Jump to main content or area navigation.

Contact Us

CADDIS Volume 2: Sources, Stressors & Responses

Physical Habitat

Sources and activities that suggest listing candidate causes associated with altered physical habitat


The most obvious sources or activities associated with altered physical habitat are direct alterations of the stream channel.  Although dredging operations to straighten and deepen stream channels (i.e., channelization) are seldom observed directly, records may have been kept by the local government agency that facilitated the dredging and the effects of past activity look like a straight or constrained channel. Observations that may indicate channelization include sinuosity near 1 (where sinuosity equals the distance along the channel thalweg between two points divided by the straight line distance between those two points), low wetted width to thalweg depth ratio, rectangular channel cross-section, dredged material (i.e., sediments) piled along and adjacent to the channel, or the dominance of run or glide habitats. The thalweg is the deepest water depth in a channel cross-section, which is usually at the center of water flow within a channel. Runs or glides are areas of the channel characterized by non-turbulent flow over a relatively flat stream bottom (Platts et al. 1983). Water depths are intermediate between riffles and pools. In glides, the surface water gradient is nearly level, and current velocities are slower than runs where the surface water gradient is greater. However, particularly in the case of alluvial channels, there may be rapid and ongoing adjustments in channel dimensions. This can include degradation (i.e., lowering) or aggradation (i.e., raising) of the bed-level and mass wasting of stream banks, which can proceed both upstream or downstream of the channelized reach (Simon 1989).

Historically, channelization was conducted to drain lands to increase areas for farming and construction or for flood control. As a result, small stream channels were lost and drainage density in many watersheds has decreased (Meyer and Wallace 2001).  Drainage density is the ratio of the total channel length in a drainage basin to the basin area (km/km2). 

In urban areas, stream channels may be buried in culverts or pipes or entirely replaced by stormwater drainage systems (Elmore and Kaushal 2008). This habitat degradation is so severe that such channels are not typically surveyed as being biologically relevant nor identified as an impaired aquatic system.


A dam either upstream or downstream of a stream reach may alter physical habitat. Effects on the channel are obvious when a downstream dam impounds a reservoir, but they may not be as obvious with smaller run-of-the-river type dams or weirs (Poff and Hart 2002). Upstream of reservoirs, operation of the dam to manage pool heights and water releases may periodically inundate the features of reaches of tributary streams. Run-of-the-river dams or weirs, including many smaller, older dams, generally increase thalweg depths  for some distance upstream and create pool habitat. However, the water remains mostly within the stream channel at normal flows. There are estimated to be over one million dams with heights less than four meters throughout the United States (Poff and Hart 2002). Dams with any significant hydraulic residence times (i.e., the length of time water remains within a reservoir) will capture bedload and suspended sediments and disrupt sediment transport in stream systems (Kondolf 1997, Poff and Hart 2002). As a result, water released from the dam has energy to move sediment but has little or no sediment load. This “hungry water” (Kondolf 1997) erodes the channel bed and banks downstream from the dam, resulting in incision of the stream bed and coarsening of bed materials; this continues at least until a new equilibrium state is reached, with the bed becoming armored and flows no longer able to move remaining bed materials (Ligon et al. 1995). 

After a dam is breached, some former slackwater sediment deposits will be transported downstream and deposited over coarser substrates (Orr et al. 2008). Deposits upstream of the dam may be incised by the stream to form steep, easily-eroded banks composed of fine sediments, and they can produce the incised, meandering gravel-bed streams observed in the Northern Piedmont ecoregion of the eastern United States (Walter and Merritts 2008). This reinterpretation of the historical effects of milldams may apply to other regions with historical construction of small, run-of-the river dams or weirs to provide water power to mills and other operations.


Another alteration of the stream channel is caused by instream mining for sand and gravel used as construction aggregate (Kondolf 1997). Gravel mining alters channel geometry and bed elevation. It also may include clearing of vegetation from sandbars and riparian zones and digging of deep pits in the channel. The result is disruption of natural channel morphology and production of a local sediment deficit. In some cases, a headcut is produced by gravel mining. In turn, this results in channel incision, bed coarsening and lateral channel instability, which can migrate both upstream and downstream from the actual site of gravel mining (Kondolf 1997).

Other types of mining also can locally alter stream channels. Mountaintop removal mining for coal completely buries headwater stream channels with valley fills (Meyer and Wallace 2001, Palmer et al. 2010). Surface subsidence associated with long-wall coal mining has been observed to reduce stream gradients or otherwise alter stream flows (PADEP 1999). Deposition of iron hydroxide and other metal hydroxides downstream from acidic mine drainage sources can embed and cement stream sediments. Dredge mining of placers in stream channels for minerals, such as gold, can alter thalweg profiles and reduce residual pool lengths and maximum pool depths (Mossop and Bradford 2006).

