Influence of Tree Proximity on Simulated Sub-Canopy Incoming Longwave Radiation to the Snow Surface in Mid-Latitude Boreal Forests
Thursday, 18 December 2014
Snowmelt in boreal forest environments contributes substantially to spring surface runoff and is largely controlled by the local surface radiation budget. In mid-latitude forested regions, the shading, absorption and emission of incoming radiation causes significant spatial and temporal variation in the surface energy balance within forests of variable canopy densities. Canopy gaps and edges in particular have been shown to have markedly different radiative regimes compared to continuous forests and open areas due to heightened exposure to solar radiation and increased emission of longwave radiation. Improved simulation of the sub-canopy energy fluxes will reduce uncertainty in snowmelt modelling in these areas. Simulated incoming longwave radiation can be partitioned between that from the sky and the canopy, i.e. a two part model, using estimates of emissivity and temperatures. Hemispherical photography allows objective partitioning into sky-view fraction (SVF) and forest fraction (1 – SVF). While air temperature may be used as a proxy for canopy temperature in these models, explicit representation of variability of canopy temperatures have been shown to improve the simulations of longwave emission from the canopy to the snow surface. Simulated incoming longwave radiation, within close proximity to tree trunks, was evaluated at edges of gaps in a discontinuous forest of predominantly spruce and larch trees in Davos, Switzerland (46°50’N 9°52’E) between January and April 2014. Measurements were made of long- and shortwave radiation at increasing distances from tree trunks, tree trunk and air temperatures, SVF using hemispherical photography, and both closed and above canopy radiation. Results show that the simulation bias of incoming longwave radiation estimates are related to proximity of the nearest tree and the sky-view fraction at the modelled point. This suggests a greater complexity to modelling incoming longwave radiation along forest gaps and edges than a simple two-part model.