Charles Simenstad
simenstd@u.washington.edu
Steve Bollens
sbollens@sfsu.edu

One of the early and more comprehensive CALFED (see below) research projects on wetland restoration that has been conducted in the San Francisco estuary (including both the Sacramento-San Joaquin Delta and San Francisco Bay) are the breached levee wetland studies (BREACH) projects being conducted by an interdisciplinary team of scientists representing universities (University of Washington, San Francisco State University, University of New Orleans), a consultant (Philip Williams, Assoc., Ltd.), a non-governmental agency (Point Reyes Bird Observatory) and a governmental agency (California Department of Water Resources). Researchers are comparing historically restored wetland sites to remnant natural wetland sites in the estuary to determine how far (if at all) the restoring wetlands are progressing toward the CALFED goal to "rehabilitate natural processes in the Bay-Delta system to support, with minimal ongoing human intervention, natural aquatic and associated terrestrial biotic communities."

Over 95% of the once-vast tidal-freshwater and brackish wetlands of the Sacramento-San Joaquin Delta have been leveed and removed from tidal and floodwater inundation; by 1988, 79% of the tidal marsh in Suisun Bay had disappeared, most of it behind levees, and tidal wetland loss in San Pablo/North Bay is comparable (70%). Restoration of a significant portion of these tidal wetlands by breaching levees of selected "reclaimed" islands is rapidly becoming one of the more common restoration strategies to recover ecosystem integrity through rehabilitation of shallow water habitat. However, not unlike coastal restoration throughout the region and country, we know very little about the patterns, rates and even endpoints of breached-levee restoration in the San Francisco Estuary.

There continues to be a high degree of uncertainty and unpredictability associated with re-establishing sustainable ecological functions in restored wetlands, such that: (1) direct measurement of function is difficult, if not impossible; (2) indirect or surrogate 'indicators' have not be developed or validated; and, (3) we don't know what independent factors promote rapid development of functions.

The "Natural" Experiment
Given the extensive changes in watershed structure and processes, complex management and manipulation of water, introductions of pollutants and non-indigenous species, and drastic declines in fish populations in the San Francisco Estuary, there are many critical questions about the feasibility and practicality of restoration. Ecosystem restoration at the scale of the San Francisco Estuary epitomizes the need for comprehensive scientific investigations in systems facing such constraints. The BREACH studies address both the feasibility of restoration of shallow water habitat, and differences inherent to different regions of the Bay-Delta continuum, as well as the contribution of that endpoint to the long-term recovery of fisheries and ecological integrity of the San Francisco Estuary.

In the BREACH investigations, we are taking advantage of the "natural" experiment of the diverse age distribution of selected breached-levee sites in a '"space-for-time substitution" to predict the patterns and rates (trajectories) of restoration that would be expected from levee breaching. This involves evaluation of the rates and patterns of restoration, and determination of sources of variability, in both naturally and artificially restoring diked wetland sites of a broad spectrum of ages as compared to relict, reference marshes (Fig. 1).

We hypothesize several significant differences among processes and biota along the Bay-Delta estuarine continuum that would influence the rates and patterns of shallow water habitat restoration. As salinity and tidal energy increase from the Delta toward the Golden Gate, the geomorphic processes driving the rates and endpoints of site evolution change. Our goals are to: (1) systematically address the present status, rates, and patterns of tidal ecosystem restoration in recognizably different Bay-Delta ecosystems; (2) evaluate factors that promote rapid restoration of shallow-water habitat versus factors that have potentially inhibited natural rates and patterns of functional development; (3) evaluate the contribution of shallow water habitats to food webs supporting Bay-Delta ecosystems; and (4) assess the overall outcome of breached-levee restoration in the different Bay-Delta regions and recommend optimum strategies and spatial distribution of future restoration initiatives. Our products are intended to provide critical information necessary to predict whether breached-dike restoration strategies proposed under CALFED would provide natural wetland functions to support tidal (shallow-water) aquatic habitat for aquatic and terrestrial species of concern and rehabilitate a robust estuarine food web.

