https://hdl.handle.net/10012/22529

Inland freshwater bodies, including lakes, reservoirs, and wetlands, are critical components of the global freshwater system. They play essential roles in water storage, flow regulation, biodiversity, nutrient cycling, and many other ecosystem services including food production and local climate regulation. However, human interventions such as dam construction, urbanization, intensive agriculture, and climate change significantly alter their hydrological and ecological functions. This underscores the importance of data integration and modeling tools to predictively understand and effectively manage these water bodies to ensure their sustainability and resilience under changing climate and environmental conditions. With this thesis, I aim to contribute to the comprehensive, global-scale analysis of inland water bodies. I present a new global reservoir water modeling database, abbreviated GRM, that can be coupled with hydrodynamic and water quality simulations. The use of GRM is illustrated by computing temperature changes in nearly 7000 large reservoirs between 1980 and 2019. I also present a database of depth-area-volume (D-A-V) relationships for over 1.4 million lakes and reservoirs worldwide. These relationships provide easy access to essential bathymetric information to users interested in carrying out modeling studies. The D-A-V database is complemented by a Python package that generates bathymetric representations for multidimensional water quality modeling. Lastly, for a reservoir in southern Ontario, I analyze how controlling water level, and the positioning of dam outflow gates can be used to reduce the outflow of total and bioavailable phosphorus, which, more generally, opens the possibility of considering dam operation strategies that help protect downstream water bodies from eutrophication impacts. Chapter 1 provides an overview of the significance of inland water bodies and the impacts of anthropogenic activities on their biogeochemical dynamics. The chapter reviews existing global databases on lakes and reservoirs, highlighting their strengths and limitations. I further argue that existing global-scale biogeochemical modeling studies of inland waters have primarily relied on simple box models and empirical relationships that lack the ability to capture the complex temporal and multidimensional physical-geochemical-biological interactions in these ecosystems. This gap sets the stage for developing the comprehensive global multidimensional model database presented in Chapter 2. Chapter 2 describes the development of the Global Reservoir Modeling (GRM) database that integrates multiple existing global datasets to facilitate reservoir hydrodynamic and water quality modeling on a global scale. The current GRM database version brings together 40 years (1980-2019) of diverse data series for nearly 7,000 reservoirs worldwide. The corresponding data are extracted from the following datasets: GRanD for reservoir attributes, ReGeom for bathymetric data, WaterGAP for streamflow, and ERA5 for meteorological parameters. With these data, GRM can generate compatible input files for hydrodynamic and water quality simulations with the popular CE-QUAL-W2 model. Thus, GRM offers researchers a practical and readily usable tool to model changes in reservoir water temperature and mixing regimes and their impacts on water quality, whether for a single or a large selection of the GRanD reservoirs. Chapter 3 offers an example of the type of global-scale assessments that can be performed with GRM by calculating the temporal trajectories of the thermal gradients in all the reservoirs included in GRM from 1980 to 2019. For each reservoir, a 30×30 depth-length bathymetry is generated by GRM, which is then used in the multithreaded, process-based CE-QUAL-W2 model to calculate the temperature distribution as a function of space and time. The results are illustrated globally by mapping both the surface-to-bottom temperature difference and the distributions of thermocline depth for 1980, 2000, and 2019. The findings confirm not only a widespread increase in surface-to-bottom temperature differences (on average by 0.39 ℃ per decade) but also a generalized deepening of the thermocline, on average by 1.2 m (around 0.3 m per decade) between 1980 and 2019. The results confirm that global reservoir thermal stratification has both intensified and migrated downward over the past four decades. Chapter 4 compiles depth–area–volume (D-A-V) relationships for over 1.4 million lakes and reservoirs by merging HydroLAKES and GLOBathy. The resulting GLRDAV database contains > 17 million equations—five polynomial functions (orders 1–5) and one power function for both depth–area and depth–volume—evaluated at 0.1 m depth increments. Validation against ReGeom, GRDL, and in-situ Texas Water Development Board surveys show that 4th- and 5th-order polynomials deliver the highest accuracy. Lower-order polynomials and the power function perform adequately for small, simple basins but not for large waterbodies. A Python package, named “Global Waterbody Calculator”, provides streamlined access to all 17 million equations and coefficients, facilitating rapid bathymetric reconstruction for hydrodynamic and water-quality models. The tool rasterizes shoreline vii polygons at 1 arcsecond (~30 m) resolution and rapidly generates full 3-D GeoTIFF bathymetry on a standard desktop, enabling immediate visualization and 3-D model-ready inputs. Chapter 5 applies the CE-QUAL-W2 model to Fanshawe Reservoir (Ontario, Canada) to test how 33 dam operation scenarios—three dam withdrawal elevations crossed with eleven water-level elevations from 0 to +10 m relative to the current conditions—alter phosphorus retention by the reservoir. Under baseline operation (normal withdrawal, 0 m water level) the reservoir retains only 13% of incoming total phosphorus (TP) annually and can become a net TP and dissolved phosphorus (DRP) source in summer. Switching to surface withdrawal alone boosts the annual TP retention to 20 %, while combining surface withdrawal with a +10 m pool-raise pushes summer retention of TP above 78% and the annual retention to 52%. These gains stem from a four-fold lengthening of the water residence time (peaking at 149 days) and the hydraulic isolation of the P-rich hypolimnion. However, these dam operation and water level conditions also prolong bottom-water hypoxia (with dissolved oxygen < 2 mg/l for around 48 days). The modeling highlights the trade-off between maximizing phosphorus retention and avoiding in-reservoir deoxygenatiom, underscoring the need for seasonally targeted, adaptive dam management. Finally, Chapter 6 synthesizes the thesis’s key insights and charts a path forward. It emphasizes how the new global datasets (GRM and GLRDAV) and the Fanshawe case study together advance our understanding of the links between hydrodynamics, nutrient cycling, and dam operation. Looking ahead, this chapter calls for coupling “big-data” archives with process-based and machine-learning models, building reservoir-scale digital twins, and incorporating sediment and groundwater interactions to assess long-term climate and management impacts. Strengthening this predictive framework will be instrumental in safeguarding lakes and reservoirs under accelerating environmental change.

https://hdl.handle.net/10012/22529