Impervious surfaces

Often, more widespread alterations of stream physical habitat result from anthropogenic activities in the watershed that increase the magnitude and frequency of stormflows. Residential, commercial and urban development are known to increase impervious surface area, and some agricultural activities, such as drainage ditches and tile drainage systems, are intended to route storm water quickly off of agricultural fields. These activities increase soil erosion and input of fine sediments to streams, but particularly increase peak discharges, which erode stream banks (Chin 2006).  This results in widening of streams, deposition of finer sediments in pools, and embedding of coarser sediments with those finer sediments.

Top of page

Site evidence that suggests listing candidate causes associated with altered physical habitat

Various localized anthropogenic activities directly alter the stream channel and affect its physical habitats. The most obvious evidence is the presence of rip-rap, gabions (i.e., usually wire mesh cages containing rocks), or concrete structures (e.g., walls and other revetments) lining the channel, usually to prevent lateral movement of the active channel and limit land loss and damage to infrastructure, such as roads. These structures usually affect stream banks, which are important transition zones within the riparian ecotone.  The stream bank is the part of the stream channel between the lateral margins of the stream bed normally wetted at baseflow and at bankfull height, where flood flows would begin to spread laterally over the floodplain surface (Florsheim et al. 2008).  Riparian vegetation on the stream bank interacts much more strongly with the wetted channel, and construction of artificial structures removes this vegetation and its functions (Rhodes and Hubert 1991, Wood and Sites 2002, Sweeney 1993).  Such structures also may increase flood velocities along stream banks and prevent reestablishment of riparian vegetation, arrest exchanges of sediment between channel and bank, and destroy habitat for fauna that use riparian zones, including birds, reptiles, and amphibians.  Moreover, the design of these structures often fails to account for bank erosion processes: consider the appropriate scale for bank erosion management, or consider secondary, long-term or cumulative effects (Piégay et al. 2005, Florsheim et al. 2008).

Other localized structures that can alter stream channels include bridge abutments, culverts, or pipes that constrict flows—particularly higher flows.  Constricted flows below bridges and culverts erode the stream bed and banks (Langendoen 2000, Johnson et al. 2002).  Moreover, pipes or culverts usually lack benthic sediments and may act as barriers to migration for aquatic fauna (Warren and Pardew 1998, Poplar-Jeffers et al. 2009). This barrier effect is enhanced when a pipe or culvert is placed above the stream channel and flows erode a scour pool below its outlet.

photo of two Mississippi creeks
Figure 4. Examples of headcuts on Hotophia Creek and Big Creek, Mississippi, respectively.
Courtesy of F.D. Shields and S. Knight, USDA, ARS, National Sedimentation Laboratory, Oxford, MS.

The presence of a headcut is another indicator of stream channel degradation. A headcut is an abrupt change in stream bed elevation marking the upstream migration of channel incision in a stream (Figure 4). Development of a headcut can result from gully development associated with poor agricultural erosion control or channelization of streams (Shields et al. 1998). In some cases, the channel incision associated with such headcuts may eliminate stream-floodplain interactions.

Logging of riparian forests may have reduced the size of woody debris falling into many forested stream ecosystems. This is particularly true if the riparian zone is no longer forested but also occurs in streams where the riparian forest has been allowed to regenerate. Initially, logging may increase large woody debris stocks, but long-term loadings of large woody debris decrease, because of increased rates of decay and transport and reduced inputs (Chen et al. 2005). Large woody debris can be important to the geomorphology of streams (Flebbe and Dolloff 1995, Abbe and Montgomery 1996, Hassan et al. 2005), and the size of the woody debris relative to channel width is an important factor affecting the movement of the wood during high discharges (Berg et al. 1998). Smaller woody debris and brush can be used for cover by fish and invertebrates (Harvey et al. 2005), but such cover is more ephemeral.

Beyond noting the obvious types of evidence discussed so far, it is also helpful to examine the channel's dimensions and classify it by geomorphological type. Channel dimensions include wetted width, thalweg depth (and their ratio), as well as bankfull channel width, depth, and their ratio. A high ratio of bankfull channel width to depth indicates stream-bank degradation or stream bed aggradation, while a low ratio may indicate stream bed entrenchment. Rosgen (1996) presents a classification system to describe channel geomorphological types and assess stream channel condition.

photo of eroded banks in the Little Miami River, OH
Figure 5. Little Miami River, Ohio. This eroded bank shows some of the characteristics of stream banks susceptible to erosion described by Rosgen (1996).

