Channel Evolution Model - Chris Babis

Introduction
Channel Evolution Models (CEM) have been developed to qualitatively describe the morphological adjustments of channels undergoing incision in various forms and stages; Where morphology can be defined as the “science of structure or form” and fluvial may be defined as “produced by the action of flowing water” (Lane, 1955). In essence, CEM helps describe a broad range of morphological adjustments including deposition, migration, enlargement, undercutting, recovering, and compound phases (Hawley, Bledsoe, Stein and Haines, 2012). One of the advantages of the CEM is its selection of relatively few key variables to represent many processes operating across a broad spatial scale. Specifically, by observing key variables such as channel width, depth, bank condition (including vegetation), and Thalweg profile, one may classify a channel network according to the CEM and thus, qualitatively predict the future course of channel evolution (Doyle and Shields, 2000). The dynamic equilibria of river systems and their response to disturbance can make discerning transition states of CEM challenging.

Channel Geometry Adjustment and the CEM Model
As cited by Simon (1989), changes imposed on a fluvial system, be they natural or human-induced, tend to be absorbed through the system by a series of channel adjustments (Gilbert 1880; Mackin 1948; Lane 1955; Heck 1960; Schumm 1973; Bull 1979). Simon (1986) and Schumm (1981) have described the adjustment of channel geometry and channel evolution based on a series of five or six process-oriented stages of morphological development. Simon (1989) described the adjustment of channel geometry in six stages: 1) premodified; 2) constructed; 3) degradation; 4) threshold; 5) aggradation; and 6) restabilization (see Figure 1 below). The processes that dominate adjustment and the ultimate stable geometries, however, are vastly different, depending on the cohesion of the channel banks and the supply of hydraulically-controlled sediment (sand) provided by bank erosion (Simon and Rinaldi, 2006). Furthermore, Hawley, Bledsoe, Stein and Haines (2012) described that departures from previous CEM are common and can include transitions of single thread to braided evolutionary endpoints, as opposed to a return to quasi-equilibrium single thread planform. Similarly, there are known limitations of the CEM, discussed in detail by Hawley, Bledsoe, Stein and Haines (2012), with respect to diverse stream responses, alternate channel states and type of environment specifically in semiarid regions.

Channel Morphology
Changes in channel morphology are the basis of channel evolution models (CEM). Schumm et al. (1984), demonstrated that disturbed channels follow a predictable pattern of adjustments through time which varies along the channel longitudinal profile, and can be described by a series of five process-oriented stages of development. There are only slight variations when comparing the Simon and Schumm models of channel evolution properties that include successive forms of bank retreat and bank slope development. Simon’s model of bank-slope development includes six-stages, wherein stages three, four and six respectively, contain one subsection each. Simon’s model, described in order by stages: 1) premodified or sometimes referred to as sinuous premodified; 2) constructed; 3) degradation; 3a) degradation with undercutting; 4) threshold stage; 4a) threshold stage with increased vertical face profile, increased upper-bank angle, and degraded channel bottom; 5) aggradation; 6) restabilization; and 6a) restabilization with increased fluvial deposition. Refer to Figure 1 in the Appendix A for a diagram depicting CEM stages.

Human Influences and Channel Morphology
Changes in channel morphology can be influenced by natural and or anthropogenic disturbances; the construction of levees, small diversion dams, concrete lined drainage ditches and other alterations to the floodplain can drastically impact channel response and flow regimes.

The Rio Grande has been extremely altered, throughout time, according to Massong, Makar and Bauer (2010):
“First documented in the 1500s by Spanish explorers, large floods played a significant role on the Rio Grande in New Mexico, creating periods of punctuated channel evolution with avulsions, migrations and other large-scale channel shifts (Scurlock 1998). So big were these floods and shifts that the Spanish simply named this river the “grand river”. As communities began to develop along this wandering river corridor, desires to control the system became paramount. Initially, small diversion dams, levees, and channelization projects were implemented to protect local communities (Scurlock 1998), however these proved inadequate. Starting in the 1950s, construction of several large dams in the watershed (Lagasse 1980, Massong and Porter 2004) concurrent with natural changes in precipitation, runoff, and sediment supply (Leopold 1951, Gutzler 2000, and Molnar and Ramirez 2001) resulted in a reduction in flood magnitudes and frequencies, with the last large flood occurring in 1942.” (Massong, Makar and Bauer, 2010, p 1).

