A comparison of the mosaic and aggregated canopy frameworks for representing surface heterogeneity in the Canadian boreal forest using CLASS: a soil perspective

Abstract

The Canadian Land Surface Scheme is employed in off-line tests to represent a 1200 km 2 patch of the Boreal Ecosystem-Atmosphere Study Northern Study Area, a heterogeneous region of boreal forest in north-central Manitoba. The soils in the region vary from rapidly draining coarse sand to poorly drained areas of peat overlying clay, including wetlands. Observed surface fluxes from four sites, a young jack pine stand, an old jack pine stand, an old black spruce stand and a FEN, are scaled by proportional weighting to represent the study area. The performance of the mosaic approach, in which the region has been represented by four distinct patches, is compared with that of the aggregated approach, in which mean surface parameters have been defined based on fractional coverage. Various soil columns are employed in the aggregated approach, in order to assess the effects of representing the heterogeneous soils of the region in a single column. While the conservative water use at all of the sites limits the effects of changes in the soil column on the turbulent fluxes, important effects are noted. Soil evaporation is overestimated when an organic or composite soil layer is represented at the surface, and underestimated when a sandy soil column is represented. Soil evaporation can be calibrated to match that of the mosaic method, however, the need for more robust methods of accomplishing this is acknowledged. The process of aggregating the soil column eliminates the dry sandy zones which experience periods of moisture stress. This results in periods of increased canopy conductance and an overestimation of the latent heat flux. A 2-patch mosaic, with one patch consisting of sandy soil and the other patch of organic soil over clay, allows moisture stress to occur and produces modelled half-hourly sensible and latent heat fluxes that agree more closely to those of the 4-patch mosaic than any of the aggregated runs, across the range of hydro-meteorological variables. Throughout the course of the model runs presented here, the 2-patch mosaic also exhibits cumulative drainage and evapotranspiration closer to that of the 4-patch mosaic than any of the aggregated soil columns.