Measuring snow depth

Autumn 1939, as the Second World War breaks out, the warden at the Mont Cenis dam, stranded at an altitude of 2,000 metres, on the border between Italy and France has a new task: to measure the falling snow. On the morning of 26 October, on the plateau near his hut, using a long measuring stick, he records a measurement of only two centimetres, the next day it’s increased to 20 and the day after that 28. The same warden repeated and noted down measurements all winter long and so did his successors, for over seventy years. The measurements recorded at Mont Cenis are one of the first historical series on snow depth in the Alps. These measurements, available on the website of the Italian Meteorological Society, which were unfortunately discontinued in 2014, are an example of the classic technique of measuring snow depth. The technique is undertaken where the ground is as flat and as protected from the wind as is possible and a graduated rod called a snowmeter probe is inserted into the ground. The 330cm rod/probe uses alternate yellow and black stripes, each ten centimetres high to enable an easily readable measurement. Since 1981 in South Tyrol, measurements have been recorded at the region’s 15 avalanche observatories. Every day during the winter season, data on snow depth and other parameters is sent to the civil protection office and some meteorological stations also carry out automatic measurements.

Although essential, measurements with snow poles are expensive and sometimes risky. Often entailing repeated climbs to high altitudes in difficult conditions. Today, some stations use more advanced systems such as ultrasonic sensors resulting in automated and more accurate measurements. Regardless of the technology, ground measurements do have a limitation: they only refer to a specific point. In the mountains this can be especially misleading, winds and highly differing solar exposure from slope to slope mean that snow is often distributed very unevenly. In order to get a more accurate picture, the areas where snow surfaces are measured would have to be pitted like Swiss cheese.
This is where satellite images, which provide data over larger areas, should be able to come to the rescue. However, this combination of ground and satellite data - which works well for other parameters such as the extent of snow cover - is more problematic here. Currently, satellite techniques use radar sensors. Researchers set the right frequency for incoming electromagnetic waves from the satellite to interact with the snow grains. It's a bit like researchers tuning the satellite radio to the snow channel. If the signal encounters water particles, it sends a recognisable signal back to the satellite. But not all snow is the same, if it’s wetter or more powdery, the flakes - technically called grains – differ in dimension. The electromagnetic waves interact intensively with the water in the snow, preventing interaction with the dry part, this means that not all the snow can be detected. When the snow is only dry, in most cases it is transparent; in the presence of liquid water, the electromagnetic waves interact essentially with the water. The result is that the snow station, even though it communicates, doesn’t always tell the whole story on snow.

For some years now, Earth observation scientists have been trying out new approaches. If they can’t measure the "what", i.e. the snow, they measure the "time", i.e. how long it takes the electromagnetic waves to reach the ground beneath the snow and return to the satellite. It’s the same principle used by bats - they launch a wave and wait for its echo to return. Depending on the time it takes, they can see whether there is an obstacle in front of them. In a similar way, electromagnetic waves from satellites pass through snow, touch the ground and return to their base. Depending on the thickness and type of snow cover that the waves encounter on their way, the time needed for a round-trip, varies. This time interval is technically called a 'phase' and this technique may allow the snow’s total mass to be assessed, indirectly giving the value of the water content and of the snow depth.