Analog, Empirical, Process and Hybrid Techniques for Channel Design

by Karen Torres

Wiki Overview

Image from SRWQIS Webpage
Stream restoration takes into account the relationship between the stream channel function, watershed processes and ecosystems with the goal of returning to a former natural or stable condition. To optimize restoration design, analog, empirical, process and hybrid design methods were developed to mimic stable stream reaches and predict changes in stream characteristics as a result change in a watershed parameter or stream configuration. Stable design considers the natural migration of a channel while maintaining equilibrium between cross section, floodplain and slope.

The objective of this wiki space is to provide an introduction to analog, empirical, process based and hybrid design techniques for channel reconstruction to meet ecological and engineering criteria. These methods estimate stable channel design parameters by adopting features of a stream that is in equilibrium and by calculating channel characteristics. Such methods allow the designer to consider channel form that can withstand maximum flows without aggregation or degradation of the streambed. Analog design adopts parameters from a similar stream or previous stream conditions as a restoration goal. Empirical techniques relate measurable processes, such as discharge to channel geometry by the application of empirical equations. Process or analytical approach uses physically derived equations of continuity, flow resistance and sediment transport to estimate specific design component. A hybrid approach considers stabilization and ecological goals in restoration design. Strengths, weakness and appropriate use of each technique are discussed as single approach does not work for all restoration projects.

Design in Stream Restoration

Steam restoration generally takes place when a function of a stream is hydrologically or ecologically impaired. Identification of a technology or design to improve stream functionality depends on site conditions and data availability. Examples of restoration include but are not limited to detention of urban runoff, dam removal, induced stream meanders, vegetative management, channel structures and bank lowering. The characterization of hydrology, hydraulics, habitat, vegetation and land use activity as individual components is the first step in understanding the complex processes that maintain stream functionality.
Restoration design methods and models are developed to predict changes in stream characteristics, such as high and low flows, erosion, sediment transport as a result of change in a watershed parameter on stream configuration. Design goals seek to maintain the natural form of a channel while maintaining equilibrium between cross section, floodplain and slope of the stream channel. Adhering to these criteria saves expense of maintenance and encourages sustained results (NEH 654-7). Depending on the type of stream and nature of the impairment, design goals may be unique even within the same drainage. A stable stream configuration may have wide meanders which, in some instances, are not possible to mimic if no deviation of the current channel shape is permitted.
In 1998 the Federal Interagency Stream Restoration Working Group (FISWG 1998) identified in-stream restoration practices as, stream bank restoration, stream habitat recovery and channel reconstruction.
Restored Meander
The goal of stream bank restoration is to reduce erosion rates by vegetation or the use of hard structures to provide stream stability. Stream habitat recovery involves the assessment of the quality and quantity of habitats provided by the proposed design. Channel reconstruction is the alteration of a channel to historic or stable conditions which include the following basic procedures:

1. Describe the physical aspects of the watershed and characterize the hydrologic response
2. Select a stream reach to be restored and compute valley length and slope
3. Determine bed material for new channel
4. Conduct a hydrologic and hydraulic analysis to select a single or a range of discharges.
5. Determine an appropriate channel configuration to meet restoration goal.

A tool for effective restoration design is the use of models to predict stream response to changes within the channel or off site. Since models are merely a gross representation of real system they may not accurately characterize every single watershed processes but can use the best available data to develop meaningful alternatives and good design. One of the benefits is saving time and costs to simulate a restoration area based on current relationships than to spend years collecting necessary design data. Once parameters such as precipitation, land cover, channel geometry and topography are described and correlated to a hydrologic response; they can be modified to reflect a proposed change caused by restoration activities and predict a new response.

Method and Model Overview

The Fleming classification (Boonstron 1980) of Methods and Models, charted below, subdivides hydrology methods into deterministic and statistical components. Deterministic models treat the hydrological processes in a physical way and make use of historical stream flow data as well as data on rainfall and other phenomena which affect the properties of runoff. This method always produces the same output from a given input and are classified as empirical or conceptual.
Figure 1: Flemming Classification of Methods and Models (after Boonstron)

As the name implies, empirical methods are based on observation or experiment to relate measureable processes to a particular response. Examples are hydraulic geometry and regime equations which take stream data and by correlation and regression methods, describe the relationship between channel geometry, discharge and slope

In the conceptual category direct models use physical properties to represent a system. Examples are miniature or small scale version of a particular watershed or where flow of electricity simulates the flow of water. (Brooks 2003) Semi direct methods use an analog approach where a channel characteristics from another stream are applied to channel design without a hydrologic or hydraulic analysis. Process design is an indirect method which relies on mathematical solutions and the description of independent variables to predict dependent variables, such as sediment load, discharge and channel geometry.

