close

Вход

Забыли?

вход по аккаунту

?

The role of geomorphology in substratum patch selection by freshwater mussels in the Hawkesbury УNepean River (New South Wales) Australia.

код для вставкиСкачать
AQUATIC CONSERVATION: MARINE AND FRESHWATER ECOSYSTEMS
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
Published online 5 June 2008 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/aqc.949
The role of geomorphology in substratum patch selection by
freshwater mussels in the Hawkesbury–Nepean River (New South
Wales) Australia
M. BRAINWOODa, SHELLEY BURGINa,* and M. BYRNEb
a
College of Health and Science, University of Western Sydney, Locked Bag 1797, South Penrith DC, Australia, 2753
b
Department of Anatomy and Histology, F-13, University of Sydney, Camperdown, Australia, 2006
ABSTRACT
1. Microhabitat preferences of freshwater mussels and associated substrate characteristics were investigated
across a range of geomorphic reaches in the Hawkesbury–Nepean River, Australia.
2. The structure of substratum patches available was strongly influenced by geomorphic reach type. In each
reach type, mussel distribution was most frequently correlated with coarse sand and a roughness element
characteristic for that reach. Roughness elements such as boulders and cobbles create a flow refuge and were
linked with mussel size.
3. Small mussels tended to be associated with boulder-stabilized habitats and medium sized mussels with
cobble habitats. Large mussels rarely co-occurred with any particular roughness element. Individual species were
strongly linked to geomorphic reach type. This association may be due to species’ differences in ability to colonize
available microhabitat types.
4. The highly tolerant Velesunio ambiguus dominated shale reaches, characterized by fine sediments and human
impacts. In contrast, Hyridella depressa dominated in gorges, utilizing small flow refuges among boulders, while
H. australis were present in low densities across a range of substrate conditions.
5. The persistence of multispecies assemblages in mussel beds throughout the Hawkesbury–Nepean River
implies similar niche utilization among species. Partitioning of habitats across species on the basis of size suggests
some degree of habitat selection, or differential survival. At the local scale, microhabitat characteristics influenced
the size distribution and densities of mussel assemblages. Continuing declines in mussel densities are likely to
result from ongoing channel modification and increased siltation resulting from changes to riparian vegetation.
Copyright # 2008 John Wiley & Sons, Ltd.
Received 25 June 2007; Revised 15 December 2007; Accepted 17 December 2007
KEY WORDS:
freshwater mussels; substrates; flow refuge; microhabitat
*Correspondence to: Shelley Burgin, College of Health and Science, University of Western Sydney, Locked Bag 1797, South Penrith DC, Australia,
2753. E-mail: s.burgin@uws.edu.au
Copyright # 2008 John Wiley & Sons, Ltd.
1286
M. BRAINWOOD ET AL.
INTRODUCTION
Very little is known of the population ecology of Australian
hyriid species (Walker et al., 2001). Like many of the better
known northern hemisphere species, Australian hyriids are
under threat from a range of factors that are detrimentally
affecting their habitat. Catchment-scale changes including
channel modification, flow alteration, clearing of riparian
vegetation, and changes in land use are all occurring with
increasing frequency, and each of these has the potential to
seriously impede the normal functioning of mussel
populations.
Several recent studies have focused on linking freshwater
mussel distribution with a range of habitat factors in the
Hawkesbury–Nepean River, a coastal river in south-eastern
Australia. Brainwood et al. (2006) found that, for 26 sites in
the lower catchment, presence of mussels was adversely
affected by increasing levels of riparian disturbance, and that
distribution of species was related to geomorphic reach type.
In a further study of 64 sites at 32 weirs throughout the greater
Hawkesbury–Nepean catchment, presence of mussels was
linked with location above or below a major impoundment,
and distribution of species and mussel abundance was related
to geomorphic reach type (Brainwood et al., in press). While
different sized mussels were associated also with land use and
level of riparian disturbance, the relationship with geomorphic
reach type and mussel distribution was consistent. Another
study assessed the relationship of mussels with water quality at
28 sites (Brainwood, 2007), and found that high dissolved
oxygen and total phosphorus were related to increased mussel
density. Inclusion of catchment based characteristics indicated
that species distributions were most closely aligned with
geomorphic reach type, and a high level of disturbance to
riparian vegetation was linked with absence of mussels.
