The Asat/E response to VPD is shown in Figure 5.6. Asat/E showed a strong response to VPD. The
comparison of modelled and measured Asat/E using estimated g1 and k parameters is shown in Figure
5.7. Estimated parameter values of field grown cotton were g1 = 4.35 (95% confidence interval (CI) =
4.24 - 4.47) and k = 0.59 (95% CI = 0.53 – 0.64). In addition, for the comparison of modelled and measured Asat/E for field data MAD= 0.546, whereas using the glasshouse data prediction MAD= 0.551.
Figure 5.6: Asat/E response to VPD of “well-watered” field-grown cotton. Black solid line represents model fit
using g1 and k estimates from field data. Blue dashed line represents g1 and k model prediction based on cotton
grown in the glasshouse.
Figure 5.7: Comparison of modelled and measured Asat/E using Eq. (2) where g1 and k parameters are from (a)
field data and (b) glasshouse data prediction from Duursma et al. (2013). Also shown are the 1:1 lines (black). (a) RMSE= 0.714; MAD= 0.546 and (b) RMSE does not apply; MAD= 0.551.
5.4. Discussion
Environmental conditions in a field can greatly influence crop physiology and yield (Pettigrew et al., 1990), and therefore it is important to assess the impact of the environment on physiology of field- grown cotton as warmer temperatures, changes in rainfall distribution and altered VPD are projected in the future for Australian cotton regions. In this study, we found that increased VPDL may reduce
stomatal conductance in field-grown cotton; that variation in stomatal conductance and photosynthetic rates can be explained by changes in growth conditions and consequently variables that describe environmental factors, such as VPD; and that the Asat/E (ITE) model developed using
cotton grown in glasshouse conditions can be used to estimate Asat/E of field-grown cotton.
In this study, a large proportion of variation in gs-sat was accounted for by the VPD environment. We
found that VPDL alone accounted for 32.3 and 39.5% of the variation in gs-sat for the complete and
ambient gas exchange measurements, respectively. Similar to numerous other studies (Duursma et al., 2013; Oren et al., 1999), our data showed a general decline in gs-sat with increased VPD. Our study
highlights that although VPDL accounts for a large proportion of the cumulative variation in gs-sat, there
were still a number of other variables that influenced variation in stomatal response, including the plant (4.3%), Tl-Ta (32.8%) and Plant x Tl-Ta (0.8%) interactions. Nonetheless, we could only account
for c. 70% of variation in gs-sat. Therefore, 30% of the variation in gs-sat is due to something that we
either did not measure or analyse. For example, Duursma et al. (2013) developed models to describe the stomatal response to environmental factors of cotton grown in the glasshouse, where conditions such as growth temperatures were highly controlled, unlike in the field. Variables that were accounted for included VPD, assimilation rate and atmospheric [CO2] (Duursma et al., 2013). Therefore, when
these models are used for field-based studies, there may be unexplained variation depending on the antecedent growth conditions of the crop, which may include factors such as nutrient status of an individual leaf, and leaf angle affecting light-interception.
For photosynthetic responses, VPDL accounted for only 16.8% of the variation in the complete dataset,
and accounted for 28.9% of the variation in photosynthetic rates for the ambient gas exchange measurements. Tl-Ta and ASH were also important factors for plant photosynthetic response. Adding
Tl-Taincreased the variation accounted for by 15.4% in the complete dataset but was not a significant
variable in the ambient dataset. In addition, adding the variable ASH increased the variation accounted for by 6.3% in the complete dataset and by 17.1% in the ambient dataset. The addition of Plant x Tl-Ta
and VPDL x ASH were significant interactions for photosynthetic response in the complete dataset,
whereas the VPDL x Tl-Ta interaction was significant in the ambient dataset. Other studies reported a
these experiments temperatures were generally held constant during the study. Similarly, Duursma et al. (2013) found that photosynthesis was relatively insensitive to VPD, with a 13% decrease in maximum photosynthesis over 1 - 4 kPa, but reported higher photosynthetic rates of cotton grown at warmer air temperatures resulting in a higher transpiration rate at a given VPD, again highlighting the impact of temperature effects on photosynthesis. Therefore, in comparison, these studies allowed a better identification of the direct effects of VPD, but not allowed the independent effect of temperature to be observed. Given that both Tl-Ta and ASH have accounted for variation in
photosynthetic rates, this highlights the importance of how warmer temperatures may affect photosynthesis of cotton grown in future, warmer climates, regardless of the small direct impact of VPD on photosynthesis.
The Asat/E model fit to the field data suggests that the g1 and k parameters used in the glasshouse can
also be used to estimate Asat/E in the field. Therefore, this indicates that the Asat/E model developed
using cotton grown in the glasshouse is also applicable to cotton grown in the field and highlights that controlled environment glasshouse studies can be successfully utilised to further our understanding of leaf-level physiological responses to environmental conditions. In addition, these studies are useful when attempting to scale from leaf to canopy level responses. Thus, this improves our ability to predict the effect of climate change on crop water use efficiency (Duursma et al., 2013). However, limitations were that although plants were grown in the field, Asat/E was measured using the cuvette of the Licor,
where wind speeds, and thus boundary layer conductance, were high (Grantz and Vaughn, 1999). Boundary layer conductance can affect leaf temperature, and transpiration rates at a given stomatal conductance, and therefore may not represent actual gas exchange in the field. Therefore, the combination of canopy and leaf-level measurements may be the most useful in describing cotton response to the environment. However, the success in using the Asat/E model in both glasshouse and
field-grown cotton is promising for the validation of other simulation models. For example, the OZCOT cotton crop simulation model currently does not account for physiological changes in canopy photosynthesis or transpiration in response to VPDL. Therefore, a better understanding of the
physiological responses may improve our predictions of growth and water use, especially with the simulation of future environments.
5.4.1. Conclusions
VPDL accounted for a large proportion of the variation in gs-sat and photosynthesis, with smaller
percentages attributed to other factors such as the individual plant, Tl-Ta, ASH and VPDL x Tl-Ta, Plant x
Tl-Ta and VPDL x ASH interactions. Using generalised linear models, c. 70% of variation in gs-sat was
variation in photosynthetic rate was accounted for by VPDL, Tl-Ta, VPDL x Tl-Ta, Plant x Tl-Ta and VPDL x
ASH interactions. However, a proportion of the variation in gs-sat and photosynthesis were not
explained by these measurements.
Data from this study can be used for Asat/E models and can potentially can be used to inform crop
simulation models to account for possible impacts of climate change on crop production. In conjunction with information of cotton canopy temperature response (Conaty et al., 2014), a better understanding of VPD may aid our understanding of physiological responses of field-grown cotton and lead to better mangement of cotton production in future environments.