Investigation of the impacts of elevated ozone on maize and soybean using the Terrestrial Ecosystem Model in R (TEMIR) version 2.0: differential responses of biomass and photosynthetic rates to cumulative stomatal ozone uptake in different crops

Surface ozone air pollution impairs carbon assimilation in terrestrial ecosystems. For crop species, ozone pollution reduces biomass and crop yield and therefore poses challenges on food security in regions with large populations such as India and China. The ozone impacts on crop yield can be examined with a mechanistic crop model, which explicitly simulates plant physiological responses (e.g., gas exchange rate, leaf area index) to changes in environmental conditions. In mechanistic crop models, ozone-induced yield loss is primarily determined by the sensitivity parameter (a_sen) of photosynthetic rate loss to cumulative stomatal ozone uptake. Derivation of a_sen follows different approaches: one based on statistical relationships between relative yield or biomass loss and cumulative ozone uptake, as described in Sitch et al. (2007); another based on relationships between gas exchange rate losses (photosynthesis and stomatal conductance) and cumulative ozone uptake, as described in Lombardozzi et al. (2015).

In this study, gas-exchange measurement data from multiple elevated ozone exposure experiments for maize and soybean are used to calibrate a_sen following Lombardozzi et al. (2015). Validation simulations are conducted using the Terrestrial Ecosystem Model in R (TEMIR) version 2.0, a mechanistic crop model akin to those in land surface models such as JULES and CLM4.5, implemented with two plant-ozone damage schemes following Sitch et al. (2007) and Lombardozzi et al. (2015).

With the newly calibrated a_sen, modeled ozone-induced relative yield loss shows good agreement with observed values for soybean, with a mean error of less than 5 percentage points across different ozone levels. Simulations using the calibrated a_sen following Lombardozzi et al. (2015) exhibit superior performance compared to those using the default a_sen from Lombardozzi et al. (2015) or the calibrated a_sen following Sitch et al. (2007), both of which have mean errors exceeding 25 percentage points in the modeled ozone-induced relative yield loss. The low mean error from the simulations using the calibrated a_sen following Lombardozzi et al. (2015) suggests the sensitivity of relative photosynthetic rate loss to ozone is similar to that for relative yield loss in soybean. In contrast, for maize, with the calibrated a_sen following Lombardozzi et al. (2015), the model overestimates relative ozone-induced yield loss by about 30 percentage points at the highest ozone concentration (~100 ppbv). Sensitivity simulations with varying values of a_sen indicate that the parameter calibrated to photosynthetic rate loss must be reduced to about one-third of its original value to align modeled and observed relative yield and biomass losses for maize. Modelers should account for these differential responses of photosynthetic rates versus yield and biomass losses among crops species, when assessing future ozone impacts on crop productivity.