Shallow Water Habitat Restoration
Initial findings from the BREACH studies in the Delta suggest that restoration of emergent tidal-freshwater and brackish wetlands of the Delta is contingent upon the interaction of tidal and fluvial processes with vegetation communities that depend upon prior (e.g., subsidence) and initial (e.g., breach locations) conditions affecting the rate of emergent marsh colonization and expansion. We have identified the importance of a minimum ground surface elevation (approximately 2 feet below mean lower low water) for the establishment of tule vegetation and restoration of marsh habitat through natural processes. Although we have been able to quantify the rates and mechanisms of site evolution that occur once this minimum "threshold" elevation is achieved, the rates and mechanisms of sub-tidal accumulation up to this minimum threshold are less well understood. Subtidal accumulation appears to be the slowest step in the natural restoration of breached-levee sites, and can prevent natural restoration altogether where rates of accumulation are slower than sea level rise. However, many factors influencing the rate and pattern of shallow water habitat restoration are significantly different between the Delta and other regions of the estuary under consideration for restoration actions by CALFED.

Differences in tidal regime, fluvial influence, suspended sediment sources, vegetation communities, exotic species and anthropogenic manipulations may result in significant divergences from our present understanding of restoration stages and endpoints originating from studies in the Delta. By identifying the corresponding biological communities and food webs associated with transitional restoration stages in different regions of the estuary, we are refining and extending our original BREACH conceptual model to address emerging questions about shallow water restoration, such as:

In addition to the fundamental data on marsh geomorphology, vegetation structure, fish, invertebrate and bird assemblages and food web structure, our primary synthesis product is a conceptual model of the patterns, rates and underlying processes that are involved in the evolution of sediment accretion and emergent marsh vegetation development, and associated biological community responses, at breached-levee restoration sites (Fig. 2). We anticipate that the expanded and refined BREACH conceptual model will become a potentially valuable tool in developing priorities and allocating restoration resources to maximize the contribution to the system's fishes and wildlife and, ultimately, the San Francisco Estuary food web.

Delta Sites Sampled
BREACH research in the Delta was conducted at ten sites (four reference and six restoring) and our continuing research in the Bay (San Pablo-Suisun bays) involves eleven sites (five reference, six restoring). As an example of the data that are emerging from the BREACH comparisons of restoring and relict wetlands, Figure 3 illustrates fish and benthic macroinvertebrate assemblages, and diet composition for one representative resident fish species, from a cluster of three BREACH sites in San Pablo Bay, in the lower Petaluma River in April 2001.

These sites include a ~16-ha restoring marsh (Carls Marsh) that underwent dike breaching in 1994, a ~20-ha "centennial" marsh, and a 1,134-ha "ancient" marsh (Petaluma), where the ancient marsh is thousands of years in age but the centennial marsh has only recently formed over the last century, as a result of the gold rush placer mining sediments entering and accreting in the estuary. Diverse fishes are already found utilizing the restoring site at Carls Marsh, dominated by Pacific staghorn sculpin (Leptocottus armatus) but also including Pacific herring (Clupea harengus pallasi), northern anchovy (Engraulis mordax), Sacramento splittail (Pogonichthys macrolepidotus), striped bass (Morone saxitalis), inland silversides (Menidia beryllina), and a few more rare species. In contrast, the centennial marsh at Greenpoint was dominated more equitably by Pacific herring, northern anchovy, bay goby (Lepidogobius lepidus), and fewer Pacific staghorn sculpin. The large ancient Petaluma marsh was almost completely dominated by Pacific herring.

Among the benthic invertebrates initially colonizing Carls Marsh, oligochaetes and non-indigenous polychaete annelids (i.e., Neanthes succinea, Polydora ligni, Streblospio benedicti) predominated, but non-indigenous crustaceans (cumacean, Nippoleucon hinumensis; gammarid amphipod, Corophium alienense), as well as nematodes, were also prevalent. Benthos composition at the centennial marsh was somewhat similar, although with a greater contribution by non-indigenous crustaceans and non-indigenous bivalves (Macoma balthica). In contrast, the benthos in the ancient marsh at Petaluma was more dominated by oligochaetes, with lesser contributions by non-indigenous crustaceans and polychaetes.

