- Email: [email protected]
Catchment-scale (dis)connectivity in sediment flux in the upper Hunter catchment, New South Wales, Australia
Geomorphology 84 (2007) 297 – 316 www.elsevier.com/locate/geomorph
Catchment-scale (dis)connectivity in sediment flux in the upper Hunter catchment, New South Wales, Australia Kirstie A. Fryirs a,⁎, Gary J. Brierley b , Nicholas J. Preston c , John Spencer a a Department of Physical Geography, Macquarie University, North Ryde, NSW 2109, Australia School of Geography and Environmental Science, University of Auckland, Box 92019, New Zealand Institute of Geography, School of Earth Sciences, Victoria University of Wellington, Box 600, Wellington, New Zealand b
c
Received 10 May 2005; received in revised form 3 January 2006; accepted 3 January 2006 Available online 24 July 2006
Abstract (Dis)connectivity within and between landscape compartments affects the extent and rate of transfer of energy and matter through catchments. Various landforms may impede sediment conveyance in a river system, whether laterally to the channel (termed buffers) or longitudinally along the channel itself (termed barriers). A generic approach to analysis of landscape (dis) connectivity using slope threshold analysis in GIS, tied to air photograph interpretation and field mapping of buffers and barriers, is tested in the upper Hunter catchment, Australia. Under simulated conditions, effective catchment area, which is a measure of the proportion of a catchment that has the potential to contribute sediment to the channel network, varies from 73% to just 3% of the total catchment area for differing subcatchments in the upper Hunter. This variability can be explained by the spatial distribution and assemblage of buffers and barriers in each subcatchment. Multiple forms of disconnectivity are evident in some subcatchments, such that when one buffer or barrier is breached, other features still impede sediment transfer within the system. The importance of the position of buffers and barriers within any given subcatchment is emphasised. Spatial variability in valley width exerts a critical control on catchment connectivity, with more efficient sediment conveyance in narrow valleys relative to wider valleys characterised by piedmonts, terraces, fans and extensive floodplains in which conveyance is impeded. This variability reflects the landscape history and geological setting of each subcatchment. The framework developed in this paper can be used to assess the impact of natural or human-induced buffers and barriers on catchment-scale sediment flux in any landscape setting, providing a physical template atop which other biogeochemical fluxes could be examined. © 2006 Elsevier B.V. All rights reserved. Keywords: Disconnectivity; Coupling; Sediment flux; Buffers; Barriers; Landscape history; GIS; Hunter catchment; Australia
1. Introduction Sediment flux within river systems has been described as a ‘jerky conveyor belt’ (Schumm, 1977; Kondolf, 1994). While this conceptualisation has been framed in ⁎ Corresponding author. Tel.: +61 2 9850 8367; fax: +61 2 9850 8420. E-mail address: [email protected] (K.A. Fryirs). 0169-555X/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.geomorph.2006.01.044
terms of variability in flow energy which determines the capacity of the system to transport sediment along the channel network, due regard must also be given to the availability and delivery of sediment into the channel network, as this drives the propagation of disturbance responses through a catchment (Brierley and Fryirs, 2005). These relationships vary in different landscape settings, reflecting the perennial/ephemeral nature of the flow regime, controls on sediment availability (e.g.
298
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
Fig. 1. Location of the upper Hunter catchment within the Hunter catchment.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
landscape history, relief and lithology) and the configuration of any given catchment. Efforts to model sediment flux in catchments must integrate these controls on energy and sediment availability, ensuring that outputs are adequately grounded by field mapping of impediments to sediment conveyance. The distribution of riverine sediment stores and sinks, and the frequency with which sediment is added or removed from compartments, reflects the degree to which a river system is (dis)connected or (de)coupled (Caine and
299
Swanson, 1989; Lang and Hönscheidt, 1999; Harvey, 2002; Hooke, 2003; Schrott et al., 2003). Any factor that impedes sediment conveyance constrains sediment delivery to the catchment outlet. Fryirs et al. (in press) differentiate between landforms that impede sediment movement into the channel network (e.g. lateral disconnectivity from slopes to channels) and features that impede sediment movement along the channel (i.e. longitudinal disconnectivity), referring to these features as buffers and barriers respectively. The nature and
Fig. 2. Methods used to model the effective catchment area and the role of buffers in controlling (dis)connectivity in the upper Hunter catchment.
300
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
extent of (dis)connectivity vary across any given catchment, and over time, as influenced by the form, composition, scale and position of buffers and barriers, and the ease with which these features are reworked and breached. Impediments to sediment movement within a catchment restrict the rate of sediment transfer from the area upstream of that point. Hence, these impediments determine the area or proportion of a catchment that has the potential to directly contribute sediment to, or transport sediment along, the channel network under given flow conditions. This is referred to here as the effective catchment area (Fryirs et al., in press). Landscape (dis)connectivity, in turn, determines how geomorphic instability is manifest throughout a river system, shaping the direction and rate of geomorphic change, and lagged and off-site (propagatory) responses (Goodbred and Kuehl, 1999; Metivier and Gaudemer, 1999; Sidorchuk, 2003). Numerous models have been used to examine hydrological or sedimentary (dis)connectivity and its effect on contributing areas (e.g. Michaelides and Wainwright, 2002; Kirkby et al., 2002). Other researchers have analysed landscape connectivity using conceptual frameworks tied to field-based applications (e.g. Harvey, 1992; Fryirs and Brierley, 1999, 2001; Hooke, 2003). In few instances, however, have these principles been merged through analyses of landscape (dis)connectivity at the catchment-scale. In this paper, we appraise the role of buffers and barriers as a control on bedload sediment flux and channel network connectivity in the upper Hunter catchment, NSW, Australia. Rather than attempting to quantify sediment delivery to the channel within a sediment budget approach, we use a simulated set of conditions, tied to catchment-scale air photograph and field analyses, to examine how buffers and barriers restrict bedload supply across this 4300 km2 catchment. 2. Regional setting The upper Hunter catchment has a shorter north-south (∼ 60 km) than east-west (∼ 80 km) axis (Fig. 1). The Hunter–Mooki fault, which defines the boundary between the Sydney Basin and the New England Fold Belt, divides the catchment into two distinct zones. The eastern side generally comprises Carboniferous and Devonian metasediments, with a Tertiary basalt cap at high elevations. The metasediments comprise marine and terrestrial sediments associated with a Devonian volcanic arc (e.g. limestone, chert, lithic sandstone, conglomerates, etc.) that have been folded, faulted, and uplifted during the Permian–Triassic orogenic episode that formed the New England Fold Belt. The western side of the Hunter–
Mooki fault consists of near-horizontal Permian and Triassic sedimentary rocks (primarily shales, sandstones, conglomerates and coal), with relatively small areas of dissected Tertiary basalt. The four landscape units of the upper Hunter roughly align with the geological characteristics of the catchment (Fig. 1). The Remnant Plateau landscape unit that encompasses the Tertiary basalt plateau has a relatively high elevation (often >1000 m asl) and is deeply dissected. The Plateau Slopes landscape unit, which occurs at the margin of plateau remnants and high elevation ridges, comprises steep slopes and confined valleys. The Rugged and Hilly landscape unit in the central part of the catchment is characterized by partly-confined valleys. The Undulating Plains landscape unit, which occurs primarily to the west of the Hunter–Mooki fault, comprises low slopes and laterally-unconfined valleys. Average annual rainfall across the upper catchment ranges from 1400 mm/yr in the Barrington Tops, to 550 mm/yr west of Murrurundi. The largest flood since European settlement in the 1820s occurred in 1870. The largest recorded flood in 1955 had a peak discharge at Muswellbrook of more than 42 000 m3 s− 1. The discharge for a 1 in 2 year flow at Muswellbrook is 379 m3 s− 1. 3. Methods 3.1. GIS modelling of effective catchment area GIS modelling of the effective catchment area is based on the simple assumption that slope angle (gradient) controls delivery of sediments to the channel. Clearly, there are many factors that influence off-slope sediment delivery. Some of these relate to the nature of the land surface itself, and include, for example, slope angle, surface roughness and infiltration capacity. Others relate to the nature of runoff, e.g. its depth, volume and duration. Together, these factors determine the amount of energy available for sediment erosion and transport. Further, there are issues of spatial scale and the topological relationships between different parts of the land surface that have a strong influence on sediment delivery. It is these that the focus on (dis)connectivity addresses. We focus on slope angle to illustrate the nature of changing connectivity for a number of reasons. First, along with discharge, slope angle is the principal determinant of stream power and the energy available for sediment transport. Other factors associated with the land surface play secondary roles, moderating the amount of runoff or its velocity. Second, in contrast to many of the other controls on runoff and sediment transport, slope angle is constant at the temporal scale under consideration and can be readily
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
301
Table 1 Definition and functionality of buffers and barriers in the upper Hunter catchment Feature Buffers Piedmont
Alluvial fan
Definition and position
How identified/mapped
Impeding/blocking effect
Gentle sloping ground lying between the steep slope of the valley margin and floodplain. Often have floodplains or fans inset against them. Sediment lobe that radiates from a small tributary network (feeder channel) where the stream emerges from confinement onto a valley floor. Often have floodplains inset against them.
Distinctly graded surface running along the valley margin. Often grades into floodplain.
Blocks sediment derived from adjacent hillslopes as sediments cannot be transported across the surface.
Terrace
Elevated palaeo floodplain surface that is rarely inundated. Occur along the valley margin and often have floodplains inset against them.
Continuous floodplain
Flat surface adjacent to the channel margin in a laterally-unconfined valley. May be inset against bedrock valley margins, terraces or piedmonts in different settings.
Discontinuous floodplain
Flat surface adjacent to the channel margin in a partly-confined valley. May be inset against bedrock valley margins or terraces.
Trapped Tributary system trapped behind a terrace tributary fill or floodplain surface.
Barriers Sediment slug
Distinctly conical feature located between bedrock spurs. Flow paths evident on the surface.
Not currently active. Impedes sediment conveyance from hillslopes and feeder tributaries to the channel, as sediments cannot be transported across the surface. Can only be reworked if incision extends through them from the trunk stream. Flat-topped, stepped features inset Not currently active. Impedes sediment against a bedrock valley margin conveyance from hillslopes and tributaries but elevated above the to the channel. Elevated surface means contemporary floodplain surface. contemporary flows cannot breach their surfaces. Can only erode via channel expansion processes. Consolidated sediments require high energy conditions for this to occur. Surface immediately adjacent to the Impedes sediment conveyance from channel margin that contains hillslopes to the channel. Disrupts floodplain geomorphic units such longitudinal conveyance of material as levees and palaeochannels. by removing it from channel conveyance. Surfaces currently active but require channel expansion for them to be breached. Impedes sediment conveyance from Surface immediately adjacent to hillslopes to the channel. the channel margin that contains Disrupts longitudinal conveyance of floodplain geomorphic units such material by removing it from channel as floodchannels. conveyance. Surfaces currently active but require channel expansion or stripping to be breached. Trapped behind other forms of Intact valley fill surface with no buffer and disconnect tributary well-defined channel that flows networks from the trunk stream valley floor. into the back of a terrace or floodplain along a tributary valley.