The Middle Rio Grande’s more recent history is rife of instances where humans have attempted to alter the riverscape to capitalize the surface hydrologic resources or “tame” the river in an attempt to reduce the risk of flooding. The record of alterations goes back thousands of years to indigenous cultures who engineered small diversions, similar in nature to acequias built by Spanish settlers to divert water for the irrigation of crops and livestock. The 20th century brought more severe alterations to the landscape, “Initially, only small channel modifications occurred, but beginning in the twentieth century, large channel-realignments, miles of levees and jetty fields, numerous diversion dams and large dams were constructed” (Massong, Makar, Bauer, 2010, p. 4).

The Flood Control Act of 1936 (FCA) was an Act of Congress and was signed into law by Franklin Delano Roosevelt on June 22, 1936. Essentially it provided the Department of War, now the Department of the Defense to authorize civil engineering projects such as dams, levees, and dykes to be carried out by the Army Corps of Engineers and other designated Federal Agencies. Section one of the FCA clearly recognizes that flooding in nature is at odds with human development within floodplains. The FCA was clearly defined to mitigate the “disruption of orderly processes” as well as the loss of life, property, land and to attenuate the impairment of critical physical infrastructure such as highways, bridges and railroads.

“Congress passed several flood control acts after the last large flood of 1942. Although the flood of 1942 was not the largest flood of record, this flood occurred on the heels of the 1941 flood, which was one of the largest and longest events recorded in the MRG. By 1942, aggradation of the Rio Grande channel in Albuquerque and at several other locations had super-elevated the channel, creating a flood situation whenever the flows increased. Several large dams were constructed within the MRG watershed with the goal of flood peak and sediment load reduction. The most influential of the large dams is Cochiti Dam; located on the main-stem Rio Grande, this dam began operations in 1973. This project and others effectively controls flood flows in the MRG and have captured large volumes of sediment” (Massong, Makar, Bauer, 2010, p. 4).

Over the past one hundred years much emphasis has been placed on structural solutions that leverage institutional arrangements such as the FCA of 1936, when it comes to managing surface water hydrology. One does not have to look deeply to discover that most waterways have been altered to some degree in an attempt to meet the growing demands of a diverse population. As our collective scientific knowledge as a society has changed, we have improved our transdiciplinary understanding of complex systems, as well as historic and environmental flow regimes with respect to surface water hydrology. That new insight has revealed much in terms of how human manipulation has altered or can potentially affect channel morphology, annual flows, flood damage, and flooding—to name a few. Many intellectual works have called for a more comprehensive flood management strategies that includes wetland restoration, as well as, stormwater capture. For example, from Hey and Phillips (1995). “Despite this nation's massive effort during the past 90 years to build levees throughout the upper Mississippi Basin, mean annual flood damage in the region has increased 140% during that time. These levees exacerbate the flood damage problem by increasing river stage and velocity” (Hey and Phillips, 1995, abstract).

Again, from a historical perspective, Hey and Phillips (1995), “History testifies to the truth of this premise: it was the rampant drainage of wetlands in the nineteenth century that gave rise to many of today's water resources management problems. The 1993 flood verifies the need for additional wetlands: the amount of excess water that passed St. Louis during the 1993 flood would have covered a little more than 13 million acres —half of the wetland acreage drained since 1780 in the upper Mississippi Basin. By strategically placing at least 13 million acres of wetlands on hydric soils in the basin, we can solve the basin's flooding problems in an ecologically sound manner” (Hey and Phillips, 1995, abstract)

Explanation of CEM Planform Stages
According to Simon (1989) premodified (Stage 1) bank conditions are assumed to be the result of natural-fluvial processes and land-use practices; banks are generally considered stable and mass-wasting processes generally do not occur. Premodified bank conditions are also characterized by low angle slopes, typically in the range of 20 to 25 degrees with established wood vegetation along the top bank and downslope towards the low flow channel; the premodified stage can be further defined by attributes of reach including channel width (Simon, 1989).

During the constructed (Stage 2) stage, Simon (1989) described the banks as being steepened, heightened and linear; this stage involves the reshaping of channel banks or a repositioning of the channel. Generally speaking, the constructed stage is considered a transitional process from stable to less stable—premodified to degradation; vegetation is often removed as the channel is widened, channel conveyance is increased.

Stage three in the CEM, the degradation (Stage 3) stage can be defined by rapid erosion of the channel bed and as a result an increase in bank heights. As cited in Simon (1989), research by others including, Skempton 1953; Carson and Kirby 1972; and Simon & Hupp 1986, downcutting is not the primary factor influencing an increase in bank slopes but is attributed to being related to bank angles and angles of friction. The amount of incision is partly controls what is referred to as the bank failure threshold, thus making the degradation stage one of the most important with respect to the magnitude of channel widening (Simon, 1989).

