Some Suggested Guidelines for Geomorphic Aspects of
Anadromous Salmonid Habitat Restoration Proposals
G. Mathias Kondolf, PhD
4 September 1998
With the passage of the Central Valley Project Improvement Act, the advent of Category III funding under the Bay-Delta Accord, and other sources of funds for salmonid habitat restoration in the Sacramento-San Joaquin River system, the US Fish and Wildlife Service (USFWS) and other agencies are now receiving proposals to fund specific habitat restoration and rehabilitation projects, such as channel and floodplain modification, and spawning gravel enhancement. Experience with habitat restoration projects already funded under the "Four-Pumps Agreement" (between the California Departments of Water Resources and Fish and Game) has demonstrated that despite good intentions, some of the these projects have been ineffective or detrimental because project planning did not adequately consider geomorphic setting on the reach or watershed scale (Kondolf et al. 1996a, 1996b).
Aquatic and riparian habitat for salmon and other organisms are, in effect, a by-product of the existing channel geomorphology. Channel geomorphology in alluvial reaches, in turn, largely reflects an adjustment to prevailing flow and sediment load, as well as effects of human modifications. Many restoration projects around the world have failed because their design did not account for geomorphic influences at the watershed1 scale, such as increased or decreased sediment loads and runoff (Iversen et al. 1993, Kondolf and Downs 1996). Downstream of reservoirs, rivers have experienced reductions in flood regime and sediment supply. It follows that these controlling factors must be understood to plan and design habitat restoration projects. Similarly, an adequate understanding of channel geomorphology at the reach scale is needed to design a restoration project that accounts for actual conditions at the project site (as opposed to generic approaches based on presumed attributes of the channel) and specifies the changes in channel form, distribution of velocities in the project reach, and sediment transport patterns anticipated from the project.
Just as geomorphic factors must be considered in project planning and design, biologically limiting factors must be understood, both on the reach and watershed scale to develop specific objectives of restoration actions. Moreover, the size of these investments in habitat restoration argues for careful evaluation of the actual effectiveness of the projects in achieving their objectives.
Purpose and Scope
The purpose of this report is to suggest some guidelines for preparation and evaluation of salmonid habitat restoration proposals in the Sacramento-San Joaquin River system. The emphasis here is on geomorphic attributes, although some other considerations are also discussed. This report is not intended as comprehensive, as there are many other considerations (geomorphic and otherwise) that might be important in preparing or evaluating restoration proposals. These suggestions are based on review of a number of salmon habitat restoration proposals for Central Valley channels over the period 1994-1997, evaluation of other river restoration projects, and recent scientific literature.
Geomorphology must be considered at both the watershed and reach scales. The issues of particular importance to a specific project depend on the local conditions, but for projects in the Sacramento-San Joaquin River system, altered flow and sediment transport regimes usually figure prominently as constraints on channel behavior. Thus, the elements needed in a sound restoration proposal would generally include basic information on alterations in flow, and sediment supply and transport (Table 1).
Gravel supply and intragravel flow must be addressed for projects to create spawning habitat, whereas bank vegetation and large woody debris might be more important for projects to create holding habitat.top
Geomorphic Setting at the Watershed Scale
To provide an adequate understanding of the geomorphic setting at the watershed scale, project proposals should show the project site in the watershed context, and indicate upstream influences (such as dams) and their effects on flow and sediment load. In addition, biologically significant effects of up- and downstream influences (such as barriers to migration) should be indicated.
The project site should be shown in the context of upstream influences. In some cases, where dams have hydrologically isolated the project reach from some upstream influences, a small-scale map of the entire watershed can be augmented by a larger-scale map of the river downstream of the dams. The watershed map(s) should indicate relevant information such as areas with particularly high erosion rates, land-use changes likely to have altered runoff patterns, reaches important for spawning and rearing habitat, gravel pits, levees, locations of stream gauges, towns, road crossings, and other features that may affect the streams and restoration opportunities.
The flow regime, and any changes resulting from land-use change or reservoir construction, should be quantified. Specifically, the high flows, which have the greatest geomorphic effect on channel form, should be described with a histogram of peaks annual flows and flood frequency analysis (Dunne and Leopold 1978).
For reaches downstream of dams, dam effects on the flood frequency regime can be quantified by comparing (1) pre- and post-dam conditions, if sufficient pre-dam and post-dam gauging records are available; (2) gauges upstream and downstream of the reservoir for simultaneous periods of operation (this may require adjustments for differences in drainage area); or (3) actual flows below the reservoir with inflows calculated from reservoir storage changes (already available for many reservoirs).