Accumulation of sediment that covers the entire channel bed and infills the channel.
As for description for morphology and sedimentology
Competencelimited channel
Small, sinuous (often tortuous) channel found on a very low slope.
As for description for morphology and sedimentology
Dam
Concrete structure that extends across the valley behind which a lake is formed.
As for description for morphology and sedimentology
Temporarily impedes upstream–downstream sediment conveyance along the channel. The volume and frequency with which sediment moves depends on the magnitude of the flow event, and the energy available to transport sediments. Tend to be transportlimited settings. Impedes upstream–downstream sediment conveyance along the channel. Channels tend to be competence limited, meaning that any bedload materials that are made available to these channels cannot be transported over the low slopes or along a sinuous channel with very low energy conditions. Total blocking of most flow and sediment transported down the valley. All bedload materials are trapped behind the dam wall.
302
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
Fig. 3. Scheme used to classify tributary-trunk stream (dis)connectivity in the upper Hunter catchment. Types 1, 2 and 3 are forms of tributary-trunk stream disconnectivity, while Types 4, 5 and 6 are forms of tributary-trunk stream connectivity.
assessed on a spatially distributed basis. Factors such as roughness or infiltration capacity are less readily mapped due to their spatial variability, while the temporal variability of runoff creates difficulties in making generalisations. Thus, slope angle is used to illustrate the variability of connectivity. Below a given slope angle, energy for transport is insufficient, creating an obstacle to connectivity. As such, the focus is principally on lateral (dis)connectivity. Adopted procedures are summarised in Fig. 2. Threshold values of slope angle were used to perform this assessment, with three critical slope angles chosen to illustrate a range of energy conditions under which buffers may be breached. Slope thresholds of 0.5°, 2° and 25° were arbitrarily chosen to gain some indicative sense of how the model would perform at these extremes. Because slope interacts with more dynamic factors (rainfall intensity, runoff velocity) in determining sediment delivery, looking at different slope angles allows generalisation about differing energy conditions. Using a high slope angle threshold, for example, is analogous to considering low intensity rainfall or a surface with high roughness and high infiltration capacity. Initially this may seem counter-intuitive in the sense that high slope angle implies high energy. In effect, however, this represents the situation where other factors are such that only high slope angle will promote sediment transport. Conversely, the use of a low slope threshold represents conditions in
which there is so much energy (or little resistance) that sediment transport meets with little obstacle. The input data for the modelling were a 25 m DEM of the study region and a data set containing streamlines digitised from 1:25,000 topographic maps (sourced from the NSW Department of Natural Resources). GIS analysis was undertaken using ArcView 8.3 and the ArcHydro extension from the Centre for Research in Water Resources at the University of Texas (ArcHydro, 2005). The first step, after filling anomalous sinks on the DEM surface, was to derive a set of streamlines. This was done using flow direction, flow accumulation and stream definition routines within ArcHydro. A channel initiation threshold catchment area of 2 km2 was chosen after several iterations suggested that this best reflected the distribution of the digitised channel network. The automated derivation of streamlines is often limited in areas of low slope, and the digitised streamline data set was used to enforce the channel position within the DEM. ArcHydro was also used to define subcatchments. Five subcatchments, incorporating the major stream networks in the upper Hunter catchment, were modelled as separate entities. These subcatchments are Dart (which includes Dart Brook, Middle Brook and Kingdon Ponds), the Upper Hunter (which includes the trunk stream and all its minor tributaries upstream of Glenbawn Dam), Rouchel (which encompasses Rouchel Brook and Davis Creek), the Pages and the Isis. In addition, the upper Hunter
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
303
Table 2 Effective catchment area of each major subcatchment in the upper Hunter catchment Subcatchment
Total catchment area (km2)
Effective catchment area (km2)
% of subcatchment connected (2° slope simulation)
% of catchment disconnected (2° slope simulation)
Degree of connectivity
Rouchel Pages Upper Hunter Isis Dart Lower Hunter Total catchment (due to buffers only) Total catchment (when barriers added)
435 642 1294 551 806 567 4295
317 348 849 258 256 17 2045
73 54 67 47 32 3 48
27 46 33 53 68 97 52
Highly connected Connected Connected Disconnected Highly disconnected Highly disconnected Disconnected
4295
334
8
92
Highly disconnected
catchment, including the trunk stream extending from Glenbawn Dam to Muswellbrook and all the major tributary subcatchments (i.e. Rouchel, Isis, Upper Hunter, Pages and Dart), was modelled as a single entity. Each modelling unit was then treated as follows. A 50 m buffer was placed around the streamlines for each subcatchment, and the outline of this buffer (polyline feature) was converted into a raster of 25 × 25 m cells. Each of the cells representing this rasterised stream edge was then converted to a point within a point feature layer. This creates a data set containing a point every 25 m on both sides of the channel. The catchment of each of these points were then derived using flow direction and flow accumulation routines within ArcHydro. In this way, the entire surface area of each subcatchment was divided into a series of smaller drainage units, with every part of the subcatchment draining into one of the points along the streamline. A slope raster layer was created from the DEM. Threshold slope layers were derived by applying a conditional statement that returns No Data values to cells with slope angles below the chosen threshold. These threshold slope layers effectively remove areas below the chosen slope threshold from further processing when the ArcHydro hydrological modelling is repeated using a revised DEM from which the cells with slopes less than the threshold value have been removed. The catchments of all cells adjacent to the channel were redefined using modified flow direction layers. Cells upslope of low angle areas are thus not connected to the channel; the effective catchment area comprises only those cells that have both angles greater than the threshold, and are connected to the channel through other cells with slopes greater than the threshold. The result is a layer that describes the catchment area connected to a channel by surfaces that have a slope greater than the threshold slope. This approach is considered to be indicative rather than absolute, allowing
across-catchment comparisons of patterns of (dis)connectivity to be examined. The output shows where there is ‘potential’ for sediment supply. It does not quantify the volumes of sediment moving or determine if sediment is available to be moved. Further work is required to tie this indicative appraisal to the discharge record to model the types of flows under which various simulations occur, and to test the sensitivity of model output to various slope threshold analyses. Of three values tested, the across-catchment variability in effective catchment area was most pronounced using a threshold slope angle value of 2°. This value reflects energy conditions that are capable of moving materials only on slopes over 2° (but may have the capability to rework sediment along channels and immediately adjacent sediment storage zones). This threshold value is thought to represent a rainfall event of some unknown but relatively high magnitude. The output from thresholds of 0.5° and 25° showed, respectively, almost entire catchment connectivity with only the very lowest slopes (<0.5°) representing an impediment to slope-channel connectivity, or no lateral connectivity (i.e. sediment transport would be limited to the channel zone). Because the 2° simulation is thought to represent conditions under which off-slope sediment transport is likely to be promoted, it was chosen as the basis for further analysis. By choosing this simulation the results only represent a single snapshot of the spatial patterns of (dis)connectivity in this catchment. 3.2. Identification and analysis of buffers and barriers, and tributary-trunk stream (dis)connectivity Air photograph interpretation was tied to field investigations and ground truthing to map the distribution of buffers and barriers in the catchment (Fig. 2). A complete aerial photograph coverage of the upper Hunter
304
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
catchment consists of 1998, 1999, 2000, and 2001 photo sets with an approximate scale of 1:25,000. The photos containing areas of interest were ortho-rectified using ERDAS Imagine 8.4. Digital versions of the appropriate CMA 1:25,000 topographic maps were used as the source of reference co-ordinates for the ortho-rectification. Buffers were identified using a stereoscope and manually digitised onto the aerial photographs. This procedure was completed along all major stream networks in the upper Hunter catchment to a point where the size of the sediment stores was beyond the resolution of the aerial photos. The digitised maps were checked in the field to ensure that features had been mapped and categorised correctly. Major barriers to sediment movement along channels were identified in the field. Six types of buffer (piedmont,
alluvial fan, terrace, continuous floodplain, discontinuous floodplain, trapped tributary fill), and three types of barrier (sediment slug, competence-limited channel, dam) have been identified in the upper Hunter catchment (Table 1). Differing patterns of buffers result in six forms of tributary-trunk stream (dis)connectivity in the upper Hunter catchment (Fig. 3). Types 1–3 are forms of tributarytrunk stream disconnectivity that are characterised by either discontinuous channels in the tributary system or the occurrence of a buffer along the trunk stream valley margin that severs the connection between the tributary channel and the trunk stream. Types 4–6 (Fig. 3) are characterised by tributaries that directly connect to the trunk stream, associated either with the lack of an impeding buffer on the valley floor (type 6), or situations
Fig. 4. Effective catchment area of subcatchments in the upper Hunter. The effective catchment area is noted as a percentage of the area that is disconnected due to buffers and barriers.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
where buffers fail to impede the connectivity between the tributary and the trunk stream (types 4 and 5; see Fig. 3). Impediments to sediment conveyance from tributary systems to the trunk stream may take the form of a single buffer (e.g. a discontinuous floodplain pocket) or multiple buffers where 2 or more features disconnect the tributary from the trunk stream (e.g. a fan or a continuous floodplain). In addition, barriers may disconnect linkages along the channel network. These ‘types’ of (dis)connectivity were categorised using the aerial photo mosaics and the digitised buffer map. These categories were added as attributes to the digitised streamline data set. In ArcView, this data set was then used to calculate subcatchment areas and measure stream lengths at each tributary-trunk stream confluence in the catchment.
305
Overlaying of these data sets was then used to assess the degree to which the distribution of buffers and barriers could explain the effective catchment area simulation, and thus whether different parts of the catchment were connected or disconnected. 4. Results 4.1. Effective catchment area of the upper Hunter catchment and its subcatchments Table 2 shows the results of the 2° slope threshold simulation in the upper Hunter catchment, and Fig. 4 shows the effective catchment area of the upper Hunter catchment draining to Muswellbrook under these
Fig. 5. Catchment area (a) and stream length (b) relationships for connected and disconnected tributaries in the upper Hunter catchment. These have been divided according to whether they occur east or west of the Hunter–Mooki fault. Note that all disconnected tributaries drain catchment areas less than 22 km2 and are less than 8 km in length.
306
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
conditions. In this simulation, the effective catchment area is 2045 km2 from a total of 4295 km2, indicating that around 52% of the catchment is disconnected and does not contribute to the overall sediment cascade as measured at Muswellbrook. This calculation reflects the role of buffers that determine lateral connectivity (i.e. sediment supply from hillslopes and the valley floor zones adjacent to the channel). It does not take account
of longitudinal disconnectivity induced by barriers (i.e. whether blockages along the channel network further disconnect areas of the catchment). There is significant variability in the degree of (dis) connectivity in different subcatchments of the upper Hunter (Table 2). Using four arbitrary classes of (dis) connectivity (0–30% = highly connected, 30–50% = connected, 50–70% = disconnected, 70–100% = highly
Fig. 6. Assemblages of tributary-trunk stream types in each subcatchment of the upper Hunter. These are presented as (a) percentages of the types of (dis)connectivity (see Fig. 3), and (b) percentages of the subcatchment total. C = connected tributaries, D = disconnected tributaries, primary = one buffer present.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
disconnected), Rouchel subcatchment is the only subcatchment that is highly connected (27%). The Upper Hunter and Pages subcatchments are classed as connected (33% and 46% respectively). The Isis subcatchment is considered to be disconnected (53%). The Dart subcatchment and the Lower Hunter subcatchment are highly disconnected (68% and 97% respectively). These results suggest that available sediments are likely to be efficiently contributed to, and conveyed through, Rouchel subcatchment while the Dart and Lower Hunter subcatchments are extremely inefficient at contributing and conveying sediment. To explain the subcatchment variability in the degree of disconnectivity, the types and distribution of impediments have been examined across the catchment.