Until recently, the role of geomorphic reach type has been
neglected in the study of freshwater mussel distribution
(Gangloff and Feminella, 2007). The relationships between
freshwater mussel distribution and macrohabitat variables
such as land use (Arbuckle and Downing, 2002; McRae et al.,
2004), riparian vegetation (Morris and Corkum, 1999; Poole
and Downing, 2004) and impoundments (Blalock and Sickel,
1996; Layzer and Scott, 2006; Haag and Warren, 2007) have
been used to predict mussel diversity and abundance in large
river systems (Strayer, 1993; Strayer et al., 1994; Di Maio and
Corkum, 1995; Hastie et al., 2003). In smaller drainages,
microhabitat factors such as current velocity, channel depth
and sediment size are thought to be important for predicting
mussel distribution (Neves and Widluk, 1987; Layzer and
Madison, 1995; Johnson and Brown, 2000). These
microhabitat factors provide a localized surrogate for
geomorphic reach type; however, these authors did not
explore the relationship from this perspective.
Copyright # 2008 John Wiley & Sons, Ltd.
Freshwater mussels occur in aggregated multispecies groups
known as mussel beds (Strayer et al., 1994). Distribution of
mussels within these beds is patchy, potentially resulting from
variations in substrate and hydrodynamic conditions at a local
scale (Strayer, 1999; Hardison and Layzer, 2001). Relationships
between mussel distribution and preferred substrates are
complicated and tend to be species-specific (Hastie et al.,
2003). The complexity of interactions among substrate and
hydrological variables makes the influence of these factors on
mussel distribution difficult to identify (Arbuckle and
Downing, 2002; Morales et al., 2006). As a result, the
mechanisms underlying the structuring of freshwater mussel
communities are among the poorest known for riverine
macrofauna, despite a plethora of studies (Haag and Warren,
1998; Hastie et al., 2000, 2003; McRae et al., 2004).
Studies relating unionid distribution with substrate
composition have obtained results ranging from inconclusive
(Brown and Banks, 2001; Box et al., 2002; Howard and Cuffey,
2003) to contradictory (Di Miao and Corkum, 1995; Box and
Mossa, 1999). Trends identified range from shared associations
among species (Hastie et al., 2000; Johnson and Brown, 2000)
to differential selection among species (Bronmark and
Malmqvist, 1982), or even among different age groups within
a species (Hastie et al., 2000). Juvenile unionids that settle into
suitable microhabitats are believed likely to be more successful
during their relatively long life-cycles (Johnson and Brown,
2000). At the river reach level, reduced availability of preferred
microhabitats is associated with reduced diversity and local
abundance of mussels (Downing et al., 2000). However, the
absence of clear microhabitat preference suggested that
freshwater mussels have a broad tolerance of a range of
substrate types, and this is believed to be a key feature of
mussel ecology (Downing et al., 2000).
That there are ecological drivers for habitat partitioning of
mussel species and age classes is well recognized but poorly
understood (Box et al., 2002). Substrate preference may be age
related, although the natural distribution of mussels may be
complicated by other factors that influence their growth or
survival (Downing et al., 2000). Important among these factors
is the need for flow refuges for mussels (Di Maio and Corkum,
1997; Cope et al., 2003). Strayer and Ralley (1993) found little
consistency in the relationship between mussels and a
granulometric description of microhabitat, and recommended
the addition of geomorphic characteristics in quantitative
microhabitat analysis.
Geomorphology gives rise to bedform, aspects of hydrology
including flow, turbulence and shear stress, derived sediment
particle sizes and sediment transport regimes (Gordon et al.,
2004). Channel bedform heavily influences flow rates and
stream energy levels, and the resulting turbulence and shear
stress of the channel bed modify the bedform. On a localized
scale, this channel bed, or substratum patch structure has been
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
GEOMORPHOLOGY ROLE IN SUBSTRATUM PATCH SELECTION BY FRESHWATER MUSSELS
described as habitat architecture (Robson and Barmuta, 1998).
Particle size is derived from the catchment geology, with shales,
mudstones and siltstones generating smaller particles, while
quartzes and sandstones produce larger particles (Hiscock,
2005). Larger particles from upper reaches of the catchment are
continually abraded as they are transported, reducing their size,
and creating a trend towards more uniform particle sizes in
lower catchment reaches (Gordon et al., 2004). Local
conditions may have a sorting effect on particles. Patterns of
granulometry are only significantly changed by floods, and may
actually more truly reflect the last flood than flow effects since
then, except in streams where a large amount of fine sediment is
generated. These high deposition streams include those with a
highly disturbed riparian zone (Gordon et al., 2004).
