10:45 〜 12:15
[ACG39-P05] Estimation of leaf chemical composition using Spatially Explicit Individual Based Dynamic Global Vegetation Model
キーワード：SEIB-DGVM、総フェノール
To protect themselves from herbivorous insect and pathogen infection, trees produce a variety of defenses as secondary metabolites from non-structural carbohydrates (NSC), a surplus photosynthetic product. The spatiotemporal distribution of the concentrations of these defensive compounds is thought to affect forest population dynamics and carbon cycling through changes in tree mortality rates, and quantitative estimation is important. Modeling is considered necessary to estimate changes in forest dynamics due to climate change expected in the future. In this study, we used SEIB-DGVM-NSC v1.0 (Ninomiya et al., submitted), a dynamic global vegetation model that reproduces forest population dynamics and carbon cycle with an NSC estimation module added to original SEIB-DGVM (Sato et al., 2007) to model leaf cover mass of deciduous broadleaf trees.
SEIB-DGVM-NSC v1.0 is a model that can handle the interactions between the climatic environment and the plant ecosystem and the feedback effects of the distribution and structure of the plant ecosystem on the climatic environment. In addition, the design of the model, in which trees grow in grid boxes, allows for the reproduction of localized interactions between individuals. This was also chosen because the nature of this study lends itself to modeling NSC and then estimating the amount of defensive material. This vegetation model uses climate and soil data as input and outputs both short-term responses of vegetation, such as photosynthesis and respiration, and long-term responses, such as biomass and ecosystem distribution, through the physical, physiological, and ecological modules. From the output photosynthetic and respiration rates, it also calculates the amount of carbon synthesized by the plants as non-structural carbon. Quercus crispula, Acer mono ver. glabrum, Ostrya japonica Sargent, and Acer amoenum Carr. Var. matsumurae in the Tomakomai Experimental Forest (42°44'N, 141°31'E) was selected for this study. The measured data of chemical defensive substances in leaves used in the model was based on leaf traits measured from 2011-2014 (Nakaji et al., 2019). Among the chemical compounds in leaves, total phenols and condensed tannins and lignin were modeled. Harmonized meteorological data from field observations (JaLTER site), AMeDAS, and NOAH global climate models for the Tomakomai Research Forest for 1901-2018 were input to SEIB-DGVM-NSC ver. 1.0 to simulate the amount of NSC in the Tomakomai Experimental Forest. We also analyzed the total phenolic and condensed tannin content of leaves measured during 2011-2014 on a leaf weight basis. Approximating and distributing these NSC amounts to total phenol and condensed tannin and lignin amounts, for example, in Quercus crispula, total phenol was about 64.3% of the NSC amount and condensed tannin was about 6.9% of the NSC amount, and lignin was about 70.8% of the NSC amount.
As a result, both measured leaf compositions showed seasonal changes, as did the modeled NSCs. Of these, for condensed tannins and total phenols, the shape of the graph of seasonal changes could be reproduced by the modeling. Lignin, on the other hand, showed the exact opposite seasonal change from these two in terms of percentage in leaves, and could not be modeled well in this study. I believe this is because lignin is used in the composition of the cell wall itself, in addition to its role as a chemical defense. In addition, in common with all substances, modeling of the entire NSC in SEIB-DGVM-NSC ver. 1.0 included a small amount of evergreen trees in the Tomakomai Research Forest, so modeling in terms of percentages in leaves did not reproduce winter defoliation. In addition, modeling during the young and dead leaf periods was also inaccurate. To improve the accuracy of this modeling, the simulation should be linked to weather factors such as temperature, precipitation, and carbon dioxide, rather than simply assigning the amount of NSC.
SEIB-DGVM-NSC v1.0 is a model that can handle the interactions between the climatic environment and the plant ecosystem and the feedback effects of the distribution and structure of the plant ecosystem on the climatic environment. In addition, the design of the model, in which trees grow in grid boxes, allows for the reproduction of localized interactions between individuals. This was also chosen because the nature of this study lends itself to modeling NSC and then estimating the amount of defensive material. This vegetation model uses climate and soil data as input and outputs both short-term responses of vegetation, such as photosynthesis and respiration, and long-term responses, such as biomass and ecosystem distribution, through the physical, physiological, and ecological modules. From the output photosynthetic and respiration rates, it also calculates the amount of carbon synthesized by the plants as non-structural carbon. Quercus crispula, Acer mono ver. glabrum, Ostrya japonica Sargent, and Acer amoenum Carr. Var. matsumurae in the Tomakomai Experimental Forest (42°44'N, 141°31'E) was selected for this study. The measured data of chemical defensive substances in leaves used in the model was based on leaf traits measured from 2011-2014 (Nakaji et al., 2019). Among the chemical compounds in leaves, total phenols and condensed tannins and lignin were modeled. Harmonized meteorological data from field observations (JaLTER site), AMeDAS, and NOAH global climate models for the Tomakomai Research Forest for 1901-2018 were input to SEIB-DGVM-NSC ver. 1.0 to simulate the amount of NSC in the Tomakomai Experimental Forest. We also analyzed the total phenolic and condensed tannin content of leaves measured during 2011-2014 on a leaf weight basis. Approximating and distributing these NSC amounts to total phenol and condensed tannin and lignin amounts, for example, in Quercus crispula, total phenol was about 64.3% of the NSC amount and condensed tannin was about 6.9% of the NSC amount, and lignin was about 70.8% of the NSC amount.
As a result, both measured leaf compositions showed seasonal changes, as did the modeled NSCs. Of these, for condensed tannins and total phenols, the shape of the graph of seasonal changes could be reproduced by the modeling. Lignin, on the other hand, showed the exact opposite seasonal change from these two in terms of percentage in leaves, and could not be modeled well in this study. I believe this is because lignin is used in the composition of the cell wall itself, in addition to its role as a chemical defense. In addition, in common with all substances, modeling of the entire NSC in SEIB-DGVM-NSC ver. 1.0 included a small amount of evergreen trees in the Tomakomai Research Forest, so modeling in terms of percentages in leaves did not reproduce winter defoliation. In addition, modeling during the young and dead leaf periods was also inaccurate. To improve the accuracy of this modeling, the simulation should be linked to weather factors such as temperature, precipitation, and carbon dioxide, rather than simply assigning the amount of NSC.