Simple Models of Light Interception in Heterogenous Canopies

Solar radiation drives biophysical processes within terrestrial ecosystems, and thus quantifying plant biophysical processes typically also involves quantifying radiation interception. Although there are an extremely wide range of models available for this task, many applications (e.g., crop models, ecosystem models, remote sensing inversion) necessitate a simple description of radiation interception, usually based on a simple one-dimensional (1D) turbid medium model (Beer's law). However, this approach results in significant over prediction of radiation interception in sparse or heterogeneous canopies (e.g., crops, savannas, coniferous forests), which can lead to large biases in predicted ecosystem fluxes. To correct for this over prediction, an empirical "clumping factor" Ω is typically used to reduce the effective leaf area index input to the model.

In this work, we argue against the clumping factor paradigm for modeling radiation interception in heterogeneous canopies, and alternatively propose a new simpler and more robust framework. Given the difficulty of collecting data to accurately characterize radiation interception in a wide range of architectures, the approach was evaluated using an array of virtually-generated canopies along with a detailed 3D leaf-resolving radiation model.

Results showed that the simple 1D model could accurately estimate radiation interception across a wide range of crop species and plant spacings/densities. Furthermore, for a given canopy, heterogeneity is described using only a single, constant parameter that can be easily estimated. We compared against other previously proposed simple geometric models that assume some particular canopy shape (e.g., hedgerow, spherical/ellipsoidal crowns), and found that the new, more general model performed better than these more specific models.