Impacts of climate change on US wetlands will add to those of historical impacts due to other causes. In the US, wetland losses and degradation result from drainage for agriculture, filling for urbanization and road construction. States that rely heavily on agriculture (California, Iowa, Illinois, Missouri, Ohio, Indiana) have lost over 80% of their historical area of wetlands, and large cities, such as Los Angeles and New York City, have retained only tiny remnants of wetlands, all of which are highly disturbed. The cumulative effects of historical and future degradation will be difficult to abate. A recent review of mitigation efforts in the US shows a net loss of wetland area and function, even though ‘no net loss’ is the national policy and compensatory measures are mandatory. US policy does not include mitigation of losses due to climate change. Extrapolating from the regulatory experience, one can expect additional losses in wetland areas and in highly valued functions. Coastal wetlands will be hardest hit due to sea-level rise. As wetlands are increasingly inundated, both quantity and quality will decline. Recognition of historical, current and future losses of wetland invokes the precautionary principal: avoid all deliberate loss of coastal wetland area in order to reduce overall net loss. Failing that, our ability to restore and sustain wetlands must be improved substantially.
A principal objective of the Clean Water Act (Federal Water Pollution Control Act, Public Law 92-500) is ‘to restore and maintain the chemical, physical, and biological integrity of the Nation's waters’. The ‘Nation's waters’ include wetlands. Thus, in general, the discharge of materials such as soil or sand into wetlands must be authorized by a permit from the US Army Corps of Engineers (CoE). Such permits typically require compensatory mitigation to fulfil a national policy that there should be no net loss of wetland area and function. That is, damage to one wetland area must be compensated by appropriate actions at another location. Permits are granted when the discharge of materials cannot be avoided or minimized; thus, compensatory actions are last in the sequence of possibilities for action on a permit application (Fig. 1, NRC 1995, 2001).
As an example, an owner of a wetland might propose to fill one hectare of wetland in order to build a housing complex. The CoE could deny the permit, perhaps suggesting that the project be moved to an upland area. If the upland supported endangered species, other legislation would apply, and the US Fish and Wildlife Service would have regulatory authority. If avoidance of the wetland were not possible, the CoE would suggest ways to reduce the impacts to much less than one hectare. Only if this were unworkable would a permit be considered and compensatory actions established. Compensation might include the restoration of a former wetland, the enhancement of an existing but degraded wetland or the creation of a new wetland from upland. Ideally, a nearby area of drained and farmed wetland would be restored, so that it would truly compensate for the area and functions lost through filling. Usually, the property owner would be required to restore a larger area than the one to be damaged, as one way of reducing risks. Risks are inevitable, as the compensation site might not replace all the functions (Table 1) of the site to be developed and the functions might take years to develop.
Table 1. Functions of wetlands that are considered highly valuable in the US (NRC 2001).
abatement of floods
Support of vegetation
diverse species of wetland-dependent plants
Support for fauna
water, food, and habitat for a diversity of wetland-dependent animals
wetland soils serve a rooting medium for plants, store seeds and rhizomes, provide water and nutrients for plants, provide habitat for mycorrhizae and symbiotic bacteria, support chemical transformations that improve water quality, sustain soil macrofauna, convey groundwater and store contaminants
Under the US Clean Water Act, the CoE regulates the discharge of fill to wetlands, and the US Environmental Protection Agency (EPA) is responsible for overseeing the CoE's regulatory activities. Under Section 401, states are required to certify that permits comply with their water quality standards; hence, states also participate in the permit-review process. In 1999, the EPA asked the National Research Council (NRC) to convene a panel of experts to review the institutions under which compensatory mitigation is accomplished and the effectiveness of mitigation projects that have been completed. We worked closely with our NRC project officer, Dr Suzanne van Drunick, to prepare a report (NRC 2001) that details our review process and identifies changes that could improve compensatory mitigation.