307
4.2. Catchment configuration and types of tributarytrunk stream (dis)connectivity Analysis of tributary subcatchment area and maximum stream length along tributaries in the upper Hunter catchment indicates that all tributaries with catchment areas >22 km2 or stream lengths >8 km are connected to the trunk stream (Fig. 5). This suggests that these tributary systems generate sufficient discharge and energy to breach any buffers that may be present, thereby maintaining their connectivity to the trunk stream. However, tributaries with catchment areas <22 km2 or stream lengths <8 km can be either connected or disconnected as determined by the presence and type of buffer along the trunk stream valley
Fig. 7. General trends in tributary-trunk stream (dis)connectivity in the upper Hunter catchment. Note the variability in the dominant forms of (dis) connectivity in different subcatchments and the dominance of various types of buffers. Black lines divide the catchment according to these dominant types of (dis)connectivity.
308 K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
Fig. 8. (a) Effective catchment area and (b) the distribution of buffers in Dart subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
margin. All but four of the small, disconnected tributaries are impeded by extensive features such as piedmonts, continuous floodplains or fans. In 41% of these cases, secondary blockages occur (i.e. 2 or more buffers are present; Fig. 6a). Seemingly, the position of the trunk stream channel on the valley floor influences whether subcatchments with small areas (and hence small discharges) have the potential to contribute bedload sediments to the trunk stream confluence, as the subcatchment may not have the capacity to form continuous channels that can breach buffer(s). There is no significant relation between
309
the type of (dis)connectivity and the catchment area or length of the tributary stream for subcatchments found on different sides of the Hunter–Mooki fault (Fig. 5). 4.3. The types and patterns of buffers and barriers across the upper Hunter catchment Significant variability in the forms of tributary-trunk stream (dis)connectivity and the associated type and assemblage of buffers is evident for differing subcatchments of the upper Hunter catchment (Fig. 7). A distinct
Fig. 9. (a) Effective catchment area and (b) the distribution of buffers in Rouchel subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
310 K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316 Fig. 10. (a) Effective catchment area and (b) the distribution of buffers in the Upper Hunter subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
pattern emerges for subcatchments east and west of the Hunter–Mooki fault. Tributary subcatchments to the east have less accommodation space in which to store sediments. The types of buffers tend to be discontinuous pockets of floodplain or terraces that are sheltered behind bedrock spurs in bedrock-controlled valleys. In contrast, valleys west of the Hunter–Mooki fault have greater accommodation space in the form of broad, open valleys
311
in which a range of buffers have formed, including piedmont zones, terraces, continuous floodplains and alluvial fans. Dart subcatchment, located West of the Hunter–Mooki fault, is one of the most disconnected subcatchments in the upper Hunter (Fig. 8). This subcatchment contains the highest prevalence of types 1–3 tributary-trunk stream disconnectivity (Fig. 6a). In almost half (47%) of these
Fig. 11. (a) Effective catchment area and (b) the distribution of buffers in Isis subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
312
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
cases, secondary impediments are evident (Fig. 6b). Most tributaries either have discontinuous channels that do not reach the trunk stream valley or they flow into broad and open valleys that are characterised by a range of buffers. The valleys store vast volumes of alluvial sediment in the form of continuous floodplains. Adjacent to the floodplains, piedmont zones extend along the western margin of Dart Brook and along the eastern side of Kingdon Ponds. These piedmont zones act as significant buffers to materials that may be eroded from the surrounding hillslopes. Even if overland flow has sufficient energy to move materials over this surface, the alluvial floodplains act as a secondary buffer to conveyance to the channel. The only materials entering the main channel networks are transported along gully networks that have incised through the floodplain into the piedmont zones. In reaches
further up catchment, alluvial fans also occur between bedrock spurs or in depressions on the piedmont. Discontinuous floodplain pockets are often inset against these fan deposits. In addition to these features, upper Kingdon Ponds contains extensive flights of terraces that extend up to 10 m above the channel bed. These terraces often trap tributary valleys behind them, forming trapped tributary fills. Discontinuous floodplain pockets are often inset against these terraces. The assemblage and distribution of buffers in Dart subcatchment produces a high degree of disconnectivity in this system, to the extent that if one form is breached a secondary feature acts as another impediment further along the sediment cascade. Topography on this side of the Hunter–Mooki fault is subdued, so that channels are very low slope, highly sinuous, competence limited and inefficient at transporting
Fig. 12. (a) Effective catchment area and (b) the distribution of buffers in the Pages subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
bedload materials. This form of barrier induces further disconnectivity in this system. Rouchel Brook, located east of the Hunter–Mooki fault, is the most connected subcatchment in the upper Hunter (Fig. 9). Relatively confined valleys along the length of Rouchel Brook and Davis Creek restrict the formation of buffers to occasional terrace pockets and discontinuous floodplains. These are typically located on the insides of sinuous valley meander bends or sheltered behind bedrock spurs (Fig. 9). Tributary-trunk stream (dis)connectivity is dominated by type 6 confluences (Fig. 6b). This topographic setting, along with the lack of any significant barriers along the channel network, ensures that any sediment that is made available to the channel can be readily conveyed to the Hunter River. Streams in the Upper Hunter subcatchment have a similar assemblage of buffers to that found in Rouchel subcatchment (Fig. 10). Tributary-trunk stream (dis) connectivity is dominated by types 5 and 6 (Fig. 6b). The valleys in this subcatchment are relatively confined and sinuous with little accommodation space. Buffer formation is restricted to terrace pockets and discontinuous floodplains that combine to produce secondary block-
313
ages. Glenbawn Dam, located at the outlet of the Upper Hunter subcatchment, traps up to 98% of bedload sediment contributed from the upper catchment (Erskine, 1985), effectively disconnecting 30% of the upper Hunter catchment from the contemporary sediment cascade. Tributary-trunk stream (dis)connectivity in the Isis subcatchment is dominated by types 5 and 6 (Fig. 11). The assemblage of buffers in this catchment is dominated by terraces and discontinuous floodplain pockets. Around half of all disconnectivity is induced by secondary blockages (Fig. 6b). The terraces in the Isis subcatchment often trap tributaries inducing significant disconnectivity along this stream. Nevertheless, any sediments that do reach the channel network are effectively conveyed to the Pages River through the bedrock-controlled valleys that dominate this system. The Pages subcatchment can be considered as two subsystems that operate upstream and downstream of the mid-catchment gorge and west-east of the Hunter–Mooki fault (Fig. 12). The upper Pages subcatchment has similar characteristics to upper Dart subcatchment, flowing through relatively broad-open valleys. Alluvial fans and
Fig. 13. Landscape configuration as a control on the forms of (dis)connectivity, and the role of various types of buffers in two contrasting subcatchments of the upper Hunter catchment, Dart subcatchment and Rouchel subcatchment.
314
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
discontinuous floodplains with occasional terraces dominate this section of the catchment (Fig. 12). Given the larger catchment areas that drain into these tributary systems, the alluvial fans have been breached by continuous channels, resulting in type 4 connectivity (Fig. 6a). If materials are transported into the gorge they are effectively flushed to the lower Pages River. Downstream of the gorge, the Pages River flows through a partly-confined valley. As a result, discontinuous floodplain pockets act as localised buffers to lateral connectivity. However, as the valley begins to widen along the lower Pages River, the buffering effect caused by continuous floodplains, which extend to the Hunter trunk stream, increases. Within this laterally-unconfined valley, a high width–depth ratio channel has formed and a large sediment slug has accumulated. This channel/slug has a relatively low slope and is acting as a temporary barrier to sediment conveyance. The slug extends for over 10 km and is several metres thick. Bedload sediment originating from upstream becomes trapped in this slug, significantly reducing the volume of bedload material that reaches the Hunter trunk stream from the Pages (and Isis) subcatchments. Except during flows large enough to source sediment from the slug and transport it downstream, the Pages and Isis subcatchments are effectively disconnected from the Hunter trunk stream. Cumulative responses to sediment contribution and conveyance along streams of the upper Hunter catchment are manifest within the Lower Hunter subcatchment around Muswellbrook (Figs. 4 and 9). While this section of the river is itself significantly buffered in terms of its lateral connectivity, owing to the extensive nature of continuous floodplains along both channel banks, there are no significant barriers to sediment conveyance along the channel itself. When the combined effect of buffers and barriers is examined, the effective catchment area of the entire upper Hunter catchment at Muswellbrook drops significantly to 334 km2 or 9% of the entire catchment area for this simulation (Table 2). 5. Discussion and implications Buffers and barriers are agents of disconnectivity in catchments. The means by which they bring about disconnectivity varies depending on their morphology, position in the landscape and juxtaposition relative to other landforms, i.e. whether primary or secondary blockages occur (Table 1). Differing patterns of these features form in particular valley settings. For example, piedmonts, alluvial fans and continuous floodplains that form in wide, open valleys with extensive accommodation space
account for a significant percentage of the disconnected catchment area of the upper Hunter, buffering slopes from channels and tributaries from trunk streams (Harvey, 2002). Along disconnected tributaries in the upper Hunter catchment, trapped tributary fill, continuous floodplains, piedmonts and terraces account for 24%, 22%, 21% and 17% of the total disconnected catchment area respectively. Fans and discontinuous floodplains each buffer 8% of the total disconnected catchment area. Discontinuous floodplains and terraces are the most commonly breached buffers in the upper Hunter catchment. Around 32% and 18%, respectively, of the total connected tributary catchment area is connected through these features. Direct connection of a tributary and trunk stream, where no buffer is present, accounts for 32% of the total connected tributary catchment area. Discontinuous floodplains or isolated terraces tend to occur in partly-confined valleys where sediment supplied from upstream is deposited behind bedrock spurs. Under these conditions, slope-channel and tributary-trunk stream linkages are only locally disconnected. Unlike more extensive landforms that are not readily breached (less than 8% each of the total connected tributary catchment area is connected through piedmont, fan, trapped tributary fill and continuous floodplains), these localised buffers can be readily breached, resulting in connectivity. Other than dams, barriers temporarily disrupt sediment conveyance along the channel network, disconnecting upstream– downstream linkages (Lane and Richards, 1997; Hooke, 2003). Catchment (dis)connectivity can be considered within a nested hierarchy at local, zonal and system scales (Harvey, 2002; Fryirs et al., in press). (Dis)connectivity within-compartments or within-landforms (e.g. on a single hillslope) that occurs at the local-scale was not examined in this study (c.f. Lang and Hönscheidt, 1999). Connectivity between landscape compartments is the first-order determinant of landscape behaviour examined. For example, slopes and channels can be disconnected where floodplains or terraces buffer the base of a slope. At this zonal scale (Harvey, 2002), the degree and effectiveness of an impediment is related to the size and sedimentary composition of the landform that acts as a buffer. At the broader system scale, (dis) connectivity relates to the behaviour of whole subcatchments or catchments. Tributary-trunk stream connectivity and the presence/character of barriers affect the operation of the sediment cascade at this scale. The position of impediments is a key determinant of effective catchment area, dictating how responses to disturbance and instability are likely to be manifest through a catchment. These relationships vary over time,