Continued degradation and basal erosion that increases the height and steepness of the channel bank are the hallmarks of the fourth stage in the CEM model— the threshold (Stage 4) stage (Simon, 1989). Succeeding and building upon the previous stage, the threshold stage is characterized when the critical bank height is exceeded and bank slopes and shapes become the product of mass wasting—a notable difference from the preceding stage response.

Simon (1989) described the aggradation (Stage 5) stage, as the onset of aggradation on the channel bed and is often identified by sand accumulating or being deposited on the bank surfaces. Bank retreat dominates the vertical face and upper bank sections; frequent wetting by rises in stage and from the added weight of deposited fluvial materials—making the upper bank subject to secondary, low-angle slides.

Morphological Adjustments
As reported in Simon and Rinaldi (2006) incision is a common response of alluvial channels that have been disturbed such that they contain excess amounts of flow energy or stream power relative to the sediment load. It is clear that there is some interplay between channel incision and denudation in the context of geomorphic evolution. Underscoring much of the understanding of fluvial systems is the idea that despite a multitude of varying conditions inherent unto them, the incision response trends to disturbances are typically of the same order of magnitude with respect to time and space (Simon and Rinaldi, 2006). A key distinction is necessary, particularly when examining large scale or catastrophic natural events which have the potential to create large imbalances between sediment delivery and transporting power. To illustrate the point further, Simon and Rinaldi (2006) reported, “Channel incision is part of denudation, drainage-network development, and landscape evolution. Rejuvenation of fluvial networks by channel incision often leads to further network development and an increase in drainage density as gullies migrate into previously non-incised surfaces. Large, anthropogenic disturbances, similar to large or catastrophic “natural” events, greatly compress time scales for incision and related processes by creating enormous imbalances between upstream sediment delivery and available transporting power” (Simon and Rinaldi, 2006, abstract).

System Imbalance
This idea of system imbalance is further explored by Bledsoe, Watson and Biedenharn (2002), “Incised channels are caused by an imbalance between sediment transport capacity and sediment supply that alters channel morphology through bed and bank erosion. Consistent sequential changes in incised channel morphology may be quantified and used to develop relationships describing quasi-equilibrium conditions in these channels” (Bledsoe, Watson and Biedenharn, 200, p. 861). Channel incision can be driven by a broad range of variables; Simon and Rinaldi (2006), noted the causes of incision have been broken down into six broad categories (Schumm 1999), geologic, geomorphologic, climatic, hydrologic, animals, and humans, with specific mechanisms associated with each cause. As cited in Simon and Rinaldi (2006), the causes of river channel incision are numerous, but the morphological effects and hazards associated with incised channels are often similar across a spectrum of physiographic environments (Parker and Andres, 1976; Elliott, 1979; Schumm et al., 1984; Williams and Wolman, 1984; Simon, 1989a, 1992; Rinaldi and Simon, 1998; Schumm, 1999). System imbalance and CEM and channel variation are intimately interconnected; riverine systems with respect to time are in a constant flux between system imbalances and equilibria(s).

CEM—Disturbance, Sediment Delivery and Transporting Power
With respect to disturbances, large, anthropogenic disturbances, similar to large or catastrophic “natural” events, greatly compress time scales for incision and related processes by creating enormous imbalances between upstream sediment delivery and available transporting power (Simon and Rinaldi, 2006). This is an important consideration, with respect to channel variation in the context of river systems and their dynamic equilibria state(s). Research has demonstrated that magnitude of vertical and lateral widening and or other geomorphological response may vary among similar types of disturbances. Responses to different types of disturbances are shown to result in similar spatial and temporal trends of incision for vastly different fluvial systems. Similar disturbances are shown to result in varying relative magnitudes of vertical and lateral (widening) processes, and different channel morphologies as a function of the type of boundary sediments comprising the bed and banks (Simon and Rinaldi, 2006). There are many important considerations when evaluating channel response to disturbance, for example, flow energy, shear stress and stream power. As referenced in Simon and Rinaldi (2006), Degradation of channel beds represents a response to a disturbance in which an excess of flow energy, shear stress or stream power (sediment-transporting capacity) occurs relative to the amount of sediment supplied to the stream. This is often expressed using the stream power proportionality popularized by Lane (1955) which is applicable to channel with mobile boundaries:
Lane’s Equation = QSQsd50
Human Influence and the Rio Grande
Riverine systems have been influenced and altered by human actions for thousands of years; the Rio Grande is no exception. Massong, Makar and Bauer (2010), described the Middle Rio Grande (MRG), in its present, altered and more confined state, from Cochiti Dam to Elephant Butte; engineering of the MRG has in some ways created a more predictable system of change largely related to a more controlled state with respect to flows—a factor that has decreased flooding and aggregation (Massong, Makar, Bauer, 2010). However, the predictability of how the MRG will respond to disturbance or change over time can by widely disputed. For example, side-channel evolution within the Isleta Reach of the MRG is complex and not well understood “The channel dimensions, including width, have greatly changed over this five-year period and more observable gravel deposits are being found. The channel has changed so dramatically in this reach since 1999 that predicting future planforms is a widely disputed topic” (Massong, Makar, Bauer, 2010, p.5). The complexity of the MRG, with respect to CEM is further illustrated by the complex nature of planform evolution despite the more systemic nature of flow on the Rio Grande. For example, from Massong, Makar and Bauer (2010):