For freshwater mussels, relationships with geomorphic
reach types are likely to be complex, and may involve
habitat architecture, granulometry, or a combination of both
aspects of reach geomorphology. As an extension of previous
research, this study focused on the distribution and community
structure of mussel species in the Hawkesbury–Nepean River,
Australia. The consistently sympatric distribution of Velesunio
ambiguus, Hyridella australis and Hyridella depressa raises
questions about habitat partitioning, resource partitioning and
synergistic relationships among these species that may enhance
their survival in adverse conditions. Given the detrimental
effects of human modification of habitat on mussel
populations in the Hawkesbury–Nepean system (Brainwood
et al., 2006), these mussels may simply be congregated in
remnant patches of suitable habitat.
The present study explores the role of geomorphically
derived microhabitat factors in determining the distribution of
mussel species in the Hawkesbury–Nepean River. The
relationships of species richness, density, and size class
distribution of mussels with granulometric characteristics and
benthic structure were investigated for sites in different
geomorphic reach types, and around small impoundments to
represent different flow regimes. The aim was to ascertain
patterns in distribution that may be directly associated with
microhabitat types. Particular attention was given to the
location of smaller, younger mussels to determine the role of
size (as a surrogate for age) and habitat partitioning, a poorly
understood aspect of freshwater mussel ecology (Bauer, 2001).
This study is the first to link habitat complexity quantitatively
with freshwater mussel density and size class distribution.
METHODS
Site description
The Hawkesbury–Nepean River is one of the longest coastal
rivers in eastern Australia, consisting of around 320 km of
Copyright # 2008 John Wiley & Sons, Ltd.
1287
main channel (Figure 1) that drains 21 730 km2 of catchment
(HNCMT, 1998). The river is important in the Sydney region,
providing 97% of potable water for the 5 million residents of
the region. By also providing water for a variety of
agricultural, industrial and recreational activities, the river
contributes significantly to the economy of the Sydney region
(Laut et al., 1995). The lower part of the catchment, where this
study is focused, is now highly urbanized. Considerable
demands on the river system have led to the widespread
opinion that the river is under increasing stress (EPA, 1993;
SCA, 2005).
For the Hawkesbury–Nepean, geology influences channel
morphology in two ways. Lithology, the actual composition of
the rock, is the main control on the nature of the sediment
delivered to the channel. This, in turn, determines the nature of
the abrasive and depositional processes associated with
sediment load and boundary sediments. Particularly in an
alluvial channel, such as the sandstone straights, the bank and
bed sediments are representative of the sediment load,
although this is not always the case (Pickup and Warner,
1986). Coarse sediments in high flow velocity channels increase
shear stress, maximizing associated bed shear. However, the
bedrock that forms the valley also constrains the channel.
Thus the highly erosive Wianamatta Shales form sinuous
channels in wide, shallow valleys, and are characterized by
considerable channel migration. In contrast, the more resistant
quartz based sandstones tend to form narrow, deep valleys
with high energy straight channels confined by bedrock
(Erskine, 1998).
Before European settlement the Hawkesbury–Nepean River
was described as creek and river systems with clear water, low
nutrient levels and clean sandy or rocky substrates (Recher
et al., 1993). The catchment was subjected to periodic large
floods, which were followed by deposition of silt and sand
from the flood waters (Howell and Benson, 2000). High flood
variability predisposes the Hawkesbury–Nepean catchment to
instability, creating a non-equilibrium river that is only stable
for short periods of time. Alternating periods of rainfallinduced highs and lows in flow activity, coupled with erodible
channel boundaries, means that an unstable channel bed is the
normal condition for much of the river (Erskine, 1998). The
implications for river management contradict the basic
philosophy that channels should either be stable or be
assisted to become stable since, in the short term, channel
changes are inconvenient for humans and viewed as damaging
in their ecological impact (Warner and Erskine, 1998).
Sites in the Hawkesbury–Nepean catchment were sampled
from a range of geomorphic reaches, and in areas with varying
levels of disturbance to riparian vegetation. Disturbance levels
were summarized as: 530% modification of vegetation
structure=low disturbance; 30–70% modification=medium
disturbance; and >70%=highly disturbed (Brainwood et al.,
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
1288
M. BRAINWOOD ET AL.
Figure 1. Map of Hawkesbury–Nepean River catchment showing distribution of geomorphic reach types, and locations of sites sampled for
freshwater mussels and substrate characteristics during 2005/2006.
