Scale influences of surface parameterization on modelled boreal carbon and water budgets

Current analyses of Boreal carbon and water budgets suggest that the Boreal ecosystem may be acting as a sink of atmospheric CO2 by accumulating soil organic matter in mesic conifer landscapes. An empirical study demonstrated that moss and soil moisture content differs significantly between sphagnum and feather moss regions within mesic conifer stands. Moss moisture content was also related to top of moss mid-infrared reflectance to facilitate remote mapping of moss type. Inversion of LANDSAT TM imagery, using a radiative transfer model relating top of moss reflectance to top of canopy reflectance, was used to map moss types within sparse, mesic conifer stands. Uncertainties in canopy cover and the co-variation of canopy cover with surface moisture limited the precision of retrieved moss maps.;A diagnostic, daily time-step, model of energy, water and carbon cycles in Boreal conifer stands, was developed based on the RHESSys model. The model included a distributed hydrologic parameterization with lateral water redistribution and a spatially variable moss surface layer. Comparison of estimated total evapotranspiration (ET) to eddy flux measurements indicated typical relative bias errors within 5% and precision errors between 13% to 22%. Modelled understory ET ranged from 31% to 50% of total ET with substantially higher average ET rates in sphagnum dominated regions compared to feather moss or lichen covered areas. Substantial bias errors in modelled ET and net ecosystem production were observed over mesic conifer sites when a uniform feather moss cover was specified. A two patch representation based on the proportion of moss cover type was sufficient to provide estimates of total ET and NEP that were not substantially different from model estimates using over 20 patches in a site. Further research should investigate the relative sensitivity of modelled fluxes to uncertainties in the joint spatial distribution of wetness, biomass and nutrients. The validated numerical model could be used to calibrate simpler, analytical, models relating remotely sensed or empirical parameterizations, of moss wetness, biomass and leaf area to spatial patterns of Boreal water and carbon dynamics.