Machine learning for long-term streamflow prediction and hydrological flood and drought analysis for the Sesan Basin under climate change | |
| Author | Nuttapon Ratchakom |
| Call Number | AIT Thesis no.WM-24-12 |
| Subject(s) | Machine learning Streamflow--Forecasting Drought forecasting Hydrologic models |
| Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Water Engineering and Management, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Thesis; no. WM-24-12 |
| Abstract | Floods and droughts are among the costliest natural disasters, causing extensive damage to people, properties, economies, and environment globally. Accurate prediction of future streamflow is crucial for assessing these events and implementing effective mitigation strategies. With advancements in technology and data availability, Machine Learning (ML) has appeared to be a promising tool for streamflow prediction. This study aims to make long-term streamflow predictions and analyze characteristics of future floods and droughts in the Sesan Basin using ML and a Reservoir Operation Model to account for dam operations. First, General Circulation Models (GCMs) undergo a bias-correction process applying a linear scaling method to project future climate scenarios for the Sesan Basin. An LSTM model was then developed to simulate historical streamflow from observed precipitation, daily minimum temperature, and daily maximum temperature. The trained LSTM model was then used to predict future streamflow under medium scenario SSP245 and extreme scenario SSP585. These flow predictions were input into the HEC-RESSIM model to assess the impact of dam operations. Flood and drought indicators were calculated to describe future flood and drought characteristics. The results indicate increased rainfall and higher temperatures in the Sesan Basin, with considerable changes in streamflow patterns. Despite higher precipitation, streamflow is expected to decrease. The Lower Sesan 3 dam will substantially alter seasonal flow patterns, resulting in higher flow during dry months and lower flow during wet months. Floods are projected to have lower magnitude, frequency, and duration, while droughts are expected to become more severe and prolonged. The presence of the dam can mitigate some drought impacts by maintaining environmental flows. These findings can be beneficial for future water resource management and policymaking for the Sesan Basin and similar regions. |
| Year | 2024 |
| Corresponding Series Added Entry | Asian Institute of Technology. Thesis; no. WM-24-12 |
| Type | Thesis |
| School | School of Engineering and Technology |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Water Engineering and Management (WM) |
| Chairperson(s) | Shrestha, Sangam; |
| Examination Committee(s) | Shanmugam, Mohana Sundaram;Natthachet Tangdamrongsub; |
| Scholarship Donor(s) | Royal Thai Government;AIT Scholarships; |
| Degree | Thesis (M. Eng.) - Asian Institute of Technology, 2024 |