| dc.description.abstract | The Earth Systems computer model CLASS has been used to predict land surface and atmospheric conditions, regionally, in the Netherlands. A portion of this model predicts vegetative response: the A-gs model – dealing with the relationship between photosynthesis and stomatal conductance – is this research’s focus. The A-gs model was parameterized with only one plant species, Vitis Vinifera, in one growing location, central Spain. This limited parameterization suggests the model may not predict plant response accurately for other species, from other Plant Functional Types (PFTs) (De Kauwe et al., 2015; Kala et al., 2016). Through this research, the model’s ability to predict experimental net photosynthesis data is considered. Experimental data were derived from gas exchange experiments on two common Dutch C3 grasses: Triticum aestivum and Alopecurus pratensis. Net photosynthesis was examined in response to changes in light intensity (PAR) and intercellular CO2 concentration (Ci), respectively.
Statistical analysis (through t-tests) revealed that the A-gs model predicted photosynthesis values which were, overall, significantly different than observed experimental photosynthesis. The A-gs model predicted the photosynthetic response to light and Ci in T. aestivum better than in A. pratensis. The model predicted observed values of T. aestivum reasonably well. To this point: the significant differences resulting from t-tests is thought to be due to low sample size rather than strong variation between data. Therefore, a more robust experimental analysis of T. aestivum is recommended, in terms of both A-PAR and A-Ci response curves, using increased sample size.
It was expected, due to the similarity of the grass species studied, that the model would generate data which either explained the photosynthetic response of both grasses, or not. However, this was not the case. This result may be mitigated by species-specific parameterization (Prentice, et al., 2015). However, due to similarity of studied grasses, this divergent result more likely points to oversensitivity in the A-gs model to certain environmental factors, i.e. light. Different light levels were used in the A-Ci experiments on T. aestivum and A. pratensis, and in the corresponding modeled environments. The low-light modeled photosynthetic response curve (used in the A-Ci A. pratensis simulation) reached Amax at a low Ci and was shaped more logistically than commonly seen in literature (During, 1991; Greer, 2012; Jacobs, 1994) or in experiments. Under saturating light conditions (used in the A-Ci T. aestivum simulation) the model generated an A-Ci curve closer to curves observed experimentally and seen in literature. | |
