parameter for any study that needs surface radiative balance Dozier 1989, Grenfell et al. 1994, Genthon 1994. But snow grain size is a parameter which is not easy to define: The sizes that are computed by the model, those which are measured in situ, and those which are deduced from the reflectance are not the same, although they are certainly related. This difficulty must not be forgotten when comparing

Surface snow temperature

The temperature gradient in the snow mantle is one of the most important factors for snow metamorphism. The determination of a good surface temperature is therefore necessary. The absorption of the thermal infrared electromagnetic waves in the snow is very high Warren 1982, Warren 1984; therefore, only the upper surface of the snow is seen by the thermal infrared channel TM6 (Table 1).

Conclusion

The Landsat Thematic Mapper data associated with a digital elevation model at a 250 m resolution can be used to provide some characteristics of the snow mantle in Alpine areas that are comparable to the output of a snow metamorphism model (CROCUS). The comparisons can be made for different snow classes according to their geographical location (range), their elevation, slope, and orientation.

The lower limit of the snow mantle was defined as a percentage of snow covered pixels between 30% and

Acknowledgements

These experiments were funded by the French Programme National de Teledetection Spatiale and the Centre National d’Etudes Spatiales (SPOT4/MIR).

References (20)

There are more references available in the full text version of this article.

Cited by (20)

  • Ice Surface Temperatures in the Arctic Region

    2019, Taking the Temperature of the Earth: Steps towards Integrated Understanding of Variability and Change

  • Reconstructing snow-avalanche extent using remote sensing and dendrogeomorphology in Parâng Mountains

    2019, Cold Regions Science and Technology
    Citation Excerpt :

    Among these, Landsat imagery and derived products were already successfully used in the study of various types of forest disturbance: windfall disturbance (Baumann et al., 2014; Rich et al., 2010), fire disturbances (White et al., 1996), forest succession (Song and Woodcock, 2002), bark beetles attacks (Berryman et al., 1989; Carroll and Safranyik, 2004; Wulder et al., 2006), leaves-feeding pathogens (Hódar et al., 2003; Hall et al., 2006; Meigs et al., 2011). But in the field of snow-avalanche research, the optical satellite remote sensing applications are generally limited to mapping the avalanche paths (Butler and Walsh, 1990; Walsh et al., 1994; Walsh et al., 2004; Barlow and Franklin, 2008;), mapping snow covered areas (Walsh et al., 1994; Rosenthal and Dozier, 1996; Lazar and Williams, 2008; Huang et al., 2011; Stigter et al., 2017), vegetation classification (Walsh et al., 1994; Dorren et al., 2003; Gruber and Haefner, 2013) and estimating snow parameters (Fily et al., 1997; Fily et al., 1999; Dozier and Painter, 2004). Dendrogeomorphology is a method widely used in snow-avalanche research.

  • Separating snow and forest temperatures with thermal infrared remote sensing

    2018, Remote Sensing of Environment
    Citation Excerpt :

    Satellite thermal infrared radiation (TIR) images—available from such sensors as Landsat TM, ETM+, and TIRS, ASTER, NOAA AVHRR, GOES, MODIS, and VIIRS—provide comprehensive spatial and temporal data (see Tomlinson et al., 2011 for history and IR specifications). Applications cover a wide range of topics, including surface heat flux (Norman et al., 1995), evapotranspiration (Wang and Dickinson, 2012), stream temperatures (Handcock et al., 2006, 2012), a surrogate for air temperature (Shamir and Georgakakos, 2014; Pepin et al., 2016), urban heat island studies (Rizwan et al., 2008), and snow surface temperature (Fily et al., 1999; Westermann et al., 2012; Pérez Díaz et al., 2015). MODIS land surface temperature (LST) and emissivity products have proven reliable over homogeneous terrain (Wang and Liang, 2009) but have greater biases and uncertainty in mountain areas (Lipton and Ward, 1997; Liu et al., 2006).

  • Measurements and modeling of the vertical profile of specific surface area of an alpine snowpack

    2013, Advances in Water Resources
    Citation Excerpt :

    In the near-infrared (NIR) and short-wave infrared (SWIR) portions of the solar spectrum (800–2500 nm), ropt controls snow albedo, so that the radiative energy balance of the snowpack depends crucially on its SSA vertical profile close to the surface [9,41]. SSA of near-surface snow is therefore accessible through optical remote sensing [14,18,24,33,35]. Snow SSA is also highly relevant to other electromagnetic properties of snow involved in microwave emissions [3,40] and radar backscatter [29].

  • Evaluation of MODIS land surface temperature with in-situ snow surface temperature from crest-safe

    2017, International Journal of Remote Sensing

  • Evaluation of VIIRS land surface temperature using crest-safe air, snow surface, and soil temperature data

    2015, Geosciences (Switzerland)
View all citing articles on Scopus
View full text
Copyright © 1999 Elsevier Science Inc. All rights reserved.