Excessive stream bank erosion is another important indicator of physical habitat alteration. Stream banks susceptible to erosion may be characterized by one or more of the following (Rosgen 1996; Figure 5):

  1. A low ratio of streambank height to bankfull height (i.e., usually less than one)
  2. A low ratio of riparian vegetation rooting depth to streambank height (i.e., usually much less than one)
  3. A low rooting density of plants (i.e., streambank relatively lacking in rooted vegetation)
  4. A stream bank composed of relatively fine materials, such as fine gravels, sand, or clay;
  5. A steep stream bank angle or slope (i.e., near to 90%)
  6. A stream bank material either stratified (i.e., showing layers of sediment) or containing soil lenses (i.e., buried soil layers)
  7. A lack of stream bank surface protection by debris or vegetation (absence of vegetation or other material) that armor the stream bank

Other structures in the stream channel are also important for stream biota, because they supply cover or form habitat, and the lack of such cover or habitat forming features is an indicator of poor habitat quality. These features can include woody debris, filamentous algae, aquatic macrophytes, overhanging vegetation, undercut banks, or artificial structures that supply cover (Jakober et al. 1998,; Lyons et al. 2000, Roni et al. 2002, Fritz et al. 2004). One needs to evaluate these on a case-by-case basis. For example, too much cover by filamentous algae or aquatic macrophytes may be indicators of impairment.

Top of page

Biological effects that suggest listing candidate causes associated with altered physical habitat

We encourage consideration of local fauna and flora as indicators of habitat disturbance.  Physical habitat is closely linked to many other stressors (e.g., sediment, flow, temperature, dissolved oxygen), and biological effects usually are not sufficiently specific to be considered symptomatic of impaired physical habitat alone. However, certain biological effects often are associated with physical habitat degradation such as changes in the composition of fish, macroinvertebrates and algal assemblages.

Changes in fish assemblage composition

Changes in substrate or cover within a stream may provide less suitable habitat for shelter or reproduction (Matthews 1998). Fish that rely on natural wood in streams for cover will be adversely affected by wood removal or a decline in the input of wood (e.g., Centrarchidae, Salmonidae) (Angermeier and Karr 1984). A wide variety or abundance of submerged structures tends to increase habitat, and thus fish diversity (Barbour et al. 1999). Fish species that seek cover in streamside vegetation may be adversely affected by the loss of this habitat or its alteration by invasive plant species. Fishes that rely on floodplain connections for reproduction and nurseries may be affected if floodplain access is restricted (Lasne et al. 2007). If channels become incised, large pool-dwelling fishes (e.g., top carnivores or piscivores, such as Centrarchidae) may be favored while fishes that require shallow riffles [e.g., darters (Etheostoma or Percina), sculpins (Cottus), or madtoms (Noturus)] or stream-margin habitats (e.g., young-of-year of Salmonidae, Centrarchidae, Cyprinidae, or Catostomatidae) may decline (Moore and Gregory 1988, Aadland 1993, LaVoie and Hubert 1996).

Dams and reservoirs may drastically alter fish assemblages. Upstream fish assemblages may be altered because the reservoir provides habitat for facultative riverine species, such as Dorosoma cepedianum and Cyprinus carpio, that require lentic reproductive habitats. They also fragment the stream system for obligate river species that migrate during the year [ex., salmonids in Pacific Northwest rivers; Atlantic sturgeon (Acipenser oxyrhynchus), Alabama shad (Alosa alabamae), and mountain mullet (Agonostomus monticola) in southern Atlantic and Gulf Coast rivers; Atlantic shad (Alosa sapidissima) in mid-Atlantic rivers] (Pringle 1997).

Changes in composition of macroinvertebrate assemblages

Macroinvertebrate assemblages reflect stream habitat complexity (Brown 2007). Channel alteration resulting from channelization or other direct modifications of stream channels, or from incisement or channel widening, generally reduces habitat complexity and abundance of habitat specialists with specific flow and substrate requirements. For example, the relative abundance of clingers [i.e., invertebrates with behavioral (e.g., fixed retreats) or morphological (e.g., long, curved tarsal caws, or dorsoventral flattening) adaptations for living attached to hard surfaces] may decrease in stream reaches with homogeneous substrate composition, current velocity, and water depth (Rabeni et al. 2005).