Statistical methods utilize information from the analysis of historical stream flow and use the theory of statistics to predict possible flow outcomes. Stochastic Models use a random element to generate different outcomes with each model simulation. The probability of the various model results are calculated and used to predict the frequency of a particular scenario. Though statistical models are valuable in areas with high variability, they will not be discussed in this wiki.

Analog Methods for Channel Design

The simplest design is the analog approach which takes the dimensions of a stable channel and replicates them in the restoration or design reach. To apply this method the bed and bank materials, sediment inflow, slope, valley type and annual discharge should be the same in both the referenced stable channel and the design reach (NEH 654-9). This is considered the simplest approach to design as hydrologic and hydraulic analysis are minimized (Skidmore 2001).
Figure 2: Rosgen Stream Classification Overview (EPA Webpage)
The reference reach approach by Rosgen (1996) is an analog method where a classification system describing morphological and
physical variables (Figure 2) with design based on

dimensionless ratios of a stream class rather than actual

An overview of this classification system can be found
on the EPA Website Rosgen Stream Classification

Another method is a cross-section analog where measurements from stable reaches are used to estimate bankfull discharge, dominant discharge and sediment transport character. (Skidmore 2001)

Benefits of the Analog Approach
This is a low cost design method which simplifies specifications. Additionally, the design goals are easily explained to stakeholders and other interested parties as an example exists for demonstrative purposes. Additionally this can not only apply channel geometry but can also apply to design for physical habitat, such floodplains and riffle pools.

Limitations of the Analog Approach
Several important design considerations, such as watershed and boundary conditions, are presumed identical and at equilibrium in the stable referenced reach, as well as the design reach. If unstable conditions exist, the cause may be systemic of the watershed as a whole, not a localized issue. In this event all stream reaches may be unstable and are suspect for eligibility for an analog approach. (Skidmore 2001)

Empirical Design Methods

In an empirical approach to design, the observation of measurable stream components, such as slope, bed material, sediment, vegetation and discharge are used to establish cause and affect relationships. The simplest examples are the hydrograph, where basin complexities are defined by a single empirical curve of the relationship between precipitation and stream discharge. (Boonstra 1980).
In the hydraulic geometry approach, proposed by Leopold and Maddock (1953), scatter diagrams of the relationship of discharge to the channel's width, depth, velocity and sediment load were created based on measurements of various streams in the United States. It was found the empirical relationships of average values were similar in nearly all river systems regardless of location and environmental setting. (Figure 3) A simple example is at specific discharge and constant velocity, an increase in stream width results in a decrease in suspended sediment load (Leopold 1953).
Figure 3: Correlation Plots of Depth, Width and Velocity vs Discharge
A basic goal of stream restoration is to develop a design to nudge an impaired system back to a state of equilibrium or what is referred to as grade. Merely looking at plotted empirical data does not indicate if a stream is in a state where it is neither aggrading nor degrading. Studying the hydraulic geometry of a graded stream section may lead to a better understanding of the factors which created this equilibrium. This can assist the engineer in the development of effective and durable design.
With the use of empirical data from regime channels, which are stable natural rivers, regime equations are derived through statistical regression (NEH 654-9) Relationships between channel geometry, discharge and slope are calculated using empirical coefficients from grade channels and applied to design channels.

Benefits of Empirical Methods

Using this approach design can be accomplished with information for only a few variables and the selection of constants. Many empirical equations exist, giving the designer a wide variety to choose from. This method provides cost savings, if applicable equations exist, as do not have to be derived for a specific project.

Limitations of Empirical Methods
The application of empirical equations is limited by location and quality of the data set. Data collected at a specific location in a stream may quantify existing relationships but does not adequately represent the processes which impact channel development. Skidmore (2001) states generally empirical equations do not directly account for sediment supply, bed material gradation, bank cohesion, vegetative character, slope or roughness, all of which influence natural channels.