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
reflecting, among many factors, the ease and frequency with which impediments are breached and reform. Hence, the geomorphic effectiveness of any given flow is influenced by the position and relative scale of impediments within the system. For example, if an impediment (e.g. a fan) forms in a headwater subcatchment, it will have a limited effect on the overall conveyance of sediment in the larger catchment, thus having little effect on the effective catchment area. However, if that same sort of impediment occurs along the trunk stream or close to the outlet of the catchment, the impact is significantly greater. Hence, the impact of a given type of impediment may vary, dependent upon its scale and position in the catchment (Beven and Wood, 1983; Brierley and Fryirs, 1999; Dalla Fontana and Marchi, 2003; Schrott et al., 2003). These relationships reflect catchment configuration and landscape history (i.e. catchment shape, patterns of valley width, tributary-trunk stream relationships), as determined by geological controls. Landscape history and geological controls on valley configuration (morphology, accommodation space and slope) constrain the pattern of sediment storage in any catchment, and hence the type and distribution of impediments to sediment conveyance. To demonstrate this in the upper Hunter, landscape configuration along Dart Brook and Rouchel Brook, located west and east of the Hunter–Mooki fault respectively, are expressed in terms of Land Systems Units (Story et al., 1963) in Fig. 13. The differing types and patterns of buffers that have formed in these two settings affect the degree of lateral and longitudinal (dis)connectivity. In Dart subcatchment, buffers such as piedmonts, fans, continuous floodplains and terraces induce disconnectivity and hence effective catchment area is low. The Dart subcatchment drains relatively erodible geologies, is relatively low relief, and has wide, open valleys in which multiple buffers have formed. Extensive disconnectivity results, and effective catchment area is low. In contrast, Rouchel subcatchment drains relatively resistant geologies, has relatively high relief and is dominated by bedrock-controlled valleys. Buffers are restricted to isolated pockets of floodplain or terraces that are preserved behind bedrock spurs in a partly-confined valley. Only localised lateral disconnectivity results, and the effective catchment area is high. The frequency with which different types of buffer and barrier are breached is analogous to switching on or off certain parts of the sedimentary cascade over different timeframes (see Fryirs et al., in press). To appraise these relationships, the evolution and age structure of buffers and barriers in a catchment must be assessed, interpreting how long these features have acted as impediments and how connectivity has
315
changed over time (e.g. in response to direct or indirect human-disturbance; Lang and Hönscheidt, 1999; Dearing and Jones, 2003; Schrott et al., 2003). The notion of effective timescales of (dis)connectivity (Harvey, 2002; Fryirs et al., in press) adds a temporal component to the spatial analyses presented in this paper. As an example, some of the piedmonts, fans and terraces along Dart Brook have resided in the landscape for several tens of thousands of years (Fryirs, unpublished OSL dates) and are currently inactive, highlighting the degree of antecedent control on the contemporary sediment flux in this catchment. This example highlights the critical role of landscape disconnectivity in analysis of catchment-scale sediment budgets in landscapes of low relative relief and long-term sediment storage where a distinct set of impediments to sediment conveyance occur. These conditions prevail across much of the Australian continent (e.g. Olive et al., 1994; Wasson et al., 1996; Fryirs et al., in press). The impact of landscape disconnectivity on the behaviour of the ‘jerky conveyor belt’ that characterises the bedload sediment flux of river systems can be extended to a range of other biophysical considerations. For example, buffers and barriers affect water conveyance and the form of hydrographs (e.g. Kirkby et al., 2002; Michaelides and Wainwright, 2002), the filtering role of alluvial sediment stores and their effect upon water quality, nutrient cycling and water chemistry (e.g. Cooper et al., 2000), fish migration pathways and the distribution of macroinvertebrate communities (e.g. Perry and Schaeffer, 1987; Jungwirth et al., 2000). As such, landscape (dis) connectivity is a critical consideration in modelling activities that examine biophysical fluxes and associated biogeochemical responses to disturbance events. Unless these concepts are appropriately grounded by field investigations, the generation of realistic ‘visions’ and prediction of the timeframes over which system responses to disturbance are likely to be manifest will be compromised (Wohl, 2004; Brierley et al., 2006). The generic principles outlined in this paper provide a landscape framework by which these relationships can be realistically appraised, enabling the impact of natural and human-induced variants of disconnectivity to be assessed in any landscape setting. Acknowledgements This research is supported by an ARC Linkage Project LP0346918 held by Kirstie Fryirs and Gary Brierley. ARC Discovery Project DP0345451 supported Nick Preston during his time at Macquarie University. We thank Amalia Short for undertaking some of the
316
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
ArcMap manipulations and characterisation of tributarytrunk stream (dis)connectivity, and Anne Ashworth for orthorectifying many of the aerial photographs. Review comments by Stephen Tooth and Janet Hooke significantly improved the manuscript. References ArcHydro, 2005. Extension for ArcView 8.3. http://www.crwr.utexas. edu/giswr/hydro/index.html. Beven, K., Wood, E.F., 1983. Catchment geomorphology and the dynamics of runoff contributing areas. Journal of Hydrology 65, 139–158. Brierley, G.J., Fryirs, K., 1999. Tributary-trunk stream relations in a cutand-fill landscape: a case study from Wolumla catchment, N.S.W., Australia. Geomorphology 28, 61–73. Brierley, G.J., Fryirs, K.A., 2005. Geomorphology and River Management: applications of the river styles framework. Blackwell Publishing, Oxford, UK. 398 pp. Brierley, G.J., Fryirs, K.A., Jain, V., 2006. Landscape connectivity: The geographic basis of geomorphic applications. Area 38.2, 165–174. Caine, N., Swanson, F.J., 1989. Geomorphic coupling of hillslope and channel systems in two small mountain basins. Zeitschrift fur Geomorphologie 33, 189–203. Cooper, D.M., Jenkins, A., Skeffington, R., Gannon, B., 2000. Catchment-scale simulation of stream water chemistry by spatial mixing: theory and application. Journal of Hydrology 233, 121–137. Dalla Fontana, G., Marchi, L., 2003. Slope–area relationships and sediment dynamics in two alpine streams. Hydrological Processes 17, 73–87. Dearing, J.A., Jones, R.T., 2003. Coupling temporal and spatial dimensions of global sediment flux through lake and marine sediment records. Global and Planetary Change 39, 147–168. Erskine, W.D., 1985. Downstream geomorphic impacts of large dams: the case of Glenbawn Dam, NSW. Applied Geography 5, 195–210. Fryirs, K., Brierley, G.J., 1999. Slope–channel decoupling in Wolumla catchment, New South Wales, Australia: The changing nature of sediment sources following European settlement. Catena 35, 41–63. Fryirs, K., Brierley, G.J., 2001. Variability in sediment delivery and storage along river courses in Bega catchment, NSW, Australia: implications for geomorphic river recovery. Geomorphology 38, 237–265. Fryirs, K.A., Brierley, G.J., Preston, N.J., Kasai, M., in press. Buffers, barriers and blankets: The (dis)connectivity of catchment-scale sediment cascades. Catena. Goodbred, S.L., Kuehl, S.A., 1999. Holocene and modern sediment budgets for the Ganges–Brahmaputra river system: Evidence for highstand dispersal of floodplain, shelf and deep-sea depocenters. Geology 27, 559–562.
Harvey, A.M., 1992. Process interactions, temporal scales and the development of hillslope gully systems: Howgill Fells, northwest England. Geomorphology 5, 323–344. Harvey, A.M., 2002. Effective timescales of coupling within fluvial systems. Geomorphology 44, 175–201. Hooke, J.M., 2003. Coarse sediment connectivity in river channel systems: a conceptual framework and methodology. Geomorphology 56, 79–94. Jungwirth, M., Muhar, S., Schmutz, S., 2000. Fundamentals of fish ecological integrity and their relation to the extended serial discontinuity concept. Hydrobiologia 422/423, 85–97. Kirkby, M., Bracken, L., Reaney, S., 2002. The influence of land use, soil and topography on the delivery of hillslope runoff to channels in SE Spain. Earth Surface Processes and Landforms 27, 1459–1473. Kondolf, G.M., 1994. Geomorphic and environmental effects of instream gravel mining. Landscape and Urban Planning 28, 225–243. Lane, S.N., Richards, K.S., 1997. Linking river channel form and process: time, space and causality revisited. Earth Surface Processes and Landforms 22, 249–260. Lang, A., Hönscheidt, S., 1999. Age and source of colluvial sediments in Vaihingen-Enz, Germany. Catena 38, 89–107. Metivier, F., Gaudemer, Y., 1999. Stability of output fluxes of large rivers in South and East Asia during the last 2 million years: implications on floodplain processes. Basin Research 11, 293–303. Michaelides, K., Wainwright, J., 2002. Modelling the effects of hillslope–channel coupling on catchment hydrological response. Earth Surface Processes and Landforms 27, 1441–1457. Olive, L.J., Olley, J.M., Murray, A.S., Wallbrink, P.J., 1994. Spatial variation in suspended transport in the Murrumbidgee River, New South Wales, Australia. In: Olive, L.J., Loughran, R.J., Kesby, J.A. (Eds.), Variability in stream erosion and sediment transport. IAHS Publ., vol. 224, pp. 241–249. Perry, J.A., Schaeffer, D.J., 1987. The longitudinal distribution of riverine benthos: a river discontinuum? Hydrobiologia 148, 257–268. Schrott, L., Hufschmidt, G., Hankammer, M., Hoffman, T., Dikau, R., 2003. Spatial distribution of sediment storage types and quantification of valley fill deposits in an alpine basin, Reintal, Bavarian Alps, Germany. Geomorphology 55, 45–63. Schumm, S.A., 1977. The Fluvial System. John Wiley and Sons, New York. Sidorchuk, A., 2003. Floodplain sedimentation: inherited memories. Global and Planetary Change 39, 13–29. Story, R., Galloway, R.W., van de Graaff, R.H.M., Tweedie, A.D., 1963. General Report on the Lands of the Hunter Valley. CSIRO, Melbourne. 152 pp. Wasson, R.J., Olive, L.J., Roswell, C.J., 1996. Rates of erosion and sediment transport in Australia. In: Walling, D.E, Webb, B.W. (Eds.), Erosion and Sediment Yield: Global and Regional Perception. IAHS Publ., vol. 236, pp. 139–148. Wohl, E., 2004. Disconnected Rivers: Linking Rivers to Landscapes. Yale University Press, New Haven, US. 301 pp.
