Graduation Semester and Year
Spring 2026
Language
English
Document Type
Thesis
Degree Name
Doctor of Philosophy in Civil Engineering
Department
Civil Engineering
First Advisor
Dr. Habib Ahmari
Second Advisor
Dr. Timothy M. Dellapenna
Third Advisor
Dr. Srinivas Prabakar
Fourth Advisor
Dr. Jessica Eisma
Abstract
Accurate quantification of suspended sediment concentration (SSC) is fundamental to understanding sediment flux and water quality across the river–coast continuum. A robust, multi-scalar remote sensing framework is developed in this dissertation by leveraging machine learning and hybrid modeling to retrieve SSC from satellite-derived surface reflectance across Texas river reaches, spanning fluvial sections, transitional zones, and coastal areas.
The study utilizes the major Texas rivers that deliver the highest sediment loads to the Texas coast, leveraging their naturally variable sediment regimes to ensure the developed models account for diverse hydrologic and optical complexities.
First, a regional prediction framework was developed for inland rivers, demonstrating that high-spatial-resolution sensors enhance operational efficiency by capturing site-specific variability across lower river basins. This regional approach offers a powerful tool for monitoring areas where field data is unavailable or limited, streamlining the SSC estimation process without the need for extensive localized tuning.
Second, for the coastal zone, a framework utilizing high-temporal resolution sensors was developed to mitigate atmospheric noise and biological interference. The model demonstrated promising cross-regional transferability by capturing sediment dynamics along the Louisiana coast and has the potential to be applicable to similar river-dominated coastal systems.
Finally, to address the data-limited nature of the river–coast transition zones, a hybrid strategy was developed using a fusion of high-spatial and high-temporal resolution satellite data. This approach outperformed existing global models by overcoming the lack of in-situ training data in dynamic river-coast transitional regions by representative environmental datasets. A critical finding is the framework’s exceptional performance during extreme flood events, proving its capacity to quantify massive sediment pulses associated with climate extremes.
This research confirms that while increased model generalization improves efficiency, integrating region-specific variables ensures that localized precision is maintained. Ultimately, this work provides a transferable methodology for establishing regional models globally, offering a scalable and efficient solution for monitoring sediment transport from river mouths to the open ocean.
Keywords
Suspended Sediment Concentration, Remote Sensing, Machine Learning, Random Forest, Regional Modeling, Sediment Transport, River–Coastal Systems
Disciplines
Civil and Environmental Engineering | Civil Engineering | Water Resources Engineering
License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.
Recommended Citation
Khandaker, Nabila, "ESTIMATION OF SUSPENDED SEDIMENT CONCENTRATION IN TEXAS RIVERINE, COASTAL, AND TRANSITION ZONES USING MULTI-SENSOR SATELLITE DATA AND MACHINE LEARNING ALGORITHM" (2026). Civil Engineering Dissertations. 4.
https://mavmatrix.uta.edu/civilengineering_dissertations2/4