In some studies of the impacts of climate change on global crop production, crop growth models were empirically adapted to improve their response to increased CO2 concentration and air temperature. This chapter evaluates the empirical adaptations of the photosynthesis and evapotranspiration (ET) algorithms used in the soybean [Glycine max (L.) Merr.] model, SOYGRO V5.42, by comparing it with a new model that includes mechanistic approaches for these two processes. The new evapotranspiration-photosynthesis sub-model (ETPHOT) uses a hedgerow light interception algorithm, a C3-leaf biochemical photosynthesis submodel, and predicts canopy ET and temperatures using a three-zone energy balance. ETPHOT uses daily weather data, has an internal hourly time step, and sums hourly predictions to obtain daily gross photosynthesis and ET. The empirical ET and photosynthesis curves included in SOYGRO V5.42 for climate change prediction were similar to those predicted by the ETPHOT model. Under extreme conditions that promote high leaf temperatures, like in the humid tropics, SOYGRO V5.42 overestimated daily gross photosynthesis response to CO2 compared with the ETPHOT model. SOYGRO V5.42 also slightly overestimated daily gross photosynthesis at intermediate air temperatures and ambient CO2 concentrations. The ETPHOT model needs to include a cold-damage function based on the minimum hourly air temperature since it appears to overpredict photosynthesis at low air temperatures. SOYGRO predicted less decrease in ET with increased CO2 than the ETPHOT model and gave a much larger response to air temperature than the ETPHOT model. The implication is that water stress and irrigation demand could be severely overestimated at high air temperature and slightly underestimated at high CO2 concentration using SOYGRO V5.42 vs. the new model. Although season-long predictions of yield, biomass, and cumulative ET showed small relative differences between the two models at Gainesville, FL, other sites and conditions could produce greater differences.
Atmospheric CO2 concentration has increased markedly from the pre-Industrial Revolution level of 270 μmol mol−1 to the current levels of ≈355 μmol mol−1 (Barnola et al., 1987; Keeling et al., 1989). There is reasonable certainty that CO2 and other greenhouse effect gas levels will continue to rise in the near future. Atmospheric general circulation models predict, with less certainty, that a doubling of these gases could cause global air temperatures to increase as much as 3 to 6° C (Grotch, 1988), depending on latitude. Precipitation is expected to be slightly greater (3–11%) on a global scale, but this effect could vary both temporally and spatially (Wigley et al., 1986).
The complexity of plant response to CO2, air temperature, and water stress (leaf biochemistry, leaf and canopy gas exchange processes, crop development, growth, and partitioning) makes a mechanistic modeling approach necessary to provide reliable predictions of response to multiple climatic factors (Reynolds & Acock, 1985). Although the beneficial effects of CO2 on the growth and productivity of crops, especially C3 plants, are well-documented (Acock & Allen, 1985; Kimball, 1983; Cure & Acock, 1986), the effects of long-term interactions of increased CO2, air temperature, and water stress have been inadequately studied (Cure & Acock, 1986). The need to quantify crop response to projected climate changes stimulated the adaptation of a family of crop models developed by the International Benchmark Sites Network for Agrotechnology Transfer project (IBSNAT, 1989). These models have been used to study the possible impacts of climate change on crop production and economics both in the USA and internationally (Ritchie et al., 1989; Rosenzweig, 1989; Adams et al., 1990; Curry et al., 1990a,b).
The soybean crop growth model used in the above studies was SOYGRO V. 5.42, developed at the University of Florida (Wilkerson et al., 1983; Jones et al., 1989). This soybean model has been extensively validated under ambient conditions (Brisson et al., 1989; Jones & Ritchie, 1991; Egli & Bruening, 1992; Nagarajan et al., 1993). Currently the model has been distributed to >500 users and >20 international institutions. SOYGRO's photosynthesis and ET algorithms were empirically adapted to better respond to increased CO2 and air temperature conditions (Peart et al., 1989; Curry et al., 1990a,b). This chapter compares these earlier adaptations to SOYGRO V5.42 for climate change with a more mechanistic model for predicting photosynthesis and ET.