Top Banner
The waterfall paradox: How knickpoints disconnect hillslope and channel processes, isolating salmonid populations in ideal habitats Christine May a, , Josh Roering b , Kyle Snow a , Kitty Griswold c , Robert Gresswell d a James Madison University, Department of Biology, MSC 7801, Harrisonburg, VA, 24401, USA b University of Oregon, Department of Geological Sciences, 325C Cascade Hall, Eugene, OR 97403, USA c Idaho State University, Department of Biological Sciences, 921 S. 8th Ave., Pocatello, ID, 83209, USA d U.S. Geological Survey, Northern Rocky Mountain Science Center, 2327 University Way, Suite 2, Bozeman, MT, 59715, USA abstract article info Article history: Received 9 July 2015 Received in revised form 22 March 2016 Accepted 25 March 2016 Available online xxxx Waterfalls create barriers to sh migration, yet hundreds of isolated salmonid populations exist above barriers and have persisted for thousands of years in steep mountainous terrain. Ecological theory indicates that small iso- lated populations in disturbance-prone landscapes are at greatest risk of extirpation because immigration and re- colonization are not possible. On the contrary, many above-barrier populations are currently thriving while their downstream counterparts are dwindling. This quandary led us to explore geomorphic knickpoints as a mecha- nism for disconnecting hillslope and channel processes by limiting channel incision and decreasing the pace of base-level lowering. Using LiDAR from the Oregon Coast Range, we found gentler channel gradients, wider valleys, lower gradient hillslopes, and less shallow landslide potential in an above-barrier catchment compared to a neighboring catchment devoid of persistent knickpoints. Based on this unique geomorphic template, above-barrier channel networks are less prone to debris ows and other episodic sediment uxes. These above-barrier catchments also have greater resiliency to ooding, owing to wider valleys with greater oodplain connectivity. Habitat preference models further indicate that salmonid habitat is present in greater quantity and quality in these above-barrier networks. Therefore the paradox of the persistence of small isolated sh popula- tions may be facilitated by a geomorphic mechanism that both limits their connectivity to larger sh populations yet dampens the effect of disturbance by decreasing connections between hillslope and channel processes above geomorphic knickpoints. © 2016 Elsevier B.V. All rights reserved. Keywords: Knickpoints Salmonid habitat Disturbance ecology Shallow landslides 1. Introduction One of the most pressing conservation questions for Pacic salmon and trout (Oncorhynchus sp.) is the role of episodic disturbances in structuring stream habitat and triggering extreme uctuations in popu- lation abundance. From an ecological perspective, disturbances can be dened as any relatively discrete event in time that disrupts ecosystem, community, or population structure and changes resources, substrate availability, or the physical environment(White and Pickett, 1985). Rivers in steep mountainous landscapes are subject to severe distur- bances such as landslides, debris ows, res, oods, and droughts. These geomorphic processes are important drivers of landscape evolu- tion but pose a signicant ecological challenge for organisms to survive in these highly variable systems. Indeed, the concept of riverscapes rec- ognizes that unique geomorphic features can have overriding effects on stream sh (Fausch et al., 2002) and that large infrequent disturbances may become the dominant force structuring the system, creating the template upon which subsequent ecological processes and interactions among species occur (Frissell et al., 1986; Reeves et al., 1995; Turner and Dale, 1998). To facilitate long-term persistence, many sh populations are dependent upon connected habitats to provide opportunities for dispersal and recolonization following localized extirpations caused by disturbances (e.g., Lamberti et al., 1991; Schlosser, 1991; Northcote, 1997; Gresswell, 1999; Roghair et al., 2002), and therefore, fragmenta- tion of the mountainous landscape (intrinsic and anthropogenic) can reduce dispersal and may render sh populations less resilient to epi- sodic disturbances (Kruse et al., 2001; Guy et al., 2008). Understanding the linkages between geomorphic disturbance pro- cesses and riverine habitat is particularly important for Pacic salmo- nids because these sh are intricately tied to Pacic Rim topography (Montgomery, 2000; Waples et al., 2008; May et al., 2013). Disturbances play such an important role in this region because hillslope and channel processes are tightly coupled in terms of routing water and sediment. This type of connectivity throughout the landscape can lead to large short-term uxes of sediment such as debris ows and post-re surface erosion (e.g., Benda and Dunne, 1997; Jackson and Roering, 2009), as well as long-term uxes driven by channel incision triggered by chang- es in base-level. Lithologic knickpoints may play a unique role in the Geomorphology xxx (2016) xxxxxx Corresponding author. E-mail addresses: [email protected] (C. May), [email protected] (J. Roering), [email protected] (K. Griswold), [email protected] (R. Gresswell). GEOMOR-05561; No of Pages 9 http://dx.doi.org/10.1016/j.geomorph.2016.03.029 0169-555X/© 2016 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Geomorphology journal homepage: www.elsevier.com/locate/geomorph Please cite this article as: May, C., et al., The waterfall paradox: How knickpoints disconnect hillslope and channel processes, isolating salmonid populations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10.1016/j.geomorph.2016.03.029
9

The waterfall paradox: How knickpoints disconnect ... · The waterfall paradox: How knickpoints disconnect hillslope and channel processes, isolating salmonid populations in ideal

Jun 01, 2020

Download

Documents

dariahiddleston
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
  • Geomorphology xxx (2016) xxx–xxx

    GEOMOR-05561; No of Pages 9

    Contents lists available at ScienceDirect

    Geomorphology

    j ourna l homepage: www.e lsev ie r .com/ locate /geomorph

    The waterfall paradox: How knickpoints disconnect hillslope and channelprocesses, isolating salmonid populations in ideal habitats

    Christine May a,⁎, Josh Roering b, Kyle Snow a, Kitty Griswold c, Robert Gresswell da James Madison University, Department of Biology, MSC 7801, Harrisonburg, VA, 24401, USAb University of Oregon, Department of Geological Sciences, 325C Cascade Hall, Eugene, OR 97403, USAc Idaho State University, Department of Biological Sciences, 921 S. 8th Ave., Pocatello, ID, 83209, USAd U.S. Geological Survey, Northern Rocky Mountain Science Center, 2327 University Way, Suite 2, Bozeman, MT, 59715, USA

    ⁎ Corresponding author.E-mail addresses: [email protected] (C. May), jroering@

    [email protected] (K. Griswold), [email protected] (R. G

    http://dx.doi.org/10.1016/j.geomorph.2016.03.0290169-555X/© 2016 Elsevier B.V. All rights reserved.

    Please cite this article as: May, C., et al., Thepopulations in ideal habita..., Geomorpholog

    a b s t r a c t

    a r t i c l e i n f o

    Article history:Received 9 July 2015Received in revised form 22 March 2016Accepted 25 March 2016Available online xxxx

    Waterfalls create barriers to fish migration, yet hundreds of isolated salmonid populations exist above barriersand have persisted for thousands of years in steepmountainous terrain. Ecological theory indicates that small iso-lated populations in disturbance-prone landscapes are at greatest risk of extirpation because immigration and re-colonization are not possible. On the contrary, many above-barrier populations are currently thrivingwhile theirdownstream counterparts are dwindling. This quandary led us to explore geomorphic knickpoints as a mecha-nism for disconnecting hillslope and channel processes by limiting channel incision and decreasing the pace ofbase-level lowering. Using LiDAR from the Oregon Coast Range, we found gentler channel gradients, widervalleys, lower gradient hillslopes, and less shallow landslide potential in an above-barrier catchment comparedto a neighboring catchment devoid of persistent knickpoints. Based on this unique geomorphic template,above-barrier channel networks are less prone to debris flows and other episodic sediment fluxes. Theseabove-barrier catchments also have greater resiliency to flooding, owing towider valleys with greater floodplainconnectivity. Habitat preferencemodels further indicate that salmonid habitat is present in greater quantity andquality in these above-barrier networks. Therefore the paradox of the persistence of small isolated fish popula-tionsmay be facilitated by a geomorphicmechanism that both limits their connectivity to larger fish populationsyet dampens the effect of disturbance by decreasing connections between hillslope and channel processes abovegeomorphic knickpoints.

