Paola Passalacqua (The University of Texas at Austin, United States); Jon Schwenk (Los Alamos National Laboratory, United States); Anastasia Piliouras (Pennslyvania State University , United States)

Coastal river deltas are dynamic hydrologic systems that present unique challenges for study and modeling. Their channel networks exhibit extremely low topographic gradients, allowing coastal forces such as winds and tides to generate bidirectional flow and create loops within channel structures. Although more difficult to model than tributary networks, recent research has increasingly focused on simulating exchanges between these loopy channel systems and adjacent deltaic islands. Delta networks are inherently leaky systems—at Wax Lake Delta (LA, USA), for example, up to 54% of water can be lost from primary channels into island complexes (Hiatt and Passalacqua, 2015). Accurately representing these exchanges is essential for predicting how sediments, solutes, and nutrients are partitioned and processed across the delta landscape.

Widely used hydrodynamic models like Delft3D and ANUGA, which solve the depth-integrated shallow-water equations, have been employed to estimate hydrologic connectivity in deltas and wetlands. However, these models typically require extensive field data to calibrate and validate parameters governing connectivity. Other recent advances in modeling approaches have utilized graph-theory to represent channel networks as a series of nodes (junctions) and links (channels). Sediments, solutes, and nutrients can then be partitioned along this extracted series of nodes and links through remotely sensed estimates of channel widths and available discharge data. This method offers advantages by relying solely on remotely sensed data, eliminating the need for intensive fieldwork. However, traditional graph-based models conserve mass along links, assuming no material loss from the delta apex to its outlets, which is inherently limited.

This research introduces a new methodology to estimate channel-island connectivity—or channel “leakiness”—in coastal river deltas using reported water surface elevations and pixel cloud classifications from the SWOT Level-2 High Resolution Pixel Cloud product. The approach will first be applied to Louisiana’s Wax Lake Delta, where extensive field data are available for validation. Ultimately, this methodology will be extended to deltas worldwide and integrated into graph-based network extraction techniques.