Catchment-scale (dis)connectivity in sediment flux in the upper Hunter catchment, New South Wales, Australia Kirstie A. Fryirs a,⁎, Gary J. Brierley b , Nicholas J. Preston c , John Spencer a a Department of Physical Geography, Macquarie University, North Ryde, NSW 2109, Australia School of Geography and Environmental Science, University of Auckland, Box 92019, New Zealand Institute of Geography, School of Earth Sciences, Victoria University of Wellington, Box 600, Wellington, New Zealand b
c
Received 10 May 2005; received in revised form 3 January 2006; accepted 3 January 2006 Available online 24 July 2006
Abstract (Dis)connectivity within and between landscape compartments affects the extent and rate of transfer of energy and matter through catchments. Various landforms may impede sediment conveyance in a river system, whether laterally to the channel (termed buffers) or longitudinally along the channel itself (termed barriers). A generic approach to analysis of landscape (dis) connectivity using slope threshold analysis in GIS, tied to air photograph interpretation and field mapping of buffers and barriers, is tested in the upper Hunter catchment, Australia. Under simulated conditions, effective catchment area, which is a measure of the proportion of a catchment that has the potential to contribute sediment to the channel network, varies from 73% to just 3% of the total catchment area for differing subcatchments in the upper Hunter. This variability can be explained by the spatial distribution and assemblage of buffers and barriers in each subcatchment. Multiple forms of disconnectivity are evident in some subcatchments, such that when one buffer or barrier is breached, other features still impede sediment transfer within the system. The importance of the position of buffers and barriers within any given subcatchment is emphasised. Spatial variability in valley width exerts a critical control on catchment connectivity, with more efficient sediment conveyance in narrow valleys relative to wider valleys characterised by piedmonts, terraces, fans and extensive floodplains in which conveyance is impeded. This variability reflects the landscape history and geological setting of each subcatchment. The framework developed in this paper can be used to assess the impact of natural or human-induced buffers and barriers on catchment-scale sediment flux in any landscape setting, providing a physical template atop which other biogeochemical fluxes could be examined. © 2006 Elsevier B.V. All rights reserved. Keywords: Disconnectivity; Coupling; Sediment flux; Buffers; Barriers; Landscape history; GIS; Hunter catchment; Australia
1. Introduction Sediment flux within river systems has been described as a ‘jerky conveyor belt’ (Schumm, 1977; Kondolf, 1994). While this conceptualisation has been framed in ⁎ Corresponding author. Tel.: +61 2 9850 8367; fax: +61 2 9850 8420. E-mail address: [email protected] (K.A. Fryirs). 0169-555X/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.geomorph.2006.01.044
terms of variability in flow energy which determines the capacity of the system to transport sediment along the channel network, due regard must also be given to the availability and delivery of sediment into the channel network, as this drives the propagation of disturbance responses through a catchment (Brierley and Fryirs, 2005). These relationships vary in different landscape settings, reflecting the perennial/ephemeral nature of the flow regime, controls on sediment availability (e.g.
298
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
Fig. 1. Location of the upper Hunter catchment within the Hunter catchment.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
landscape history, relief and lithology) and the configuration of any given catchment. Efforts to model sediment flux in catchments must integrate these controls on energy and sediment availability, ensuring that outputs are adequately grounded by field mapping of impediments to sediment conveyance. The distribution of riverine sediment stores and sinks, and the frequency with which sediment is added or removed from compartments, reflects the degree to which a river system is (dis)connected or (de)coupled (Caine and
299
Swanson, 1989; Lang and Hönscheidt, 1999; Harvey, 2002; Hooke, 2003; Schrott et al., 2003). Any factor that impedes sediment conveyance constrains sediment delivery to the catchment outlet. Fryirs et al. (in press) differentiate between landforms that impede sediment movement into the channel network (e.g. lateral disconnectivity from slopes to channels) and features that impede sediment movement along the channel (i.e. longitudinal disconnectivity), referring to these features as buffers and barriers respectively. The nature and
Fig. 2. Methods used to model the effective catchment area and the role of buffers in controlling (dis)connectivity in the upper Hunter catchment.
300
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
extent of (dis)connectivity vary across any given catchment, and over time, as influenced by the form, composition, scale and position of buffers and barriers, and the ease with which these features are reworked and breached. Impediments to sediment movement within a catchment restrict the rate of sediment transfer from the area upstream of that point. Hence, these impediments determine the area or proportion of a catchment that has the potential to directly contribute sediment to, or transport sediment along, the channel network under given flow conditions. This is referred to here as the effective catchment area (Fryirs et al., in press). Landscape (dis)connectivity, in turn, determines how geomorphic instability is manifest throughout a river system, shaping the direction and rate of geomorphic change, and lagged and off-site (propagatory) responses (Goodbred and Kuehl, 1999; Metivier and Gaudemer, 1999; Sidorchuk, 2003). Numerous models have been used to examine hydrological or sedimentary (dis)connectivity and its effect on contributing areas (e.g. Michaelides and Wainwright, 2002; Kirkby et al., 2002). Other researchers have analysed landscape connectivity using conceptual frameworks tied to field-based applications (e.g. Harvey, 1992; Fryirs and Brierley, 1999, 2001; Hooke, 2003). In few instances, however, have these principles been merged through analyses of landscape (dis)connectivity at the catchment-scale. In this paper, we appraise the role of buffers and barriers as a control on bedload sediment flux and channel network connectivity in the upper Hunter catchment, NSW, Australia. Rather than attempting to quantify sediment delivery to the channel within a sediment budget approach, we use a simulated set of conditions, tied to catchment-scale air photograph and field analyses, to examine how buffers and barriers restrict bedload supply across this 4300 km2 catchment. 2. Regional setting The upper Hunter catchment has a shorter north-south (∼ 60 km) than east-west (∼ 80 km) axis (Fig. 1). The Hunter–Mooki fault, which defines the boundary between the Sydney Basin and the New England Fold Belt, divides the catchment into two distinct zones. The eastern side generally comprises Carboniferous and Devonian metasediments, with a Tertiary basalt cap at high elevations. The metasediments comprise marine and terrestrial sediments associated with a Devonian volcanic arc (e.g. limestone, chert, lithic sandstone, conglomerates, etc.) that have been folded, faulted, and uplifted during the Permian–Triassic orogenic episode that formed the New England Fold Belt. The western side of the Hunter–
Mooki fault consists of near-horizontal Permian and Triassic sedimentary rocks (primarily shales, sandstones, conglomerates and coal), with relatively small areas of dissected Tertiary basalt. The four landscape units of the upper Hunter roughly align with the geological characteristics of the catchment (Fig. 1). The Remnant Plateau landscape unit that encompasses the Tertiary basalt plateau has a relatively high elevation (often >1000 m asl) and is deeply dissected. The Plateau Slopes landscape unit, which occurs at the margin of plateau remnants and high elevation ridges, comprises steep slopes and confined valleys. The Rugged and Hilly landscape unit in the central part of the catchment is characterized by partly-confined valleys. The Undulating Plains landscape unit, which occurs primarily to the west of the Hunter–Mooki fault, comprises low slopes and laterally-unconfined valleys. Average annual rainfall across the upper catchment ranges from 1400 mm/yr in the Barrington Tops, to 550 mm/yr west of Murrurundi. The largest flood since European settlement in the 1820s occurred in 1870. The largest recorded flood in 1955 had a peak discharge at Muswellbrook of more than 42 000 m3 s− 1. The discharge for a 1 in 2 year flow at Muswellbrook is 379 m3 s− 1. 3. Methods 3.1. GIS modelling of effective catchment area GIS modelling of the effective catchment area is based on the simple assumption that slope angle (gradient) controls delivery of sediments to the channel. Clearly, there are many factors that influence off-slope sediment delivery. Some of these relate to the nature of the land surface itself, and include, for example, slope angle, surface roughness and infiltration capacity. Others relate to the nature of runoff, e.g. its depth, volume and duration. Together, these factors determine the amount of energy available for sediment erosion and transport. Further, there are issues of spatial scale and the topological relationships between different parts of the land surface that have a strong influence on sediment delivery. It is these that the focus on (dis)connectivity addresses. We focus on slope angle to illustrate the nature of changing connectivity for a number of reasons. First, along with discharge, slope angle is the principal determinant of stream power and the energy available for sediment transport. Other factors associated with the land surface play secondary roles, moderating the amount of runoff or its velocity. Second, in contrast to many of the other controls on runoff and sediment transport, slope angle is constant at the temporal scale under consideration and can be readily
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
301
Table 1 Definition and functionality of buffers and barriers in the upper Hunter catchment Feature Buffers Piedmont
Alluvial fan
Definition and position
How identified/mapped
Impeding/blocking effect
Gentle sloping ground lying between the steep slope of the valley margin and floodplain. Often have floodplains or fans inset against them. Sediment lobe that radiates from a small tributary network (feeder channel) where the stream emerges from confinement onto a valley floor. Often have floodplains inset against them.
Distinctly graded surface running along the valley margin. Often grades into floodplain.
Blocks sediment derived from adjacent hillslopes as sediments cannot be transported across the surface.
Terrace
Elevated palaeo floodplain surface that is rarely inundated. Occur along the valley margin and often have floodplains inset against them.
Continuous floodplain
Flat surface adjacent to the channel margin in a laterally-unconfined valley. May be inset against bedrock valley margins, terraces or piedmonts in different settings.
Discontinuous floodplain
Flat surface adjacent to the channel margin in a partly-confined valley. May be inset against bedrock valley margins or terraces.
Trapped Tributary system trapped behind a terrace tributary fill or floodplain surface.
Barriers Sediment slug
Distinctly conical feature located between bedrock spurs. Flow paths evident on the surface.
Not currently active. Impedes sediment conveyance from hillslopes and feeder tributaries to the channel, as sediments cannot be transported across the surface. Can only be reworked if incision extends through them from the trunk stream. Flat-topped, stepped features inset Not currently active. Impedes sediment against a bedrock valley margin conveyance from hillslopes and tributaries but elevated above the to the channel. Elevated surface means contemporary floodplain surface. contemporary flows cannot breach their surfaces. Can only erode via channel expansion processes. Consolidated sediments require high energy conditions for this to occur. Surface immediately adjacent to the Impedes sediment conveyance from channel margin that contains hillslopes to the channel. Disrupts floodplain geomorphic units such longitudinal conveyance of material as levees and palaeochannels. by removing it from channel conveyance. Surfaces currently active but require channel expansion for them to be breached. Impedes sediment conveyance from Surface immediately adjacent to hillslopes to the channel. the channel margin that contains Disrupts longitudinal conveyance of floodplain geomorphic units such material by removing it from channel as floodchannels. conveyance. Surfaces currently active but require channel expansion or stripping to be breached. Trapped behind other forms of Intact valley fill surface with no buffer and disconnect tributary well-defined channel that flows networks from the trunk stream valley floor. into the back of a terrace or floodplain along a tributary valley.
Accumulation of sediment that covers the entire channel bed and infills the channel.
As for description for morphology and sedimentology
Competencelimited channel
Small, sinuous (often tortuous) channel found on a very low slope.
