Radiation and transpiration partitioning in a maize-sorghum intercrop: Test and evaluation of two models

Two models - one mechanistic and one more empirical - for partitioning of radiation and transpiration within a maize-sorghum intercrop were evaluated against measurements of canopy, radiative balance, and micrometeorological and ecophysiological parameters. The mechanistic model is based on a detailed two-dimensional description of canopies, and solves the energy balance by simulating the distribution of sunlit and shaded leaves. The semi-empirical model refers to an hypothetical 'two-layer' canopy, and simulates the sharing of transpiration using a modified form of the Penman-Monteith equation, including an estimation of the fractional radiation intercepted by each species. Interception of photosynthetically active radiation (PAR) was predicted with good accuracy by the two models, but differences were observed in the computed fractional intercepted radiation, particularly for the sorghum-dominated crop. Based on the same light interception model, partitioning of maximum transpiration was compared with sap-flow measurements. Transpiration was predicted with relatively good accuracy with the two models when non-limited by soil water conditions. Sensitivity analyses revealed that the semi-empirical model was highly sensitive to uncertainties in the calculated fraction of intercepted radiation, but less sensitive than the mechanistic model to uncertainties in stomatal conductance, vapor pressure deficit, and wind speed. When soil water is limited, coupling separated climatic demand with a suitable soil-root water uptake model is a promising way to predict both transpiration and soil water sharing in mixed crops.