<p>Topography is one of the primary drivers of physical processes in the rivers and floodplains. Advances in remote-sensing and survey techniques have provided high-resolution representation of the floodplains but information regarding the 3D representation of river channels (commonly known as river bathymetry) is sparsely available. Field surveys along an entire river network in a watershed remains infeasible and algorithms for estimating simple but effective characterization of river channel geometry are hindered by an incomplete understanding of the role of river bathymetry in surface and subsurface processes. </p> <p> The first objective of this dissertation develops an automated framework – System for Producing RIver Network Geometry (SPRING) for improving the geospatial descriptors of a river network. The tool takes as input the DEM and erroneous river centerline to produce spatially consistent river centerlines, banks, and an improved representation of river channel geometry. SPRING can process entire river networks and is not limited single reach applications. The proposed framework is flexible in terms of data requirements, resolution of output datasets and user preferences. It has a user-friendly graphic user interface (GUI) and is appropriate for large-scale applications since it requires minimal user input.</p> <p> A better understanding of the role of bathymetric characteristics in surface-subsurface hydrology and hydrodynamics can facilitate an efficient incorporation of river bathymetry in large river networks. The second objective explores the level of bathymetric detail required for accurately simulating surface and subsurface processes by developing four bathymetric representations using SPRING with reducing level of detail. These bathymetric configurations are simulated using a physically based tightly coupled hydrologic and hydrodynamic model to estimate surface and subsurface fluxes in the floodplains. Comparison of fluxes for the four bathymetric configurations show that the impact of river bathymetry extends beyond surface routing to surface water – groundwater interactions. Channel conveyance capacity and thalweg elevation are the most important characteristics controlling these interactions followed by channel side slope and channel asymmetry. </p> <p> The final objective aims to develop benchmarks for bathymetric characteristics for accurately simulating flooding related physical processes. The sensitivity of surface and subsurface fluxes to error in channel conveyance capacity is investigated across reaches with varying geomorphological characteristics. SPRING is used to create six bathymetric configurations with varying range of error in channel conveyance capacity (ranging from 25% to 300%). They are simulated using a tightly coupled physically distributed model for a flood event and the estimates of water surface elevation, infiltration and lateral seepage are compared. Results show that incorporating channel conveyance capacity with an error of within 25% significantly improves the estimates of surface and subsurface fluxes as compared to those not having any bathymetric correction. For certain reaches, such as those with high drainage area (>1000km²) or low sinuosity (< 1.25), errors of up to 100% in channel conveyance capacity can still improve H&H modeling.</p>