Process Based Design

The process or analytical method for channel design assigns a mathematical operation to each major physical process such as discharge, sediment inflow, geology, bank vegetation and bed material; which are all independent variables. By understanding of the interrelation of these processes on channel form and function, potential changes in channel geometry can be estimated to assist in a stable design (Langendoen, 2001). Soar (2001) shows the equations that use independent variables to solve for dependant variables through the solution of physically based governing equations as shown below :
Figure 4: Lower Santa Fe River Basin HEC-HMS Model (after Gant 2012)

Conservation of Mass is used to solve for discharge and sediment transport where Flow of Water: -Discharge (Q)
= Velocity x Cross Section
= Velocity x Width x Depth
= Velocity x Wetted Perimeter x Mean Hydraulic Radius

Sediment Concentration (C) = Discharge of Sediment / Flow of Water

Fluid Motion describes the flow of water by the fluid momentum relationship (flow resistance expression). This is also used for
bed sediment through a bed material load transport equation which incorporates bed load and suspended load. Bank sediment is accounted for through bank erodability/mass failure models.

For channel geometry, the solution to one dependent variable of width, slope or depth based on stream characteristics is used to solve for the two remaining unknown variables. (NEH 654-9) In stable design, the channel depth, slope, roughness and plan form are selected to allow the sediment load to pass with no aggregation or degradation. Through the simultaneous solution of three hydraulic equations the mean depth (or maximum depth in a trapezoidal cross section), slope and width of a stable channel can be estimated for a straight channel with specified discharge and known boundary roughness. This does not consider plan form or bed form characteristics of natural channels, assumes uniform, steady and essentially one-dimensional flow and conservation of mass. (Soar 2001)These hydraulic equations are summarized as follows:

Flow Resistance: Mean Depth = f (Slope, Width, energy losses
Sediment Transport: Slope = f (Mean Depth , Width, flow resistance)
Bank Erosion: Width = f (Mean Depth , Slope, bank angle, flow resistance, bank sediment character and other factors

A wide range of equations exist to solve for discharge, sediment transport, maximum channel slope and base width. For each specified combination of discharge, sediment transport rate and transport grain size, unique values of slope and depth are calculated (NEH 654 9-34). The numerous calculations involved in the method requires the use of a hydraulic model, Commonly used programs, developed by the U.S. Army Corp of Engineers, are HEC (Hydraulic Engineering Center) - Geo-HMS (Geospatial Hydrologic Modeling Extension) and SAM (Stable Channel Analytical Method).

A HEC- GeoHMS model consists of three major components. A basin model where the watershed is delineated by sub-basins through GIS software. A precipitation model specifying how precipitation will be generated for each sub-basin. Finally a control specification model, which uses an synthetic unit hydrograph method (empirical) to estimate the time base of the hydrograph and width at a specific peak flow. Ultimately peak flow calculations performed are imported back into the GIS software to create flood plain maps. (Gant 2010) Figure 4 is an image of the basin model for the Lower Santa Fe Watershed where sub-basins are delineated from DEM data.

Advantages of a Process Based Approach
Figure 5: Pools and Riffles

This method is well suited for complex urban runoff issues that involve complex land use patterns and stationary channel constraints. Depending on the size of the project area, this approach can save time and money by estimating historic conditions and calibrating to current conditions to project various future scenarios. Software can easily solve for multiple unknown variables iteratively offering a range of solutions.

Process Based Approach Limitations

The level of analysis required makes this method impractical for small projects. Shields (2003) admits the use of analytical equations are generally better than empirical formulas but warns this assumes expertise in the selection of resistance coefficients or appropriate sediment transport relations. Models based on data with a low level of confidence produce poor un-defendable results, which will save neither time or budget.

Hybrid Channel Design

The hybrid design blends physical and life sciences to balance the goals of stabilization and habitat enhancement. (Shields 2001) Stable design to improve habitat quality may actually involve destabilizing a stream section to restore pre-impairment function. Design criteria to improve stream function consider the availability of in stream physical habitat where flow regime is an important factor in ecological integrity. Physical properties of pools and riffles are illustrated in Figure 6 in regard to location and flow regime. The engineer has to choose a stable design taking into consideration, for instance, over-bank flooding or in stream flow discharge.

Rock Weir: Santa Cruz River, New Mexico
Selecting an Approach

Design method driven by size and scope of project. Small projects call for an analog or empirical approach, where projects with a large area or complex flows would call for a more extensive design method. Figure 4 is an example of a small project where bank erosion was causing damage to a bridge footing. Design criteria for a Bendway Weir was used to select size and placement of boulders in a streambed which required and evaluation of the peak hydrograph and channel geometry.

