5:15 PM - 6:30 PM
[ACG37-P09] Development of process-based model for estimating wheat production MATCRO-Wheat
Keywords: Process-based model, climate change, vernalization
Crop yield has been significantly influenced by climate change (Schindeler et al .2001). In recent years, there is a concern that crop yields will decline due to rising temperatures and extreme weather events caused by global warming. Under the assumption that rapid population growth could cause food shortages in near future, therefore, many crop models have been developed though most of them are statistical models or site-based models. Statistical model is generally not able to show each process in the interaction between climate change and crops because they do not take into account the material cycle based on plant physiology. On the other hand, though site-based model has the elaborate representation of crop growth, they rarely have been applied at global scales (Zhang et al .2017), probably due to lack of high precision input data and representative parameter sets. Therefore, this study aims to develop a process-based crop model, extended version of Minimal Advanced Treatment of Surface Interaction and Runoff (MATSIRO; Takata et al., 2003) for wheat yield estimation at a global scale, named here as MATCRO-Wheat.
To estimate the wheat yields, MATCRO-Rice model (Masutomi et al. 2016) for rice yield, was modified and added new function specialized for wheat. First, the model parameters were newly set for wheat and a vernalization process was newly introduced. Second, the model was validated with observational data in 5 field sites: Memuro, Hokkaido in Japan (Hokkaido Agricultural Research Center), Nanjing in China (Nanjing Agricultural University), Wongan Hills in Australia, Valcarce in Argentina, and Wanningen in Netherland (AgMIP wheat for last three sites; Benchmark data set for wheat growth models: field experiments and AgMIP multi-model simulations). Model parameters were adjusted to match the modeled biomass evolution and yield to the measured ones as much as possible. Finally, MATCRO-Wheat was run at global scale with adjusted parameter with vernalization process with the spatial resolution of 0.5 degree x 0.5 degree and with the simulation period from 2009 to 2010. Yield per unit land area was calculated by weighting rain-fed and irrigated yields according to the fraction of irrigated area which is drive from AQUASTAT (FAO. AQUASTAT – Global Map of Irrigation Areas), and were validated with the national yields survey according to the GDHY-aligned version v1.2+v1.3 dataset (Iizumi et al .2020).
Results showed that wheat yields were generally well represented of field observations. In the Netherland and Japan, the overestimation of crop growth was improved by introducing vernalization treatment during winter. The global simulation indicated that wheat yields were overestimated in higher latitudinal zones and in summer wet regions. These would be caused as much photosynthate was produced because MATCRO-Wheat does not consider vernalization and growth restriction due to high humidity. Therefore, further improvement of MATCRO-Wheat is required for accurate representation of wheat growth and yield.
To estimate the wheat yields, MATCRO-Rice model (Masutomi et al. 2016) for rice yield, was modified and added new function specialized for wheat. First, the model parameters were newly set for wheat and a vernalization process was newly introduced. Second, the model was validated with observational data in 5 field sites: Memuro, Hokkaido in Japan (Hokkaido Agricultural Research Center), Nanjing in China (Nanjing Agricultural University), Wongan Hills in Australia, Valcarce in Argentina, and Wanningen in Netherland (AgMIP wheat for last three sites; Benchmark data set for wheat growth models: field experiments and AgMIP multi-model simulations). Model parameters were adjusted to match the modeled biomass evolution and yield to the measured ones as much as possible. Finally, MATCRO-Wheat was run at global scale with adjusted parameter with vernalization process with the spatial resolution of 0.5 degree x 0.5 degree and with the simulation period from 2009 to 2010. Yield per unit land area was calculated by weighting rain-fed and irrigated yields according to the fraction of irrigated area which is drive from AQUASTAT (FAO. AQUASTAT – Global Map of Irrigation Areas), and were validated with the national yields survey according to the GDHY-aligned version v1.2+v1.3 dataset (Iizumi et al .2020).
Results showed that wheat yields were generally well represented of field observations. In the Netherland and Japan, the overestimation of crop growth was improved by introducing vernalization treatment during winter. The global simulation indicated that wheat yields were overestimated in higher latitudinal zones and in summer wet regions. These would be caused as much photosynthate was produced because MATCRO-Wheat does not consider vernalization and growth restriction due to high humidity. Therefore, further improvement of MATCRO-Wheat is required for accurate representation of wheat growth and yield.