2026-05-24T16:20:52Zhttps://keep.lib.asu.edu/oai/request

oai:keep.lib.asu.edu:node-1982412024-12-23T18:01:48Zoai_pmh:alloai_pmh:repo_items

198241 https://hdl.handle.net/2286/R.2.N.198241 http://rightsstatements.org/vocab/InC/1.0/ All Rights Reserved 2024 133 pages Doctoral Dissertation Academic theses Text eng Wei, Shiqi Xu, Tianfang TX Garcia, Margaret MG Niu, Guo-Yue GN Wang, Zhihua ZW Zeng, Ruijie RZ Arizona State University Partial requirement for: Ph.D., Arizona State University, 2024 Field of study: Civil, Environmental and Sustainable Engineering Global population growth and rapid urbanization are intensifying water demand. With a large portion of freshwater already allocated for agricultural and urban irrigation, even more water will be required in the future. On the other hand, irrigation—an activity that redistributes water across the Earth's surface—disrupts the natural hydrological cycle by increasing evapotranspiration, altering temperatures, and affecting atmospheric moisture, ultimately leading to abnormal local and regional climate patterns. Despite this, challenges remain in monitoring and optimizing irrigation practices. Machine learning, known for its ability to analyze complex, high-dimensional data and capture non-linear relationships, has gained particular attention for studying interactions between natural and human systems in recent years. This dissertation focuses on leveraging machine learning approaches to monitoring irrigation water use and optimize irrigation practices, supporting climate-resilient irrigation and sustainable water resources management. The first and second sections address the lack of large-scale, ground-truth, and detailed information on the timing and quantity of agricultural irrigation. By utilizing remote sensing data associated with irrigation operations, along with ground-truth records such as groundwater levels and annual pumping amounts, this study presents a remote-sensing-informed machine learning approach to estimate daily irrigation water use. This information can not only support policymakers in local water management but also benefit the modeling community by improving the representation of human activities in models. The third section builds on the previous two chapters to explore irrigation decision-making. Using urban irrigation in arid areas as an example, this section proposes a coupled simulation-optimization framework, with machine learning as a fast surrogate model of a process-based urban canopy model, to support localized, climate-resilient irrigation practice, aimed at mitigating heat while optimizing water conservation. This flexible framework can be tailored to specific needs, lowering barriers to support the irrigation related planning and decision-making. Water resources management Agriculture Irrigation Machine learning Optimization Remote Sensing Towards Sustainable Irrigation Management: a Machine Learning Approach to Monitoring and Optimization