Date of Award
2022
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Atmospheric and Earth Science
Committee Chair
Robert Griffin
Committee Member
Vikalp Mishra
Committee Member
Sundar Christopher
Committee Member
Amanda Markert
Subject(s)
Flood control, Flood damage, Machine learning, Floods--Mekong River Valley--Case studies
Abstract
Cambodia suffers from flooding and flash-flooding events every year, and thus better flood management and forecasting are crucial in taking early action. The Streamflow Prediction Tool (SPT) provides forecast discharge 15 days in advance globally. This thesis applies data-driven and Machine Learning (ML) techniques to improve Cambodia’s hydrological applications. Two case studies are presented – the first predicts the river discharge from the water level using time-series information; the second uses the predicted discharge from the first to bias-correct the SPT’s historical reanalysis data using several ML techniques. The historical Bias Correction (BC) is the first step toward the forecasted BC. ML methods can produce a generalized model that can be used to BC the SPT historical reanalysis in the ungauged station. The BC products have other applications: analyze long-term trends, generate local BC forecasts, and use during the design phase of the forecast-based early action.
Recommended Citation
Bhandari, Biplov, "Applying data science and machine learning techniques for hydrological applications : case study in Cambodia" (2022). Theses. 604.
https://louis.uah.edu/uah-theses/604