Date of Award
2026
Degree Type
Dissertation
Degree Name
Doctor of Philosophy in Civil and Environmental Engineering
Specialization
Environmental Engineering
Department
Civil and Environmental Engineering
First Advisor
Ali S. Akanda
Abstract
Land-to-ocean carbon (C) transport is an important pathway in the global C cycle, particularly in river-estuary systems where terrestrial inputs are transported and transformed exported to the ocean. Hydroclimatic change and anthropogenic disturbances increasingly shape coastal water-quality patterns, yet consistent and continuous monitoring remains limited in many developing regions, particularly for C-related biogeochemical parameters. This limitation is especially important in Asia, where large river systems contribute substantially to land-to-ocean C transport. In this dissertation, organic C (OC), including dissolved OC (DOC) and particulate OC (POC), was used as C-cycling indicators to evaluate how hydroclimatic variability and anthropogenic disturbance can influence river-estuary C transport across Asia.
A synthesis of 1,593 DOC observations from 40 published studies showed that DOC concentrations were the highest in tropical rainforest regions and above 40°N, highlighting that DOC behavior in Asia cannot be explained by uniform assumptions across basins and climate zones. While most existing observations are based on short-term datasets, the impacts of hydroclimatic change on C transport remain uncertain and require longer-term observation frameworks. To extend this analysis, six major Asian river basins spanning temperate (Yellow River, YL; Yangtze River, YG), subtropical monsoon (Ganges-Brahmaputra, GB; Mekong, MK), and tropical rainforest climates (Barito, BR; Rikomanggar, RM) were selected.
The estuarine zone was selected as the study area due to its larger water surface for remote sensing while representing the final interface where riverine C enters and interacts with the marine environment. MODIS-Aqua ocean-color products were used to retrieve estuarine DOC and POC within fixed plume regions defined from concentration gradients. Validation showed underestimation biases of -25% for DOC and -11% for POC, likely due to coastal optical complexity. Across basins, total OC (TOC) flux rankings were largely controlled by river discharge, while TOC concentrations showed catchment-specific lag responses, indicating contrasting hydroclimatic effects such as dilution in some systems and flushing in others.
Using two decades of satellite-derived TOC, this dissertation further evaluated the influence of climate extremes across different climate regimes using World Meteorological Organization (WMO) climate-extreme indices. Our random forest (RF) model constructed with the selected climate extremes and discharge, where the dominant predictors varied by region. The temperate catchment was mostly influenced by general warming temperature and increased summer days, subtropical monsoon catchment by antecedent wet days and rainfall intensity, and tropical rainforest catchment by discharge and rainfall intensity. Seasonal models showed the best predictability during the dry season for tropical rainforest catchments (test R² = 0.494; cross-validation R² = 0.529) and during the wet season for subtropical monsoon catchments (test R² = 0.376; cross-validation R² = 0.204), while temperate catchments performed moderately in the all-year model. These results indicate that future climate extremes may produce different OC responses across catchments depending on their hydroclimatic regimes and the evolution of their climate extremes.
Finally, land-cover change and forest disturbance were evaluated in four Kalimantan catchments: Kapuas, Barito, Mahakam, and Rikomanggar, which contains large C stocks and are undergoing rapid land conversions. GLAD forest-loss data were coupled with TOC at the plume, where plume extent was determined using the light attenuation coefficient (Kd490), generating dynamic plume geometry. Our results showed that forest disturbance affected plume geometry more clearly than plume optical intensity, and higher forest disturbance was generally associated with lower TOC concentration. This result contradicts the initial expectation that disturbance would produce a direct TOC pulse in estuarine waters, which suggests that forest disturbance may modify C sources and hydrological pathways, while the estuary may not capture the signal of in-stream forest disturbance directly.
Overall, this dissertation highlights the importance of comparative catchment-scale satellite observation across diverse hydroclimatic settings. Long-term ocean-color observations provide a useful baseline for evaluating estuarine OC variability in underrepresented catchments, but satellite-derived TOC should be interpreted as an optical indicator of concentration and flux rather than a direct measurement of C source, composition, or fate. Future work should be paired with in-situ observations and laboratory analysis of OC composition, to validate the satellite derived OC products, as well as to determine whether OC captured by satellite is river- or marine-derived. Integrating higher-spatial-resolution satellites, such as Landsat and Sentinel-2, with hyperspectral observations from missions such as PACE and EMIT could improve monitoring of estuarine regions, where optical complexity can limit retrieval accuracy.
Recommended Citation
Septiani, Retno Wulan, "UNDERSTANDING HYDROCLIMATIC AND ANTHROPOGENIC IMPACTS ON BIOGEOCHEMICAL COASTAL WATER QUALITY INDICATORS IN ASIA: A REMOTE SENSING-BASED INVESTIGATION" (2026). Open Access Dissertations. Paper 4599.
https://digitalcommons.uri.edu/oa_diss/4599