River Restoration From Boundary Layer to Watershed: Integrating Physical and Biological Science Over Space and Time II
Presiding: P Wilcock, Applied Physics Laboratory, Johns Hopkins University; M Power, University of California, Berkeley
NB21G-01 INVITED 08:30h
An Experimental Approach for Restoration of Salmon River Ecosystems
River ecosystem theory predicts that dynamic, nonlinear physical and biological processes linking water, heat and materials (biota, sediment, plant-growth nutrients) flux and retention to fluvial landscape change in a habitat mosaic context drive salmon life histories and productivity in freshwater. Multidisciplinary studies and cross-site comparisons within a network of pristine salmon river observatories around the north Pacific Rim support these predictions. Billions of dollars have been spent on salmon-river restoration worldwide to little avail, mainly because salmon biology, rather than ecosystem process boundaries and bottlenecks, is driving restoration goals. I argue that entire river catchment restoration, in relation to these dynamic processes and bottlenecks and also coherent with the estuarine and marine implications of salmon life history parameters, is the only possibility for sustaining or restoring natural productivity and life history (genetic) diversity in salmon rivers. This can be done only in a few places owing to the continual press of human demands on river ecosystems, the morass of legal challenges to proactive salmon river restoration strategies and insufficient understanding of freshwater and marine linkages. The Elwha and Yakima Rivers in Washington, among a few others that I will name, offer real opportunities to restore entire watersheds for wild salmon. These restorations should be viewed as experimental manipulations in which outcomes may be evaluated against norms measured in the salmon river observatory network. Bias from hatcheries and harvest, among other anthropogenic interferences, must be eliminated for such experiments to be evaluated in light of contemporary river ecosystem theory. And, a much more synthetic understanding of freshwater and marine linkages must be forthcoming in concert with a much more robust general theory of river restoration.
NB21G-02 INVITED 08:45h
Searching for an Integrated Watershed Salmonid Population Model
In proposing to restore a stream or watershed, we imply that we know what we are restoring and why. However, in many cases, restoration proceeds without having the tools to adequately assess the efficacy of a project-both in terms of its likely local success and what effects it may have at a larger watershed or regional scale. In salmonid ecology and restoration, our approach has been to "step back" and investigate the degree to which geology, tectonics, and climate determine the relative abundance and temporal variability of the species present in a watershed or reach. In other words, does the habitat provided by a landscape prior to European disturbance (the so-called historical reference condition) allow us to predict, using salmon life history theory, the historical population dynamics of all salmonid species for a watershed or region? Likewise, as changes in physical processes have occurred due to human disturbance, can we predict the differential effects on the abundance of different salmonid species historically present in the watershed? To explore these questions, we have developed the "reference model". The reference model is based on a set of desktop watershed analyses tools that estimate a number of landscape attributes including channel slope, drainage area, stream temperature, and shallow landslide sensitivity. The model uses these tools to predict the spatial distribution of habitat and its quality and quantity. Then, by relating the habitat to life stage specific survival, the model predicts population dynamics under reference and current conditions, and under proposed restoration scenarios. In applying the reference model, we explicitly link salmon restoration targets and plans to an understanding of the role of physical processes on the historical and current population dynamics. As is true with most models, the reference model's predictions have the greatest value when they can be treated as testable hypotheses. Ongoing restoration projects that demonstrate the model and explicitly test model predictions will be presented.