Historically, compensatory mitigation projects have been undertaken by permittees and their consultants, with preference for on-site, in-kind wetland restoration efforts. Typically, the mitigator would hire a consultant to restore a wetland of the same kind as that being filled, and preferably nearby. More recently, third-party approaches have arisen to fill the growing need for mitigation. Today, it is common for a party other than the permittee to develop a ‘mitigation bank’ by restoring or constructing wetland in a relatively large area away from land undergoing development. Once the wetland is certified as a bank, the mitigation banker can sell credits (hectares) to those who obtain a CoE permit to damage wetlands. Alternatively, a permittee might pay a fee to a third party in lieu of developing a project or buying mitigation bank credits. That third party (a governmental body or a non-profit organization) applies the money to their wetland restoration and creation programme. In our NRC review, we found that mitigation banks and in-lieu fees have high potential for providing compensation in advance of damage and also have potential for long-term stewardship. However, we were unable to provide a critique of the effectiveness of actual projects developed by third parties, owing to their newness and the paucity of evaluations in the literature. Because permittee-responsible projects dominate the review literature, these formed the basis of the NRC review of compensatory mitigation. In this paper, I focus on the ecological evaluation of compensatory mitigation projects, rather than institutional mechanisms. Of relevance to climate change and sea-level rise is the general question, ‘can losses of wetland quantity and quality be mitigated?’.
THE COMPENSATORY MITIGATION RECORD
Determining how well the compensation process is working requires information on the area and functions lost at fill sites in relation to gains at mitigation sites. The NRC panel determined that records were inadequate to document all changes in wetland area and functions in the US. This is because the CoE tracks only the area of wetland that they require as mitigation, not the area that is actually restored or created. For every hectare for which fill is permitted, an average of 1.78 ha of mitigation is required (NRC 2001). To determine how much of the required compensation is realized, and whether or not functions are restored, the panel had to rely on a small subset of mitigation projects for which compliance and ecological functions have been quantified (NRC 2001: appendix A, which summarizes the findings of 25 reviews). More recently, Turner et al. (2001) concluded from the limited data available that, of 100 ha of wetland permitted to be filled and 178 ha of mitigation required, the actual mitigation projects would be initiated on only 134 ha. That is, many mitigation projects are never implemented. Of the 134 ha of projects initiated, 77–104 ha would be judged as ‘in compliance’ with CoE criteria. Finally, if the area in compliance were evaluated by qualified ecologists, only about 19 ha would be judged functionally equivalent to appropriate reference sites. These projections from the available data suggest both a net loss in area and a net loss in functions, with functions lost at a higher rate than area.
Mitigation proposals have several common shortcomings.
1The permitted project often causes a net loss of wetland, because the compensatory actions are not undertaken or they fail to produce wetland.
2Mitigators claim that improvements to wetland remaining on the property will compensate for damage to the area that would be filled. In such plans, ‘mitigation’ consists of remodelling existing marshes. If an existing wetland is modified, there is no new wetland. Hence, the filling of a wetland for development produces a net loss in area.
3Some functions of existing wetlands are lost with the conversion of one wetland type to another. A common mitigation practice in New Jersey is to convert Phragmites-dominated marsh to Spartina alterniflora-dominated marsh. Whereas the former provides habitat for the Yellow-headed Blackbird Xanthocephalus xanthocephalus, the latter supports rails (Rallus spp.). A loss of one function is not replaced by the gain of another. Plans to enhance existing wetlands rarely specify the existing functions that might be lost during enhancement.
4The functional value of the ‘enhanced wetlands’ cannot be predicted. If the mitigation site is developed to function as a sediment-retention basin, the wetland might or might not support native plants and animals. Benefits, such as ‘enhanced ecosystem function’, are often hollow promises, neither linked to reference systems nor backed by performance bonds.
5Reference systems are not identified as models for planning or assessing the ecological/functional values of enhanced wetlands. If targets for restoration are not clarified, any outcome might be pronounced to be in compliance with permit conditions. In order to determine the baseline for comparison, the functions of the site to be filled or a reference wetland (or group of wetlands) must be assessed. Only then can targets be quantified and achievements at the mitigation site compared for determination of net loss or gain in function.
6Plantings are rarely detailed or requirements for survival are not indicated. Simply listing species to be planted is not sufficient, especially if the list includes pest plants (e.g. Typha angustifolia, Phalaris arundinacea). Performance standards could concentrate on the degree to which a range of functions (hydrological, soil, vegetation, faunal) must be provided at the site or the degree to which selected functions should match those of a reference site.