As for description for morphology and sedimentology
Dam
Concrete structure that extends across the valley behind which a lake is formed.
As for description for morphology and sedimentology
Temporarily impedes upstream–downstream sediment conveyance along the channel. The volume and frequency with which sediment moves depends on the magnitude of the flow event, and the energy available to transport sediments. Tend to be transportlimited settings. Impedes upstream–downstream sediment conveyance along the channel. Channels tend to be competence limited, meaning that any bedload materials that are made available to these channels cannot be transported over the low slopes or along a sinuous channel with very low energy conditions. Total blocking of most flow and sediment transported down the valley. All bedload materials are trapped behind the dam wall.
302
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
Fig. 3. Scheme used to classify tributary-trunk stream (dis)connectivity in the upper Hunter catchment. Types 1, 2 and 3 are forms of tributary-trunk stream disconnectivity, while Types 4, 5 and 6 are forms of tributary-trunk stream connectivity.
assessed on a spatially distributed basis. Factors such as roughness or infiltration capacity are less readily mapped due to their spatial variability, while the temporal variability of runoff creates difficulties in making generalisations. Thus, slope angle is used to illustrate the variability of connectivity. Below a given slope angle, energy for transport is insufficient, creating an obstacle to connectivity. As such, the focus is principally on lateral (dis)connectivity. Adopted procedures are summarised in Fig. 2. Threshold values of slope angle were used to perform this assessment, with three critical slope angles chosen to illustrate a range of energy conditions under which buffers may be breached. Slope thresholds of 0.5°, 2° and 25° were arbitrarily chosen to gain some indicative sense of how the model would perform at these extremes. Because slope interacts with more dynamic factors (rainfall intensity, runoff velocity) in determining sediment delivery, looking at different slope angles allows generalisation about differing energy conditions. Using a high slope angle threshold, for example, is analogous to considering low intensity rainfall or a surface with high roughness and high infiltration capacity. Initially this may seem counter-intuitive in the sense that high slope angle implies high energy. In effect, however, this represents the situation where other factors are such that only high slope angle will promote sediment transport. Conversely, the use of a low slope threshold represents conditions in
which there is so much energy (or little resistance) that sediment transport meets with little obstacle. The input data for the modelling were a 25 m DEM of the study region and a data set containing streamlines digitised from 1:25,000 topographic maps (sourced from the NSW Department of Natural Resources). GIS analysis was undertaken using ArcView 8.3 and the ArcHydro extension from the Centre for Research in Water Resources at the University of Texas (ArcHydro, 2005). The first step, after filling anomalous sinks on the DEM surface, was to derive a set of streamlines. This was done using flow direction, flow accumulation and stream definition routines within ArcHydro. A channel initiation threshold catchment area of 2 km2 was chosen after several iterations suggested that this best reflected the distribution of the digitised channel network. The automated derivation of streamlines is often limited in areas of low slope, and the digitised streamline data set was used to enforce the channel position within the DEM. ArcHydro was also used to define subcatchments. Five subcatchments, incorporating the major stream networks in the upper Hunter catchment, were modelled as separate entities. These subcatchments are Dart (which includes Dart Brook, Middle Brook and Kingdon Ponds), the Upper Hunter (which includes the trunk stream and all its minor tributaries upstream of Glenbawn Dam), Rouchel (which encompasses Rouchel Brook and Davis Creek), the Pages and the Isis. In addition, the upper Hunter
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
303
Table 2 Effective catchment area of each major subcatchment in the upper Hunter catchment Subcatchment
Total catchment area (km2)
Effective catchment area (km2)
% of subcatchment connected (2° slope simulation)
% of catchment disconnected (2° slope simulation)
Degree of connectivity
Rouchel Pages Upper Hunter Isis Dart Lower Hunter Total catchment (due to buffers only) Total catchment (when barriers added)
435 642 1294 551 806 567 4295
317 348 849 258 256 17 2045
73 54 67 47 32 3 48
27 46 33 53 68 97 52
Highly connected Connected Connected Disconnected Highly disconnected Highly disconnected Disconnected
4295
334
8
92
Highly disconnected
catchment, including the trunk stream extending from Glenbawn Dam to Muswellbrook and all the major tributary subcatchments (i.e. Rouchel, Isis, Upper Hunter, Pages and Dart), was modelled as a single entity. Each modelling unit was then treated as follows. A 50 m buffer was placed around the streamlines for each subcatchment, and the outline of this buffer (polyline feature) was converted into a raster of 25 × 25 m cells. Each of the cells representing this rasterised stream edge was then converted to a point within a point feature layer. This creates a data set containing a point every 25 m on both sides of the channel. The catchment of each of these points were then derived using flow direction and flow accumulation routines within ArcHydro. In this way, the entire surface area of each subcatchment was divided into a series of smaller drainage units, with every part of the subcatchment draining into one of the points along the streamline. A slope raster layer was created from the DEM. Threshold slope layers were derived by applying a conditional statement that returns No Data values to cells with slope angles below the chosen threshold. These threshold slope layers effectively remove areas below the chosen slope threshold from further processing when the ArcHydro hydrological modelling is repeated using a revised DEM from which the cells with slopes less than the threshold value have been removed. The catchments of all cells adjacent to the channel were redefined using modified flow direction layers. Cells upslope of low angle areas are thus not connected to the channel; the effective catchment area comprises only those cells that have both angles greater than the threshold, and are connected to the channel through other cells with slopes greater than the threshold. The result is a layer that describes the catchment area connected to a channel by surfaces that have a slope greater than the threshold slope. This approach is considered to be indicative rather than absolute, allowing
across-catchment comparisons of patterns of (dis)connectivity to be examined. The output shows where there is ‘potential’ for sediment supply. It does not quantify the volumes of sediment moving or determine if sediment is available to be moved. Further work is required to tie this indicative appraisal to the discharge record to model the types of flows under which various simulations occur, and to test the sensitivity of model output to various slope threshold analyses. Of three values tested, the across-catchment variability in effective catchment area was most pronounced using a threshold slope angle value of 2°. This value reflects energy conditions that are capable of moving materials only on slopes over 2° (but may have the capability to rework sediment along channels and immediately adjacent sediment storage zones). This threshold value is thought to represent a rainfall event of some unknown but relatively high magnitude. The output from thresholds of 0.5° and 25° showed, respectively, almost entire catchment connectivity with only the very lowest slopes (<0.5°) representing an impediment to slope-channel connectivity, or no lateral connectivity (i.e. sediment transport would be limited to the channel zone). Because the 2° simulation is thought to represent conditions under which off-slope sediment transport is likely to be promoted, it was chosen as the basis for further analysis. By choosing this simulation the results only represent a single snapshot of the spatial patterns of (dis)connectivity in this catchment. 3.2. Identification and analysis of buffers and barriers, and tributary-trunk stream (dis)connectivity Air photograph interpretation was tied to field investigations and ground truthing to map the distribution of buffers and barriers in the catchment (Fig. 2). A complete aerial photograph coverage of the upper Hunter
304
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
catchment consists of 1998, 1999, 2000, and 2001 photo sets with an approximate scale of 1:25,000. The photos containing areas of interest were ortho-rectified using ERDAS Imagine 8.4. Digital versions of the appropriate CMA 1:25,000 topographic maps were used as the source of reference co-ordinates for the ortho-rectification. Buffers were identified using a stereoscope and manually digitised onto the aerial photographs. This procedure was completed along all major stream networks in the upper Hunter catchment to a point where the size of the sediment stores was beyond the resolution of the aerial photos. The digitised maps were checked in the field to ensure that features had been mapped and categorised correctly. Major barriers to sediment movement along channels were identified in the field. Six types of buffer (piedmont,
alluvial fan, terrace, continuous floodplain, discontinuous floodplain, trapped tributary fill), and three types of barrier (sediment slug, competence-limited channel, dam) have been identified in the upper Hunter catchment (Table 1). Differing patterns of buffers result in six forms of tributary-trunk stream (dis)connectivity in the upper Hunter catchment (Fig. 3). Types 1–3 are forms of tributarytrunk stream disconnectivity that are characterised by either discontinuous channels in the tributary system or the occurrence of a buffer along the trunk stream valley margin that severs the connection between the tributary channel and the trunk stream. Types 4–6 (Fig. 3) are characterised by tributaries that directly connect to the trunk stream, associated either with the lack of an impeding buffer on the valley floor (type 6), or situations
Fig. 4. Effective catchment area of subcatchments in the upper Hunter. The effective catchment area is noted as a percentage of the area that is disconnected due to buffers and barriers.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
where buffers fail to impede the connectivity between the tributary and the trunk stream (types 4 and 5; see Fig. 3). Impediments to sediment conveyance from tributary systems to the trunk stream may take the form of a single buffer (e.g. a discontinuous floodplain pocket) or multiple buffers where 2 or more features disconnect the tributary from the trunk stream (e.g. a fan or a continuous floodplain). In addition, barriers may disconnect linkages along the channel network. These ‘types’ of (dis)connectivity were categorised using the aerial photo mosaics and the digitised buffer map. These categories were added as attributes to the digitised streamline data set. In ArcView, this data set was then used to calculate subcatchment areas and measure stream lengths at each tributary-trunk stream confluence in the catchment.
305
Overlaying of these data sets was then used to assess the degree to which the distribution of buffers and barriers could explain the effective catchment area simulation, and thus whether different parts of the catchment were connected or disconnected. 4. Results 4.1. Effective catchment area of the upper Hunter catchment and its subcatchments Table 2 shows the results of the 2° slope threshold simulation in the upper Hunter catchment, and Fig. 4 shows the effective catchment area of the upper Hunter catchment draining to Muswellbrook under these
Fig. 5. Catchment area (a) and stream length (b) relationships for connected and disconnected tributaries in the upper Hunter catchment. These have been divided according to whether they occur east or west of the Hunter–Mooki fault. Note that all disconnected tributaries drain catchment areas less than 22 km2 and are less than 8 km in length.
306
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
conditions. In this simulation, the effective catchment area is 2045 km2 from a total of 4295 km2, indicating that around 52% of the catchment is disconnected and does not contribute to the overall sediment cascade as measured at Muswellbrook. This calculation reflects the role of buffers that determine lateral connectivity (i.e. sediment supply from hillslopes and the valley floor zones adjacent to the channel). It does not take account
of longitudinal disconnectivity induced by barriers (i.e. whether blockages along the channel network further disconnect areas of the catchment). There is significant variability in the degree of (dis) connectivity in different subcatchments of the upper Hunter (Table 2). Using four arbitrary classes of (dis) connectivity (0–30% = highly connected, 30–50% = connected, 50–70% = disconnected, 70–100% = highly
Fig. 6. Assemblages of tributary-trunk stream types in each subcatchment of the upper Hunter. These are presented as (a) percentages of the types of (dis)connectivity (see Fig. 3), and (b) percentages of the subcatchment total. C = connected tributaries, D = disconnected tributaries, primary = one buffer present.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
disconnected), Rouchel subcatchment is the only subcatchment that is highly connected (27%). The Upper Hunter and Pages subcatchments are classed as connected (33% and 46% respectively). The Isis subcatchment is considered to be disconnected (53%). The Dart subcatchment and the Lower Hunter subcatchment are highly disconnected (68% and 97% respectively). These results suggest that available sediments are likely to be efficiently contributed to, and conveyed through, Rouchel subcatchment while the Dart and Lower Hunter subcatchments are extremely inefficient at contributing and conveying sediment. To explain the subcatchment variability in the degree of disconnectivity, the types and distribution of impediments have been examined across the catchment.