NB21G-03 09:00h
A Conceptual Model for Floodplains in California's Central Valley and a Method for Identifying Representative Floods and Floodplains
Currently, significant resources are being invested in restoring native species and ecosystems in California's Central Valley and the Sacramento-San Joaquin Delta, led by the California Bay-Delta Authority (CBDA). Functioning floodplains provide numerous ecological benefits and floodplain restoration is emerging as important component of ecosystem restoration in this region. We developed a conceptual model that describes the linkages between physical (hydrologic and geomorphic) processes and ecosystem processes and responses on Central Valley floodplains. Central to this model is the role of hydrological variability in driving topographic diversity, ecosystem heterogeneity and ecological processes. We attempt to capture the extremely complex linkages between hydrological variability and ecosystem response through `representative floods.' A representative flood encompasses a set of hydrological variables, such as frequency and duration, which produce a characteristic suite of ecological benefits. For example, frequent, long duration flooding in the spring provides spawning and rearing habitat for native fish and promotes high phytoplankton productivity which can be exported to riverine and delta ecosystems. Less frequent, higher magnitude floods drive extensive geomorphic change upon the floodplain, creating topographic and, ultimately, ecological heterogeneity. Here we describe a process to define, map, and quantify the area inundated by a particular representative flood in the Sacramento River valley. To illustrate we identify the area inundated by a frequent (exceedance probability of 67%), long duration (> 7 days) flood that occurs in the spring. We used paired gauges to find the stage corresponding to the representative flood parameters and compared a plane connecting the gauges to topography in the intervening reach of river. We found that this type of representative flood inundates very little area in the Sacramento Valley; primary areas of inundation are within the Yolo Bypass, an engineered floodplain that flanks the Sacramento River. This analysis can be used to identify areas of floodplain that potentially provide the ecological benefits described in the conceptual model and can guide restoration programs seeking to increase these benefits.
NB21G-04 09:15h
"Desktop Watersheds" in River Restoration: Static to Dynamic Digital Terrain-based Modeling
A "Desktop Watershed" is the goal of developing a process-based model, which uses digital topographic and surface attributes (such as geology, vegetation, and land use) to predict the linkages between land use and ecosystem function; such predictions would then guide management decisions. The central hypothesis is that with accurate topography, spatially referenced information on material and biological properties, and simple process-based models, digital terrain-based analyses can assist resource managers in three ways. First, a Desktop Watershed model could be used to develop broad predictions about the expected spatial distribution of resource properties (e.g., landslide location, river bed grain size, and stream temperature) before going to the field. These predictions then become expected states that field observations can test; the results can then be used to gain deeper insight on how management activities control these properties. Second, a Desktop Watershed model could be used to extrapolate local field measurements to entire watersheds. Third, a Desktop Watershed digital terrain model could model the dynamic linkages between land use and resource state. We propose that much can be learned from the use of an analytical reference state calculation, in which an idealized quantitative statement of an expected condition is made based on simple observable properties such as topography and general climate setting. A key assumption is that relatively immutable properties of a specific watershed (such as topography, drainage area to a point, local slope, aspect, and climatic setting) can be quantified and used in simple mechanistic models to calculate an expected condition in the system. This reference watershed model becomes a theoretical condition that can be calculated for any watershed, but will differ between watersheds depending on the watershed's intrinsic properties. Importantly, the reference model also becomes a null hypothesis by which we can guide fieldwork: it is a state against which to detect and measure deviations from predictions that are caused by processes not included in the model. An example of an analytical reference state is the prediction of the median grain size of river bed throughout a channel network. Other examples include channel and bed morphology estimates, shallow landslide locations, and stream temperatures. These physical attributes can then be linked to habitat conditions that allow an estimate of potential fish abundance. What is now needed-and many groups are working on this--are dynamic models that route water, sediment, wood, heat, and nutrients through the watershed; the models must also link these attributes to ecosystem processes. Gaps in current knowledge make this linkage difficult and currently necessarily crude. Nonetheless, such models would generate hypotheses that would guide further fieldwork.