7Sustainability of the enhanced wetlands is questionable. For example, the need to manage sedimentation and control weeds is rarely addressed. The typical requirement of a 3- to 5-year monitoring programme does not ensure that monitoring is performed, that measurements are meaningful or that shortcomings in progress are addressed.
DETERMINING NET LOSSES OR GAINS IN AREA AND FUNCTION
On many occasions, changes in wetland area are hard to assess, because the condition of the mitigation site is not fully described. If a permit applicant is allowed to fill 5 ha of wetland with a requirement to compensate this filling by restoring 10 ha, the effect would appear to be a net gain of 5 ha. However, if the mitigation site is an existing 10 ha wetland that is ‘remodelled’, there would be a net loss in area of 5 ha. By remodelled, I mean changed in quality, not quantity, for example an area dominated by an invasive species might be controlled using herbicide as compensatory mitigation. Even if the mitigator modifies the mitigation site's topography and installs native species, there would still be a net loss in wetland area. Assessing changes in functional capacity is far more difficult. Some functions performed by the 5 ha that are filled and the 10 ha that are treated with herbicide might be lost. Only if the 10 ha that were remodelled could be made functional enough to make up for all the processes of the 5 ha fill site and the 10 ha before remodelling could wetland functions be sustained by this trade-off. If this is possible, it has not been demonstrated, owing to lack of adequate functional assessment protocols.
Net changes in function are difficult to determine because the criteria for compliance are limited in scope relative to ecological assessments. Requirements are often a rate of survival of plantings or some level of plant cover (Streever 1999; Table 2). Rarely are attributes of hydrology, soils, fauna or microbes considered, let alone the processes they perform (Table 1). Hydrological/ecological processes are rarely quantified at fill sites or mitigation sites prior to modification. Denitrification is a vital process provided by wetlands (Table 1), yet we lack low-cost methods for quantifying nitrogen-removal and, even if they existed, long-term sampling would be needed to describe this function. Lacking specific knowledge of wetland functions, the conservative conclusion is that a net loss in wetland area carries with it a net loss in function.
Table 2. Examples of criteria used by the CoE to judge mitigation compliance (taken from Streever 1999, full text reprinted in NRC 2001, appendix E).
Site must be jurisdictional wetland (in Florida, this required that the soil must be saturated to the surface for 31 days/year for 5 years)
50% survival of planted trees for 3 years (an example from Mississippi)
85% of site vegetated by planted species for 5 years (from Massassachusetts)
80% survival of plantings (from Illinois)
< 5% cover of nuisance species; 80% cover of planted species by year 5 (from Florida)
85% vegetated; no Phragmites infestation; ‘animal use’ (from New Jersey)
< 5% cover of exotics in 5 years; increasing numbers of bird species, 75% cover of riparian scrub vegetation (from California)
The restoration of out-of-kind wetlands further complicates the determination of net change in functions. Permittees often prefer to create open water ponds ringed with cattails (Typha spp.) instead of replacing more complex wetlands. Cattail-lined ponds are popular compensation targets, because they are easily constructed and they are clearly identifiable as wetlands. Cattail-lined ponds have been permitted in various states, even where such permanent-water ponds are not natural landscape features, as in Oregon and Pennsylvania (NRC 2001). Indiana recently reviewed its compensatory mitigation programme and a random selection of 31 mitigation sites revealed that losses of forested wetlands were being compensated in large part by producing emergent marsh and open water (Robb 2001), which function differently from forested wetlands. If trees are planted, the functions of mature woody vegetation still take decades to develop.
Even if in-kind wetlands are restored, some functions might still be lacking. In studies of California saltmarshes, the restoration of nesting habitat for the endangered Light-footed Clapper Rail Rallus longirostris levipes was not possible on the coarse soil (sandy dredge spoil) that was available at the mitigation site (Zedler 1998). This subspecies of Clapper Rail needs tall vegetation to hide its nest from raptors and to allow vertical space for the nest to float when the tide comes in. The Pacific Cordgrass Spartina foliosa grew too short at the mitigation site, and the entire plant canopy was often inundated by high tides, providing no habitat for nesting. The right vegetation was present, so the wetland creation project mitigated the loss of wetland ‘area’, but the plant canopy did not offer the required nesting ‘function’. In this case, a typical assessment of plant species or cover would not have shown the site's shortcomings.