In sandy, lowland streams or rivers that otherwise lack stable substrates, submerged large woody debris and snags may support high invertebrate productivity, particularly of collector-filterers (e.g., Hydropsychidae) that feed on fine particulate organic matter suspended in the water column (Benke et al. 1984). In other streams, woody debris locally alters stream depth, gradient and substrate, affecting adjoining microhabitats. This may change the abundance and biomass of functional feeding groups, for example by decreasing collector-filterers and grazers, increasing collector-gatherers and predators, and changing dominant shredder groups (e.g., increased Trichoptera and Diptera, decreased Plecoptera) (Wallace et al. 1995).

Bank alterations may reduce the amount of streamside habitat available for invertebrate species (Rankin 1989). Connection to the floodplain is also important for invertebrate productivity (Benke 2001). If the input and retention of organic material in a stream are reduced due to loss of riparian vegetation or channel incisement, then the invertebrates that rely on this food source (i.e., detritivores, such as shredders and collector-gatherers) will decline in abundance. This in turn can affect the invertebrate food resources for fish.

Macroinvertebrate community traits, such as maximum size, body form, feeding habits, mode of reproduction, lifespan, and strategies of dissemination are related to habitat variables—especially hydraulics. For example, Lamoroux et al. (2004) found the percentage of macroinvertebrates with maximum body size greater than 20 mm was inversely correlated and the percentage of macroinvertebates with maximum body size between 5 and 10 mm was positively correlated with the reach-scale Froude number (R-Fr, a measure of the hydraulic force exerted on macroinvertebrates clinging to benthic substrata) and reach-scale substrate roughness (R-Kv). Moreover, the percentage of macroinvertebrates with streamlined body form, lifespan less than 1 year, univoltine (i.e., 1 generation per year) reproductive cycle, egg masses that are cemented or fixed to substrates, aerial active dispersal, respiration through the tegument (i.e., exoskeleton), a temporarily or permanently attached life style, or scraper feeding habitats were positively correlated with R-Fr and R-Kv, while the percentage of macroinvertebrates with spherical body form, lifespan less than 1 year, multivoltine (i.e., more than 1 generation per year) reproductive cycle, ovoviviparous (i.e., eggs develop within the parent) reproduction, aquatic passive dispersal, respiration by gills, swimmer or burrower life style, or shredder feeding habitats were negatively correlated with R-Fr and R-Kv. Beyond these measurements of channel hydraulics, this paper suggests that trait metrics could be sensitive measures of assemblage responses to alterations in physical habitat.

photo of an incised channel
Figure 6. An incised channel, Bobo Bayou, Mississippi.
Courtesy of F.D. Shields, USDA, ARS, National Sedimentation Laboratory, Oxford, MS.

Changes in algae assemblages

  • Large inorganic substrates or wood in streams provide stable surfaces for algae, a food source for invertebrates. Factors affecting the stability of such substrates can affect algal standing crops (Myers et al. 2007), and loss of these surfaces reduces the area available for algae growth. Channel incision (Figure 6) may reduce connection with the floodplain, the areal extent of substrates for algal growth, and the retentiveness for nutrients (Coleman and Dahm 1990). Hill et al. (2003) showed that changes in algae were related to channel substrate size, canopy cover, stream width, channel slope, and the proximity-weighted sum of human disturbances in the riparian zone. Algae also respond to the pattern of riffles and pools (Stevenson 1996) by influencing shear stress and delivery of nutrients to periphytic assemblages.

Top of page

Site evidence that supports exclusion of candidate causes associated with altered physical habitat as a candidate cause

General advice on excluding candidate causes from your initial list is provided in Step 2 of the Step-by-Step Guide and in Tips for Listing Candidate Causes. Exclusion of altered physical habitat as a candidate cause should be based on high quality in-stream measurements and the absence of evidence of sources or activities that alter stream physical habitats. However, few watersheds across the United States have not experienced at least some human activities that disturb the dynamic quasi-equilibrium of stream channels (Vannote et al. 1980, Thrush et al. 2000). This includes the historical elimination of beaver, which were once a keystone species in many stream ecosystems across the northern United States (Naiman et al. 1988). You could demonstrate that the identified impairment does not occur at reference sites with similar physical habitat. Therefore, excluding physical habitat usually requires at least a screening-level assessment.

Site observations or measurements may at least support deferring consideration of physical habitat to focus on other candidate causes. For example, if the stream channel is characterized by an alternating sequence of riffles and pools, the stream channel is not incised and lacks apparent physical alteration, and the stream banks and riparian zone lack erosion scars and are well-vegetated with plants native to and characteristic of the ecoregion, physical habitat is unlikely to be a cause. For alluvial channels, a more detailed description of good physical habitat may be found in Thrush et al. (2000).

Top of page

Jump to main content.