Research - Category I: Hydrology, Climatology & Hydraulics

Hydrodynamic Design in Coastal Wetland Restoration: Topography Optimization and Stability Assessment by Adjoint Sensitivity Method
(Funded 2000-2001)

Principal Investigator: Brett Sanders
Department of Civil and Environmental Engineering
UC Irvine
(949) 824-4327
bsanders@uci.edu

Executive Summary:
Problem Statement: Promoted by conservation-banking practices initiated in 1995, wetland restoration and enhancement projects have emerged in southern California s the primary means to mitigate the impact of coastal development. For example, the enhancement of the Bolsa Chica wetland now underway in Huntington Beach is tied tot he expansion of the Ports of Los Angeles and long Beach. Constructed wetlands reclaim habitat for fish and wildlife, provide nutrients to surrounding coastal waters, mitigate the impact of pollutants in urban runoff, and create recreational opportunities (Mitsch and Gosselink 1986).

Successful enhancement and restoration designs hinge on a reliable prediction of physical processes including flow, circulation, and transport of nutrients, salinity, sediments, and pollutants. A poor prediction can cause a design to fail. Talbert Marsh in Huntington Beach serves as an excellent example. It was constructed in 1990, but by 1991 the marsh had yet to achieve the intertidal zone of the design. By 1995, longshore sediment-transport led to a partial blocking of the Marsh outlet, sedimentation lead to 4-6 feet of silt on the marsh bottom, and the expected level of diversity had not materialized (Jones & Stokes 1997).

Engineers use hydrodynamic models to predict depth of inundation, salinity, and velocity regimes-information that is important to wetland planners, ecologists and biologists. When properly utilized, numerical models can reliably predict circulation in complex estuarine systems (Cheng et al. 1993, Abbott 1997). However, engineers face a second and much more difficult challenge in the numerical modeling process: assessing the stability of a design. The Talbert Marsh experience clearly elucidates the consequences of a poor stability assessment.

The challenge in assessing the stability of a wetland design is rooted in our primitive understanding of the suspension, bed load transport, suspended load transport, and deposition of sediments. Models capable of simulating, let alone predicting, these processes for the wide range of particle class sizes present in estuarine systems can only be greeted with the utmost suspicion. This situation has rendered empiricism to be the only viable mechanism to aid engineers in stability assessments. Accordingly, the reliability of stability assessments depends highly on inexpensive high-speed computers, many unqualified engineers are now able to participate in the wetland design process. To avoid future wetland failures, a systematic design process that is less reliant on interpretive skills is needed.

Research Approach: In this study, a novel approach used in parameter identification and hydrodynamic control applications will be developed and tested as a new paradigm for computer aided coastal wetland design. The bed elevation as a function of location will be optimized to (i) maximize inter-tidal area, a common indicator of productivity and (ii) achieve a hydrodynamic condition characteristic of self-flushing systems. The goal of (ii) is avoid hydrodynamic conditions that have been shown to promote the sedimentation and filling of wetlands. Namely, the analysis of hundreds of wetlands has lead to a strong association between flood-dominance (a short and strong flood tide followed by a long and weak ebb tide) and wetland sedimentation (Aubrey 1986). By embracing the opposite condition (ebb-dominance) as the optimization goal, it will be possible to test whether this simple characteristic of wetland designs represents a reliable measure of stability to be embraced in the design process.

Optimization will be performed using a gradient-based method. An objective function reflecting goals (i) and (ii) above will be formulated. To efficiently obtain the necessary gradient vector, an adjoint sensitivity of the objective function to the parameter space describing the bed elevation at all locations. The proposed methodology will be tested first on a series of hypothetical wetlands, and second on the Talbert Marsh.

Research Benefits: Substantial resources continue to be invested in coastal wetland enhancement and restoration efforts by numerous public and private organizations. Reliable estimates of the physical processes driving wetland ecosystems are essential for designs that maximize long-term productivity. This research project will lead to a field-tested wetland design tool to aid the wetland restoration process. In addition, this project will address whether a simple hydrodynamic characteristic of wetland systems can be used as an indicator of wetland stability in lieu of complex and poorly understood sedimentation processes.