Copyright # 2008 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
GEOMORPHOLOGY ROLE IN SUBSTRATUM PATCH SELECTION BY FRESHWATER MUSSELS
2006). Sites were selected from each of the three main
geomorphic reaches present in the lower Hawkesbury–
Nepean River (Erskine, 1998). These comprised four
sandstone gorges, four sandstone straights, and seven shale
reach sites, including three pairs of sites above and below
weirs. Additional sites sampled were subsequently deleted from
the data set as no mussels were found at these sites.
Substrate sampling
Data were recorded from 33 1 m2 quadrats (11 per transect,
three transects per site) at 15 sites in the lower catchment.
Substrate characteristics were determined for each quadrat
prior to sampling mussels (Table 1), and recorded by one of
two divers to reduce variation. Divers compared particle sizes
with a standardized grain size comparator field chart.
Classification of particle sizes was adapted from this
standard, with particles 5500 mm grouped as fine sand, 500–
4000 mm grouped as coarse sand, 5–250 mm classed as cobbles,
and >250 mm as boulders. Macrophytes were identified to
species, and a percentage total cover recorded for the quadrat.
Algal species were not identified. Fine particulate organic
matter (FPOM) was defined as organic material less than 2 mm
in diameter, while coarse particulate organic matter (CPOM)
was larger. Woody debris was classified as plant material with
a diameter greater than 25 mm. A generalized description of
the substratum patch structure was recorded for each quadrat
sampled (Table 1). Of the 11 generalized benthic structures
identified, six were only recorded in one of the three
geomorphic reaches sampled. It was hypothesized that, given
the absence of clear relationship patterns for mussels with a
particular substrate type, the role of patch structure was more
important in creating suitable microhabitat conditions.
Mussel sampling
Quadrats were sampled exhaustively for mussels by divers
using snorkel or SCUBA. Sediments were disturbed to a depth
at which no further mussels were found, generally around
15 cm. Cracks and crevices, undercut banks and around logs
were all searched, and cobbles and smaller woody debris were
moved to gain access to any additional mussels. Divers were
directed to sample each quadrat exhaustively, so that smaller
mussels could be located. The smallest mussels recorded had
valves 17 mm in length. Quadrats were regularly checked by
other divers for additional mussels before they were identified,
weighed, measured, and returned to the same location. Four
species of mussels have been recorded from the Hawkesbury–
Nepean River: V. ambiguus, H. australis, H. depressa and
H. drapeta (McMichael and Hiscock, 1958), although
H. drapeta were not found at sites sampled in this survey.
Copyright # 2008 John Wiley & Sons, Ltd.
1289
Data analysis
Sites were classified based on mean local mussel density as high
density sites with more than 4 mussels m 2 (averaged over
150 m2), or low density sites with less than 4 mussels m 2
(Table 1). Two high density and two low mussel density sites
were analysed from each geomorphic reach, including
sandstone gorges, sandstone straights, and shale straight/
sinuous reaches. Data were analysed to address a number of
questions:
(1) Is benthic structure or granulometry linked with
presence/absence or density of mussels?
(2) Does benthic structure or granulometry dominate in
observed distribution patterns?
(3) Is habitat usage consistent across geomorphic reach
types?
(4) Are observed habitat preferences consistent across
species?
(5) Are observed habitat preferences consistent across size
classes?
Mussels were allocated to size classes based on a range
of shell characteristics, with classes summarized as:
small=540 mm valve length; medium=40–60 mm; and
large=>60 mm length (see Brainwood et al., 2006). When
populations were described by density within size classes
per species, this resulted in nine faunal groups, and enabled
differentiation between ageing populations with mussels
all at, or close to, maximum size, and recruiting populations
with mussels encompassing a range of sizes. Other researchers
have found a significant trend of increasing sampling efficiency
with mussel size, and have attributed an observed lack of
small or juvenile mussels to this (Hastie and Cosgrove,
2002). In this survey, the combination of exhaustive
searching and an enthusiasm for locating recruitment,
resulting in a desire for search effectiveness rather than
sampling efficiency, reduced the likelihood of missing mussels
520 mm length, and was evidenced by the regular discard
of pea clams (Corbicula australis) 510 mm during searching.
Based on Hastie and Cosgrove’s (2002) 50% recovery in the
youngest age class, it is likely that a consistent proportion of
the juvenile mussels present were found, and that this
constituted a representative sample of recruitment presence.
This, however, remains an area for future research in
Australian rivers.
Data were assessed graphically, using residual plots, for
normality, outliers, and compliance with the underlying test
assumptions of the statistical tests used. Statistical significance
was defined as P50.05. Density of mussels was not normally
distributed, and density of mussel size classes within these sites
showed a right skewed distribution characteristic of mussel
populations examined in the Hawkesbury–Nepean River
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
Copyright # 2008 John Wiley & Sons, Ltd.
Variables correlated
Scale
Land use, level of
riparian disturbance,
geomorphic
reach type
Biii, Bix, Bx, Bxi, Gv, Gvii
Level of riparian
disturbance,
geomorphic reach type,
Gv, Gvi, Gvii
Microhabitat versus
catchment
scale (benthic structure)
Microhabitat scale
(benthic structure
vs granulo-metry and
biotic components)
Microhabitat versus
catchment scale
(granulometry and
biotic)
Gv, Gvii
Biii, Bix, Bx
Microhabitat
(benthic structure)
Velesunio ambiguus
Microhabitat scale
(granulometry and biotic
components)
Variables not correlated
(i) Boulder
(ii) Deep undercut
(iii) Shallow undercut
(iv) Fine sand
(v) Fine silt
(vi) Rocky slope with sand
(vii) Sand in boulders
(viii) Silt in boulders
(ix) Silt on clay
(x) Silt on cobbles
(xi) Steep clay slope
Substratum patch
description
Gi, Gii, Giii, Giv, Gvi,
Gviii, Gix
Land use, Gi, Gii, Giii,
Giv, Gviii, Gix
Bi, Bii, Biv, Bv, Bvi,
Bvii, Bviii, Gi, Gii, Giii,
Giv, Gvi, Gviii, Gix
Bi, Bii, Biii, Biv, Bv, Bvi,
Bvii, Bviii, Bix, Bx, Bxi
Bi, Bii, Biv, Bv, Bvi, Bvii,
Bviii, Bxi
Whole mussel assemblage
(v) % algae surface film
(vi) % macrophytes
Macrophyte species
(vii) % FPOM (52 mm)
(viii) % CPOM (>2 mm)
(ix) % woody debris
Organic overlayer (biotic)
(i) % boulders
(5250 mm)/bedrock
(ii) % cobbles and
pebbles (5–250 mm)
(iii) % coarse sand
(500–4000 mmm)
(iv) % fine sand or silt
(5500 mm)
Granulometric
characteristics (G)
Channel bedform
(granulometric)
*
*
*
*
*
*
*
Model correctly predicts mussels
91.6% absence, 36.5% presence
Cox and Snell R2=0.105**
Model correctly predicts mussels
67.1% absence, 76.3% presence
Model correctly predicts mussels
59.2% absence, 76.4% presence
Cox and Snell R2=0.288**
Model correctly predicts mussels
62.3% absence, 83.2% presence
Cox and Snell R2=0.119**
Model correctly predicts mussels
44.9% absence, 76.4% presence
Cox and Snell R2=0.275**
Cox and Snell R2=0.058**
R2, Model predictions
*
*
Sandstone
gorge
*
*
*
Sandstone
gorge
*
*
*
*
*
*
*
*
*
Sandstone
gorge
Geomorphic reach type
Substratum patch structure (B)
Table 1. Results of hypotheses testing using forward stepwise logistic regression for presence/absence of mussel assemblages in the Hawkesbury–Nepean River with
substrate characteristics
1290
M. BRAINWOOD ET AL.
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
Copyright # 2008 John Wiley & Sons, Ltd.