Depending on the biological role of the project reach, other aspects of the river hydrology may be appropriate for presentation, such as changes in the frequency duration of flood recession flows or baseflows. Potential implications of changes in the flow regime should be discussed to provide insight into the existing conditions.
Sediment budgets can vary widely in their level of detail and in the components measured and reported. The appropriate scale and scope of the sediment budget depends upon the reason for developing it to begin with. Many sediment budgets in the geomorphic literature were developed to better understand the relative importance of different erosional and sediment transport processes operating in the watershed. These budgets have typically included direct measurements or historical assessments of geomorphic processes such as fluvial erosions of hillslopes, landslide initiation and movement, earthflow movement, bank erosion, and floodplain sedimentation (e.g., Swanson et al. 1982).
For most habitat restoration projects, we not concerned so much with the specific geomorphic processes upstream but rather the runoff and sediment produced by those processes, and the temporal and spatial patterns of its transport and deposition. The purpose of the sediment budget for our restoration projects is to place the proposed project reach in a larger context of sediment supply and transport.
Gravel Supply. For spawning habitat enhancement projects, the sediment budget should include estimates of gravel supply from upstream. For rivers in the Sacramento-San Joaquin River system, the supply of gravel from upstream changed dramatically this century with widespread construction of dams (which trapped the gravels), so pre- and post-dam yields are needed to understand the channel and its evolution. If reservoir sedimentation data are available (from surveys of the reservoir bottom), they can be used to estimate pre-dam sediment supply. Most of the sediment deposited in the reservoir is usually fine-grained, so the percentage of spawning-sized gravel must be estimated from the total. If no reservoir sedimentation data are available, probable rates of natural sediment supply from the watershed can be estimated using data from other rivers in the region, bearing in mind differences in rock type and other factors that might affect gravel yield.
Gravel can also be supplied from bank erosion and tributaries. These sources should be identified and, to the extent possible, quantified. For example, rates of bank erosion can often be estimated from changes visible on aerial photographs or estimated from changes relative to known landmarks such as buildings or fences. By estimating the proportion of gravel in the sediments exposed in the cut bank it is then possible to estimate the amount of gravel that would have been contributed by bank erosion. Along many rivers below dams, bank erosion is an important source of gravel (largely because upstream supply has been eliminated), and the supply of gravel from bank erosion has been reduced by bank protection works.
Gravel Losses and Transport. The sediment budget should include estimates of losses, notably direct losses to aggregate extraction and subsequent trapping of gravel in pits in the river upstream. Potential transport rates under pre- and post-dam flow regimes should also be estimated for the project reach, based on field measurements, observations of tracer gravel movement, or calculations of sediment transport competence and capacity. (Competence refers to the largest sizes of gravel the river can transport, while capacity refers to the total mass of sediment it can transport.)
Fine Sediment Sources and Transport. If fine sediment deposition in gravel and cobble substrates is an issue, the sediment budget can be expanded to address fine sediment (sand, silt, and clay) sources and transport. The seasonal timing of fine sediment delivery to the channel is particularly important because fine sediment delivered during summer base flows (from agricultural erosion and irrigation return flow) is likely to deposit on the bed, while fine sediment contributed during high flows will likely be washed downstream without depositing.
Spatial Relations. The sediment budget should be presented in part in map format, showing spatial relationships of upstream dam, tributaries, instream or captured gravel pits, other potential sediment sources and sinks, and important spawning and rearing habitats. Potential implications of recent changes in the sediment budget supply should be indicated.
Large Woody Debris Supply and Transport
As the importance of large woody debris in aquatic habitat on the local scale has become increasingly recognized, so too has the importance of large woody debris supply and transport at the river basin scale, both to maintain the continuity of supply and transport to downstream reaches, and in regulating sediment transport through the river system (Abbe and Montgomery 1996, Malanson and Butler 1990, Nakamura and Swanson 1993, Piegay et al. in review). Many artificial habitat structures are constructed of logs and, in effect, are attempts to replicate some of the functions of natural large woody debris in increasing channel roughness, providing high flow refuge and cover for fish along channel margins, inducing formation of scour pools, and regulating the transport of gravel through the river system. It only makes sense to analyze the current supply and transport of large woody debris, to understand how this may have changed due to human influences, and to consider whether some of the functions for which artificial habitat structures are proposed might not be achieved on a more sustainable basis by permitting (or encouraging) large woody debris to enter the channel and to transport through the river system.
