Scaling BVOC Emissions from Leaf to Canopy and Landscape: How Different Are Predictions Based on Contrasting Emission Algorithms?

A variety of leaf-level models has been embedded in a canopy model and used to predict monoterpene emissions from canopies and landscapes, but there is no objective basis of choice between different models. Here we analysed the capacity of four leaf-level models and their variations, yielding altogether eight models, for predicting diurnal and seasonal variations in canopy monoterpene emissions. The main models tested were Guenther et al. model with fixed light and temperature dependencies or with optimally adjusted dependencies, two models linking emissions to foliage photosynthetic rate, one to electron transport rate (ETR model) and the other to gross assimilation rate (C-ratio model), and a dynamic model considering non-specific monoterpene storage in leaves. Once parameterized in a consistent manner, all models showed similarly high performance, assessed by explained variance, modelling efficiency and average model deviations for homogeneous canopies. Simulations suggested potentially stronger deviations for landscapes with fragmented vegetation. This analysis indicates that the choice among the models cannot be based on model validation statistics alone, but depends on whether only BVOC emissions need to be simulated (Guenther et al. model) or both photosynthesis and BVOC fluxes are needed (ETR or C-ratio model) or whether one needs data on night atmospheric reactivity (dynamic model).