One of the common goals design for restoration projects involves inducing meanders to return stream stability and function . An example of empirical design considerations for a meandering stream is outlined in NEH and summarized in the table below. A combination of field measurements, standard calculations and empirical formulas are suggested as a basis of design.

Meander Design Considerations
Selected Method (after NEH 653 8-31)
Determine meander geometry and channel alignment
Empirical formulas for meander wavelength are used
along with measurement from undisturbed section
Compute sinuosity, channel length and slope
Use field measurements of length, sinuosity and valley slope.
Channel length = sinuosity x valley length
Channel slope = valley slope / sinuosity
Compute mean flow width and depth at design discharge.
Empirical Regime or hydraulic geometry formulas or analytical methods.
Compute riffle spacing
Empirical formulas, observation of similar streams


Stream restoration design takes into account the relationship between the stream channel function, watershed processes and ecosystems with the goal of returning to a former natural condition. Design method is driven by the scope, goals, size and budget of a project but whether analog, empirical , process or hybrid are used they all have same underlying theme of equilibrium. Stable design considers the natural migration of a channel while maintaining equilibrium between cross section, floodplain and slope.

All the methods discussed are attempts to find the closest channel configuation for stability. These methods can be used in conjunction with field work or as stepping stones to a more complex design, if necessary. Just as the processes involved in the watershed are complex so is restoration.

Boonstra, J. and de Ridder, N.A., 1980, Land Reclamation and Water Management Developments, Problems and Challenges International Institute for Land Reclamation and Improvement Publication 27

Brooks, K.N., et al., 2003 Hydrology and the Management of Watersheds 3rd ed., Iowa State Press, Ames Iowa.

Federal Interaagency Stream Restoration Working Group (1998) Stream Corridor Restoration: Principles, Processes and Practices. National Technical Information Service, U.S. Department of Commerce, Springfield, VA.

Gant, S., Shoaff, L (2010) Draft Hydrologic and Hydraulic Assessment Appendix- Santa Fe River, New Mexico Watershed Management Plan Study, U.S. Army Corps of Engineers. Unpublished
Langendoen, Eddy J., Simon, A. and Thomas, R.E. ,2001 CONCEPTS- A Process-Based Modeling Tool to Evaluate Stream-Corridor Restoration Designs, USDA-ARS National Sedimentation Laborator, ASCE Publication ( American Society of Civil Engineers), Proceedings of Wetlands Engineering and River Restoration Conference Reno Nevada Aug 27 - 31 2001
Leoplod, L.B. and Maddock, T. 1953. The Hydraulic Geometry of Stream Channels and Some Physiographic Implications. U.S. Geological Survey Professional Pater 252, Washington, DC.
Rosgen, D.L., 1996. Applied River Morphology. Wildland Hydrology Books, Pagosa Springs, Co.

Shields F.Douglas,C.M Cooper,Scott S Knight,M.T Moore , October 2003. Stream Corridor Restoration Research: A Long and Winding Road. Ecological Engineering Vol 20 (441), Elsevier B.V.

Shields, F.D, Copeland, R.R., Klingeman P.C., Doyle M.W. and Simon A. (August 2003) Design for Stream Restoration. Journal of Hydraulic Engineering , Vol 129, No 8 (575) ACE (American Society of Civil Engineers) Publication, Reston, Va.

Skidmore, P.D., Shields D.,Doyle M., Miller D., 2001 A Categorization of Approaches to Natural Channel Design. In Wetlands Engineering and River Restoration, ed. D.Heyes,ASCE Conf. Proc. doi:, ASCE Publication ( American Society of Civil Engineers), Proceedings of Wetlands Engineering and River Restoration Conference Reno Nevada Aug 27 - 31 2001

Soar, P.J., Thorn, C.R. (2001) Channel Restoration Design for Meandering Rivers, CR-01-1 Research and Develop Center / Coastal and Hydraulics Laboratory, US ACE Washington DC.

U.S. Department of Agriculture (2001), National Engineering Handbook (NEH) Part 653, Stream Corridor Restoration: Principles, Processes and Practices. National Technical Information Service, U.S. Department of Commerce, Springfield, VA.( 1998- Revised 2001 by Natural Resources Conservation Service.

U.S. Department of Agriculture (2007), National Engineering Handbook (NEH) Part 654, Stream Restoration Design, Natural Resources Conservation Service, 201-VI-NEH