    © 2016 Elsevier B.V. All rights reserved.

    Keywords:KnickpointsSalmonid habitatDisturbance ecologyShallow landslides

    1. Introduction

    One of the most pressing conservation questions for Pacific salmonand trout (Oncorhynchus sp.) is the role of episodic disturbances instructuring stream habitat and triggering extreme fluctuations in popu-lation abundance. From an ecological perspective, disturbances can bedefined as ‘any relatively discrete event in time that disrupts ecosystem,community, or population structure and changes resources, substrateavailability, or the physical environment’ (White and Pickett, 1985).Rivers in steep mountainous landscapes are subject to severe distur-bances such as landslides, debris flows, fires, floods, and droughts.These geomorphic processes are important drivers of landscape evolu-tion but pose a significant ecological challenge for organisms to survivein these highly variable systems. Indeed, the concept of riverscapes rec-ognizes that unique geomorphic features can have overriding effects onstream fish (Fausch et al., 2002) and that large infrequent disturbancesmay become the dominant force structuring the system, creating the

    uoregon.edu (J. Roering),resswell).

    waterfall paradox: How knicky (2016), http://dx.doi.org/10

    template upon which subsequent ecological processes and interactionsamong species occur (Frissell et al., 1986; Reeves et al., 1995; Turner andDale, 1998). To facilitate long-term persistence, many fish populationsare dependent upon connected habitats to provide opportunities fordispersal and recolonization following localized extirpations caused bydisturbances (e.g., Lamberti et al., 1991; Schlosser, 1991; Northcote,1997; Gresswell, 1999; Roghair et al., 2002), and therefore, fragmenta-tion of the mountainous landscape (intrinsic and anthropogenic) canreduce dispersal and may render fish populations less resilient to epi-sodic disturbances (Kruse et al., 2001; Guy et al., 2008).

    Understanding the linkages between geomorphic disturbance pro-cesses and riverine habitat is particularly important for Pacific salmo-nids because these fish are intricately tied to Pacific Rim topography(Montgomery, 2000;Waples et al., 2008;May et al., 2013). Disturbancesplay such an important role in this region because hillslope and channelprocesses are tightly coupled in terms of routing water and sediment.This type of connectivity throughout the landscape can lead to largeshort-term fluxes of sediment such as debris flows and post-fire surfaceerosion (e.g., Benda and Dunne, 1997; Jackson and Roering, 2009), aswell as long-term fluxes driven by channel incision triggered by chang-es in base-level. Lithologic knickpoints may play a unique role in the

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://dx.doi.org/10.1016/j.geomorph.2016.03.029mailto:[email protected] logohttp://dx.doi.org/10.1016/j.geomorph.2016.03.029http://www.sciencedirect.com/science/journal/0169555Xwww.elsevier.com/locate/geomorphhttp://dx.doi.org/10.1016/j.geomorph.2016.03.029

  • 2 C. May et al. / Geomorphology xxx (2016) xxx–xxx

    landscape by functioning to disconnect hillslope and channel processesthat are driven by baselevel lowering (Crosby and Whipple, 2006;Ouimet et al., 2009; Hurst et al., 2012).

    Wehypothesize that the restriction of base-level lowering imposed bywaterfalls (i.e., geologic knickpoints) inhibits channel incision, resulting inwider than anticipated valleys, lower than anticipated channel slopes,lower hillslope angles, and reduced shallow landslide potential. Thesemorphologic conditions would in turn lead to the development of idealfish habitat and dampen the incidence of landslides and debris flows inabove-barrier channel networks.

    The Oregon Coast Range is an ideal landscape to explore these ideasbecause a large portion of this humid, steep, soil-mantled landscape isunderlain by the Tyee sandstone formation and presents a uniformridge-and-valley landscape where sediment transport is dominated byshallow landslides and debris flows (e.g., Stock and Dietrich, 2003).Within the Tyee Formation; however, resistant igneous dikes crop outlocally, often forming persistent waterfalls (Oxford, 2006). Variation intributary networks and slopemorphology above and below knickpointsindicate that the rate of knickpointmigration is slow comparedwith thetimescale of landform adjustment (Sweeney et al., 2012). This providestime for the above-barrier channel and valley network to adjust to base-level lowering that proceeds more slowly than the surrounding terrain.

    Within a stream network the place for greatest potential overlap be-tween debris flow disturbance and fish habitat is in small headwaterstreams, which provide habitat for resident and anadromous salmonids(May and Lisle, 2012). In the Oregon Coast Range, Coastal CutthroatTrout (Oncorhynchus clarkii clarkii) populations are present above nu-merous waterfalls that limit connectivity with downstream popula-tions, with an estimated 269 barrier isolated populations in westernOregon (Gresswell et al., 2006). Similarly, waterfalls created by isostaticrebound in Alaska have isolated numerous populations of coastal cut-throat for thousands of years (Whiteley et al., 2010). Ecological theoryindicates that small isolated populations in disturbance-prone ecosys-tems are at greatest risk of extirpation because immigration and recolo-nization are not possible and because small populations are subject togenetic drift, inbreeding, and basic demographic factors that reducetheir survival potential (e.g., Lande, 1988; Purvis et al., 2000; Woffordet al., 2005). In the Oregon Coast Range; however, above-barrier popu-lations of coastal cutthroat trout have persisted over long timescales,potentially millennia (Guy et al., 2008).

    The location of suitable habitat for salmonids in the Pacific North-west is predicted for habitat conservation plans (e.g., Agrawal et al.,2005) using Intrinsic Potential (IP) modeling (Burnett et al., 2007).Based upon topographic analysis, this approach identifies streamreaches of high potential for developing productive habitat, specificallyareas with broad valleys and low gradient channels within the range ofvarious salmonid species. We hypothesize that a network of unusuallybroad valleys will be present upstream of knickpoints in the OregonCoast Range; and we further hypothesize that when normalized fordrainage area, channel slopes will be lower because of reduced verticalincision and increased lateral erosion. An overall reduction in channelslope would be associated with more productive habitat (e.g., reducedwater velocity and increased nutrient retention) and would allow foran increased spatial distribution of fish by making more of the networkaccessible as a result of slope-induced limits to their upstreamextent. Inaddition, we hypothesize that hillslope angles will be lower above theknickpoint due to the decreased pace of base-level lowering, reducingthe frequency of landslide and debris flow disturbance to channels.

    Our approach to test these hypotheses about knickpoint effects onchannel development and disturbance potential uses high resolutionlight detection and ranging (LiDAR) derived topography in a paired wa-tershed comparison. The pattern of valleywidening, drainage area depen-dent channel slope, hillslope angles, and shallow landslide potential arequantified for a basin above a knickpoint (Kentucky Falls) and a neighbor-ing basin absent of knickpoints (Harvey Creek). This type of pairedwater-shed comparison holds great promise in the Oregon Coast Range (OCR)

    Please cite this article as: May, C., et al., The waterfall paradox: How knickpopulations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10

    because of the uniformity of the landscape (Gresswell et al., 2004;Sweeney et al., 2012; May et al., 2013; Marshall and Roering, 2014); inthe absence of deep-seated landslides and gabbro dikes that regulatebase-level lowering and upstream geomorphic processes for a small frac-tion (b20%)of catchments in theOCR. Past studies haveobserved remark-ably consistent hillslope angles, slope-area channel network metrics, andshallow landslide potential for OCR catchments (Montgomery, 2001;Kobor and Roering, 2004; Montgomery and Dietrich, 1994). In fact, theobserved uniformity of steep, dissected terrain in western Oregon in-spired a suite of geomorphic studies that sought to minimize local varia-tions in lithology, climate, vegetation, and the pace of uplift/erosion inorder to isolate the influence of geomorphic processes on topography(Reneau and Dietrich, 1991; Seidl and Dietrich, 1992; Roering et al.,1999; Heimsath et al., 2001; Stock and Dietrich, 2003, 2006). The limitedavailability of LiDAR derived topography prevented us from investigatingnumerous above-waterfall basins; therefore, the results are limited to apaired watershed study where the generality of the results remain to betested.

    2. Methods

    Paired basins were selected based on the availability of LiDAR data forthe entire drainage network (101 km2). TheNorth Fork Smith River (basinarea 27.6 km2) resides above an ~100 mwide gabbroic dike that forms a70 m high knickpoint at Kentucky Falls (Fig. 1). Of the estimated 269 ba-sins above waterfalls in western Oregon, basin area ranged from 5.4 to57.5 km2with an average size of 14.2 km2 (Gresswell et al., 2006). NearbyHarvey Creek (basin area 21.7 km2) contains no sizable knickpoints(Marshall and Roering, 2014) and was selected as the reference basin torepresent typical ridge-valley topography in the absence of variationfrom lithology or baselevel lowering.

    2.1. Measuring valley width

    To map valley width, we used the approach by May et al. (2013)where the gridded bare earth LiDAR data was smoothed with a movingwindow algorithm (Wood, 1996). At each grid node, we fit a second-order weight polynomial to a 15x15 node matrix of neighboring pointsand calculated slope gradient from the polynomial coefficients. Fromthese slope maps, flat valley floors can be clearly distinguished fromthe adjacent steep hillslopes, which transition abruptly from the valleyfloor. Next we measured cross sections on the slope map perpendic-ular to the valley along the mainstem and several major tributariesin the drainage network of Harvey Creek to construct the baselinerelation between drainage area and valley width. Measurementspacing corresponded to individual stream reaches in a synthetic channelnetwork developed by Clarke et al. (2008) and described below. AboveKentucky Falls, valley width was also measured at cross sections on theslope map along the mainstem and major tributaries.

    We tested for a difference in the regression coefficients between thetwo basins using multiple linear regression of the log-transformed var-iable. Differences in intercepts of the regression lines were evaluated byusing a variable that identified the basin; differences in slope of theregression lines were evaluated by adding an interaction term to the re-gression (basin ∗ drainage area). Effect size of thewaterfall on upstreamvalley widthwas calculated from the difference between Kentucky FallsandHarvey Creek,whichwas divided by the value fromHarvey Creek asa reference point. These values were calculated over a range of drainageareas using the regression equations for the width-area relations. Wehypothesized that the above-waterfall basin would have broader val-leys for a given drainage area compared to the ambient landscape, asrepresented byHarvey Creek. Data from basins b0.1 km2were excludedfrom the Harvey Creek data set because of the scaling break identifiedby May et al. (2013).

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://dx.doi.org/10.1016/j.geomorph.2016.03.029

  • Fig. 1. Slope maps based on LiDAR-derived topography for the basin above Kentucky Falls on the North Fork Smith River and nearby Harvey Creek, which contains nomajor knickpoints.

    3C. May et al. / Geomorphology xxx (2016) xxx–xxx

    2.2. Slope-area relation

    Geomorphologists have long recognized the relationship betweenchannel slope (S) and drainage area (A) as an important predictor ofriver profile characteristics (Hack, 1957). Empirical observations fromriver systems around the world reveal a consistent power-law scalingof the slope-area relation (e.g., Hack, 1973; Flint, 1974; Howard andKerby, 1983). Reach-scale slope-area datawere obtained from syntheticchannel networks developed by Clarke et al. (2008). The resulting linearrelation was restricted to stream reaches with drainage areas N1 km2 tolimit observations to the fluvial process domain, based on the scalingbreak observed in previous studies (Stock and Dietrich, 2003). Differ-ences in the regression coefficients between the two basinswere exam-ined using multiple linear regression of the log-transformed variable.The difference in intercepts of the regression lines was tested by addinga variable that identified the basin; a difference in slope of the regres-sion lines was tested by adding an interaction term to the regression(basin ∗ drainage area). Effect size of the waterfall on upstream channelslope was calculated from the difference between Kentucky Falls andHarvey Creek and dividing by the value from Harvey Creek as a refer-ence point. These values were calculated over a range of drainageareas using the regression equations for the width-area relations. Weused a steepest descent algorithm to generate longitudinal profiles foreach mainstem channel. We hypothesized that the above-waterfallbasin would have a lower channel slope for a given drainage area com-pared to the basin adjusting to a fluctuating base level, as representedby Harvey Creek.

    Please cite this article as: May, C., et al., The waterfall paradox: How knickpopulations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10

    2.3. Hillslope angles and shallow landslide potential

    Because the propensity for hillslopes to produce shallow landslidesdepends on the topography of headwall regions, including colluvial hol-lows and sideslopes, we analyzed the average angle of hillslopes. In ad-dition, we employed a commonly-used coupled slope stability-hydrology algorithm to identify sites of potential shallow landslidingin the paired basins and used the resulting data as a proxy for landslidedisturbance potential (Montgomery and Dietrich, 1994). The actuallikelihood of channel disturbance depends on the details of predictedlandslide runout pathways (e.g., May and Gresswell, 2004), but suchan analysis is beyond the scope of this contribution. Instead, the long-term persistence of knickpoints at our study site implies that hillslopesabove and below have adjusted to the pace of local base-level loweringsuch that sediment production processes can be inferred frommorphol-ogy all else equal.

    We calculated values of q/T, where q is effective rainfall and T is soiltransmissivity according to Montgomery and Dietrich (1994) using pa-rameters characteristic of the Oregon Coast Range. The topographic vari-ables required for this computation include local hillslope gradient, whichwegeneratedusing the LiDAR smoothing procedure described above, anddrainage area per unit contour width, which we generated using the Dinfalgorithm as implemented in Matlab by Perron et al. (2008). Followingthe procedures described in Montgomery and Dietrich (1994), we de-fined unconditionally stable terrain as having slopes less than the criticalvalue for full soil saturation (in this case 0.38 or 38%). Slopes greater than1.0 (or 100%) were deemed unconditionally unstable and these areas

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://dx.doi.org/10.1016/j.geomorph.2016.03.029

  • Table 1Knickpoint effect size on valley width (m), calculated from the regression equations inFig. 2 for a range of drainage areas.

    Drainage area (km2) Harvey Creek Kentucky Falls Effect size

    0.1 4.8 7.8 63%1 18.1 32.5 80%10 67.8 134.4 98%

    4 C. May et al. / Geomorphology xxx (2016) xxx–xxx

    commonly have thin or absent soil mantle based on our field observa-tions. As such, the q/T values define the relative propensity for shallowlandslide failure for slopes between 0.38 and 1.0. Low values of q/T reflectfailure-prone siteswith a high propensity for failure because low effectiveprecipitation is required to induce landsliding.

    2.4. Fish habitat modeling

    We modeled juvenile trout habitat potential using a reach-scaleintrinsic potential (IP) value:

    IP ¼ IMAD � IBFW � IGRADIENTð Þ1=3

    where IMAD is an index of mean annual discharge, reflecting the need forchannels large enough to supply adequate perennial streamflow for sal-monid persistence (Burnett et al., 2007). Suitability curves developedfor coastal cutthroat trout indicate that habitat suitability is maximizedin mid-sized streams (0.05–0.6 m3/s). The second parameter, IBFW, is anestimate of the bank-full channel width from local hydraulic geometryrelations (based on field data from the Oregon Department of Fish andWildlife and the U.S. Environmental Protection Agency, as cited byClarke et al., 2008). Suitability curves suggest that habitat is maximizedfor coastal cutthroat trout in channel reaches with an estimated bank-full width between 1and 8 m. The third parameter, IGRADIENT, is anindex of reach-scale channel gradient estimated from synthetic streamnetwork routed through 10-m DEMs using elevations interpolatedfrom contour lines (Clarke et al., 2008). The habitat suitability curvefor coastal cutthroat trout developed by Griswold and Reeves (2014)originated from below-barrier streams where cutthroat are in competi-tion with other salmonids, most notably Coho Salmon (Oncorhynchuskisutch). Competition leads to habitat segregationwith coastal cutthroattrout occupying riffles and coho salmon occupying pools, resulting inthe competitively inferior trout occupying less favorable habitatwhen dominated by salmon (Glova, 1987). Similarly, Sabo andPauley (1997) concluded that the advantages of size and experienceallow salmon to dominate and that trout populations that haveevolved with salmon are less competitive than those that haveevolved in isolation. Therefore, we compared the coastal cutthroatsuitability curve developed by Griswold and Reeves (2014), wherehabitat suitability is maximized in channel reaches with 4–8%channel gradient with a revised curve where habitat suitability ismaximized in all reaches with b8% channel gradient to reflect thelack of competitive exclusion from low gradient habitats in above-barrier populations.

    Fig. 2. The power-law relation between valley floor width and drainage area for the abovewaterfall basin (Kentucky Falls) compared to the ambient landscape (Harvey Creek) forthe full extent of the channel network.

    Please cite this article as: May, C., et al., The waterfall paradox: How knickpopulations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10

    The IP value is based on reach-scale habitat suitability scores, whichconvert the value of each variable to a score between 0 and 1. The com-ponent indices are equally weighted and IP value is calculated at thegeometric mean of each of the three parameters. The IP values close toone indicate great potential for high-quality habitat; whereas, IP valuesclose to zero indicate almost no potential for habitat. The proportion ofthe channel network length in each of four IP categories (negligiblehabitat b0.25, low quality habitat 0.25–0.49, moderate quality habitat0.5–0.74, and high quality habitat N0.75) was calculated for each basinand compared using a Chi-square (χ2) test.

    In addition to determining the overall quality of habitat using theIP value, the quantity of habitat can be calculated as the proportionof the drainage network that is accessible to fish. This calculation isbased on a slope-induced limit for coastal cutthroat trout of ~10% for thereach-scale channel gradient (unpublished data by R. Gresswell). Paststudies have shown that slope-induced limits to fish distributioncan be a strong predictor for trout occurrence and their presencetends to decrease linearly with increasing channel slope up to the10% limit (Kruse et al., 1997).

    3. Results and discussion

    3.1. Valley width

    Regression analysis revealed a significant difference in the rela-tion between drainage area and valley floor width in the two basins(Fig. 2). The intercept of the regression lines was greater for KentuckyFalls (p b 0.001); however, the slope of the regression lines did not differ(p = 0.142). Because the regression lines diverged at larger drainageareas, the effect size increased with drainage area and ranged from 63to 98% (Table 1). These results indicate that valley floorwidthwas consis-tently broader than anticipated in the above-waterfall basin, suggestingthat lateral river erosion that broadens valleys is playing a larger rolethan in actively incising valleys that are keeping pacewith base-level low-ering. This in turn can lead to lower hillslope angles (Penserini, 2015),

    Fig. 3. Comparison of the slope-area relation for stream reaches in the fluvial processdomain (N1 km2) in the above waterfall basin (Kentucky Falls) and the ambientlandscape (Harvey Creek).

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://dx.doi.org/10.1016/j.geomorph.2016.03.029

  • Fig. 4. Comparison of hillslope steepness in the above-knickpoint basin (Kentucky Falls)and the basin devoid of knickpoints (Harvey Creek).

    Table 2Knickpoint effect on channel slope (m/m), calculated from the regression equations inFig. 3 for a range of drainage areas.

    Drainage area (km2) Harvey Creek Kentucky Falls Effect size

    1 0.07 0.03 −57%5 0.02 0.01 −32%10 0.01 0.01 −18%

    Fig. 5.Maps of q/T values (Montgomery and Dietrich, 1994) for Harvey Creek and the basin aboindicate highpropensity for shallow landsliding;whereas highvalues (slightlynegative, cool colstable and deep red colors show terrain that is unconditionally unstable (or too steep to retain

    5C. May et al. / Geomorphology xxx (2016) xxx–xxx

    Please cite this article as: May, C., et al., The waterfall paradox: How knickpopulations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10

    longer residence times for soils (Sweeney et al., 2012), reduced erosionrates (Hurst et al., 2012), and is consistent with the effect of knickpointsformed by resistant rock layers in the study by Marshall and Roering(2014).

    Exponents for the power-function relation of drainage area andvalley floor width ranged from 0.57 for Harvey Creek to 0.62 aboveKentucky Falls, which are both within the range reported by Snyderet al. (2000) of 0.50–0.67but notably higher than the value of 0.4 reportedby Snyder et al. (2003). Surprisingly, no scaling breakwas observed in thewidth-area relation for the North Fork Smith River above Kentucky Falls.A previous study by Snyder et al. (2003) and our research inHarvey Creekdocumented a scaling break at a drainage area of 0.1 km2 (May et al.,2013), below which no corresponding decrease in valley width withdrainage areawas observed. This constant width is likely the signal of de-bris flows carving valleys of small drainage areas, with downstream areasbeing carved by fluvial incision (Montgomery and Foufoula-Georgiou,1993; Stock and Dietrich, 2003). The lack of a robust scaling breakabove Kentucky Falls suggests that some headwater channels may betoo broad and low gradient to transport debris flows.

    Broad valley floors are a critical driver of fish habitat because theyallow for floodplain development (e.g., Naiman et al., 2010), increasedcarbon storage (Wohl et al., 2012), enhanced hyporheic exchange(e.g., Baxter and Hauer, 2000), and provide off-channel habitats duringfloods that increase juvenile salmonid survival (e.g., Solazzi et al., 2000;Bell et al., 2001). A previous study by Belmont (2011) observed a similarpattern of relatively wide and mostly unconfined floodplains above asteep incising knickpoint. Broad valley floors also intercept debris flows

    ve Kentucky Falls on the N.F. Smith River. Low values of q/T (highly negative, warm colors)ors) reflect lowpropensity for failure.Darkblue pixels reflect terrain that is unconditionallya persistent soil mantle).

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://dx.doi.org/10.1016/j.geomorph.2016.03.029

  • Fig. 6. The distribution of IP values from habitat suitability curves developedwhere coastalcutthroat trout are in competition with coho salmon and habitat utilization is maximizedin reaches with 4–10% gradient (A), and the distribution of IP values from habitatsuitability curves where cutthroat are expected to use lower gradient channels (allreaches b8%) in the absence of competition (B).

    6 C. May et al. / Geomorphology xxx (2016) xxx–xxx

    and thus dampen episodic inputs of sediment (May andGresswell, 2004),rendering fish populations more resilient to disturbance.

    3.2. Channel slope

    Reach-scale channel slopewas lower for a given drainage area in theabove-barrier catchment (Fig. 3). Regression analysis revealed a differ-ence in the slope-area relation between basins, where the intercept(p b 0.001) and slope (p b 0.001) of the regression was lower in theabove-waterfall basin compared to the ambient landscape. Becausethe regression lines converged at larger drainage areas, the effect sizedecreased with drainage area and ranged from 57 to 18% reduction inslope with increasing drainage area (Table 2). With lower channelslopes extending farther up into the drainage network, the potentialfor pool-riffle morphologies was greatly expanded. Pool-riffle morphol-ogies typically occur in reaches ≤2% slope (Montgomery and Buffington,1997;Wohl andMerritt, 2005), which extend as high up in the networkas 1.8 km2 inNorth Fork Smith River above Kentucky Falls,whereas theyare limited to drainage areas N4.0 km2 in Harvey Creek.

    3.3. Hillslope angles and shallow landslide potential

    While some subcatchments above Kentucky Falls are as steep asthose observed in Harvey Creek many are systematically gentler im-plying decreased rates of sediment production through soil creepprocesses and shallow soil slips. The mean hillslope gradient forHarvey Creek is 0.74±0.20, whereas the mean value above KentuckyFalls is 0.53±0.25 (Fig. 4). The coupled hydrology-slope stability modelresults demonstrate that a substantial fraction of the basin upstream ofKentucky Falls is unconditionally stable (slopes b 0.38) in contrast toHarvey Creek where only the valley floor and debris fan deposits exhibitsuch low slope angles (Fig. 5). The origin of the numerous and contiguousgentle hillsides in the basin above Kentucky Falls is unclear, although itmay result from decreased steepness owing to reduced base-level lower-ing. The proportion of unconditionally unstable slopes (slope N 1.0) ismuch higher in Harvey Creek than above Kentucky Falls, where steepslopes are scattered throughout the catchment. The q/T values aresystematically lower in Harvey Creek than above Kentucky Falls aswell, suggesting that the frequency of shallow landsliding and de-bris flows may be higher in Harvey given that lower values of effectiveprecipitation are required for failure. Taken together, the topographic dif-ferences between Harvey Creek and the hillslopes above Kentucky Fallsare profound despite their proximity, implying a substantial knickpointcontrol on the rate and pattern of sediment production in the basinabove Kentucky Falls.

    3.4. Habitat quality and quantity

    Because broad valleys and low-gradient channels extend higher inthe network in the above-waterfall basin, a greater proportion of thenetwork exhibits high IP values compared to the reference landscapeof Harvey Creek (Fig. 6). This finding is consistent with the original IPvalues developed for coastal cutthroat trout where habitat use wasmaximized in channels with 4–8% gradient because of competitiveexclusion (p-value b 0.01; Fig. 6A), and the revised IP value for coastalcutthroat trout residing in the absence of competition (habitat utiliza-tion maximized in all reaches b8%) provided similar results (p b 0.01;Fig. 6B). Correspondingly, we observed an increase in habitat quantity,with 59% of the total channel length above Kentucky Falls occurring inreaches that were less than the slope-induced limit for cutthroat trout(b10% slope); whereas Harvey Creek had only 27% of the network pro-viding habitat (Fig. 7). This is a direct result of the less steep river profileaboveKentucky Falls (Fig. 8). However, the areawhere there is potentialfor direct overlap between debris flows and fish habitat (S = 3–10%:May and Lisle, 2012) is greater in the above-barrier stream network(Table 3).

    Please cite this article as: May, C., et al., The waterfall paradox: How knickpopulations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10

    The resulting effect of the differences in river profiles is an increasein potential habitat of 2.3 times in the above-barrier basin, with morelongitudinally and latitudinally connected habitat. This high connectiv-ity of habitat and consistently lower gradient reaches in the above-barrier network can increase gene flow and reduce genetic differen-tiation (Wofford et al., 2005; Guy et al., 2008; Kanno et al., 2011).Greater expansion of the fish-bearing channel network also createsan opportunity for a broader spatial distribution of fish, thus reduc-ing competition for limited resources and enhancing the opportuni-ty for the spatial spreading of risk. Previous studies have also foundthat isolated populations occurring in larger stream networkshave retained substantial genetic variation, which suggests thatthe amount of habitat in headwater streams is an important consid-eration for maintaining the evolutionary potential of isolated popu-lations (Whiteley et al., 2010).

    4. Conclusions

    Ecogeomorphology is emerging as an interdisciplinary field that ex-plores the concept of connectivity and places particular importance ondescribing the structural and functional linkages between the flow ofwater, landforms, and the dispersal of organisms (e.g., With and Crist,1995; Pringle, 2003; Wainwright et al., 2011). The opposing conceptof disconnectivity in landscape processes is also being recognized asplaying some important functional roles in riverine systems (e.g., Laneet al., 2004; Jackson and Pringle, 2010). One of the drivers on land-scape connectivity is the geomorphic response of hillslopes andchannels to base-level lowering (e.g., Faulkner, 2008). By exploringhow knickpoints disconnect small watersheds from base-level lowering,and thereby the surrounding landscape, we can reveal previously unrec-ognized patterns. This is particularly insightful when investigating head-water streams, which compose a large portion of the total channelnetwork length and directly connect the upland and riparian landscapeto the rest of the stream ecosystem (Freeman et al., 2007).

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://dx.doi.org/10.1016/j.geomorph.2016.03.029

  • Fig. 7. The spatial distribution and quantity of habitat in the above-barrier catchment (Kentucky Falls) and the ambient landscape (Harvey Creek).

    7C. May et al. / Geomorphology xxx (2016) xxx–xxx

    Results of our study explicitly linked the unique geomorphic settingof headwater streams in above-knickpoint basins with the unanticipat-ed ecological outcome of barrier-isolated populations inhabiting themost resilient and ideal habitat in mountainous landscapes. Althoughthe geomorphic setting is unique, it is not an isolated incidence asknickpoints are common features in otherwise steady state landscapes.Specifically, this study posed a fundamental question about landscapeconnectivity and the role of knickpoints in channel development,while also addressing a pressing conservation question in aquatic ecol-ogy about disconnected fish populations. We hypothesize that limited

    Fig. 8. Longitudinal profiles for the mainstem of the North Fork Smith River aboveKentucky Falls (which corresponds with the steep reach at ~12,500 m) and themainstem of Harvey Creek.

    Please cite this article as: May, C., et al., The waterfall paradox: How knickpopulations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10

    incision imposed by a relatively stationary knickpoint has led to lateralexpansion of the valley and less steep river profile that provides for agreater expansion of low-gradient stream reaches throughout the fluvi-al network and reduced sediment flux owing to lower gradienthillslopes. From an ecological perspective, the novelty of our findingsis quantifying the geomorphic controls on the habitat template thatleads to the long-term persistence of small disconnected populations,whereas the majority of past studies have focused on factors leadingto extirpation. Future studies are needed to test the generality ofthese results as the availability of LiDAR-derived topography becomesmore spatially extensive and provides the opportunity for replication.Our current paired watershed approach is limited, but our novel hy-pothesis warrants testing. Qualitative examination of other small wa-tersheds situated above knickpoints in western Oregon revealssimilar topographic characteristics. Because our analysis derivesfrom a process-based perspective on how channel incision regulatesvalley morphology, hillslope form, and sediment production; ourapproach should have application in other mountainous settings.

    Table 3Proportion of the channel network in low gradient fish habitat (b3%), habitats wheredebrisflows andfish habitat directly overlap (3–10%), and nonfish bearing stream reachesin the debris flow process domain (N10%).

    Stream gradient Harvey Cr. Kentucky Falls

    b3% 0.08 0.273–10% 0.18 0.32N10% 0.73 0.41

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://dx.doi.org/10.1016/j.geomorph.2016.03.029

  • 8 C. May et al. / Geomorphology xxx (2016) xxx–xxx

    The combination of broader than anticipated valley floors and thegreater extent of low gradient stream reaches interact to develop agreater abundance of high quality fish habitat in above-barrier catch-ments. Our results indicate that the effect of knickpoints on valley widthincreases with drainage area, with a near doubling of valley width at adrainage area of 101 km2. Concomitantly, the effect of knickpoints onchannel slope decreases with drainage area, resulting in low gradientstream reaches with favorable fish habitat extending further up in thechannel network. In addition to differences in physical habitat that can in-crease fish production, these disconnected above-barrier catchments arealso protected from invasion by nonnative fish, hybridizationwith hatch-ery fish, and spread of disease (Rahel, 2013). These populations also havereduced competition from other species (Ross, 1991), which can increaseproductivity.

    Channel networks above barriers experience relatively constantbase-level, and thus, hillslope and channel processes that are drivenby incision are effectively disconnected. These disconnected hillslopeshave less potential for debris flows (Penserini, 2015), longer residencetimes for soils (Sweeney et al., 2012), and more rounded hillslopeswith reduced erosion rates (Hurst et al., 2012). With less potential forepisodic disturbance, fish populations may undergo fewer populationbottlenecks andmaymaintain consistently more abundant populationsthan their downstream counterparts. A primary focus in conservationgenetics is to understand factors that influence small populations, andas the rate of population fragmentation and isolation increases, it be-comes increasingly important to examine factors that influence themaintenance of genetic diversity and therefore the likelihood of persis-tence (e.g., Letcher et al., 2007; Whiteley et al., 2010; Toterotot et al.,2014). From this topographically based analysis of fish habitat we hy-pothesize that above-barrier fish populations reside inmore temporallystable and productive habitat and that populations will be consistentlylarger and more stable through time. Future research can use populationgenetics to test this hypothesis, with the prediction that above-barrierpopulations have fewer population bottlenecks and larger effective popu-lation size. In contrast, populations isolated by anthropogenic factors,such as dams and culverts, may be at greatest risk of extirpation becausethe co-evolved geomorphic template for habitat formation and reduceddisturbance is not present.

    Acknowledgements

    The authors would like to thank NCALM (National Center forAirborne Laser Mapping) and the Oregon LiDAR Consortium (operatedby the Oregon Department of Geology and Mineral Industries) for pro-viding topographic data. We also wish to thank the CLAMS project andBrett Holycross with Pacific States Marine Fisheries Commission forsharing stream layers and IP parameters. Statistician Lihua Chen gener-ously assisted on the analyses. We thank the Binghamton symposiumorganizers for the opportunity to present this study. We are grateful totwo anonymous reviewers and Frank Magilligan, who provided thor-ough reviews that improved the manuscript. Any use of trade, firm, orproduct names is for descriptive purposes only and does not imply en-dorsement by the U.S. Government.

    References

    Agrawal, A., Schick, R.S., Bjorkstedt, E.P., Szerlong, R.G., Goslin, M.N., Spence, B.C.,Williams, T.H., Burnett, K.M., 2005. Predicting the potential for historical coho,Chinook and steelhead habitat in Northern California. NOAA Technical MemorandumNOAA-TM-NMFS-SWFSC-379, Santa Cruz, Calif.

    Baxter, C.V., Hauer, R.F., 2000. Geomorphology, hyporheic exchange, and selection ofspawning habitat by bull trout (Salvelinus confluentus). Can. J. Fish. Aquat. Sci. 57(7), 1470–1481.

    Bell, E., Duffy, W.G., Roelofs, T.D., 2001. Fidelity and survival of juvenile coho salmon in re-sponse to a flood. Trans. Am. Fish. Soc. 130 (3), 450–458.

    Belmont, P., 2011. Floodplain width adjustments in response to rapid base level fall andknickpoint migration. Geomorphology 128, 92–102. http://dx.doi.org/10.1016/j.geomorph.2010.12.026.

    Please cite this article as: May, C., et al., The waterfall paradox: How knickpopulations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10

    Benda, L., Dunne, T., 1997. Stochastic forcing of sediment supply to channel networksfrom landsliding and debris flow. Water Resour. Res. 33 (12), 2849–2863.

    Burnett, K.M., Reeves, G.H., Miller, D.J., Clarke, S., Vance-Borland, K., Christiansen, K., 2007.Distribution of salmon-habitat potential relative to landscape characteristics and im-plications for conservation. Ecol. Appl. 17, 66–80. http://dx.doi.org/10.1890/1051-0761(2007)017[0066:DOSPRT]2.0.CO;2.

    Clarke, S.E., Burnett, K.M., Miller, D.J., 2008.Modeling streams and hydrogeomorphic attri-butes in Oregon from digital and field data. J. Am. Water Resour. Assoc. 44 (2),459–477. http://dx.doi.org/10.1111/j.1752-1688.2008.00175.x.

    Crosby, B.T., Whipple, K.X., 2006. Knickpoint initiation and distribution within fluvial net-works: 236waterfalls in theWaipaoa River, North Island, New Zealand. Geomorphol-ogy 82 (1), 16–38.

    Faulkner, H., 2008. Connectivity as a crucial determinant of badlandmorphology and evo-lution. Geomorphology 100, 91–103.

    Fausch, K.D., Torgerson, C.E., Baxter, C.V., Li, H.W., 2002. Landscapes to riverscapes: bridg-ing the gap between research and conservation of stream fishes. Bioscience 52 (6),483–498. http://dx.doi.org/10.1641/0006-3568.

    Flint, J.J., 1974. Stream gradient as a function of order, magnitude, and discharge. WaterResour. Res. 10 (5), 969–973.

    Freeman, M.C., Pringle, C.M., Jackson, C.R., 2007. Hydrological connectivity and the contri-bution of stream headwaters to ecological integrity at regional scales. J. Am. WaterResour. Assoc. 43 (1), 5–14.

    Frissell, C.A., Liss, W.J., Warren, C.E., Hurley, M.D., 1986. A hierarchical framework forstream habitat classification: viewing streams in a watershed context. Environ.Manag. 10, 199–214.

    Glova, G.J., 1987. Comparison of allopatric cutthroat trout stocks with those sympatricwith coho salmon and sculpins in small streams. Environ. Biol. Fish 20 (4), 275–284.

    Gresswell, R.E., 1999. Fire and aquatic ecosystems in forested biomes of North America.Trans. Am. Fish. Soc. 128 (2), 193–221.

    Gresswell, R.E., Bateman, D.S., Lienkamper, G.W., Guy, T.J., 2004. Geospatial TechniquesFor Developing A Sampling Frame Of Watersheds Across A Region. In: Nishida, T.,Kailola, P.J., Hollingworth, C.E. (Eds.), GIS/Spatial Analyses In Fishery And Aquatic Sci-ences vol. 2. Fishery-Aquatic GIS Research Group, Saitama, Japan, pp. 517–530.

    Gresswell, R.E., Torgersen, C.E., Bateman, D.S., Guy, T.J., Hendricks, S.R., Wofford, J.E.B.,2006. A spatially explicit approach for evaluating relationships among coastal cut-throat trout, habitat, and disturbance in small Oregon streams. Am. Fish. Soc. Symp.48, 457–471.

    Griswold, K., Reeves, G., 2014. Coastal Cutthroat Trout Assessment, Final Report to Nation-al Fish and Wildlife Foundation (NFWF grant #2010-0055-022).

    Guy, T.J., Gresswell, R.E., Banks, M.A., 2008. Landscape-scale evaluation of genetic struc-ture among barrier-isolated populations of coastal cutthroat trout (Oncorhynchusclarkii clarkii). Can. J. Fish. Aquat. Sci. 65 (8), 1749–1762. http://dx.doi.org/10.1139/F08-090.

    Hack, J.T., 1957. Studies of longitudinal stream profiles in Virginia andMaryland. U.S. Geo-logical Survey Professional Paper 294-B.

    Hack, J.T., 1973. Streamprofile analysis and stream-gradient index. U.S. Geol. Surv. J. Res. 1(4), 421–429.

    Heimsath, A.M., Dietrich, W.E., Nishiizumi, K., Finkel, R.C., 2001. Stochastic processes ofsoil production and transport: erosion rates, topographic variation and cosmogenicnuclides in the Oregon Coast Range. Earth Surf. Process. Landf. 26 (5), 531–552.

    Howard, A.D., Kerby, G., 1983. Channel changes in badlands. Geol. Soc. Am. Bull. 94,739–752.

    Hurst, M.D., Mudd, S.M., Walcott, R., Attal, M., Yoo, K., 2012. Using hilltop curvature to de-rive the spatial distribution of erosion rates. J. Geophys. Res. Earth Surf. 117,2003–2012. http://dx.doi.org/10.1029/2011JF002057.

    Jackson, C.R., Pringle, C.M., 2010. Ecological benefits of reduced hydrologic connectivity inintensively developed landscapes. Bioscience 60 (1), 37–46.

    Jackson, M., Roering, J.J., 2009. Post-fire geomorphic response in steep, forested land-scapes: Oregon Coast Range, USA. Quat. Sci. Rev. 28 (11), 1131–1146. http://dx.doi.org/10.1016/j.quascirev.2008.05.003.

    Kanno, Y., Vokoun, J.C., Letcher, B.H., 2011. Fine-scale population structure and riverscapegenetics of brook trout (Salvelinus fontinalis) distributed continuously along headwaterchannel networks. Mol. Ecol. 20 (18), 3711–3729.

    Kobor, J.S., Roering, J.J., 2004. Systematic variation of bedrock channel gradients in thecentral Oregon Coast Range: implications for rock uplift and shallow landsliding.Geomorphology 62 (3), 239–256.

    Kruse, C.G., Hubert, W.A., Rahel, F.J., 1997. Geomorphic influences on the distribution ofYellowstone Cutthroat Trout in the Absaroka Mountains, Wyoming. Trans. Am. Fish.Soc. 126, 418–427.

    Kruse, C.G., Hubert, W.A., Rahel, F.J., 2001. An assessment of headwater isolation as aconservation strategy for cutthroat trout in the Absaroka Mountains of Wyoming.Northwest Sci. 75, 1–11.

    Lamberti, G.A., Gregory, S.V., Ashkenas, L.R., Wildman, R.C., Moore, K., 1991. Streamecosystem recovery following a catastrophic debris flow. Can. J. Fish. Aquat. Sci. 48,196–208.

    Lande, R., 1988. Genetics and demography in biological conservation. Science 241,1455–1460.

    Lane, S.N., Brookes, C.J., Kirkby, A.J., Holden, J., 2004. A network-index based version ofTOPMODEL for use with high-resolution digital topographic data. Hydrol. Process.18 (1), 191–201.

    Letcher, B.H., Nislow, K.H., Coombs, J.A., O'Donnell, M.J., Dubreuil, T.L., 2007. Populationsresponse to habitat fragmentation in a stream-dwelling brook trout population.PLoS One 2 (11), e1139. http://dx.doi.org/10.1371/journal.pone.0001139.

    Marshall, J.A., Roering, J.J., 2014. Diagenetic variation in the Oregon Coast Range: implica-tions for rock strength, soil production, hillslope form, and landscape evolution.J. Geophys. Res. Earth Surf. 119, 1395–1417. http://dx.doi.org/10.1002/2013JF003004.

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0005http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0005http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0005http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0010http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0010http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0010http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0015http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0015http://dx.doi.org/10.1016/j.geomorph.2010.12.026http://dx.doi.org/10.1016/j.geomorph.2010.12.026http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0025http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0025http://dx.doi.org/10.1890/1051-0761(2007)017[0066:DOSPRT]2.0.CO;2http://dx.doi.org/10.1890/1051-0761(2007)017[0066:DOSPRT]2.0.CO;2http://dx.doi.org/10.1111/j.1752-1688.2008.00175.xhttp://refhub.elsevier.com/S0169-555X(16)30124-6/rf0040http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0040http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0040http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0045http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0045http://dx.doi.org/10.1641/0006-3568http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0055http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0055http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0060http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0060http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0060http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0065http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0065http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0065http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0070http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0070http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0075http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0075http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0080http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0080http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0080http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0080http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0085http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0085http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0085http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0090http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0090http://dx.doi.org/10.1139/F08-090http://dx.doi.org/10.1139/F08-090http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0100http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0100http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0105http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0105http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0110http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0110http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0110http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0115http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0115http://dx.doi.org/10.1029/2011JF002057http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0125http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0125http://dx.doi.org/10.1016/j.quascirev.2008.05.003http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0135http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0135http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0135http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0140http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0140http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0140http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0145http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0145http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0145http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0150http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0150http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0150http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0155http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0155http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0155http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0160http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0160http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0165http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0165http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0165http://dx.doi.org/10.1371/journal.pone.0001139http://dx.doi.org/10.1002/2013JF003004http://dx.doi.org/10.1016/j.geomorph.2016.03.029

  • 9C. May et al. / Geomorphology xxx (2016) xxx–xxx

    May, C.L., Gresswell, R.E., 2004. Spatial and temporal patterns of debris-flow deposition inthe Oregon Coast Range, USA. Geomorphology 57, 135–149.

    May, C.L., Lisle, T.E., 2012. River profile controls on channel morphology, debris flow dis-turbance, and the spatial extent of salmonids in steep mountain streams. J. Geophys.Res. 117, F00A03. http://dx.doi.org/10.1029/2011JF002324.

    May, C.L., Roering, J., Eaton, L.S., Burnett, K.M., 2013. Controls on valleywidth inmountainouslandscapes: the role of landsliding and implications for salmonid habitat. Geology 41(4), 503–506. http://dx.doi.org/10.1130/G33979.1.

    Montgomery, D.R., 2000. Coevolution of the Pacific salmon and Pacific Rim topography.Geology 28, 1107–1110. http://dx.doi.org/10.1130/0091-7613(2000)28b1107:COTPSAN2.0.CO;2.

    Montgomery, D.R., 2001. Slope distributions, threshold hillslopes, and steady-state topog-raphy. Am. J. Sci. 301 (4-5), 432–454.

    Montgomery, D.R., Buffington, J.M., 1997. Channel-reach morphology in mountain drain-age basins. Geol. Soc. Am. Bull. 109 (5), 596–611.

    Montgomery, D.R., Dietrich, W.E., 1994. A physically based model for the topographiccontrol. Water Resour. Res. 30 (4), 1153–1171.

    Montgomery, D.R., Foufoula-Georgiou, E., 1993. Channel network source representationusing digital elevation models. Water Resour. Res. 29, 3925–3934. http://dx.doi.org/10.1029/93WR02463.

    Naiman, R.J., Bechtold, J.S., Beechie, T.J., Latterell, J.J., Pelt, R.V., 2010. A process-based viewof floodplain forest patterns in coastal river valleys of the Pacific Northwest. Ecosys-tems 13, 1–31. http://dx.doi.org/10.1007/s10021-009-9298-5.

    Northcote, T.G., 1997. Potamodromy in Salmonidae – living and moving in the fast lane.N. Am. J. Fish Manag. 17, 1029–1045.

    Ouimet,W.B.,Whipple, K.X., Granger, D.E., 2009. Beyond threshold hillslopes: channel ad-justment to base-level fall in tectonically active mountain ranges. Geology 37 (7),579–582.

    Oxford, J., 2006. Early Oligocene Intrusions In The Central Coast Range of Oregon M.S.thesis Oregon State University (237 pp.).

    Penserini, B., 2015. Debris Flow Networks And The Role Of Baselevel In Steep HeadlandLandscapes M.S. thesis University of Oregon (89 pp.).

    Perron, J.T., Dietrich, W.E., Kirchner, J.W., 2008. Controls on the spacing of first-ordervalleys. J. Geophys. Res. Earth Surf. http://dx.doi.org/10.1029/2007JF000977.

    Pringle, C., 2003. What is hydrologic connectivity and why is it ecologically important?Hydrol. Process. 17 (13), 2685–2689.

    Purvis, A., Gittleman, J.L., Cowlishaw, G., Mace, G.M., 2000. Predicting extinction risk in de-clining species. Proc. R. Soc. Biol. Sci. 267, 1947–1952. http://dx.doi.org/10.1098/rspb.2000.1234.

    Rahel, F.J., 2013. Intentional fragmentation as a management strategy in aquatic systems.Bioscience 63 (5), 362–372. http://dx.doi.org/10.1525/bio.2013.63.5.9.

    Reeves, G.H., Benda, L.E., Burnett, K.M., Bisson, P.A., Sedell, J.R., 1995. A disturbance-basedecosystem approach to maintaining and restoring freshwater habitats of evolution-arily significant units of anadromous salmonids in the Pacific Northwest. Am. Fish.Soc. Symp. 17, 334–349.

    Reneau, S.L., Dietrich, W.E., 1991. Erosion rates in the southern Oregon Coast Range:evidence for an equilibrium between hillslope erosion and sediment yield. EarthSurf. Process. Landf. 16 (4), 307–322.

    Roering, J.J., Kirchner, J.W., Dietrich, W.E., 1999. Evidence for nonlinear diffusive sedimenttransport on hillslopes and implications for landscape morphology. Water Resour.Res. 35 (3), 853–870.

    Roghair, C.N., Dollof, C.A., Underwood, M.K., 2002. Response of a brook trout populationand instream habitat to a catastrophic flood and debris flow. Trans. Am. Fish. Soc.131, 718–730.

    Please cite this article as: May, C., et al., The waterfall paradox: How knickpopulations in ideal habita..., Geomorphology (2016), http://dx.doi.org/10

    Ross, S.T., 1991. Mechanisms structuring stream fish assemblages: are there lessons fromintroduced species? Environ. Biol. Fish 30 (4), 359–368.

    Sabo, J.L., Pauley, G.B., 1997. Competition between stream-dwelling cutthroat trout(Oncorhynchus clarki) and coho salmon (Oncorhynchus kisutch): effects of relativesize and population origin. Can. J. Fish. Aquat. Sci. 54, 2609–2617.

    Schlosser, I.J., 1991. Stream fish ecology: a landscape perspective. Bioscience 41, 704–712.Seidl, M.A., Dietrich, W.E., 1992. The problem of channel erosion into bedrock. Catena

    Suppl. 23, 101–124.Snyder, N.P., Whipple, K.X., Tucker, G.E., Merritts, D.J., 2000. Landscape response to

    tectonic forcing: DEM analysis of stream profiles in the Mendocino triple junction re-gion, northern California. Geol. Soc. Am. Bull. 112 (8), 1250–1263.

    Snyder, N.P., Whipple, K.X., Tucker, G.E., Merritts, D.J., 2003. Channel response to tectonicforcing: field analysis of stream morphology and hydrology in the Mendocino triplejunction region, northern California. Geomorphology 53, 97–127.

    Solazzi, M.F., Nickelson, T.E., Johnson, S.L., Rodgers, J.D., 2000. Effects of increasing winterrearing habitat on abundance of salmonids in two coastal Oregon streams. Can. J. Fish.Aquat. Sci. 57, 906–914.

    Stock, J., Dietrich, W.E., 2003. Valley incision by debris flows: evidence of a topographic sig-nature. Water Resour. Res. 39 (4), 1089. http://dx.doi.org/10.1029/2001WR001057.

    Stock, J., Dietrich, W.E., 2006. Erosion of steepland valleys by debris flows. Geol. Soc. Am.Bull. 118 (9), 1125–1148.

    Sweeney, K.E., Roering, J.J., Almond, P., Reckling, T., 2012. How steady are ‘steady-state’landscapes? Using soil spectroscopy to quantify erosional variability. Geology 40,807–810. http://dx.doi.org/10.1130/G33167.1.

    Toterotot, J.B., Perrier, C., Bergeron, N.E., Bernatchez, L., 2014. Influence of forest road cul-verts and waterfalls on the fine-scale distribution of Brook Trout genetic diversity in aboreal watershed. Trans. Am. Fish. Soc. 143 (6), 1577–1591.

    Turner, M.G., Dale, V.H., 1998. Comparing large, infrequent disturbances: what have welearned? Ecosystems 1, 493–496.

    Wainwright, J., Turnbull, L., Ibrahim, T.G., Lexartza-Artza, I., Thornton, S.F., Brazier, R.E.,2011. Linking environmental regimes, space and time: interpretations of structuraland functional connectivity. Geomorphology 126 (3), 387–404.

    Waples, R.S., Pess, G.R., Beechie, T., 2008. Evolutionary history of Pacific salmon in dynam-ic environments. Evol. Appl. 1, 189–206. http://dx.doi.org/10.1111/j.1752-4571.2008.00023.x.

    White, P.S., Pickett, S.T.A., 1985. Natural Disturbance And Patch Dynamics: An Introduc-tion. In: Pickett, S.T.A., White, P.S. (Eds.), The Ecology of Natural Disturbance andPatch Dynamics. Academic, New York, pp. 3–13.

    Whiteley, A.R., Hastings, K., Wenburg, J.K., Frissell, C.A., Martin, J.C., Allendorf, F.W., 2010.Genetic variation and effective population size in isolated populations of coastal cut-throat trout. Conserv. Genet. 11, 1929–1943.

    With, K.A., Crist, T.O., 1995. Critical thresholds in species' responses to landscape struc-ture. Ecology 76, 2446–2459.

    Wofford, J.E.B., Gresswell, R.E., Banks, M.A., 2005. Influence of barriers to movementon within-watershed genetic variation of coastal cutthroat trout. Ecol. Appl. 15,628–637.

    Wohl, E., Merritt, D., 2005. Prediction of mountain stream morphology. Water Resour.Res. 41, W08419. http://dx.doi.org/10.1029/2004WR003779.

    Wohl, E., Dwire, K., Sutfin, N., Polvi, L., Bazan, R., 2012. Mechanisms of carbon storage inmountainous headwater rivers. Nat. Commun. 3, 1263.

    Wood, J., 1996. The Geomorphological Characterization Of Digital Elevation Models Ph.D.dissertation University of Leicester, Leicester, UK.

    points disconnect hillslope and channel processes, isolating salmonid.1016/j.geomorph.2016.03.029

    http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0180http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0180http://dx.doi.org/10.1029/2011JF002324http://dx.doi.org/10.1130/G33979.1http://dx.doi.org/10.1130/0091-7613(2000)282.0.CO;2http://dx.doi.org/10.1130/0091-7613(2000)282.0.CO;2http://dx.doi.org/10.1130/0091-7613(2000)282.0.CO;2http://dx.doi.org/10.1130/0091-7613(2000)282.0.CO;2http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0200http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0200http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0205http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0205http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0210http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0210http://dx.doi.org/10.1029/93WR02463http://dx.doi.org/10.1007/s10021-009-9298-5http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0225http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0225http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0230http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0230http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0230http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0235http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0235http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0240http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0240http://dx.doi.org/10.1029/2007JF000977http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0250http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0250http://dx.doi.org/10.1098/rspb.2000.1234http://dx.doi.org/10.1098/rspb.2000.1234http://dx.doi.org/10.1525/bio.2013.63.5.9http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0265http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0265http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0265http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0265http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0270http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0270http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0270http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0280http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0280http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0280http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0285http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0285http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0285http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0290http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0290http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0295http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0295http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0295http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0300http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0305http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0305http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0310http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0310http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0310http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0315http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0315http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0315http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0320http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0320http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0320http://dx.doi.org/10.1029/2001WR001057http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0330http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0330http://dx.doi.org/10.1130/G33167.1http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0340http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0340http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0340http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0345http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0345http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0350http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0350http://dx.doi.org/10.1111/j.1752-4571.2008.00023.xhttp://dx.doi.org/10.1111/j.1752-4571.2008.00023.xhttp://refhub.elsevier.com/S0169-555X(16)30124-6/rf0360http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0360http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0360http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0365http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0365http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0370http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0370http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0375http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0375http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0375http://dx.doi.org/10.1029/2004WR003779http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0385http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0385http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0390http://refhub.elsevier.com/S0169-555X(16)30124-6/rf0390http://dx.doi.org/10.1016/j.geomorph.2016.03.029

    The waterfall paradox: How knickpoints disconnect hillslope and channel processes, isolating salmonid populations in ideal ...1. Introduction2. Methods2.1. Measuring valley width2.2. Slope-area relation2.3. Hillslope angles and shallow landslide potential2.4. Fish habitat modeling

    3. Results and discussion3.1. Valley width3.2. Channel slope3.3. Hillslope angles and shallow landslide potential3.4. Habitat quality and quantity

    4. ConclusionsAcknowledgementsReferences