“As a first step towards predicting future changes based on the reduced hydrology and sediment regime, an empirically based planform model is proposed that describes two sequences of planform evolution found throughout the MRG. Although changes in the planforms initially appear similar throughout the study area (first three stages of the model), planform evolution splits into two distinct tracks after the channel begins abandoning medial and point bars located within the active channel (Stage 4). At this point the relative sediment transport capacity becomes a deciding factor in the channel’s future form: those channels that are deficient in transport capacity remain deficient in capacity throughout their evolution, while those channels that have excess capacity actively shape their channels by eroding the channel bed and banks. Channels that are deficient in transport capacity evolve towards avulsion while the channels that have excess sediment transport capabilities evolve towards a migrating bend planform” (Massong, Makar and Bauer, 2010, abstract).

Rio Grande Case Study: The Isleta Reach—Belen, New Mexico
Although fluvial systems are dynamic in nature, for many years a section of the MRG near Belen, New Mexico, specifically near the confluence of the Rio Puerco and Rio Grande, has been regarded as stable and relatively static (Massong, Makar and Bauer, 2010, p.5). This state has largely been attributed to the engineering of the river, specifically floodway construction that occurred between the years of 1950-1960, as well as, cultural changes in river flow management and enhanced bank armoring from riparian vegetation. CEM with respect to large-scale disturbance such as global-climate-change-type-drought can be used to illustrate the complexity of understanding channel evolution. Cumulative changes to climate, precipitation etc. can have unanticipated or not well understood effects on of sediment transport and avulsion—among other things. Refer to Figure 2 in Appendix B for aerial photographs illustrating channelization of the Middle Rio Grande near Belen, New Mexico.

“Prior to 2000, large dune fields composed of sand could be found moving through the active channel at most times of the year with a slightly aggrading channel bed and active floodplain throughout. However, during a recent drought, 1999-2004, downstream transport of the large macro-dunes stalled; encroachment of vegetation was quick as herbaceous vegetation initially covered these bars giving way to saplings within 2 years. Although water was minimal during this time, the woody vegetation grew quickly, creating a stabilized surface prior to the return of more ‘normal’ flows in 2004” (Massong, Makar and Bauer, 2010, p.5). Channel narrowing has been occurring near this stretch of river, downstream from the San Acacia Dam (South of the Isleta reach) as a result of many factors. For example, riverbank construction, specifically channel straightening and the disconnection of natural bends that occurred in the 1930s through the 1960s. Massong, Makar and Bauer, (2010) reported that channel narrowing has continued since the 1960’s; this dynamic phenomena has been coupled with channel incision and sediment coarsening.

“Combined, these three changes (narrowing, incising and sediment coarsening) altered how the Rio Grande functions in this section of the MRG. The narrowing channel coupled with incision has created a deeper channel throughout this reach. With the high incision, the historical floodplains are now acting as terraces and less of the flood flows are leaving the active channel, creating deeper flows during normal events. In combination with these hydraulic changes and the naturally decreasing supply of sediment, the emergence of the gravel-bed is predictable. The gravel material adds stability to the channel bed during the non-channel forming flows, which essentially further reduces the channel’s supply of sediment as it no longer can erode the bed during this smaller flows” (Massong, Makar and Bauer, 2010, p.7).