The determination of gains and losses requires a careful match between functions and methods used to evaluate them. In saltmarshes, histograms of Spartina foliosa heights work well for evaluating Clapper Rail nesting potential, but canopy layering is a better indicator for overall canopy architecture in species-rich habitats (Keer & Zedler 2002). Neither of these assessment tools appears in Streever's (1999) lists of criteria used by the CoE (Table 2). Nor are any details of assessment spelled out in monitoring protocols. For example, it is necessary to sample at different times of year to characterize bird nesting potential (i.e. during the nesting season) vs. plant growth (in spring, when Carex spp. are fruiting and identifiable, or in late summer to assess peak biomass). Protocols are also lacking for the number and types of reference sites. Functions need to be assessed at both the mitigation site and the damage site (or other appropriate reference sites) in order to determine gains or losses.
CAN MITIGATION PREVENT NET LOSSES?
The many shortcomings of historical mitigation efforts were noted in the NRC panel's review, and our overall conclusion was that the permitting process does not fulfil the policy of ensuring no net loss. Given net losses in area and the likelihood of even greater losses in functions, the precautionary principle is appropriate: all deliberate damage to wetland and upland habitats should be avoided. Although desirable from a conservation perspective, such a policy would restrict new developments to previously degraded lands, and such restrictions to the rights of landowners would not be tolerated in the US. Nor would a moratorium on wetland loss prevent future losses due to climate change and sea-level rise. If we cannot halt losses, we must improve our ability to restore and sustain wetlands.
In the past, compensatory mitigation has been limited more by policy and resolve than by a lack of technology. Although we do not know how to restore all wetland types to full service, we can restore some to at least partial capacity. For example, saltmarshes dominated by Spartina alterniflora along the Atlantic and Gulf of Mexico coasts are readily restored (relative to those of the Pacific coast). The former have lower vegetation diversity and fewer threatened species. If tidal flushing is provided to sites with the appropriate elevation and protection from waves, native dominant species can be re-established (NRC 2001). Perhaps the best examples are in Delaware Bay, where thousands of hectares of enclosed saltmarsh are being restored by breaching levees. If enough money is provided, past and future losses of coastal wetlands can be mitigated. Indeed, the US's most wetland-rich states, Florida and Louisiana, have both launched multibillion-dollar wetland restoration programmes.
WAYS TO IMPROVE OUTCOMES
Although wetlands with a few dominant plant species are considered restorable, concerns remain that such marshes will be slow to assume the functions of reference ecosystems and might not be sustainable. Hence, the NRC panel provided guidance for improving restoration and creation efforts (Table 3) so that the chances of the resulting wetlands becoming self-sustaining would improve. The panel further noted that sustainability is not ensured if stewardship ends with the usual 3- to 5-year monitoring period. Public agencies, non-governmental organizations and private land managers are all potentially identifiable as long-term stewards of mitigation sites. The time frames required are likely to differ with the different types of project, but the panel was clear that stewardship should continue for time periods ‘typically accorded to other publicly valued natural assets, like parks, which are to be managed in perpetuity. This time frame emphasizes the importance of developing mitigation wetlands that are self-sustaining so that the long-term costs are not unmanageable’ (NRC 2001, p. 157).
Table 3. Guidelines for improving wetland restoration and creation of self-sustaining wetlands (from NRC 2001).
