A SIMPLE SOIL-PLANT-ATMOSPHERE TRANSFER MODEL (SISPAT) DEVELOPMENT AND FIELD VERIFICATION

When examining the various soil-plant-atmosphere models proposed in the literature, it becomes obvious that, according to the speciality of their authors, one or several compartments of the model are generally very detailed, whereas the other compartments remain crude. The aim of this work was first, to build a model, including the main physical processes, but with equivalent degrees of simplification for all the compartments and, second, to provide a validation as complete as possible for the various compartments. The resulting model, driven by meteorological forcing at a reference level (incoming solar and long-wave radiation, air temperature, humidity and wind speed, and rainfall), can be divided into four main compartments. In the soil, coupled heat and mass transfer equations, including liquid and vapour phase transfer, are solved. In the atmosphere, stability is taken into account in the calculation of the aerodynamic resistances. At the soil-plant-atmosphere interface, one vegetation layer is considered, with two energy budgets: one for the bare soil fraction of the plot and one for the vegetated fraction. In the soil, root uptake is modelled using an electrical analogue scheme with various resistances (soil, root, xylem). Finally, in the case of rainfall (or irrigation), interception, infiltration and runoff is calculated. The model is first described and then compared with field data collected on a soybean plot of 0.72 ha. The soil is composed of three horizons, the hydraulic and thermal properties of which were determined experimentally. The atmospheric forcing and the net radiation were measured. The sensible heat flux above the canopy was deduced from wind speed and temperature profiles. In the soil, water pressure, water content and temperature were measured at several depths. Temperature profiles also allowed for the derivation of the soil heat flux at the ground surface and the latent heat flux was obtained from the energy budget. Plant height, leaf area index and leaf water potential were also recorded on several days. Seven days of complete measurements were available: 2 days were under dry conditions (19-20 August 1991) and 5 days under wet conditions (24-28 August 1991) following a rainfall of 45 mm on 22 August 1991. Missing parameters were calibrated using the first 3 days of the wet period (24-26 August 1991) and the model was validated on the remaining days. A fair agreement between the model and the data was observed for both atmospheric fluxes, for soil variables (water content and temperature) and for leaf water potential, provided only an accurate determination of the parameters was made.