Written By Leena Leppänen, Finnish Meteorological Institute, Arctic Research, Sodankylä, Finland. 2017 Winning Goetz Instrument Support Program Project Title: "Spectral reflectance of snow in Sodankylä, Finland"
ASD's QualitySpec® Trek (QST) was used for reflectance measurements of natural snowpack in Sodankylä. A total of 163 spectrums were measured, including 6 test measurements in the beginning. Usually, measurements were made in Intensive Observation Area (IOA) where snow pit measurements are made weekly in addition to several automatic measurements of snow and meteorological properties (Figure 1).
Figure 1. IOA measurement field. Location of snow pit area is marked with red circle. Snow pits were made to the same area but different spot every measurement occasion. Snowpack is suspected to be homogenous in that area.
The manual snow pit measurements includes definition of layers and for each measurement of grain size and type, hardness and wetness, and density profile and specific surface area (SSA) profile with IceCube, snow depth and bulk snow water equivalent (SWE). Automatic snow measurements at the site are snow depth, SWE, snow temperature profile every 10 cm, and snowfall intensity with particle size. More information about snow pit measurements and snow conditions in Sodankylä can be found from Leppänen et al. 2016.
Vertical profiles from snowpack were measured with QST to research vertical variability of reflectance in snowpack (Figure 3a). Measurements were made approximately every 2-3 cm, and 4-5 cm in depth hoar layer where coarse grains existed. Snowpack surface was measured also from top (Figure 3b). Forest snow pit profiles were measured for research of horizontal spatial variability (Figures 2a, 2b).
Figure 2a. Three snow pits on forest in 22 March 2017 were made in area marked with red circle. In the front of the photo is the radiometer tower with two radiometers used to develop interpretation algorithms for satellite observations of snow and soil.
Figure 2b. Snow pits in the forest after the measurements in 22 March 2017.
Figure 3a. Snowpack vertical profile was measured every 2-3 cm from top of the snowpack and 4-5 cm from bottom part of the snowpack. Lowest 10-15 cm was not possible to measure because of instrument metrics. This part was also difficult to measure because of large coarse grains.
Figure 3b. Snowpack surface measurement from top. This was difficult to perform during light new snow at top (density approximately 100 kg/m3), when instrument easily sink through it.
QST reflectance from IceCube instrument samples were measured for comparing reflectance of IceCube with reflectance of QST (Figure 3c). IceCube measures hemispherical reflectance of the 1310 nm laser with integrating sphere from snow sample surface, and the values are converted to SSA which describes microstructure of snow (Figure 3e). The sample is collected from snowpack to sample holder with spatula (Figure 3d). Sample density should be around 250-350 kg/m3 and snow is packed to the sample holder with spatula if needed. The instrument is commercial, and recently one of the most important instruments to measure SSA of snow.
Figure 3c. Measurement of IceCube sample was made by pushing QST window against the sample surface.
Figure 3d. Preparation of IceCube sample, snow is collected from snowpack to the sample holder with spatula. Sample is a cylinder with 2.5 cm height and 6 cm diameter.
Figure 3e. Measurement with the IceCube instrument. The snow sample is set inside the instrument and value is read directly from the screen.
The QST measurement protocols
11 February 2017
This was the first test of QST outside. Measurements were made on river ice from top of surface of hard layer in snowpack, ice, and artificially added organic material on ice, dust on ice, and soot on ice.
22 February 2017
Vertical profile was measured from snowpack with approximately 60 cm height of snow. In addition, snow surface was measured from top.
7 March 2017
QST measurements were made from every IceCube instrument sample which were taken every 3 cm from approximately 75 cm deep snowpack.
21 March 2017
Vertical profile from 80 cm high snowpack was measured. In addition, QST measurements from every IceCube instrument sample, which were taken every 3 cm from the snowpack, were made. Density and grain size of all samples were also recorded. In addition, sampling method was tested so that surface snow and depth hoar were sampled several times.
16 March 2017
Vertical profile from snowpack with approximately 75cm height of snow was measured. In addition, snow surface was measured from top.
22 March 2017
Three different snowpack profiles were measured from forested area close to IOA with snow depths 57 cm, 57 cm and 41 cm. Aim of the experiment was to define effect of forest to snowpack structure.
3 April 2017
Vertical profile from snowpack with 72 cm height of snow was measured, and measurements from every IceCube instrument sample were made.
Preliminary results show that the reflectance varies in snowpack vertical profile (Figure 4a). Comparison between open area and forest needs further analysis because height of snow varied so much that direct comparison does not give usable results.
Figure 4 a. Reflectance from vertical profile measurements of the snowpack. Total 34 measurements were made with 2 cm interval. Total height of the snowpack was approximately 80 cm, the last measurement was made from 10 cm height. Measurements were made 21 March 2017.
Reflectance from 1310 nm is compared to reflectance values from IceCube instrument (Figure 4b). Frequently QST reflectance is smaller than IceCube reflectance. However, bias between the reflectances is quite constant so that shape of the curves are similar. Source of that bias will be researched further. One assumption is that laser reflects inside integrating sphere of IceCube without touching the snow, which causes positive bias to reflectance.
QST reflectance measured directly from snowpack and reflectance measured from IceCube samples correlates well. Therefore, it can be assumed that sampling process for IceCube measurement is not effecting too much to snow microstructure, and it is possible to think that IceCube measurements are reliable in that sense.
Figure 4b. QST reflectance of 1310 nm from vertical profile measurements and QST reflectance from IceCube samples is compared to vertical profile measured with IceCube. Total height of the snowpack was approximately 72 cm, the last QST profile measurement was made from 7 cm height and Icecube measurement from 5 cm height. Measurements were made 3 April 2017.
Both of those results will be researched further, and preliminary results show good base for the publication. A good addition would be QST measurements from the IceCube reference spectralon plates, which would allow better comparison with the instruments, however, that was not made.
Comments and error sources
QST was handy and easy to use in field. However, it was too heavy to hold stable if the snow was very thin and I couldn’t lean the instrument on snow; this was the case with newly fallen surface snow (density around 100 kg/m3) and depth hoar layer which is coarse and therefore fragile. That might cause some measurement error. In addition, some part of the light scattered from the depth hoar so that it was visible around instrument. In the surface snow, sun light penetrates snowpack and possibly causes some error to the measurements. That could be avoided by covering the snow surface during direct sunshine during vertical profile measurements.
The instrument size limited measurements close to the ground; approximately 10-15 cm was lowest possible to measure. During sunshine and positive air temperature, warming of the instrument window and grey area around it caused some melting of the snow, and probably some measurement inaccuracy.
The QST software was functional and simple to use. Storing and exporting data was easy.
I would like to thank ASD for Goetz Instrument Support, which gave a unique opportunity to test QualitySpec Trek at Sodankylä. I also would like to thank Quinghuan Li and Hanne Suokanerva for the photos. Work was partly funded by scholarship from the Vilho, Yrjö and Kalle Väisälä Foundation on Finnish Academy of Science and Letters.
Leppänen, A. Kontu, H-R. Hannula, H. Sjöblom, J. Pulliainen, Sodankylä snow survey program, Geosci. Instrum. Method. Data Syst., 5, 163-179, doi:10.5194/gi-5-163-2016, 2016. http://www.geosci-instrum-method-data-syst.net/5/163/2016/gi-5-163-2016.pdf
Thanks to Leena for providing us with this recap! For more information on ASD's snow and ice research solutions, visit our Ice Characteristics Research and Snow Research web page
We are now accepting Goetz research proposals! Borrow an ASD instrument for up to 3 months to conduct your academic research!!