Evaluating plant root capacity in extrating water from the soil is important for transpiration modeling and to understand crop growth and yield and soil water and nutrient balance. Aiming to test a macroscopic root water extraction model based on the microscopic process description, an experiment was described in which the root system of plants penetrated different soil layers with contrasting hydraulic properties. Four lysimeters containing two physically divided compartments were built and filled with material of two soils with different texture (a medium textured soil - AR and a clayey soil - AG). During a month and a half a water regime was imposed alternating the irrigation among the compartments. The soil water content in the compartments was measured with TDR and tensiometers. Soil hydraulic properties - retention and conductivity - were analyzed by standard methods. Root density was determined by weighing at the end of the experiment, resulting in values twice as high in AR than in AG soil. It was observed that water extraction occurred preferentially from the lysimeter compartments with the highest matric flux potential. Occasionally, water transfer from the compartment with higher matric flux potential to the lower one was observed, releasing water from root to soil (hydraulic lift). To compensate for the effect of heterogeneity of root distribution and root activity and soil-root contact, an empirical factor f was added to the model. Its value was determined by a numerical fitting procedure aiming at the highest correlation between model and observation in the four lysimeters. The model described 80% of the observations satisfactorily by using these f values, which were 0.01506 and 0.003713, respectively, for AR and AG. Model predictions indicated a much more frequent water release from roots to soil than observed in the experiment. This may suggest internal root resistance, not considered by the model, may play an important role in root-water relations.
Modeling; water stress; root density; transpiration