Consider the hydrogeomorphic and ecological landscape and climate
Adopt a dynamic landscape perspective
Restore or develop naturally variable hydrological conditions
Whenever possible, choose wetland restoration over creation
Avoid over-engineered structures in the wetland's design
Pay particular attention to appropriate planting elevation, depth, soil type and seasonal timing
Provide appropriately heterogeneous topography
Pay attention to subsurface conditions, including soil and sediment geochemistry and physics, groundwater quantity and quality, and infaunal communities
Consider complications associated with wetland creation or restoration in seriously degraded or disturbed sites
Conduct early monitoring as part of adaptive management
The first two guidelines (Table 3) relate to the need for a landscape or watershed approach when planning mitigation projects. By this we meant that planners should survey watersheds for needs, opportunities and constraints on wetland restoration and creation; identify and map the best possibilities; match mitigator needs to opportunities; and seek to retain wetland structure and function within watersheds. We hoped to catalyse the development of watershed-based restoration plans, and we gave the example of North Carolina, which has plans for the state's 17 watersheds (NRC 2001, appendix B). The key was the development of watershed improvement plans in advance of damage, so that subsequent proposals to mitigate wetland impacts could fit into a broader and more comprehensive evaluation of existing habitat resources and critical restoration needs. Watershed-scale plans that identify the best opportunities for restoring wetlands are urgently required, along with strategies that prioritize sites for restoration of critical wetland functions (Zedler 2003).
RECENT CHANGES IN MITIGATION GUIDANCE
The ten guidelines of Table 3 were reiterated by the CoE in their October 2001 and December 2002 Regulatory Guidance Letters, which also adopted parts of the watershed concept. However, the CoE (2002) does not include a requirement for a watershed plan in advance of identifying mitigation sites. Mitigation will thus continue to drive the reconfiguration of watersheds, rather than having watershed needs and functioning drive the positioning and design of mitigation projects.
Under the current Regulatory Guidance, the CoE (2002) encourages districts to undertake functional assessments. They are also asked to ensure no net loss of functions and to require more than 1 : 1 compensation areas where the impacted functions are high and the replacement wetland has limited functions. The desired monitoring period is extended to ‘an adequate period of time, normally 5 to 10 years, to ensure the project meets performance standards’. However, the guidelines still allow less than 1 : 1 compensation area if the impacted site has limited function and the mitigation site provides high levels of function. In addition, the guidelines also allow preservation of existing wetlands and the designation of upland, including ‘buffers’ (upland areas adjacent to wetlands that are designated as protecting the wetland from harm). There is no guarantee that uplands or buffers will provide those functions that are lost when wetlands are filled. The standards for buffers are that they ‘typically consist of native plant communities (i.e. indigenous species)’ and ‘enhance aquatic functions’. Further guidance on buffers is forthcoming (CoE 2002). Obviously, it will be hard to track net loss of area and function unless areas of preserved wetland, uplands and buffers that are designated as mitigation sites are separately recorded in the compensation database. Contrary to our call for improved tracking of mitigation requirements and outcomes (Table 4), the CoE's Regulatory Guidance Letter lacks a requirement for a comprehensive national database.
Table 4. Selected conclusions and recommendations of the NRC panel (extracted and paraphrased from NRC 2001).
Wetland functions are not being sustained by the mitigation programme, despite progress in the last 20 years
1. CoE should develop regional reference manualsto help design projects that can best fulfil permit requirements
2. Regulators should not permit filling of wetlands that are difficultor impossible to restore (e.g. fens, bogs)
Gaps in data (e.g. failure to quantify functions of wetlandsto be filled) prevent a complete evaluation of mitigationprojects nationwide
The CoE should improve data collection (compliancemonitoring) and reporting
A watershed approach would improve permit decision-making
1. States, in cooperation with federal agencies, should prepareplans to protect, acquire, restore, enhance, and create prioritieson a watershed basis
2. Planners should incorporate hydrological variability intowetland mitigation designs and evaluations
Performance expectations are often unclear and compliance is often not attained
1. Performance criteria should be clearly stated in permits,and requirements for compliance must be enforced
2. Permittees should provide a stewardship organization withauthority over the compensatory wetland site
3. Permittees should establish a stewardship fund that isappropriate for the long-term monitoring, management and maintenance of the site
Third-party compensation approaches (mitigation banks, in-lieu fee programmes) offer some advantages over permittee-responsible mitigation
Compensatory mitigation institutions should promote compensatory mitigation sites that result in a matrix of protected,restored and created wetlands, each of which contributes to thephysical, chemical and biological integrity of the waters of eachwatershed
In Wisconsin, mitigation negotiations have typically been resisted, with state regulators concentrating on the need for a water quality certification under Section 401 of the federal Clean Water Act. In May 2000, however, the Wisconsin legislature passed its first wetland mitigation law, requiring the Department of Natural Resources (DNR) to write rules for considering mitigation projects in the state decision process and setting requirements for mitigation projects including banks (off-site restoration or creation projects developed by third parties). Mindful of the political pressures and controversies over wetland mitigation (Lewis 2001), the legislature played an active role in reviewing and modifying the rules developed by agency staff. The new rules (Table 5) became effective in February 2002 and a set of guidelines (DNR 2002) has been developed for use by both the DNR and the CoE in the review and approval of mitigation proposals in Wisconsin. The state regulators now have the primary responsibility for permits to fill wetlands, as authorized by the federal government.
