Research Projects

How do rivers behave during high flows and interact with their floodplains?

Hurricanes Harvey (2017), Irma (2017), and Florence (2018) have recently demonstrated the tremendous damage to communities and infrastructure produced by flooding from extreme rainfall. This project seeks to develop and benchmark a modeling framework for predicting river flooding and overland flow as it specifically applies to coastal lowland rivers of the Gulf Coast. The resulting numerical model and benchmarking data sets will be used to predict the consequences of flooding associated with tropical cyclones, estimate to effectiveness of proposed remediation projects following flood events, and examine how anthropogenic modifications to the landscape might impact flood magnitude. Development of the model and guiding data sets will be driven by the floods on the coastal San Jacinto and Trinity rivers associated with Hurricane Harvey rain.

Collaborators: David Mohrig (Jackson School of Geosciences)

Students: Shazzad Rahman

Hydrodynamic modeling results and mechanisms of channel-floodplain connectivity in the Trinity River, TX. Modeling results by Kyle Wright.

Can we model the subsurface from surface network information?

Surface connections, such as those among channels in river networks, are important for understanding the development and evolution of landscapes, such as densely populated coastal river deltas. Connections in the subsurface are critical in understanding groundwater flow and solute transport. Preferential flowpaths, in fact, can quickly deliver contaminants to water supply wells, a particularly important problem in densely populated coastal areas. Establishing a quantitative link between surface and subsurface patterns will greatly advance our capability to predict the movement of contaminants in groundwater, thus improving access to clean water and limiting pollution and health risks. We propose to investigate quantitatively how the dynamics of surface networks create subsurface networks, and thus determine how surface information can be used to predict properties of the subsurface. This will enable us to better predict sustainability and manage water resources in densely populated deltas such as the Ganges-Brahmaputra Delta, where high concentrations of arsenic are widespread in the groundwater of the upper delta, and salinity problems are pervasive in the lower delta.

Funding: National Science Foundation

Collaborators: Holly Michael (U. Delaware), Chris Paola (U. Minnesota)

Students and post-docs: Jay Hariharan, Mariela Perignon

Numerically simulated delta using the model DeltaRCM [Man Liang -- see Liang's papers in Publication list] and two 'subsurface' realizations created by repeating the surface network or randomly shuffling the surface information.

How do the river network and the road network interact during a flood?

During rainfall events, flooding can be fluvial and/or pluvial. In both scenarios we are dealing with two networks overlaid on each other: the river network and the road network, on which emergency vehicles, private vehicles, and public transports move. These networks are overlaid on the terrain, with its depressions being the first portion to be flooded during rainfall events. During emergencies the interdependencies between the fluvial and the road network, the characteristics of the networks and of the terrain come into play and affect the capability of a city or rural area to respond to an emergency. We are developing a framework for analyzing these interdependencies and predict what portions of a landscape or State would be at risk. Our framework uses the estimated flows continuously provided at the continental scale by the National Water Model.

Collaborators: Steven Boyles (transportation)

Students: Isha Deo (water resources) and Cesar Yahia (transportation)

Multi-layer network composed of topography, river network, storm sewers, and road network. Our framework uses information from the National Water Model and network characteristics to identify areas most at risk during storms and routes for emergency response.

The Delta Connectome: Structure and Transport Dynamics of Delta Networks across Scales and Disciplines

Human-induced activities and climatic shifts are significantly impacting deltas around the world. A quantitative description of the form and structure of deltas and their dynamics is fundamental to address how they react to changes in climatic forces and human pressure. As part of this project, we are creating a research/educational framework, the Delta Connectome, based on the general idea of a delta as a directed network of connected paths (physical, functional, and conceptual paths of process coupling) which interact continuously at a broad range of space and time scales and dictate system response to change. Specific goals of this project are (i) the objective quantification of delta morphologic features (channel and island properties) to identify the signature of vegetation, anthropogenic disturbance, and processes responsible for delta formation and evolution; (ii) the development of an automatic image processing-based tool for the extraction of relevant information from remotely sensed data in order to apply the analysis in (i) to a wide range of delta morphologies; (iii) the identification of the environmental controls on channel network and shoreline dynamics through coupled analysis of the extracted features through time and time series of environmental forcings; (iv) the quantification of strength, directionality, statistical significance, and scale of couplings among key variables and the effect of anthropogenic disturbance and change on such couplings.

Funding: National Science Foundation

Students: Kyle Wright

Structural, functional, (process-based) and process connectivity in river deltas (reproduced from Passalacqua [2017]).

