Methods of estimating CO2, latent heat and sensible heat fluxes from estimates of land cover fractions in the flux footprint

We present a description of the process of estimating surface fluxes of CO2, latent heat and sensible heat from estimates of fractions of satellite-based land cover types in the flux footprint. The study is conducted at two heterogeneous sites in the boreal forest of Central Canada. Using a Twin Otter aircraft, fluxes were measured in a grid pattern during three Intensive Field Campaigns (IFCs) and Landsat thematic mapper data were used for land cover classification. Using a footprint function developed from tracer gas release experiments in the boreal forest, the fractions of cover types within the footprint were determined, and used in a regression analysis against observed fluxes. The results showed that the surface cover types within the flux footprint accounted for about 90% of the variations in the measured airborne fluxes of CO2, sensible heat and latent heat, at two different study sites. The attempted validation of the regression models, by comparing flux estimates over regional transects outside the grid area for which the regression model had been developed or over site-specific runs within the grid area against observed fluxes, based on fractional distributions of surface cover types, were encouraging. They indicate the potential for extrapolating models developed for a given location to another location, based simply on the fractions of cover types, at least for similar land cover types.