307
4.2. Catchment configuration and types of tributarytrunk stream (dis)connectivity Analysis of tributary subcatchment area and maximum stream length along tributaries in the upper Hunter catchment indicates that all tributaries with catchment areas >22 km2 or stream lengths >8 km are connected to the trunk stream (Fig. 5). This suggests that these tributary systems generate sufficient discharge and energy to breach any buffers that may be present, thereby maintaining their connectivity to the trunk stream. However, tributaries with catchment areas <22 km2 or stream lengths <8 km can be either connected or disconnected as determined by the presence and type of buffer along the trunk stream valley
Fig. 7. General trends in tributary-trunk stream (dis)connectivity in the upper Hunter catchment. Note the variability in the dominant forms of (dis) connectivity in different subcatchments and the dominance of various types of buffers. Black lines divide the catchment according to these dominant types of (dis)connectivity.
308 K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
Fig. 8. (a) Effective catchment area and (b) the distribution of buffers in Dart subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
margin. All but four of the small, disconnected tributaries are impeded by extensive features such as piedmonts, continuous floodplains or fans. In 41% of these cases, secondary blockages occur (i.e. 2 or more buffers are present; Fig. 6a). Seemingly, the position of the trunk stream channel on the valley floor influences whether subcatchments with small areas (and hence small discharges) have the potential to contribute bedload sediments to the trunk stream confluence, as the subcatchment may not have the capacity to form continuous channels that can breach buffer(s). There is no significant relation between
309
the type of (dis)connectivity and the catchment area or length of the tributary stream for subcatchments found on different sides of the Hunter–Mooki fault (Fig. 5). 4.3. The types and patterns of buffers and barriers across the upper Hunter catchment Significant variability in the forms of tributary-trunk stream (dis)connectivity and the associated type and assemblage of buffers is evident for differing subcatchments of the upper Hunter catchment (Fig. 7). A distinct
Fig. 9. (a) Effective catchment area and (b) the distribution of buffers in Rouchel subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
310 K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316 Fig. 10. (a) Effective catchment area and (b) the distribution of buffers in the Upper Hunter subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
pattern emerges for subcatchments east and west of the Hunter–Mooki fault. Tributary subcatchments to the east have less accommodation space in which to store sediments. The types of buffers tend to be discontinuous pockets of floodplain or terraces that are sheltered behind bedrock spurs in bedrock-controlled valleys. In contrast, valleys west of the Hunter–Mooki fault have greater accommodation space in the form of broad, open valleys
311
in which a range of buffers have formed, including piedmont zones, terraces, continuous floodplains and alluvial fans. Dart subcatchment, located West of the Hunter–Mooki fault, is one of the most disconnected subcatchments in the upper Hunter (Fig. 8). This subcatchment contains the highest prevalence of types 1–3 tributary-trunk stream disconnectivity (Fig. 6a). In almost half (47%) of these
Fig. 11. (a) Effective catchment area and (b) the distribution of buffers in Isis subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
312
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
cases, secondary impediments are evident (Fig. 6b). Most tributaries either have discontinuous channels that do not reach the trunk stream valley or they flow into broad and open valleys that are characterised by a range of buffers. The valleys store vast volumes of alluvial sediment in the form of continuous floodplains. Adjacent to the floodplains, piedmont zones extend along the western margin of Dart Brook and along the eastern side of Kingdon Ponds. These piedmont zones act as significant buffers to materials that may be eroded from the surrounding hillslopes. Even if overland flow has sufficient energy to move materials over this surface, the alluvial floodplains act as a secondary buffer to conveyance to the channel. The only materials entering the main channel networks are transported along gully networks that have incised through the floodplain into the piedmont zones. In reaches
further up catchment, alluvial fans also occur between bedrock spurs or in depressions on the piedmont. Discontinuous floodplain pockets are often inset against these fan deposits. In addition to these features, upper Kingdon Ponds contains extensive flights of terraces that extend up to 10 m above the channel bed. These terraces often trap tributary valleys behind them, forming trapped tributary fills. Discontinuous floodplain pockets are often inset against these terraces. The assemblage and distribution of buffers in Dart subcatchment produces a high degree of disconnectivity in this system, to the extent that if one form is breached a secondary feature acts as another impediment further along the sediment cascade. Topography on this side of the Hunter–Mooki fault is subdued, so that channels are very low slope, highly sinuous, competence limited and inefficient at transporting
Fig. 12. (a) Effective catchment area and (b) the distribution of buffers in the Pages subcatchment. Tributary-trunk stream (dis)connectivity classes are noted on each tributary stream analysed.
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
bedload materials. This form of barrier induces further disconnectivity in this system. Rouchel Brook, located east of the Hunter–Mooki fault, is the most connected subcatchment in the upper Hunter (Fig. 9). Relatively confined valleys along the length of Rouchel Brook and Davis Creek restrict the formation of buffers to occasional terrace pockets and discontinuous floodplains. These are typically located on the insides of sinuous valley meander bends or sheltered behind bedrock spurs (Fig. 9). Tributary-trunk stream (dis)connectivity is dominated by type 6 confluences (Fig. 6b). This topographic setting, along with the lack of any significant barriers along the channel network, ensures that any sediment that is made available to the channel can be readily conveyed to the Hunter River. Streams in the Upper Hunter subcatchment have a similar assemblage of buffers to that found in Rouchel subcatchment (Fig. 10). Tributary-trunk stream (dis) connectivity is dominated by types 5 and 6 (Fig. 6b). The valleys in this subcatchment are relatively confined and sinuous with little accommodation space. Buffer formation is restricted to terrace pockets and discontinuous floodplains that combine to produce secondary block-
313
ages. Glenbawn Dam, located at the outlet of the Upper Hunter subcatchment, traps up to 98% of bedload sediment contributed from the upper catchment (Erskine, 1985), effectively disconnecting 30% of the upper Hunter catchment from the contemporary sediment cascade. Tributary-trunk stream (dis)connectivity in the Isis subcatchment is dominated by types 5 and 6 (Fig. 11). The assemblage of buffers in this catchment is dominated by terraces and discontinuous floodplain pockets. Around half of all disconnectivity is induced by secondary blockages (Fig. 6b). The terraces in the Isis subcatchment often trap tributaries inducing significant disconnectivity along this stream. Nevertheless, any sediments that do reach the channel network are effectively conveyed to the Pages River through the bedrock-controlled valleys that dominate this system. The Pages subcatchment can be considered as two subsystems that operate upstream and downstream of the mid-catchment gorge and west-east of the Hunter–Mooki fault (Fig. 12). The upper Pages subcatchment has similar characteristics to upper Dart subcatchment, flowing through relatively broad-open valleys. Alluvial fans and
Fig. 13. Landscape configuration as a control on the forms of (dis)connectivity, and the role of various types of buffers in two contrasting subcatchments of the upper Hunter catchment, Dart subcatchment and Rouchel subcatchment.
314
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
discontinuous floodplains with occasional terraces dominate this section of the catchment (Fig. 12). Given the larger catchment areas that drain into these tributary systems, the alluvial fans have been breached by continuous channels, resulting in type 4 connectivity (Fig. 6a). If materials are transported into the gorge they are effectively flushed to the lower Pages River. Downstream of the gorge, the Pages River flows through a partly-confined valley. As a result, discontinuous floodplain pockets act as localised buffers to lateral connectivity. However, as the valley begins to widen along the lower Pages River, the buffering effect caused by continuous floodplains, which extend to the Hunter trunk stream, increases. Within this laterally-unconfined valley, a high width–depth ratio channel has formed and a large sediment slug has accumulated. This channel/slug has a relatively low slope and is acting as a temporary barrier to sediment conveyance. The slug extends for over 10 km and is several metres thick. Bedload sediment originating from upstream becomes trapped in this slug, significantly reducing the volume of bedload material that reaches the Hunter trunk stream from the Pages (and Isis) subcatchments. Except during flows large enough to source sediment from the slug and transport it downstream, the Pages and Isis subcatchments are effectively disconnected from the Hunter trunk stream. Cumulative responses to sediment contribution and conveyance along streams of the upper Hunter catchment are manifest within the Lower Hunter subcatchment around Muswellbrook (Figs. 4 and 9). While this section of the river is itself significantly buffered in terms of its lateral connectivity, owing to the extensive nature of continuous floodplains along both channel banks, there are no significant barriers to sediment conveyance along the channel itself. When the combined effect of buffers and barriers is examined, the effective catchment area of the entire upper Hunter catchment at Muswellbrook drops significantly to 334 km2 or 9% of the entire catchment area for this simulation (Table 2). 5. Discussion and implications Buffers and barriers are agents of disconnectivity in catchments. The means by which they bring about disconnectivity varies depending on their morphology, position in the landscape and juxtaposition relative to other landforms, i.e. whether primary or secondary blockages occur (Table 1). Differing patterns of these features form in particular valley settings. For example, piedmonts, alluvial fans and continuous floodplains that form in wide, open valleys with extensive accommodation space