Open source tools for automatic delta network extraction from remotely sensed data

As part of the Delta Connectome project, we have developed fully automatic tools for the extraction of channel networks from remotely sensed imagery. RivaMap is based on a multi-scale singularity index, while DeepWaterMap and DeepRiver are Convolutional Neural Network models.

RivaMap: An Automated River Analysis and Mapping Engine

DeepWaterMap: A Deep Learning Based Surface Water Mapper

DeepRiver: A Deep Learning Based River Network Extractor

Student: Tess Jarriel

Collaborator: Al Bovik

Coastal SEES Collaborative Research: Multi-scale modeling and observations of landscape dynamics, mass balance, and network connectivity for a sustainable Ganges-Brahmaputra delta

This project combines innovative quantitative tools (numerical modeling, network and connectivity analysis) with new and existing observational data to analyze the coupled human-natural system and long-term sustainability of the GBMD. Specifically, we will (i) develop a detailed mass balance for delta-wide sediment dispersal; (ii) quantitatively analyze the connectivity of the delta-system network that disperses this sediment; (iii) integrate this knowledge through numerical modeling at local to global scales; (iv) use observational data of landscape and channel dynamics to understand coupled land-sea interactions; (v) evaluate the quality of regional soil and water resources and their links with physical and anthropogenic processes; (vi) assess the impact of these delta dynamics on the human environment and transportation, and finally (vii) disseminate this knowledge through a variety of educational activities and opportunities for students, researchers, and professionals. The Passalacqua group will focus on the network and connectivity analysis.

Funding: National Science Foundation

Student: Tess Jarriel

Connectivity analysis of the Ganges-Brahmaputra-Meghna Delta (reproduced from Passalacqua [2017]).

Feature extraction from high resolution topography data

The advent of meter-resolution topographic data is revolutionizing the study of geomorphic processes. For the first time, the topographic patterns of surface flow, channelization, and landsliding can be resolved over large areas at resolutions commensurate with the scales of the governing processes. We apply new methods of geomorphic feature extraction and morphological analysis (GeoNet) to a large inventory of lidar DTMs in combination with detailed ground validation at key watersheds. We use the extracted information to improve inundation mapping over the continental US. We additionally use GeoNet as technique for hydro-flattening of high resolution topography data.

Funding: National Science Foundation and United States Department of Agriculture

Students: Xing Zheng

Collaborators: David Maidment

Improving flood inundation mapping at continental scale

A large scale effort is ongoing to improve flood inundation mapping at continental scale by relying on the National Water Model and the HAND method (Height Above Nearest Drainage). The National Water Model continually forecasts flows on 2.7 million stream reaches in the continental United States. These discharge forecasts are then translated into forecasts of water depths above minimum channel elevation with the HAND method. So far this work has relied on the 10m National Elevation Dataset topographic data. Our research group is exploring the use of GeoNet on high resolution topographic data to translate this procedure into lidar terrain data to increase the accuracy of the estimated flood extent.

Funding: Texas Division of Emergency Management; Texas Department of Transportation

Collaborator: David Maidment and many others

Student and post-docs: Andrew Austin-Petersen, Leah Huling, Xing Zheng

Height Above Nearest Drainage (HAND) computed over the continental United States.

Dynamics, resiliency, and sustainability of river deltas

River deltas represent a major Earth-surface system with societal need. Low-lying, ecologically productive, and inhabited by millions of people, deltas also lie directly in the path of a confluence of ongoing changes: nutrient overloading from agriculture; accelerated subsidence and sea-level rise; effects of land use and navigation; and changing hydrology and sediment supply. The overall objective of this project is to develop tested, high-resolution, quantitative models incorporating morphodynamics, ecology, and stratigraphy to predict river delta dynamics over engineering to geologic time-scales, and to specifically address questions of system dynamics, resiliency, and sustainability. The collaboratory comprises two main work centers: a field observatory and a virtual modeling center, together with supporting experimental facilities. The observatory is at Wax Lake Delta, an actively growing delta about 100 km west of the main Mississippi Delta birdsfoot. The virtual modeling center is hosted by the Community Surface Dynamics Modeling System (CSDMS) at University of Colorado, where it can contribute to an evolving library of modules for computation and visualization of geomorphic and sedimentary systems, including access to many of the existing delta models. This project is in collaboration with Dr. David Mohrig and Dr. Wonsuck Kim (UT Austin) and other 11 PIs at several US institutions.

Our research group is focusing on the flow exchange between channels and islands and the controls exerted by vegetation, topography, and environmental forcing on channel-island hydrological connectivity.

Funding: National Science Foundation

Student: Kyle Wright

Example of field observations and numerical modeling performed at the Wax Lake Delta to quantify channel-island and process connectivity.