Comparison of models for daily light partitioning in multispecies canopies

A simulation model of light partitioning in horizontally homogeneous multispecies canopies is proposed. The model is based on the Kubelka–Munk equations (KM) applied to a mixture of N vegetation components. Only two hemispherical fluxes, i.e. downwards and upwards, are considered. The exact solution of KM equations was then simplified in such a way that the model can be easily extended to multispecies canopies including several vegetation layers. The simplified KM model (KMS) was compared to two other light models dealing with mixed canopies: the more detailed model SIRASCA [Sinoquet, H., Moulia, B., Gastal, F., Bonhomme, R., Varlet-Grancher, C., 1990. Modeling the radiative balance of the components of a well-mixed canopy: application to a white clover–tall fescue mixture. Acta Oecol. 11, 469–486], and the simpler model ERIN [Wallace, J.S., 1997. Evaporation and radiation interception by neighbouring plants. Q. J. R. Meteorol. Soc. 123, 1885–1905]. All three models were applied to theoretical two-species monolayer canopies, and to actual mixed canopies, the geometry of which was retrieved from the literature. In the PAR waveband, the model KMS gave simulation results very similar to those of SIRASCA in case of contrasted canopy structures. In conditions of high leaf and soil scattering, deviations between SIRASCA and KMS outputs were higher and reached maximum values of –0.08 for erectophile species. Comparison between SIRASCA and ERIN outputs showed that ERIN largely underestimated light competition in a two-component canopy in several conditions, due to light partitioning only based on height differences between components. Simulations also showed the significant effect of the vertical distribution of leaf area on light partitioning in the case of mixtures where components had equal or different heights. Finally it appears that the model KMS could be a candidate for inclusion in growth models for multispecies canopies, since all KMS parameters have physical meaning and it is very easy to implement.