Investigation of the Sensitivity of Snow and Energy-Flux Conditions to Surface and Vegetation Parameters for Forest Sites Using the Multi-Energy Balance Option in SURFEX

English Abstract:

Interaction between vegetation and snow is usually modelled in a quite simplified manner in most Numerical Weather Prediction (NWP) and climate models. This simplified manner may have a few consequences for the model performance: (i) wrong timing in spring snow melt and river discharge peaks (usually too early in forest dominated regions). (ii) wrong energy flux exchange between surface and atmosphere (radiation and sensible and latent heat fluxes). (iii) wrong soil temperatures (too much cooling of soil in autumn and winter, too much warming in spring and summer). (iv) wrong snow area fraction (usually underestimated in forest regions). Here we will illustrate a few of these issues.

We do that by the introduction of explicit canopy energy balance in the externalized surface model SURFEX where turbulent and radiation fluxes within the canopy layer are parameterized. This Multi-Energy Balance (MEB) parameterization is part of latest release of SURFEXv8.

The snow accumulation is related to the height of the canopy (forest, shrubs, grass,…). The simulated turbulent and radiation fluxes, the snow accumulation and the soil heat flux will all be sensitive for surface and vegetation characteristics as e.g. Leaf-Area Index, surface albedo, surface roughness, vegetation extinction coefficient, vegetation heat capacity. We optimise the values of these parameters by setting up SURFEX offline for a number of flux-tower sites for which we correlate observed and simulated variables. Also, since SURFEX makes it easy to switch between different physical options we can quantify how the classical ISBA land-surface option performs compared to MEB.

We show that with the MEB option in SURFEX we get quite a good correspondence between simulated and observed snow depth and heat fluxes at forest sites. However, with the classical ISBA option it is not possible to get both snow depth and heat fluxes to correspond well with observations.