NB21G-05 INVITED 09:30h
Conflicts in Performance Criteria and a Prioritization Methodology for Stream Restoration at the Watershed Scale
River restoration has become a significant consulting, agency and academic endeavor during the past decade with the driving force for these management activities including mitigation, total maximum daily load concerns, the preservation of critical habitat or recovery of endangered species. Review of restoration activities undertaken by agencies and academia in the 1980s showed that frequently it was difficult to identify the specific project goals or quantifiable metrics. Further, many projects lacked the resources for pre- and post- monitoring to establish project performance and to develop a feed-back mechanism into future design or management strategies. Now, a greater emphasis is being placed on the articulation of clear restoration objectives and post-implementation monitoring to evaluate the performance of individual projects. This trend has coincided with new advances in technologies that are allowing not only the success of individual projects to be quantified, but also how different projects interact. The objectives of specific restoration activities undertaken for different purposes may conflict at the reach or watershed scale. Through examples in Idaho and California, the difficulties in quantifying trends of both physical processes and linkages to ecological response will be illustrated. We look at some of the potential conflicts and benefits in restoration objectives within an individual project and the cumulative effects of a series of projects that may enhance or diminish some individual benefits. An analysis framework is outlined and the temporal and spatial frequency of sampling to detect the consequences of restoration actions is described. For example, two common actions in a channelized reach might be to (1) create a more natural channel section (reducing the width-depth ratio and enhancing geomorphic diversity), and (2) allow a more natural plan-form to restore the floodplain connectivity and sustain the geomorphic diversity. One performance metric that could be used is water temperature. The objective could be cooler conditions in the low flow summer months and larger areas of viable habitat in the winter months when sections of the stream freeze. The first action will reduce thermal gain as the width to depth ratio is reduced. However, the second may reduce low flow velocities, increase the residence time of water and tend to increase the thermal gain in the reach. The relative importance of each process (together with other factors such as shade and hyporheic exchange) will govern the magnitude of the change. A prototype tool for prioritizing restoration actions at the watershed scale that integrates a predictive model of the physical processes (flow, sediment transport and water quality) with the expected ecological benefits to quantify differences in alternatives and to prioritize potential management actions will be described.
NB21G-06 INVITED 09:45h
A General Approach for Establishing National Priorities for River Restoration
The effort to restore the nation's rivers necessarily involves the interplay among physical and biological science, engineering design and implementation, and policy science. Policy science helps decision-makers prioritize where to undertake restoration and what the goal or objective of remediation should be. Some commentators have advocated that restoration be pursued on streams that constitute the "low hanging fruit," that is the streams where the greatest gain can be obtained for the least investment. There is a general need for a unifying concept with which to evaluate these policy tradeoffs. The restoration response function defines the relationship between the costs of flow and sediment alteration and the results of those actions in reversing undesired channel change. Development of these functional relations should be an essential element in stream restoration, especially where there are many environmental management programs in a basin and they all affect regional water delivery and hydroelectric power production. The nation's regulated streams - those that are extensively dammed or diverted - are typically proposed for restoration. These streams can be distinguished into three categories: (1) physically transformed, (2) essentially wild and in sediment deficit, and (3) essentially wild and in sediment surplus. Physically transformed streams are those that have been channelized or leveed. Streams in sediment deficit are those whose sediment delivery has been decreased much more than has the ability to transport the available supply. Those in surplus are those whose sediment delivery now exceeds the ability to transport the available supply. Major efforts at riverine ecosystem rehabilitation are underway in all types of streams. The Lower Colorado River Multi-Species Conservation Program and efforts on the lower Missouri River and the Columbia Rivers are focused on transformed rivers. The Grand Canyon Adaptive Management Program is focused on a river in sediment deficit. The Upper Colorado River Endangered Fish Recovery Program and the Trinity River Restoration Program are focused on streams in sediment surplus. Restoration response functions differ significantly in transformed segments and in conditions of sediment deficit and conditions of sediment surplus. In conditions of severe sediment deficit, large costs are sustained without significant results, but larger benefits are potentially possible for smaller costs in conditions of sediment surplus. Policy makers would be well served if restoration response functions were generally available and regional priorities were made based on an assessment of the comparative costs and benefits of altering flow regimes in different parts of the basin.