As discussed in Massong, Makar and Bauer (2010), subsequent to the 1990s, an evolution of river bends has begun; the significance is that just below the San Acacia diversion, bend evolution, migration and channel abandonment can happen quickly (less than 10-years) due to fluvial dynamics. These migrations and specific bend sequences occurring in the late 1990s have been monitored and documented in Massong, Makar and Bauer (2010). The Rio Grande provides and interesting study with respect to CEM as a myriad of human and natural influences have created varied evolutionary tracks creating river channels that are nearly opposite in function. From Massong, Makar and Bauer (2010):

“Generally the Rio Grande’s channel throughout the middle section is changing due to a variety of influences; even though the initial channel changes were a simple ubiquitous narrowing, we recently have been able to discern two distinct tracks of change that although subtle in the beginning, lead in two distinct directions that create channels that are nearly opposite in function. The defining process dividing these two tracks is the channel’s relative sediment transport capacity: those channels that are deficient in transport capacity remain deficient in capacity throughout their evolution, while those channels that have excess capacity actively re-activate stored sediments by eroding the channel bed and banks” (Massong, Makar and Bauer, 2010, p.7).

Channel Recovery
With respect to policy, restoration and or hydro-modification management, channel recovery is an important concept in the context of channel incision, instability and morphological responses to disturbance. As cited in Hawley, Bledsoe, Stein, and Haines (2012), the Watson et al. (2012) concept underscores that it is critical to arrest channel instability before incision has reached critical bank height; geotechnically unstable banks become disproportionately more expensive to rehabilitate—an important consideration in urban basins where infrastructure may be at risk. Channel recovery to some degree may be related to channel instability or periods where dynamic equilibria(s) of the channel system are out of balance. Instability is a function of channel gradient and discharge—among other things and is common to Western U.S. watersheds, especially in the arid and semi-arid Southwest. From Kondolf (1996), channel instability is a particularly common problem in western North America, where streams that were formerly desirable, narrow meandering channels, have been converted into wide, shallow, braided channels in response to destabilizing influences such as ill-advised channel works or land use changes in stream, as illustrated on the Blanco River, Colorado (NRC 1992)” (Kondolf, 1996, p.1).

Bank Instability
Changes in landscape ecology and hydrologic conditions over time can have an impact on bank instability. For example, alluvial aquifer losses to adjacent surface water bodies may have an impact on hydraulic head values; depth to ground water is an important factor that drives vegetative species diversity, composition and abundance among riparian corridors. Riparian vegetation is one important component of stabilized channel banks that provides habitat and ecosystem structure. There is an important connection between ground water hydrology, riparian vegetation and biological assemblages with depth to groundwater, base flow, erosion and or channelization. For example, Kondolf (1996) described, “The Carmel River, California, provides an example where dewatering of the alluvial aquifer by water-supply wells killed bank-stabilizing willows and locally lowered the banks’ resistance to erosion (Kondolf and Curry 1986). As a result, floods that had passed through many times previously without causing significant erosion (return period 6-8 years) caused locally massive bank erosion” (Kondolf, 1996, p. 1). Restoration on the Carmel River focused on stabilizing channel geometry through the creation of bankfull channel of appropriate dimensions and then stabilizing with riparian vegetation (Kondolf, 1996).

Channel Restoration and Streambank Stabilization

Channel restoration or streambank stabilization efforts will vary in nature, contingent upon a broad range of factors, including but not limited to budget constraints, stakeholder values, and key infrastructure assets. This is especially true in instances where overarching goals aim to control bank protection from bank erosion. From Kondolf (1996), “Some stream “restoration” projects do not tackle system-wide instability, but only local bank erosion that threatens structures, agricultural land, or other resources. Bank erosion along the outside of a meander bend and deposition inside the bend on the point bar is a part of natural channel migration. This disturbance regime creates new surfaces on which tree seedlings become established, thus determining patterns of riparian vegetation succession. However, even when the bank erosion results from natural migration, resources threatened by erosion or flooding may create the demand for a bank protection project” (Kondolf, 1996, p1).

Similarly, restoration approaches and techniques will vary depending on goals of the project. For example in instances where key infrastructure is determined to be at risk, a “hard” or more “engineered” approach may be employed. As cited in Kondolf (1996), from Henderson (1986), “Harder, more “engineered” bank protection measures include emplacement of gabion baskets (preferably with interstitial soil and willow shoots), riprap, masonry walls, concrete walls, boulder walls. Informal bank protection has utilized materials such as concrete rubble, automobile tires, and automobile bodies. For example, "hard" engineering materials could include gabions, riprap, and concrete rubble can be designed to accommodate willows or other riparian plantings in the interstices of the large elements” (Kondolf, 1996, p.1). Restoration as a discipline is complex, and evolving, this paper has not adequately covered river restoration as topic related to CEM, further research is necessary.


