An Integrated Conceptual Model Towards Sustainable Rural Water Management Based Remote Sensing and Machine Learning
- Publication Type:
- Thesis
- Issue Date:
- 2022
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
Recently, there have been some improvements in agricultural water supply systems. However, rural areas still face serious water deficiencies including droughts, poor water quality and floods due to inappropriate of water management systems and climate change. This critical issue points out the urgent need for developing an effective integrated rural water model, which can improve water monitoring in rural regions. This study therefore aims to develop the integrated conceptual model for rural sustainable water monitoring to help rural communities overcome the issues of water run-off, water pollution and lack of water for agricultural production.
This thesis presents a novel conceptual model framework including three sub-models (water vulnerability quantity assessment model, soil moisture prediction model, and agricultural soil organic carbon model for supporting rural water modelling using the integration of free-of-charge satellite images including MODIS, Sentinel 1, Sentinel 2, and ALOS DSM imagery and different advanced machine learning algorithms. The framework firstly demonstrates a new approach of water quantity vulnerability assessment based on reliable and updated spatial-temporal datasets (soil water stress, aridity index, rain use efficiency and leaf area index), and the incorporation of the GIS-based model. Notably, this research devises a state-of-the-art machine-learning model for monitoring agricultural drought via predicting soil moisture (SM) using active and ensemble-based decision tree learning combined with multi-sensor data fusion at a national and world scale. This work explores the use of Sentinel-1, Sentinel-2, and an innovative machine learning (ML) approach using an integration of active learning for land-use mapping and advanced Extreme Gradient Boosting Regression (XGBR) for robustness of the SM estimates. The collected soil samples from a field survey in Western Australia were also used for the model validation and indicators including the coefficient of determination (R²) and root - mean – square - error (RMSE) were applied to evaluate the model’s performance. The proposed model XGBR with 21 optimal features obtained from GA was yielded the highest performance (R² = 0.891, RMSE = 0.875%). A combination of S1 and S2 sensors could also effectively estimate SOC in farming areas by using ML techniques. Satisfactory accuracy of the proposed XGBoost with optimal features was achieved the highest performance (R² = 0.870; RMSE = 1.818 tonC/ha) which outperformed random forest and support vector machine.
Conclusively, the described conceptual model can further support precision agriculture, water management, and drought resilience programs via water use efficiency, green infrastructure, and smart irrigation management for agricultural production.
Please use this identifier to cite or link to this item: