A morpho-statistical classification of mountain stream reach types in southeastern Australia

A morpho-statistical classification of mountain stream reach types in southeastern Australia
of 23
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A morpho-statistical classification of mountain stream reachtypes in southeastern Australia C.J. Thompson  a, ⁎ , J. Croke  a  , R. Ogden  b , P. Wallbrink   c a  School of Physical, Environmental and Mathematical Sciences, University of New South Wales@ADFA, Canberra, Australia  b University of Canberra, ACT 2601, Australia c CSIRO Land and Water, PO Box 1666, Canberra, ACT 2601, Australia Received 26 October 2005; received in revised form 27 March 2006; accepted 29 March 2006Available online 19 May 2006 Abstract The concepts of sediment transport capacity ( Q c ) and sediment supply ( Q s ) have shown promise in broadly differentiatingmountain streams. The important role of lithology in determining reach characteristics is also noted but as yet not fully included inexisting process domain frameworks. This study uses topographic and grain size surveys undertaken over 42 channel reaches withslopes between 0.01 and 0.20 m/m to describe mountain stream morphology in southeastern Australia. Reaches with two different lithologies are surveyed to specifically address the role of lithology in resultant channel morphology. Spatial autocorrelationanalysis based on Moran's  I   is used to detect the frequency and occurrence of in-stream features, such as bars and pools. These dataare combined with the grain size data to describe eight dominant reach morphologies. Most channel morphologies display broadcharacteristics similar to previous accounts of mountain streams, although a number of intermediate morphologies affecting subtlevariations in form and process where evident. There was no significant longitudinal arrangement of the observed morphologies, particularly with respect to catchment area and slope. This is related to the strongly segmented longitudinal profiles characteristic of          these streams where increasing catchment area does not result in a commensurate decrease in slope. Five of the eight channelmorphologies identified are lithology dependent with respect to grain size and shape. Regime diagrams applied to quantify the physical controls on the different reach morphologies identified different   Q c – Q s  domains for most morphologies. However, somelithology-dependent morphologies could not be differentiated supporting the need to incorporate the effects of dominant grainshape (or sphericity) in assigning dimensionless critical shear stress values in determining  Q s . Overall, the differences in mountainstream reach morphologies described in this study reflect the characteristic segmentation of the longitudinal profiles of thesestreams and contrasts in the supply and breakdown of parent material.© 2006 Elsevier B.V. All rights reserved.  Keywords:  Mountain stream; Sediment supply; Transport capacity; Reach morphology; Spatial autocorrelation 1. Introduction Mountain streams exhibit a diversity of channel typesformed over a range of hydrological regimes andgeologies (Wohl, 2000a). Overall, such diversity has primarily been explained by variables related to therelationship between sediment transport capacity ( Q c )and sediment supply ( Q s ). Previous studies, includingthat of           Montgomery and Buffington (1997) whodeveloped a reach scale model based on this concept, propose that a particular ratio between  Q c  and  Q s  resultsin specific bedform configurations, each exhibiting a Geomorphology 81 (2006) 43 – 65www.elsevier.com/locate/geomorph ⁎  Corresponding author. Tel.: +61 2 6268 8317; fax: +61 2 62688786.  E-mail address:  chris.thompson@adfa.edu.au (C.J. Thompson).0169-555X/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.geomorph.2006.03.007  unique resistance to flow.  Q c  was defined as total shear  stress, which typically decreases downstream, asopposed to effective shear stress which generallyincreases downstream.  Q s  was defined as the absolutesediment supply which increases with catchment area.Hence, a generalized sequence of channel morphologiesfrom steep upstream areas to gentle downstream areas isexpected (e.g., Fig. 1). Any temporal change in thesecontrolling variables may affect bedform compositionand result in a change in reach type. Channel responsesto changing variables range from complete destructionof bedforms (Harvey, 1987) and wholesale change of          channel type to minor adjustment of channel character-istics (e.g., width, depth, and grain size) without changein overall channel type (Lisle and Hilton, 1999).Subsequent refinement of the  Q c – Q s  model (Buf-fington et al., 2003) introduced the concept of different supply – transport domains which can be differentiatedusing Parker's (1990) regime framework. This work indicated that different channel morphologies havecharacteristic combinations of channel features such asslope ( S  ), depth ( d  ) and median bed particle size (  D 50 ),which are believed to represent the morphologicaladjustment to imposed combinations of discharge andsediment supply. This interpretation is based on twoassumptions with respect to  Q s . (1) the channelmorphologies are competent to transport   D 50  into, andout of, the reach at bankfull discharge ( Q  bf  ), and (2) thedimensionless entrainment threshold (i.e.,  τ  c ⁎ ) is thesame for all morphologies. These assumptions areunlikely to be applicable to all mountain river channelsand morphologies. For example, flat grains have beenshown experimentally (Li and Komar, 1986) and innatural channels (Church, 1978; Laronne and Carson,1976) to create tighter bed packing by interlocking,which produces greater resistance to flow. Lithologycontrols grain shape (Browne and Thomas, 2001;Huddart, 1994) and its potential influence on bedformconfiguration and resultant reach morphology has also been noted (Buffington et al., 2003; Duckson andDuckson, 2001; Montgomery, 1999; Wohl, 2000b;Wohl and Achyuthan, 2002). Lithology may therefore be expected to be a significant factor influencing reachmorphology. However, the inclusion of lithology intoexisting classifications has been difficult, because thereis no single parameter that can adequately reflect therange of channel attributes associated with different rock types (Kondolf et al., 2003).Given our understanding of Australia's geologichistory, we seek to explore the relationship betweenmountain channel morphologies and lithology, specif-ically the influence of geology on both sediment transport and sediment supply. Lithological differencesand the resultant effects on bedrock resistance and particle size are likely to play a key role in controllingsediment supply and the resultant reach morphology. Sofar, few studies have looked at the morphological andsedimentological attributes of mountain streams inAustralia (e.g., Young et al., 2002). The questionremains, therefore, how well existing classificationsdepict and explain reach scale variations in channelmorphologies outside their region of reference?The major objectives of this study are threefold.Firstly, we test if surveyed reaches meet the generalmorphological characteristics, with respect to bedformconfiguration and bedform spacing, as defined by theexisting classification of           Montgomery and Buffington(1997). Secondly, we investigate the specific relation-ship between channel morphologies and local catchment lithology. Thirdly, we seek to explore the application of          the  Q c – Q s  model to differentiate southeast Australianmountain stream morphology types. We postulate that channel-reach morphologies may be strongly controlled by geologic history, and that lithology is a dominant factor influencing sediment supply, caliber and shape aswell as resultant sediment transport conditions. Theapproach we have adopted here is to apply an objective,statistical criterion to determine channel-reach morphol-ogies, the transition zones between morphologies, andthe thresholds of controlling variables. 2. Existing descriptions of dominant reachmorphology in mountain streams A brief summary of the existing reach morphologyfor mountain streams is presented below using the Fig. 1. A generic model of reach morphology in mountain rivers withconcave longitudinal profiles. Slope is a main factor of channelcompetence, and catchment area represents absolute sediment supply.Catchment area scales are approximate and infer an increasing areafrom cascade through to pool-riffle while hillslope connectivitydecreases.44  C.J. Thompson et al. / Geomorphology 81 (2006) 43  –  65  terminology of           Montgomery and Buffington (1997).Channel characteristics associated with particular  morphologies are outlined in Table 1. Bedrock reaches  contain 10% or less alluvial cover  and typically have steeper slopes than alluvial reaches of          the same catchment area, indicating that   Q c  exceeds  Q s (Montgomery et al., 1996). Bedrock reaches havemorphologies similar to alluvial channels with ana- branching (Tooth and McCarthy, 2004), meandering andstraight planforms with pool – riffle, planebed and step –  pool longitudinal profiles (Wohl and Merritt, 2001).Bedrock channels may also exhibit potholes andlongitudinal grooves (Wohl and Merritt, 2001). Cascades  are the steepest clast-containing channels.They are not strictly alluvial channels due to therelative immobility of large boulders (Grant et al.,1990). Their immobility results from a lack of flowcompetence to reorganize the boulders which displayrandom positioning in longitudinal profiles and planform. Sediment is typically supplied from debrisflows, which require valley slopes  ≫ 0.02 m/m for  transport (Benda, 1990), and from mass movementson adjacent hillslopes with gradients >25° (Rutherfurdet al., 1994). Step – pools  exhibit large clasts (or woody debris)that form discrete transverse clusters, resulting invertically oscillating longitudinal profiles (Chin, 1989,1999). This structure indicates the fluvial sorting of          available sediment. While a single mechanism for step –  pool formation is not universally accepted, theoriesinclude maximum flow resistance (Abrahams et al.,1995) and antidune formation (Grant, 1997) mechan- isms. Other work postulates that the stochastic positionof immobile keystone elements forces step locations(Zimmermann and Church, 2001). Nevertheless, stepspacing is generally less than, or equal to, 4 channelwidths (cw) (Grant et al., 1990; Whittaker, 1987). It isgenerally agreed that a low  Q s  and slopes of 0.03 m/m or  greater are required for the formation of step –  pools(Chartrand and Whiting, 2000). Planebeds  are generally gravel/cobble channelsdisplaying a featureless bed without marked lateral andvertical oscillations. Bed surfaces typically appear  armored indicating supply limited conditions (i.e., Q c > Q s ). However, beds may become transport limitedat flows exceeding the transport threshold (MontgomeryandBuffington,1997).Unlike step –  poolandpool – rifflemorphologies, little work has been conducted on thefluvial processes and sediment interactions required tomaintain this planar reach form. However, Montgomeryand Buffington (1997) suggest planebed channels lack sufficient lateral flow convergence required to scour   pools, due to the low width to depth ratios and greater  relative roughness which diffuse lateral flow. Pool – riffle  reaches have received the greatest attention in fluvial geomorphology and are often usedas a generic term to describe any bar   –  pool morphology(Knighton, 1998). Besides exhibiting laterally oscillat-ing form between pool and bar features, pools arefrequently associated with meander bends with a lateral bar producing an asymmetrical cross-section (Knighton,1998). A25%upstream phase shift inshear stress during bankfull flow is the most likely maintenance mechanismof pool – riffle morphology (Wilkinson et al., 2004). Other morphologies  with coarse beds are alsorecognized in Montgomery and Buffington's classifica-tion. These include forced and intermediate morpholo-gies, and are typically under-represented in mountainriver studies. Forced step –  pool and forced pool – rifflemorphologies have slope ranges extending to steeper  values, and can have shorter and more irregular pool-to- pool spacing compared to the equivalent free-formedmorphologies (Montgomery et al., 1995; Myers andSwanson, 1997). Grain size for forced pool – rifflereaches is typically finer than that for equivalent self          formed reaches (Buffington and Montgomery, 1999). Table 1Characteristics of mountain reach morphologies (after   Montgomery and Buffington, 1997)Reach attributes Bedrock Cascade Step –  pool Planebed Pool – riffleSlope (m/m) >Alluvial reach of          equal catchment area0.04 – 0.30 0.03 – 0.10 0.01 – 0.03 <0.015Typical grain size NA Boulder Cobble –  boulder Gravel – cobble GravelGrain size a   D 50  (mm) NA 80 65 50 17  D 84  (mm) NA 250 200 140 57Bar length (cw) Variable NA  ≤ 1 NA >1Pool-to-pool spacing (cw) Variable <1 1 – 4 0 5 – 7Bedform pattern Variable Random Vertical oscillations Uniform Lateral oscillations  NA=not applicable. a  Grain size percentiles are an example from Finney Creek in Washington (Montgomery and Buffington, 1997).45 C.J. Thompson et al. / Geomorphology 81 (2006) 43  –  65  Forcing elements are typically bedrock and large woodydebris (Buffington et al., 2002; Gomi et al., 2003).Intermediate morphologies contain attributes of morethan one of the dominant morphologies. Because thedominant morphologies are suggested to represent members along a continuum (Fig. 1), intermediatemorphologies may be expected to represent combinedcascade and step –  pool, or step –  pool and planebed, or   planebed and pool – riffle characteristics.The above terminology and characteristics are nowwidely used to describe the array of channel morphol-ogies found in mountain streams. As such, we do not seek to introduce a new terminology which merely ‘ splits ’  existing classifications based on small scalevariations in reach characteristics. Rather, we seek toexplore the usefulness of process domains and lithotopounits, areas with similar topography and geology withinwhich similar suites of geomorphic processes may occur  (cf. Kondolf et al., 2003), to define areas wheremountain channels may share common characteristics. 3. Study area The continental margin topography of southeasternAustralia experienced little tectonic modification duringthe Cenozoic era (Young and McDougall, 1993). Highelevation areas are restricted to the southeastern marginof the continent where the Great Dividing Range (meanelevation of            ∼ 1000 m a.s.l.) represents a remnant Palaeozoic highland belt into which fluvial systemshave slowly incised (5 to 18 m/Ma; Young andMcDougall, 1993). Based on inferences from valleyfilling basalts, Young and McDougall (1993) suggest the longitudinal profile shapes of rivers draining thehighland belt have remained relatively constant sincethe Tertiary. Longitudinal profile discontinuities werecaused by spatial changes in lithology and geologicstructure during orogenic activity between the LateOrdovician and the Late Cenozoic (Bishop et al., 1985;Van der Beek and Bishop, 2003). During this period, theregion was affected by four episodes of folding, strongcompression and uplift (O'Sullivan et al., 2000) leavinga composite post-orogenic landscape called the LachlanFold Belt within which the study area is located (Fig. 2).The study region has a temperate climate with a meanwinter temperature of 5 °C and a mean summer  temperature of 20 °C (Anon, 2004). Mean annualrainfall is 650 mm in the northwest, 1000 – 1100 mm onthe escarpment and coastal mountains, and 750 – 900 mm across the coastal lowlands (Anon, 2004).The southeastern Australian highlands subjected to LatePleistocene glaciation are not included in the study area.Periglacial activity is limited to areas above 2000 m(Barrows et al., 2001) which are also outside the regionof the present study. Snowfalls can occur down toelevations of 600 m a.s.l. during winter but typically thesnow melts within several days.The region is dominated by granitoid plutonic rocksconsisting mainly of granite – adamellite, adamellite andquartz diorite – granodiorite. Along the contact margin,outcrops of intensely folded and deformed Ordovicianmetasediment occur as north – south running bands.They consist of fine-grained, low-grade metamorphics,mainly greenish-grey phyllites and slates interbeddedwith greywacke, siltstones, micaceous sandstone andquartzite (Reinson, 1977).Mountain streams in this region are generally perennial for 3-order (Strahler) and higher, represented by single thread channels with a high degree of lateralconfinement. Most reaches contain both bedrock exposures and alluvial sediment. The depth of sediment in alluvial reaches is largely unknown, although channel bed excavations to a depth of 1 m at a number of sitesdid not intersect bedrock. 4. Methods 4.1. Site selection Potential study reaches were determined using aGeographical Information System (GIS). Reaches withslopes of           ≥ 0.01 m/m and source areas having a uniformlithology of either Ordovician metasediments or granite,were selected using a 25-m digital elevation model(DEM) and 1:250000 digital geology maps. A reach isdefined as a stream section lying between breaks inchannel slope, with relatively uniform valley gradient,width, lithology, and channel bank material over aminimum length of 10 cw (Bisson and Montgomery,1996; Frissell et al., 1986; Montgomery and Buffington,1997). It excludes tributary confluences that increasedischarge and sediment supply. A total of 42 reachesfrom 28 streams with a median gradient of 0.03 m/mwere then selected. The range in elevation andcatchment area for the selected study reaches is similar  for both the lithologies, although average hillslopegradients and drainage densities are notably different (Table 2). Surveyed reach lengths range from 50 – 230 m, corresponding to ca. 10 – 30 cw. 4.2. Field techniques The reaches were systematically surveyed for  channel bed elevation for a length of 10 cw using a 46  C.J. Thompson et al. / Geomorphology 81 (2006) 43  –  65  total station and stadia. Surface water height was not included because flow was predominantly well belowresidual pool volume due to prevailing drought condi-tions. Survey points were recorded at regular intervalsmarked on a tape along the thalweg of the channel bed.Channel units (homogenous morphological featuresdisplaying similar grain size, flow depth, velocity andgradient; such as a pool) in free-formed, low gradient reaches have an average spacing of 5 – 7 cw (Leopold et al., 1964) while high gradient channel units may havespacing intervals less than 4 cw (Whittaker, 1987).Interval length was set so that each channel unit  Table 2Mean catchment characteristics by lithologyCatchment lithology Catchment area (km 2 ) Mean catchment hillslope (deg) Max. elevation (m a.s.l.) Drainage density (km/km 2 )Granite ( n =22) 19.0 (3.0 – 98.0) 11.0 (4 – 21) 995 (770 – 1460) 2.9 (2.1 – 3.7)Metasediment ( n =20) 14.8 (2.0 – 57.5) 18.5 (6.6 – 26.4) 973 (363 – 1430) 3.8 (2.9 – 5.3)Probability 0.05 0.001 0.05 0.001Significantly different   ×  ✓  ×  ✓ Ranges are in parentheses. Significant differences were determined by  t  -tests applied to log-transformed data.Fig. 2. Map of study region in Australia showing study sites, major rivers and elevation. The Lachlan Fold Belt is shown in grey in the country map at the top.47 C.J. Thompson et al. / Geomorphology 81 (2006) 43  –  65
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks