A canopy scale model is presented that utilises Lagrangian dispersal theory to describe the relationship between source distribution and concentration within the canopy. The present study differs from previous studies in three ways: (1) source/sink distributions are solved simultaneously for CO2, 13CO2, H2O and sensible heat to find a solution consistent with leaf-level constraints imposed by photosynthetic capacity, stomatal and boundary layer conductance, available energy and carbon isotopic discrimination during diffusion and carboxylation; (2) the model is used to solve for parameters controlling the nonlinear source interactions rather than the sources themselves; and (3) this study used plant physiological principles to allow the incorporation of within- and above-canopy measurements of the 13C/12C ratios of CO2 as an additional constraint. Source strengths of CO2, H2O, sensible heat and 13CO2 within a Siberian mixed-coniferous forest were constrained by biochemical and energy-balance principles applied to sun and shaded leaves throughout the canopy. Parameters relating to maximum photosynthetic capacity, stomatal conductance, radiation penetration and turbulence structure were determined by the optimisation procedure to match modelled and measured concentration profiles, effectively inverting the concentration data. Ground fluxes of CO2, H2O and sensible heat were also determined by the inversion. Total ecosystem fluxes predicted from the inversion were compared to hourly averaged above-canopy eddy covariance measurements over a ten-day period, with good agreement. Model results showed that stomatal conductance and maximum photosynthetic capacity were depressed due to the low temperatures experienced during snow melt; radiation penetrated further than simple theoretical predictions because of leaf clumping and penumbra, and stability effects were important in the morning and evening. The inversion was limited by little vertical structure in the concentration profiles, particularly of water vapour, and by co-dependence of canopy parameters.DOI: 10.1034/j.1600-0889.2002.01356.x