APPENDIX A

CEM Diagram.png
Figure 1
From Hawley, R. J., Bledsoe, B. P., Stein, E. D., & Haines, B. E. (2012). Channel Evolution Model of Semiarid Stream Response to UrbanInduced Hydromodification1. JAWRA Journal of the American Water Resources Association, 48(4), 722-744. Diagram depicts the five-stages of CEM for incised single thread channels, bifurcations from conventional five-stage CEM and CEM for Braided Channels




APPENDIX B

MRG Channel.png

Figure 2
Near Belen, NM. Ariel photography illustrating channelization of the Rio Grande. A: 2000, B: 2002, C: 2005, D: 2006 (Massong, Makar and Bauer, 2010). These images show channelization over time and are of the same stretch of river.




Sources

Bledsoe, B. P., Watson, C. C., & Biedenharn, D. S. (2002). Quantification of Incised Channel Evolution and Equilibrium. JAWRA Journal of the American Water Resources Association, 38(3), 861-870.
Bunn, S. E., & Arthington, A. H. (2002). Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity. Environmental management, 30(4), 492-507.
Darby, S. E., Rinaldi, M., & Dapporto, S. (2007). Coupled simulations of fluvial erosion and mass wasting for cohesive river banks. Journal of Geophysical Research: Earth Surface (2003–2012), 112(F3).
Dietrich, W. E., Wilson, C. J., Montgomery, D. R., & McKean, J. (1993). Analysis of erosion thresholds, channel networks, and landscape morphology using a digital terrain model. The Journal of Geology, 259-278.
Dietrich, W. E., Wilson, C. J., Montgomery, D. R., McKean, J., & Bauer, R. (1992). Erosion thresholds and land surface morphology. Geology, 20(8), 675-679.
Doyle, M. W., & Shields Jr, F. D. (2000). Incorporation of bed texture into a channel evolution model. Geomorphology, 34(3), 291-309.

Flood Control Act. (n.d.). //Flood Control Act//. Retrieved April 7, 2014, from http://www.ccrh.org/comm/cottage/primar

Hawley, R. J., Bledsoe, B. P., Stein, E. D., & Haines, B. E. (2012). Channel Evolution Model of Semiarid Stream Response to UrbanInduced Hydromodification1. JAWRA Journal of the American Water Resources Association, 48(4), 722-744.
Hey, D. L. and Philippi, N. S. (1995), Flood Reduction through Wetland Restoration: The Upper Mississippi River Basin as a Case History. Restoration Ecology, 3: 4–17. doi: 10.1111/j.1526-100X.1995.tb00070.x
Kondolf, G. M. (1996). A cross section of stream channel restoration. Journal of Soil and Water Conservation, 51(2), 119-125.
Lane, E. W. (1955). Design of stable channels. Transactions of the American Society of Civil Engineers, 120(1), 1234-1260.
Langendoen, E. J., & Simon, A. (2008). Modeling the evolution of incised streams. II: Streambank erosion. Journal of hydraulic engineering, 134(7), 905-915.
Lytle, D. A., & Poff, N. L. (2004). Adaptation to natural flow regimes. Trends in Ecology & Evolution, 19(2), 94-100.
Montgomery, D. R., & Dietrich, W. E. (1988). Where do channels begin?.Nature, 336(6196), 232-234.
Montgomery, D. R., & FoufoulaGeorgiou, E. (1993). Channel network source representation using digital elevation models. Water Resources Research,29(12), 3925-3934.
Massong, T., Makar, P., & Bauer, T. (2010). Planform Evolution Model for the Middle Rio Grande, NM. In Joint Federal Interagency Conference, Las Vegas, NV.
Schumm, S. A., Harvey, M. D., & Watson, C. C. (1984). Incised channels: morphology, dynamics, and control.
Seelbach P.W., Wiley M.J., Kotanchik J.C. & Baker M.E. (1997) A Landscape-Based Ecological Classification System for River Valley Segments in Lower Michigan (MIVSEC Version 1.0). Fisheries Research Report 2036. Michigan Department of Natural Resources, Ann Arbor, MI.
Simon, A., & Rinaldi, M. (2006). Disturbance, stream incision, and channel evolution: The roles of excess transport capacity and boundary materials in controlling channel response. Geomorphology, 79(3), 361-383.