Table 5. Provisions of Wisconsin's rules for compensatory mitigation.
Compensatory mitigation has not replaced the need to look at alternatives. Proposed projects must first undergo an analysis ofalternative options (seek to avoid first, then minimize, damage to wetlands)
Wetland functions and values must be evaluated; the applicant must show that there will be no significant adverse impacts
Permitting damage and allowing compensation will only be allowed in limited circumstances and cannot be considered if the wetland isan area of special natural resource interest (under federal, state, or local rules)
Permitees and bankers must provide an ‘as-built’ report to document how the project was implemented
Mitigation site attributes (hydrology, soil, vegetation, animals, problems) must be monitored, typically for 5 years, to provide data toevaluate whether or not clearly defined performance criteria have been met
All compensatory mitigation (including use of a bank) must be within the same major watershed as the damage site
Deep ponds ringed with cattails and areas used primarily for storm water treatment will not be given credit as mitigation
Financial assurances must be provided by the applicant; these will be released by DNR when the site is complete,i.e. monitoring has been completed, the site has met performance criteria, and the management activities spelled out in the plan havebeen carried out
PROSPECTS FOR THE FUTURE
Despite my emphasis on the shortcomings of mitigation in this summary of the NRC (2001) findings, I believe that good things can happen through mitigation. In the ideal mitigation project, former wetland is restored, and wetland species, structure and functions are recovered; research goes hand in hand with the restoration process, and the outcome is a self-sustaining system. However, achieving this rosy view of mitigation will depend on the proactive efforts of citizens and the continued vigilance of national organizations like Environmental Defense, the National Resources Defense Council, the Audubon Society and the National Wildlife Federation. Interdisciplinary research is needed to fulfil the need (1) to develop and test strategies for restoring wetland functions within watersheds, identifying locations and types of wetlands that are most needed; and (2) to improve restoration tactics at the site level.
Strategic wetland restoration will require the assessment of historical and current wetlands within key watersheds, followed by the selection of watershed-scale objectives. For example, if a watershed is particularly prone to flooding, wetlands that could abate flooding might take precedence. Likewise, a watershed that is notorious for discharging nitrates into local rivers might have water quality improvement as a priority, and a watershed with large blocks of restorable wetlands near healthy wetland remnants might offer particular opportunities to restore biodiversity. Planners could then map potentially restorable wetlands and suggest the best sites to achieve each priority (Zedler 2003). Researchers need to rise to the challenge by developing conceptual models or simulation models that specify the general location, area and qualities of wetland needed to (1) improve water quality, (2) abate flooding, (3) support high biodiversity and (d) combine 1 and 3. Until then, planners might need to rely on brainstorming sessions with stakeholders and experts.
To improve site-based tactics, I recommend the ‘adaptive restoration’ approach (Zedler 2001, Zedler & Callaway 2003), which begins by admitting what remains to be determined in order to restore a particular site. In adaptive restoration, mitigation sites would be designed to test alternative ways to achieve the desired restoration or creation target. A research programme would go hand in hand with the implementation, and restoration would occur with frequent and comprehensive assessments by local researchers, perhaps funded by a 10% surcharge on the mitigation project. One mitigation bank near Chicago, Illinois, has a surcharge of this kind. Adaptive restoration allows problems to be detected and understood so that corrective measures can be prescribed and implemented. The process is iterative, with restoration informing research and research informing restoration. Although it is not practical for every mitigation project to include a research component, the failure to assess alternative restoration actions is, in the long run, extremely costly. A recent project at Tijuana Estuary required the salvage, holding and transplantation of Spartina foliosa plugs at a cost of $100 000, but only a few small patches of this desired plant persist to date. Lacking an experimental approach, no one knows why the plantings failed or how survival might have been improved and nesting habitat provided for the endangered Light-footed Clapper Rail. If, as Turner et al. (2001) calculate, only 19 ha out of 100 ha of restored or created wetlands develop the functions of the mitigation target, it is impractical not to conduct restoration in an adaptive manner. Using knowledge gained in early restoration projects makes later efforts more likely to achieve desired outcomes.
