On the simulation of bulk foliar temperature of a well-watered wheat crop
Well-watered crops growing in very dry environments show radiometric temperatures that are often a few degrees below air temperature. Using this latter temperature for simulating plant processes may compromise the performance of crop models. This study showed that simple parameterizations of canopy resistance for the 10:00-18:00 h interval resulted in predictions of bulk foliar temperature of a well-watered wheat crop that were within ±2°C of radiometric measurements 95% of the time. The assumption of constant canopy resistance (30 s m-1) for the period when 'LAI ' > 3.6 was particularly attractive given its extreme simplicity. A multi-layer model, with inclusion of a dispersion matrix based on Lagrangian theory for describing the turbulent transfer within and immediately above the canopy, was used in numerical simulations to assess the leaf-to-canopy scale translation of surface conductances. The model converged rapidly and uniformly towards solution, producing unique profiles of fluxes and scalar concentrations that satisfied both the dispersion and leaf models. The dispersion matrix proved to be neither complex nor difficult to implement, in contrast to higher-order closure models or Lagrangian stochastic models. Various environmental factors were varied for different runs as well as canopy architecture; stomatal conductance was simulated either using a simple relationship with net radiation or the Ball and Berry model which accounts for feedback mechanisms. Results supported the scaling up procedure for the range of scenarios simulated. It is suggested that large discrepancies between bulk stomatal and canopy resistances often reported in the literature may be attributed to the Penman-Monteith model failing to properly describe the situation under consideration, one example being the inaccurate parameterization of the bulk aerodynamic resistance. In particular, attention focused on the roughness length for heat and momentum, ' zoH' and 'zoM'; values of ' zoH' seemed to be strongly influenced by canopy architecture. Results from these numerical simulations, and those from comparing predicted and observed canopy temperatures of a well-irrigated wheat crop, question the widespread use of 'zoH' << 'z oM' when applying the P-M equation.