Diet composition of the Pacific staghorn sculpin illustrates the response of a resident demersal fish that typically feeds on benthos, which we consider a good integrator of benthic prey resource availability within the marshes. In this case, the contrast between the restoring and both reference marshes is more striking, where non-indigenous polychaete annelids (e.g., Neanthes succinea) dominated in the former and non-indigenous crustaceans (e.g., Corophium alienense, Nippoleucon hinumensis, Grandidierella japonica) dominated in the latter cases. Bivalve siphons were also a unique diet component at the Carls Marsh restoring site, as compared to more fish larvae evident in the diets at the two reference marsh sites.

Results such as these tend to support our impression that breaching dikes in the Bay region of the estuary results in rapid development of tidal wetland communities and ecological processes, much of which are comparable to relict wetland sites. Given the extensive "invasion" of San Francisco Estuary by non-indigenous species, it is not surprising that we're also finding pervasive occurrence of non-indigenous species in fish and invertebrate assemblages and food webs of both restoring and natural marshes. Similar samples and data from our other BREACH sites in San Pablo and Suisun Bays, spanning all different seasons, are currently being analyzed and will be compared with earlier results from the Delta in more formal and comprehensive tests of our conceptual model.

More information on the BREACH studies is available on two websites (BREACH I, in the Delta: http://depts.washington.edu/calfed/calfed.htm; BREACH II, in the northern Bay: http://depts.washington.edu/calfed/breachii.htm).

Figure legends (figures not shown):

Figure 1. Aerial photographs of BREACH study sites representing restoring wetland sites in the Sacramento-San Joaquin Delta (a; Mildred Island) and San Pablo Bay (b; Carls Marsh) as compared to centennial (c; Greenpoint) and ancient (d; China Camp) marsh reference sites.

Figure 2. Example of application of BREACH conceptual model to establishment of shallow water fish habitat in Sacramento-San Joaquin Delta, where extensive subsidence of diked wetland upon breaching of dikes initially results in "tidal lakes" (top) with submerged aquatic vegetation (SAV), compared to natural accretion or artificial introduction of sediments, such as dredged material, that results in expanding emergent vegetation (bottom). Whereas more non-indigenous (American and threadfin shads, centrarchids) and some indigenous fishes (prickly sculpin, tule perch) are enhanced by the deeper water and SAV habitats, indigenous chinook salmon and non-indigenous rainwater killifish are found more often in shallow water and emergent vegetation habitats

Figure 3. Fish (top panel) and macroinvertebrate (middle) assemblages and Pacific staghorn sculpin, Leptocottus armatus, numerical diet composition (bottom) from a restoring, breached dike site (Carls Marsh), and centennial (Greenpoint) and ancient reference (Petaluma) reference marshes in April 2001. Fish were captured by tidal channel fyke block nets (total fish n above each histogram bar), where underlined species are non-indigenous species; macroinvertebrates were sampled with five replicate benthic cores; and fish diet composition included systematic processing of stomach contents (acquired by gastric lavage) from five individual fish for frequency of occurrence, and numerical and gravimetric composition.

CALFED Bay-Delta Program

The CALFED Bay-Delta Program is a cooperative effort of more than 20 state and federal agencies working with local communities to improve the quality and reliability of California's water supplies and revive the San Francisco Bay-Delta ecosystem. The mission of the CALFED Bay-Delta Program is to develop and implement a long-term comprehensive plan that will restore ecological health and improve water management for beneficial uses of the Bay-Delta.
More specifically, the CALFED Ecosystem Restoration Program has the following six goals:

In addition to providing extensive funding of freshwater and estuarine wetland restoration projects in the Bay-Delta, as part of CALFED's efforts to achieve these goals they have supported various scientific research projects charged with elucidating the basic ecosystem dynamics of the Bay-Delta, while at the same time providing evaluations of the relative success and failure of various restoration actions. On such research project is BREACH (see accompanying article). For more information on the CALFED Bay-Delta Program, see http://calfed.water.ca.gov.