14:00 〜 15:15
[AGE03-P12] 土壌の物理的性質データベースを用いた気候変動下の農地土壌水分状態予測
キーワード：気候変動, 土壌水分, 数値シミュレーション, SolphyJ, ELPIS-JP
Soil is foundation of agriculture and ecosystems. Soil physical condition such as soil moisture and temperature directly and indirectly affects yields and quality of crop production. Therefore predicting soil moisture of arable lands under climate change is important and valuable for yield prediction and adaptation under climate change. For predicting soil moisture condition of agricultural lands in arbitrary areas, use of soil database and datasets of General Circulation Model (GCM) projections should be useful since physical properties of soils and meteorological condition vary with location. Most of available GCM projections have spatial and temporal resolution of 100 km and a month. However, using GCM projections as input data for soil moisture and temperature prediction, temporal and spatial scale of the input data is favorable to be small since effective surface soil layer of agricultural production is generally shallow. In this study we investigated possibility of predicting soil moisture of arable lands in arbitrary areas with local-scale (approximately 20km×20km) daily GCM projection dataset “ELPIS-JP” (Iizumi et al., 2012) and the agricultural soil-profile physical properties database, Japan, “Solphy” (Eguchi et al., 2010).
In this study, soybean fields of Yoshioka and Ookubo, which are located in neighbors (approximately 2∼3km), in Toyama city were chosen as experimental sites and scenario studies were done for predicting soil moisture condition with HYDRUS model (Simunek et al., 2008) under climate change in the future (2071∼2090). Soil physical properties of each site were determined with water retention data in the SolphyJ database by using RETC program (Yates et al., 1992). Before the scenario studies, validation of HYDRUS model and soil physical properties which are obtained with SolphyJ database was conducted by comparing observed and simulated soil moisture of the Yoshioka field. The projection of MIROC-hires 3.2 A1B scenario was chosen among 26 (10 GCMs × 3 Special Report of Emission Scenario) ELPIS-JP scenario datasets. For preparing input data for numerical simulation of soil water movement, daily ELPIS-JP datasets were temporally downscaled to hourly or minutes scale by using weather generator “CLIGEN” (Nicks et al., 1995)
Simulated results suggested that the duration of excess soil moisture condition following heavy rainfall events are more likely at Ookubo than Yoshioka even though they are located in neighbors and have similar soil textures. Increase in surface runoff fluxes is possible to be larger in Ookubo than in Yoshioka as well. These results imply that even in a small watershed it is important to consider soil spatial distribution in predicting effects of climate change on agricultural production. Also, combination of temporally downscaled GCM projection dataset and agricultural soil-profile physical properties database may be useful for predicting soil moisture in arbitrary areas.
References: Eguchi et al., 2011, Proceedings of Annual Meeting of JSDIRE, 302-303; Iizumi et al., 2012, Phil. Trans. R. Soc. A, 370, 1121-1139; Nicks et al., 1995, NSERL Report ♯10, pp.2.1-2.22.; Simunek et al., 2008, Vadose Zone J. 7, 587-600; Yates et al., 1992, Soil Sci. Soc. Am. J, 56, 347-354
In this study, soybean fields of Yoshioka and Ookubo, which are located in neighbors (approximately 2∼3km), in Toyama city were chosen as experimental sites and scenario studies were done for predicting soil moisture condition with HYDRUS model (Simunek et al., 2008) under climate change in the future (2071∼2090). Soil physical properties of each site were determined with water retention data in the SolphyJ database by using RETC program (Yates et al., 1992). Before the scenario studies, validation of HYDRUS model and soil physical properties which are obtained with SolphyJ database was conducted by comparing observed and simulated soil moisture of the Yoshioka field. The projection of MIROC-hires 3.2 A1B scenario was chosen among 26 (10 GCMs × 3 Special Report of Emission Scenario) ELPIS-JP scenario datasets. For preparing input data for numerical simulation of soil water movement, daily ELPIS-JP datasets were temporally downscaled to hourly or minutes scale by using weather generator “CLIGEN” (Nicks et al., 1995)
Simulated results suggested that the duration of excess soil moisture condition following heavy rainfall events are more likely at Ookubo than Yoshioka even though they are located in neighbors and have similar soil textures. Increase in surface runoff fluxes is possible to be larger in Ookubo than in Yoshioka as well. These results imply that even in a small watershed it is important to consider soil spatial distribution in predicting effects of climate change on agricultural production. Also, combination of temporally downscaled GCM projection dataset and agricultural soil-profile physical properties database may be useful for predicting soil moisture in arbitrary areas.
References: Eguchi et al., 2011, Proceedings of Annual Meeting of JSDIRE, 302-303; Iizumi et al., 2012, Phil. Trans. R. Soc. A, 370, 1121-1139; Nicks et al., 1995, NSERL Report ♯10, pp.2.1-2.22.; Simunek et al., 2008, Vadose Zone J. 7, 587-600; Yates et al., 1992, Soil Sci. Soc. Am. J, 56, 347-354