Main content area

Responses to atmospheric CO2 concentrations in crop simulation models: a review of current simple and semicomplex representations and options for model development

Vanuytrecht, Eline, Thorburn, Peter J.
Global change biology 2017 v.23 no.5 pp. 1806-1820
acclimation, carbon, carbon dioxide, carbon dioxide enrichment, case studies, crop models, nitrogen, nutritive value, photosynthesis, simulation models, stomatal conductance, stomatal movement, uncertainty
Elevated atmospheric CO₂ concentrations ([CO₂]) cause direct changes in crop physiological processes (e.g. photosynthesis and stomatal conductance). To represent these CO₂ responses, commonly used crop simulation models have been amended, using simple and semicomplex representations of the processes involved. Yet, there is no standard approach to and often poor documentation of these developments. This study used a bottom‐up approach (starting with the APSIM framework as case study) to evaluate modelled responses in a consortium of commonly used crop models and illuminate whether variation in responses reflects true uncertainty in our understanding compared to arbitrary choices of model developers. Diversity in simulated CO₂ responses and limited validation were common among models, both within the APSIM framework and more generally. Whereas production responses show some consistency up to moderately high [CO₂] (around 700 ppm), transpiration and stomatal responses vary more widely in nature and magnitude (e.g. a decrease in stomatal conductance varying between 35% and 90% among models was found for [CO₂] doubling to 700 ppm). Most notably, nitrogen responses were found to be included in few crop models despite being commonly observed and critical for the simulation of photosynthetic acclimation, crop nutritional quality and carbon allocation. We suggest harmonization and consideration of more mechanistic concepts in particular subroutines, for example, for the simulation of N dynamics, as a way to improve our predictive understanding of CO₂ responses and capture secondary processes. Intercomparison studies could assist in this aim, provided that they go beyond simple output comparison and explicitly identify the representations and assumptions that are causal for intermodel differences. Additionally, validation and proper documentation of the representation of CO₂ responses within models should be prioritized.