Both snow distribution and the snow melt pattern are much more variable in forests than in open terrain. In this project we examine the influence of the forest on water availability in subalpine drainage basins or catchment zones.
Forests cover large portions of the northern hemisphere, including around 30% of Switzerland's surface area. Processes relating to changes in the snowpack in forested areas therefore play a key role as regards both the weather and hydrology. The way in which snow is distributed in a forest depends on a variety of factors. The structure and melting of the snowpack adhere to different patterns in forests than in open terrain, partly because the crowns of trees are capable of absorbing radiation, attenuating energy fluxes, and capturing precipitation. Given that crowns have highly diverse structures and, in some cases, change in the course of a year, moreover, both the distribution of snow and its energy balance are far more complex in forests than in open terrain.
Applying the results of local measurements to large catchment zones presents an enormous challenge. New remote sensing techniques, such as LiDAR (measuring distances with a pulsed laser light), are now enabling scientists to examine the structure of forests – even those that cover large areas – in great detail. Data collected in this way allow the tree population in relatively large catchment zones to be described realistically for the first time, or at least in a less simplistic way. We are currently seeking to integrate the newly acquired information concerning the structure of forests in existing snow models (snow melt and forest snow interception models). We are using the LiDAR data to reproduce hemispherical images showing gaps in the canopy as seen from the forest floor (Fig. 1). Such images are generally used to determine characteristic properties of the tree crowns, such as foliage area or canopy. We have verified that the images produced with the LiDAR data allow these tree properties to be assessed accurately over large areas. The data can likewise be used to calculate the potential solar radiation reaching the forest floor. We also developed a new method of calculating other forest parameters on the basis of the LiDAR data, such as the open area around a point and gaps between the tree crowns. The method has additionally been applied to create a new forest snow interception model that is 30% more accurate than previous descriptions.
A key driver of snow melt in forests is the amount of longwave radiation that reaches the snow surface – in other words, the amount of heat emitted by the sky or the surface of the trees. In connection with a doctoral thesis, we examined how accurately existing computer models are able to calculate the heat radiation. The study was facilitated by a large data set – for around ten years we kept a record of the longwave and shortwave radiation reaching the snow surface in three different coniferous forest locations in the Swiss Alps (Fig. 2). We also measured the air temperature both above the canopy at a height of 35 m and within the forest at heights of 10 m and 2 m.
Data analysis showed that certain deficiencies in the models were attributable to the air temperatures measured above the canopy being higher than those measured below it – in other words, the forest was behaving rather like a cold sink. Existing models were further found to be unreliable in particular in the springtime, when the amount of shortwave solar radiation penetrating the forest and heating the tree trunks increases as a consequence of the sun rising higher and higher. As illustrated by thermal imaging data, trunk surface temperature can be as much as 30 °C higher than the ambient air temperature (Figs. 3 and 4), which demonstrates that trees emit a lot more heat than previously assumed. We have built this effect into a new model and are now achieving significantly better results.