Using Bayesian network to build a participatory framework for sustainable agriculture water management in Yunlin, Taiwan
- Department of Bioenvironmental Systems Engineering, National Taiwan University, Taipei, Taiwan
Agricultural water use comprises the major part of the total water consumption in many countries, and Taiwan is no exception. However, urbanization and industrialization have triggered the competition for water among different sectors. Water is transferred to satisfy the daily need and industrial need, especially the need of high-tech industries, from the agricultural sector. Groundwater hence becomes an alternative water resource for agriculture, but the over-exploitation of groundwater resources also leads to some problems such as environmental degradation and land subsidence, and climate change has worsened the situation in the recent years.
In Taiwan, groundwater is one of the vital water resources for irrigation, especially when the first crop rice begins being cultivated in the late dry season in central Taiwan. Yunlin County located in central Taiwan is chosen as the study area, which is now facing severe issues about groundwater over-exploitation and suffering from land subsidence threatening the safety of Taiwan High Speed Rail. Because of the high water consumption, groundwater extraction from agriculture is deemed to be the major cause of the land subsidence and should be well monitored and reduced. However, farmers’ pumping behaviors are highly related to the national water allocation policy, food policy and the socioeconomic factors in the rural area. The top-down agricultural water management might not be sufficient and sustainable. Hence, in this study, we propose a participatory framework for agricultural water management using a Bayesian network. The framework tries to incorporate the main factors that affect decision making among different stakeholders, including the Water Resources Agency, Irrigation Agency, Agriculture and Food Agency, farmers, etc., and represent the causal relationship among factors through Bayes’ theorem, or the conditional probability tables (CPTs). The CPTs are constructed based on data, literature reviews and interviews with stakeholders. The key issues concerning different stakeholders are considered in the framework as well, such as surface water shortage for agriculture, land subsidence, and sustainability of agriculture in Yunlin. The network can be used to hold discussions with stakeholders and show the interactions of their decisions among others. The aim of this framework is to facilitate the discussions and formulate the strategies for sustainable agricultural water management with the aid of the intuitive and transparent structure of the Bayesian network.
How to cite: Lee, S.-Y. and Yu, H.-L.: Using Bayesian network to build a participatory framework for sustainable agriculture water management in Yunlin, Taiwan, EGU General Assembly 2023, Vienna, Austria, 24–28 Apr 2023, EGU23-15459, https://doi.org/10.5194/egusphere-egu23-15459, 2023.