account for a significant percentage of the disconnected catchment area of the upper Hunter, buffering slopes from channels and tributaries from trunk streams (Harvey, 2002). Along disconnected tributaries in the upper Hunter catchment, trapped tributary fill, continuous floodplains, piedmonts and terraces account for 24%, 22%, 21% and 17% of the total disconnected catchment area respectively. Fans and discontinuous floodplains each buffer 8% of the total disconnected catchment area. Discontinuous floodplains and terraces are the most commonly breached buffers in the upper Hunter catchment. Around 32% and 18%, respectively, of the total connected tributary catchment area is connected through these features. Direct connection of a tributary and trunk stream, where no buffer is present, accounts for 32% of the total connected tributary catchment area. Discontinuous floodplains or isolated terraces tend to occur in partly-confined valleys where sediment supplied from upstream is deposited behind bedrock spurs. Under these conditions, slope-channel and tributary-trunk stream linkages are only locally disconnected. Unlike more extensive landforms that are not readily breached (less than 8% each of the total connected tributary catchment area is connected through piedmont, fan, trapped tributary fill and continuous floodplains), these localised buffers can be readily breached, resulting in connectivity. Other than dams, barriers temporarily disrupt sediment conveyance along the channel network, disconnecting upstream– downstream linkages (Lane and Richards, 1997; Hooke, 2003). Catchment (dis)connectivity can be considered within a nested hierarchy at local, zonal and system scales (Harvey, 2002; Fryirs et al., in press). (Dis)connectivity within-compartments or within-landforms (e.g. on a single hillslope) that occurs at the local-scale was not examined in this study (c.f. Lang and Hönscheidt, 1999). Connectivity between landscape compartments is the first-order determinant of landscape behaviour examined. For example, slopes and channels can be disconnected where floodplains or terraces buffer the base of a slope. At this zonal scale (Harvey, 2002), the degree and effectiveness of an impediment is related to the size and sedimentary composition of the landform that acts as a buffer. At the broader system scale, (dis) connectivity relates to the behaviour of whole subcatchments or catchments. Tributary-trunk stream connectivity and the presence/character of barriers affect the operation of the sediment cascade at this scale. The position of impediments is a key determinant of effective catchment area, dictating how responses to disturbance and instability are likely to be manifest through a catchment. These relationships vary over time,
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
reflecting, among many factors, the ease and frequency with which impediments are breached and reform. Hence, the geomorphic effectiveness of any given flow is influenced by the position and relative scale of impediments within the system. For example, if an impediment (e.g. a fan) forms in a headwater subcatchment, it will have a limited effect on the overall conveyance of sediment in the larger catchment, thus having little effect on the effective catchment area. However, if that same sort of impediment occurs along the trunk stream or close to the outlet of the catchment, the impact is significantly greater. Hence, the impact of a given type of impediment may vary, dependent upon its scale and position in the catchment (Beven and Wood, 1983; Brierley and Fryirs, 1999; Dalla Fontana and Marchi, 2003; Schrott et al., 2003). These relationships reflect catchment configuration and landscape history (i.e. catchment shape, patterns of valley width, tributary-trunk stream relationships), as determined by geological controls. Landscape history and geological controls on valley configuration (morphology, accommodation space and slope) constrain the pattern of sediment storage in any catchment, and hence the type and distribution of impediments to sediment conveyance. To demonstrate this in the upper Hunter, landscape configuration along Dart Brook and Rouchel Brook, located west and east of the Hunter–Mooki fault respectively, are expressed in terms of Land Systems Units (Story et al., 1963) in Fig. 13. The differing types and patterns of buffers that have formed in these two settings affect the degree of lateral and longitudinal (dis)connectivity. In Dart subcatchment, buffers such as piedmonts, fans, continuous floodplains and terraces induce disconnectivity and hence effective catchment area is low. The Dart subcatchment drains relatively erodible geologies, is relatively low relief, and has wide, open valleys in which multiple buffers have formed. Extensive disconnectivity results, and effective catchment area is low. In contrast, Rouchel subcatchment drains relatively resistant geologies, has relatively high relief and is dominated by bedrock-controlled valleys. Buffers are restricted to isolated pockets of floodplain or terraces that are preserved behind bedrock spurs in a partly-confined valley. Only localised lateral disconnectivity results, and the effective catchment area is high. The frequency with which different types of buffer and barrier are breached is analogous to switching on or off certain parts of the sedimentary cascade over different timeframes (see Fryirs et al., in press). To appraise these relationships, the evolution and age structure of buffers and barriers in a catchment must be assessed, interpreting how long these features have acted as impediments and how connectivity has
315
changed over time (e.g. in response to direct or indirect human-disturbance; Lang and Hönscheidt, 1999; Dearing and Jones, 2003; Schrott et al., 2003). The notion of effective timescales of (dis)connectivity (Harvey, 2002; Fryirs et al., in press) adds a temporal component to the spatial analyses presented in this paper. As an example, some of the piedmonts, fans and terraces along Dart Brook have resided in the landscape for several tens of thousands of years (Fryirs, unpublished OSL dates) and are currently inactive, highlighting the degree of antecedent control on the contemporary sediment flux in this catchment. This example highlights the critical role of landscape disconnectivity in analysis of catchment-scale sediment budgets in landscapes of low relative relief and long-term sediment storage where a distinct set of impediments to sediment conveyance occur. These conditions prevail across much of the Australian continent (e.g. Olive et al., 1994; Wasson et al., 1996; Fryirs et al., in press). The impact of landscape disconnectivity on the behaviour of the ‘jerky conveyor belt’ that characterises the bedload sediment flux of river systems can be extended to a range of other biophysical considerations. For example, buffers and barriers affect water conveyance and the form of hydrographs (e.g. Kirkby et al., 2002; Michaelides and Wainwright, 2002), the filtering role of alluvial sediment stores and their effect upon water quality, nutrient cycling and water chemistry (e.g. Cooper et al., 2000), fish migration pathways and the distribution of macroinvertebrate communities (e.g. Perry and Schaeffer, 1987; Jungwirth et al., 2000). As such, landscape (dis) connectivity is a critical consideration in modelling activities that examine biophysical fluxes and associated biogeochemical responses to disturbance events. Unless these concepts are appropriately grounded by field investigations, the generation of realistic ‘visions’ and prediction of the timeframes over which system responses to disturbance are likely to be manifest will be compromised (Wohl, 2004; Brierley et al., 2006). The generic principles outlined in this paper provide a landscape framework by which these relationships can be realistically appraised, enabling the impact of natural and human-induced variants of disconnectivity to be assessed in any landscape setting. Acknowledgements This research is supported by an ARC Linkage Project LP0346918 held by Kirstie Fryirs and Gary Brierley. ARC Discovery Project DP0345451 supported Nick Preston during his time at Macquarie University. We thank Amalia Short for undertaking some of the
316
K.A. Fryirs et al. / Geomorphology 84 (2007) 297–316
ArcMap manipulations and characterisation of tributarytrunk stream (dis)connectivity, and Anne Ashworth for orthorectifying many of the aerial photographs. Review comments by Stephen Tooth and Janet Hooke significantly improved the manuscript. References ArcHydro, 2005. Extension for ArcView 8.3. http://www.crwr.utexas. edu/giswr/hydro/index.html. Beven, K., Wood, E.F., 1983. Catchment geomorphology and the dynamics of runoff contributing areas. Journal of Hydrology 65, 139–158. Brierley, G.J., Fryirs, K., 1999. Tributary-trunk stream relations in a cutand-fill landscape: a case study from Wolumla catchment, N.S.W., Australia. Geomorphology 28, 61–73. Brierley, G.J., Fryirs, K.A., 2005. Geomorphology and River Management: applications of the river styles framework. Blackwell Publishing, Oxford, UK. 398 pp. Brierley, G.J., Fryirs, K.A., Jain, V., 2006. Landscape connectivity: The geographic basis of geomorphic applications. Area 38.2, 165–174. Caine, N., Swanson, F.J., 1989. Geomorphic coupling of hillslope and channel systems in two small mountain basins. Zeitschrift fur Geomorphologie 33, 189–203. Cooper, D.M., Jenkins, A., Skeffington, R., Gannon, B., 2000. Catchment-scale simulation of stream water chemistry by spatial mixing: theory and application. Journal of Hydrology 233, 121–137. Dalla Fontana, G., Marchi, L., 2003. Slope–area relationships and sediment dynamics in two alpine streams. Hydrological Processes 17, 73–87. Dearing, J.A., Jones, R.T., 2003. Coupling temporal and spatial dimensions of global sediment flux through lake and marine sediment records. Global and Planetary Change 39, 147–168. Erskine, W.D., 1985. Downstream geomorphic impacts of large dams: the case of Glenbawn Dam, NSW. Applied Geography 5, 195–210. Fryirs, K., Brierley, G.J., 1999. Slope–channel decoupling in Wolumla catchment, New South Wales, Australia: The changing nature of sediment sources following European settlement. Catena 35, 41–63. Fryirs, K., Brierley, G.J., 2001. Variability in sediment delivery and storage along river courses in Bega catchment, NSW, Australia: implications for geomorphic river recovery. Geomorphology 38, 237–265. Fryirs, K.A., Brierley, G.J., Preston, N.J., Kasai, M., in press. Buffers, barriers and blankets: The (dis)connectivity of catchment-scale sediment cascades. Catena. Goodbred, S.L., Kuehl, S.A., 1999. Holocene and modern sediment budgets for the Ganges–Brahmaputra river system: Evidence for highstand dispersal of floodplain, shelf and deep-sea depocenters. Geology 27, 559–562.
Harvey, A.M., 1992. Process interactions, temporal scales and the development of hillslope gully systems: Howgill Fells, northwest England. Geomorphology 5, 323–344. Harvey, A.M., 2002. Effective timescales of coupling within fluvial systems. Geomorphology 44, 175–201. Hooke, J.M., 2003. Coarse sediment connectivity in river channel systems: a conceptual framework and methodology. Geomorphology 56, 79–94. Jungwirth, M., Muhar, S., Schmutz, S., 2000. Fundamentals of fish ecological integrity and their relation to the extended serial discontinuity concept. Hydrobiologia 422/423, 85–97. Kirkby, M., Bracken, L., Reaney, S., 2002. The influence of land use, soil and topography on the delivery of hillslope runoff to channels in SE Spain. Earth Surface Processes and Landforms 27, 1459–1473. Kondolf, G.M., 1994. Geomorphic and environmental effects of instream gravel mining. Landscape and Urban Planning 28, 225–243. Lane, S.N., Richards, K.S., 1997. Linking river channel form and process: time, space and causality revisited. Earth Surface Processes and Landforms 22, 249–260. Lang, A., Hönscheidt, S., 1999. Age and source of colluvial sediments in Vaihingen-Enz, Germany. Catena 38, 89–107. Metivier, F., Gaudemer, Y., 1999. Stability of output fluxes of large rivers in South and East Asia during the last 2 million years: implications on floodplain processes. Basin Research 11, 293–303. Michaelides, K., Wainwright, J., 2002. Modelling the effects of hillslope–channel coupling on catchment hydrological response. Earth Surface Processes and Landforms 27, 1441–1457. Olive, L.J., Olley, J.M., Murray, A.S., Wallbrink, P.J., 1994. Spatial variation in suspended transport in the Murrumbidgee River, New South Wales, Australia. In: Olive, L.J., Loughran, R.J., Kesby, J.A. (Eds.), Variability in stream erosion and sediment transport. IAHS Publ., vol. 224, pp. 241–249. Perry, J.A., Schaeffer, D.J., 1987. The longitudinal distribution of riverine benthos: a river discontinuum? Hydrobiologia 148, 257–268. Schrott, L., Hufschmidt, G., Hankammer, M., Hoffman, T., Dikau, R., 2003. Spatial distribution of sediment storage types and quantification of valley fill deposits in an alpine basin, Reintal, Bavarian Alps, Germany. Geomorphology 55, 45–63. Schumm, S.A., 1977. The Fluvial System. John Wiley and Sons, New York. Sidorchuk, A., 2003. Floodplain sedimentation: inherited memories. Global and Planetary Change 39, 13–29. Story, R., Galloway, R.W., van de Graaff, R.H.M., Tweedie, A.D., 1963. General Report on the Lands of the Hunter Valley. CSIRO, Melbourne. 152 pp. Wasson, R.J., Olive, L.J., Roswell, C.J., 1996. Rates of erosion and sediment transport in Australia. In: Walling, D.E, Webb, B.W. (Eds.), Erosion and Sediment Yield: Global and Regional Perception. IAHS Publ., vol. 236, pp. 139–148. Wohl, E., 2004. Disconnected Rivers: Linking Rivers to Landscapes. Yale University Press, New Haven, US. 301 pp.