For example, on a large active, gravel-bed river, much of the instream habitat may be created by large woody debris, but the debris moves downstream with each flood. Thus, the overall area of habitat may remain constant, but the actual locations of the habitat units may change. In such active system, artificial habitat structures installed to improve habitat on a local, reach scale are unlikely to remain stable. However, if the supply and transport of large woody debris from upstream can be maintained, aquatic habitat can be enhanced in the local reach and throughout the river system (Piegay et al. in review). Similarly, by increasing channel roughness, large woody debris can increase the potential retention of spawning gravel within the channel (Buffington et al. 1997) naturally.
Large woody debris has been reduced in many channels through direct removal (out of concern for channel capacity or potential debris trapping on downstream bridges) and through reduced ed by high flows, the flows at which the bed is expected to mobilize and rearrange itself should be stated (and supporting calculations summarized). Similarly, proposals to import gravel to create or enhance spawning habitat should indicate the critical shear stress for the imported gravel, the shear stresses anticipated under post-project conditions, and the likelihood of the imported gravels washing away in a given year or other period of time. The planned management response to loss of the gravels should also be stated. In some cases, the short term benefit of spawning habitat created may be such that it is worthwhile to continue adding gravel every year or two; this is perfectly reasonable, so long as the reach geomorphology is considered, the losses are anticipated, and a decision is consciously made to add gravel on a frequent basis. Thus, continued addition of gravel in the future can be anticipated and so stated, and the project can then be evaluated relative to predictions.
Bank, Floodplain, and Terrace Revegetation
If the project involves establishing riparian vegetation on an existing or newly created surface, the factors influencing vegetation success should be quantified and stated. Hydrologic and geomorphic factors include inundation frequencies of the surfaces and depths to water table during fall baseflow, and soil texture. Biological aspects should also be described, such as source of plantings, depth to which cuttings will be planted (and relation of this elevation to the fall water table), seasonal timing of plantings, and strategies to control weeds. The potential for long-term recruitment of large woody debris to the channel should be considered.
Intragravel Flow in Redds
For riffle reconstruction projects, it is not only important that the project create the substrate, water depth and velocity conditions suitable for spawning nt of the study reach will depend on the site and project purpose, but as a very general rule, a study reach length of 20-50 channel widths (width at bankfull), with 10-15 cross sections spaced two-to-five channel widths apart is sufficient. If systematic undulations of the bed (such as pools and riffles) are present, the cross sections should be located to reflect those features, with replicates for similar morphological units. For example, on a meandering channel, cross sections could be located on the apex and crossover of each bend. Channel surveys should be tied to permanent benchmarks and the survey data retained in such form as to provide adequate information that new staff can replicate the surveys.
Bed Material Size. Appropriate measures for bed material size depend upon the purpose of the project. To enhance pool-riffle morphology and provide riffle substrate for juvenile holding and invertebrate production, surficial bed material sampling is adequate. If spawning gravel quality is a concern, subsurface sampling is probably needed to determine the percentage of fine sediment within the gravel. To measure sediment size on the bed surface, the pebble count (Wolman 1954, Kondolf 1997) is a tried and tested method. A recent variant termed the zig-zag count (Bevenger and King 1995) is not recommended because it does not yield an adequate sample size (n) or reproducible particle size distributions, and it mixes sample points from a variety of habitat units. More guidance on assessing salmon spawning gravel quality is presented by Kondolf (in preparation) in Appendix A.
Streamflow Data. An accurate record of streamflow through the project reach is essential to understand project performance and biological response. For most anadromous salmon restoration sites in the Central Valley, gauging data are probably available. If not, establishment of a recording gauge should be explored. At the very least, flood crests should be recorded with a crest stage gauge at the site, and the peak flows calculated using the slope-area method (Rantz et al. 1982).
Depth to Water Table and Groundwater Interactions are key controls on riparian vegetation establishment. Shallow monitoring wells can be installed in the banks and floodplain to document the water table conditions experienced by the plants and the relationship of the water table to the stream. Installation of monitoring wells requires drilling wells, which is easy in sandy substrates, difficult in cobbles and boulders (MacDonald 1988). In addition, because seepage of water into or out of the streambed can be an important attribute of salmon spawning habitat, seepage meters (Lee and Cherry 1978), dye studies (Stuart 1953), or standpipes (Terhune 1958) may be useful in documenting intragravel circulation.
Period of Monitoring. The channel should be surveyed before project construction to establish pre-project, baseline conditions, immediately after construction to establish as-built conditions, and afterwards over as long a period a possible, preferably at least a decade. The post-project channel surveys need not be done every year, but over a long period of time with flexibility built into the schedule so that the channel is surveyed after large floods. This is sometimes referred to as "pulsed monitoring", in that the observations are triggered by floods. Even with pulsed monitoring, it is useful to have a default monitoring schedule in the event that high flows do not occur. For example, the decision could be made to resurvey the channel after the as-built survey five times over the decade, in years 1, 2, 4, 7, and 10, or following each flow exceeding some threshold such as the three-year flood. If a flood occurred in year 2, the channel would be resurveyed anyway. If a flood occurred in year 6, the channel would be surveyed then and not again until year 10 (Kondolf 1995).
Table 1. Geomorphic Elements in Salmonid Habitat Restoration Proposals
Geomorphic setting at reach scale
Post-project evaluation of geomorphic conditions
A number of other considerations arise in evaluating habitat restoration proposals, some providing ecological context for the geomorphic issues, some dealing with project implementation and future management of the reach.
Existing information on the target species and run, and the role of different reaches in life history of target fish need to be summarized as a basis for stating how the proposed project will improve overall survival and natural reproduction of salmon. Implicit in this limiting factors analysis is an evaluation of alternatives, at least at the level of stating how the proposed project actually addresses a limiting factor, and thus why this project should be undertaken rather than a project addressing a different stage in the species' life history.
For proposals to import gravel into the channel or rip existing gravels, the proposal should provide reasonable justification that the existing gravels are actually unsuitable, either because the framework sizes are too large for the salmon to move, because too much fine sediment is present, or because gravels have become compacted and immobile. Size distributions for existing gravels should be presented and the proposal should compare (1) framework size with the maximum sizes movable by the fish present, and (2) fine sediment content (e.g., percentage finer than 1 mm) with maximum acceptable levels of fine sediment. If the gravel is believed to be compacted (i.e., the framework size does not appear to be excessive, but gravels are interlocking in such a way as to be immovable), this should be stated and supported by field observation to the extent possible. Unfortunately, no straightforward method exists to measure compaction.top
Clear Statement of Objectives
While the ultimate, long-term goal is to increase natural reproduction of the target fish population, the specific objectives of the project need to spelled out, not only in biological terms, but also in terms of the specific, physical channel changes anticipated. This can be justified on two grounds. First, the aquatic ecology depends upon physical channel conditions, so if natural geomorphic conditions can be recreated, there is a high probability that the associated organisms will return (Brookes and Shields 1996). Secondly, most habitat restoration projects are directly affecting only the habitat itself, that is to say, the channel (and/or floodplain) form or process. Especially with anadromous species, a project that successfully created ideal habitat conditions might not result in increased populations for reasons completely unrelated to the quality of the physical habitat at the site, but due to factors such as passage problems downstream, over-harvesting, or predation (Kondolf and Micheli 1995). Thus, the physical effects of the project are best treated as distinct from the biological effects.
In any case, general restoration goals must be translated into specific, measurable objectives to evaluate the performance of the project and to gain insights for the design of future projects. For example, the general goal of "restoring spawning habitat" should be expressed in terms of specific channel conditions and resulting habitat features that are to be created.top
Given the uncertainties of the physical and ecological behavior of the complex riverine system, there is a growing consensus that habitat management and restoration needs to be approached with flexibility to allow modifications in response to observed system responses. The concept of adaptive management (Walters 1986) relies on good monitoring data and ongoing evaluation of project performance. Moreover, the concept implies that some management decisions be deliberately taken to test the system response.top
Available Information versus Original Data Collection
Some of the components recommended here can be drawn from existing geomorphic studies. In the Sacramento-San Joaquin River system, relevant watershed-scale information can be obtained from existing studies on the Merced River (Vick 1995), the American River (Vyverberg et al. 1997), the Sacramento (Parfitt and Buer 1980, Buer 1984), the San Joaquin River (Cain 1997) and its tributaries (CDWR 1994), and other reports. In general, adequate description of reach-level conditions will typically require original work (including field surveys) to develop a credible base upon which the project can be planned and evaluated.top
Funding for Project Planning
Pre-project research to establish the site's larger context and to establish baseline conditions at the site can be undertaken prior to submission of a proposal for project implementation as a contribution of the agency proposing the project, or it can be funded as a separate Phase I planning study, whose results would later be used to develop a specific project proposal. It should be recognized that such planning studies might result in no specific project proposal because the available information suggests that a project is inappropriate at the site. For example, the planning study might indicate that additions of gravel to the channel (for spawning enhancement) would likely wash out in relatively modest floods, or that spawning habitat was not the factor limiting a fish population. Or with better understanding of a still-dynamic river, it may be recognized that the best restoration strategy for a human-modified reach of channel would be to leave the channel alone and allow natural recovery to proceed. In either case, the funds for the planning study would be well spent because they prevented larger expenditures on construction of projects that would prove to be ineffective or unnecessary.
There are fundamental problems with granting funds to large package-deal projects in which project planning and design are included as part of the proposal because the desirability of any project at the site has yet to be demonstrated. The site may be inappropriate for geomorphic or biological reasons, but once funds have been granted, there may be pressure within the grantee agency to justify some sort of project at the site. Similarly, it may be difficult for the granting agency to withdraw funds and cancel the project even if the appropriateness of the project appears increasingly doubtful as planning proceeds.
In general, the conceptual design for a project must be worked out before the suitability and potential effectiveness of the project can truly be evaluated. Specific information is needed on how the project proposes to modify channel conditions and processes, and how the modified channel is likely to interact with future flows.top
In addition to evaluating the scientific rationale for the proposed project, proposal reviewers must evaluate the suitability of the proposed project budget. This requires that the budget be sufficiently detailed that the reviewer can ascertain whether the time commitments (and rates) for various tasks are reasonable. Lump sum values attached to various tasks provide little basis for evaluating the reasonableness of the budget items, especially as a large project can be broken into smaller tasks in many different ways.
Abbe, T.B., and D.R. Montgomery. 1966. Large woody debris jams,
Bevenger, G.S., and R.M. King. 1995. A pebble count procedure for
Brookes, A., and F.D. Shields. 1996. River channel restoration:
Buer, K. 1984. Middle Sacramento River spawning gravel study.
Buffington, J.M., D.R. Montgomery, and H.M. Greenberg. 1997.
Burner, C.J. 1951. Characteristics of spawning nests of
Cain, J. 1997. Hydrologic and geomorphic changes to the San
CDFG (California Department of Fish and Game). 1996. Request for
CDWR (California Department of Water Resources). 1994. San Joaquin tributaries spawning gravel assessment:
Dunne, T., and L.B. Leopold. 1978. Water in environmental
Frissell, C. A. and Nawa, R. K. 1992. Incidence and causes of
Harrelson, C.C., C.L. Rawlins, and J.P. Potyondy. 1994. Stream
Healey, M.C., 1991. Life history of chinook salmon. p. 311-394 in
Iversen, T.M., Kronvang, B., Madsen, B.L., Markham, P., and
Kondolf, G.M. 1995. Five elements for effective evaluation of
Kondolf, G.M. 1997. Application of the pebble count: reflectionson purpose, method, and variants.
Kondolf, G.M., and P. Downs. 1996. Catchment approach to channel
Kondolf, G.M., and E.M. Micheli. 1995. Evaluating stream
Kondolf, G.M. Assessing salmonid spawning gravels.
Kondolf, G.M., J.C. Vick, and T.M. Ramirez 1996a. Salmon spawning
Kondolf, G.M., J.C. Vick, and T. Ramirez 1996b. Salmonid spawning
Lee, D.R., and J.A. Cherry. 1978. A field exercise on groundwater
MacDonald, L. 1988. An inexpensive, portable system for drilling
MacDonald, L., with A. Smart, and R. Wissmar. 1991. Monitoring
Malanson, G.P., and D.R. Butler. 1990. Woody debris, sediment,
Nakamura, F., and F.J. Swanson. 1993. Effects of course woody
Parfitt, D., and K. Buer. 1980. Upper Sacramento river spawning
Piegay, H., A. Thevenet, G.M. Kondolf, and N. Landon. Physical
Rantz, S.E. and others. 1982. Measurement and computation of
Stuart, T.A. 1953. Water currents through permeable gravels and
Swanson, F.J., R.J. Janda, T. Dunne, and D.N. Swanston. 1982.
Terhune, L.B.D. 1958. The Mark VI groundwater standpipe for
Vick, J. 1995. Habitat rehabilitation in the Lower Merced
Vyverberg, K., W. Snyder, and R.G. Titus. 1997. Lower American
Walters, C. 1986. Adaptive management of renewable resources.
Wolman, M.G. 1954. A method of sampling coarse river-bed