Delta-scale solutions for human-scale needs.

Global studies inform river management needed for landscape sustainability.

Global rates and patterns of channel migration in river deltas.

River deltas are dynamic systems whose channels can widen, narrow, migrate, avulse, and bifurcate to form new channel networks through time. With hundreds of millions of people living on these globally ubiquitous systems, it is critically important to understand and predict how delta channel networks will evolve over time. Although much work has been done to understand drivers of channel migration on the individual channel scale, a global-scale analysis of the current state of delta morphological change has not been attempted. In this study, we present a methodology for the automatic extraction of channel migration vectors from remotely sensed imagery by combining deep learning and principles from particle image velocimetry (PIV). This methodology is implemented on 48 river delta systems to create a global dataset of decadal-scale delta channel migration. By comparing delta channel migration distributions with a variety of known external forcings, we find that global patterns of channel migration can largely be reconciled with the level of fluvial forcing acting on the delta, sediment flux magnitude, and frequency of flood events. An understanding of modern rates and patterns of channel migration in river deltas is critical for successfully predicting future changes to delta systems and for informing decision makers striving for deltaic resilience.

System wide channel network analysis reveals hotspots of morphological change in anthropogenically modified regions of the Ganges Delta.

The Ganges Brahmaputra Meghna Delta (GBMD) is a large and complex coastal system whose channel network is vulnerable to morphological changes caused by sea level rise, subsidence, anthropogenic modifications, and changes to water and sediment loads. Locating and characterizing change is particularly challenging because of the wide range of forcings acting on the GBMD and because of the large range of scales over which these forcings act. In this study, we examine the spatial variability of change in the GBMD channel network. We quantify the relative magnitudes and directions of change across multiple scales and relate the spatial distribution of change to the spatial distribution of a variety of known system forcings. We quantify how the channelization varies by computing the Channelized Response Variance (CRV) on 30 years of remotely sensed imagery of the entire delta extent. The CRV analysis reveals hotspots of morphological change across the delta. We find that the magnitude of these hotspots are related to the spatial distribution of the dominant physiographic forcings in the system (tidal and fluvial influence levels, channel connectivity, and anthropogenic interference levels). We find that the anthropogenically modified embanked regions have much higher levels of geomorphic change than the adjacent natural Sundarban forest and that this change is primarily due to channel infilling and increased rates of channel migration. Having a better understanding of how anthropogenic changes affect delta channel networks over human timescales will help to inform policy decisions affecting the human and ecological presences on deltas around the world.

Transfer Entropy as a Tool for Hydrodynamic Model Validation.

The validation of numerical models is an important component of modeling to ensure reliability of model outputs under prescribed conditions. In river deltas, robust validation of models is paramount given that models are used to forecast land change and to track water, solid, and solute transport through the deltaic network. We propose using transfer entropy (TE) to validate model results. TE quantifies the information transferred between variables in terms of strength, timescale, and direction. Using water level data collected in the distributary channels and inter-channel islands of Wax Lake Delta, Louisiana, USA, along with modeled water level data generated for the same locations using Delft3D, we assess how well couplings between external drivers (river discharge, tides, wind) and modeled water levels reproduce the observed data couplings. We perform this operation through time using ten-day windows. Modeled and observed couplings compare well; their differences reflect the spatial parameterization of wind and roughness in the model, which prevents the model from capturing high frequency fluctuations of water level. The model captures couplings better in channels than on islands, suggesting that mechanisms of channel-island connectivity are not fully represented in the model. Overall, TE serves as an additional validation tool to quantify the couplings of the system of interest at multiple spatial and temporal scales.

Multilevel modeling of retention and disinfection efficacy of silver nanoparticles on ceramic water filters.

This research examined how variations in synthesis methods of silver nanoparticles affect both the release of silver from ceramic water filters (CWFs) and disinfection efficacy. The silver nanoparticles used were stabilized by four different molecules: citrate, polyvinylpyrrolidone, branched polyethylenimine, and casein. A multilevel statistical model was built to quantify if there was a significant difference in: a) extent of silver lost, b) initial amount of silver lost, c) silver lost for water of different quality, and d) total coliform removal. Experiments were performed on location at Pure Home Water, a CWF factory in Tamale, Ghana using stored rainwater and dugout water (a local surface water). The results indicated that using dugout vs. rainwater significantly affects the initial (p-value 0.0015) and sustained (p-value 0.0124) loss of silver, but that silver type does not have a significant effect. On average, dugout water removed 37.5µg/L more initial silver and had 1.1µg/L more silver in the filtrate than rainwater. Initially, filters achieved 1.9 log reduction values (LRVs) on average, but among different silver and water types this varied by as much as 2.5 LRV units. Overall, bacterial removal effectiveness was more challenging to evaluate, but some data suggest that the branched polyethylenimine silver nanoparticles provided improved initial bacterial removal over filters which were not painted with silver nanoparticles (p-value 0.038).

Subordinated Brownian motion model for sediment transport.

Following the introduction of the Brownian motion model for sediment transport by Einstein, several stochastic models have been explored in the literature motivated by the need to reproduce the observed non-Gaussian probability density functions (PDFs) of the sediment transport rates observed in laboratory experiments. Recent studies have presented evidence that PDFs of bed elevation and sediment transport rates depend on time scale (sampling time), but this dependence is not accounted for in any previous stochastic models. Here we propose an extension of Brownian motion, called fractional Laplace motion, as a model for sediment transport which acknowledges the fact that the time over which the gravel particles are in motion is in itself a random variable. We show that this model reproduces the multiscale statistics of sediment transport rates as quantified via a large-scale laboratory experiment.