COMBATING SEA-LEVEL RISE
A similar strategy of prioritizing wetland restoration sites within coastal regions and using adaptive restoration approaches at priority sites should also serve stakeholders who seek to abate the loss of wetlands due to sea-level rise. In southern California, some 24 coastal wetlands comprise the entire region's saltmarsh resource, and all of these sites are diminished in size and degraded in quality. The only pristine coastal wetlands in the ‘Californian Biogeographical Region’ are in Mexico, along the Pacific Coast of Baja California. The northernmost of these, at Bahía de San Quintín and Estero de Punta Banda, are threatened with development for tourism, largely for US visitors. Protection efforts have begun for Bahía de San Quintín. In the US, most coastal wetland area is already under federal, state or local ownership, but environmental degradation and declines in biodiversity continue along with human population growth (Zedler et al. 2001). Highest priority in this region should therefore go to protecting the Baja California sites, and indeed efforts have begun with help from citizen action groups in both Mexico and the US. Next, the area of coastal wetland should be increased at the few sites where there is still space to do so (i.e. adjacent to the few wetlands that are not entirely surrounded by urban land uses). Finally, the quality of habitat should be improved and endangered species reintroduced where possible.
In response to the region's critical needs, including some 24 wetland-dependent species that are endangered or sensitive (Zedler et al. 1992), California established and funded the Southern California Coastal Recovery Project. In addition, various mitigation and restoration projects have been initiated. So far, the restoration efforts have increased habitat through dredging of historical fill sites, improved tidal flushing and enhanced populations of the more common and broadly tolerant plant and animal species, but these have not led to full recovery of endangered species. The region's most promising site is the Tijuana Estuary, where 200 ha are available for restoration (via excavation of up to 2 m of accumulated sediments) and where the remnants of natural saltmarsh still support most of the region's threatened species (Zedler et al. 1992).
Because sedimentation is a key problem at Tijuana Estuary and other sites in the region, restoration can take sea-level rise into account by contouring sites to increase wetland now, reserving later modules for future wetland restoration. At the high cost of restoring tidal flushing in this region, the 200-ha restoration programme will probably take 50 years, coordinating saltmarsh restoration with sedimentation-control practices within the watershed. On-site efforts at Tijuana Estuary already follow adaptive restoration concepts (Zedler & Callaway 2003). Each project or module is designed as an experiment, and results from the first module were used to plan the next, and so on. The main constraint is coordinating the availability of research personnel, research funding and restoration funding. For both the first and the second modules, research funds became available about 6 months after the sites were opened to tidal flushing. Ideally, restoration and research efforts would be funded simultaneously as part of a coordinated adaptive restoration programme.
This summary of the NRC review is based on the report (NRC 2001) of a 13-member panel. For their extensive and careful work, I am indebted to: Suzanne van Drunick, project officer for the National Research Council; Leonard Shabman, Vice Chair, Virginia Polytechnic Institute and State University; Victoria Alvarez, California Department of Transportation; Robert Evans, North Carolina State University; Royal Gardner, Stetson University, Florida; J. Whitfield Gibbons, Savannah River Ecology Laboratory, South Carolina; J. Wendell Gilliam, North Carolina State University; Carol Johnston, University of Minnesota-Duluth; William Mitsch, Ohio State University; Karen Prestegaard, University of Maryland; Ann Redmond, WilsonMiller, Inc, Florida; Charles Simenstad, University of Washington; R. Eugene Turner, Louisiana State University. I thank Mark Rehfisch for inviting this article, Chris Feare for many improvements to the manuscript and Dave Siebert of Wisconsin DNR for information on the state guidelines. The Management Authority of the Tijuana River National Estuarine Research Reserve guides the restoration of Tijuana Estuary and ensures that research findings are used in planning and implementing restoration.