Bxi, Gii, Gv, Gvii, Gix
Level of riparian
disturbance, land use,
geomorphic reach type
Gi, Gv, Gvii, Gix
Microhabitat (benthic
structure) versus reach
scale
Medium (all species
combined)
Microhabitat scale
(granulometry and biotic
components)
Gi, Gii, Giv, Gv, Gvii,
Gviii, Gix
Level of riparian
disturbance, land use,
geomorphic reach type, Bxi
Gi, Gv, Gvii, Gviii, Gix
Level of riparian
disturbance, land use,
geomorphic reach type, Bxi
Gv, Gvii
Level of riparian
disturbance,
land use, geomorphic reach
type
Microhabitat (benthic
structure versus
granulometry and biotic
components)
Small (all species
combined)
Microhabitat scale
(granulometry and biotic
components)
Microhabitat (benthic
structure) versus reach
scale
Hyridella depressa
Microhabitat scale
(granulometry and biotic
components)
Microhabitat (benthic
structure) versus reach
scale
Hyridella australis
Microhabitat scale
(granulometry and biotic
components)
Microhabitat (benthic
structure)
versus reach scale
Gii, Giii, Giv, Gvi, Gviii
Bi, Bii, Biii, Biv, Bv, Bvi,
Bvii, Bviii, Bix, Bx, Bxi
Bi, Bii, Biii, Biv, Bv, Bvi,
Bvii, Bviii, Bix, Bx, Gi, Giii,
Giv, Gvi, Gviii
Giii, Gvi
Bi, Bii, Biii, Biv, Bv, Bvi,
Bvii, Bviii, Bix, Bx
Gii, Giii, Giv, Gvi
Bi, Bii, Biii, Biv, Bv, Bvi,
Bvii, Bviii, Bix, Bx
Gi, Gii, Giii, Giv, Gvi,
Gviii, Gix
Bi, Bii, Biii, Biv, Bv, Bvi,
Bvii, Bviii, Bix, Bx, Bxi
Cox and Snell R2=0.089**
Model correctly predicts mussels
97.6% absence, 27.2% presence
Model correctly predicts mussels
98.5% absence, 19.6% presence
Cox and Snell R2=0.197**
Model correctly predicts mussels
99.8% absence, 12.0% presence
Cox and Snell R2=0.158**
Cox and Snell R2=0.112**
Model correctly predicts mussels
88.2% absence, 67.4% presence
Model correctly predicts mussels
92.2% absence, 39.3% presence
Cox and Snell R2=0.424**
Cox and Snell R2=0.215**
Model correctly predicts mussels
96.4% absence, 14.1% presence
Model correctly predicts mussels
97.6% absence, 11.7% presence
Cox and Snell R2=0.116**
Cox and Snell R2=0.094**
Model correctly predicts mussels
80.7% absence, 70.5% presence
Cox and Snell R2=0.293**
GEOMORPHOLOGY ROLE IN SUBSTRATUM PATCH SELECTION BY FRESHWATER MUSSELS
1291
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
Copyright # 2008 John Wiley & Sons, Ltd.
Level of riparian
disturbance, land use,
geomorphic reach type,
Bvi, Bxi
Microhabitat (benthic
structure) versus reach
scale
Level of riparian
disturbance, land use,
geomorphic reach type, Bxi
Biii, Bix, Bx, Bxi, Gii, Gv,
Gvii
Bi, Bii, Biii, Biv, Bv, Bvi,
Bvii, Bviii, Bix, Bx
Bi, Bii, Biv, Bv, Bvi, Bvii,
Bviii, Gi, Giii, Giv, Gvi,
Gviii, Gix
Gi, Gii, Giii, Giv, Gvi, Gviii,
Gix
Bi, Bii, Biii, Biv, Bv, Bvii,
Bviii, Bix, Bx
Bi, Biii, Biv, Bv, Bvi, Bvii,
Bviii, Giii, Giv, Gvi, Gviii
Model correctly predicts mussels
81.1% absence, 50.4% presence
Model correctly predicts mussels
85.1% absence, 39.7% presence
Cox and Snell R2=0.226**
Model correctly predicts mussels
94.3% absence, 31.3% presence
Cox and Snell R2=0.156**
Cox and Snell R2=0.090**
Model correctly predicts mussels
75.1% absence, 76.5% presence
Model correctly predicts mussels
74.3% absence, 56.1% presence
Cox and Snell R2=0.293**
Model correctly predicts mussels
76.6% absence, 51.6% presence
Cox and Snell R2=0.127**
R2, Model predictions
(B=substratum patch structure, G=granulometry, * indicates substratum patch type is present in this reach; n=495 quadrats; **P50.01, *P50.05).
Microhabitat (benthic
structure) versus reach
Scale
Microhabitat (benthic
structure versus
granulometry and biotic
components)
Gv, Gvii
Bii, Bix, Bx, Bxi, Gi, Gii,
Giii, Gvii, Gix
Large (all species
combined)
Microhabitat scale
(granulometry and biotic
components)
Variables not correlated
Whole mussel assemblage
Variables correlated
Microhabitat (benthic
structure versus
granulometry and biotic
components)
Scale
Table 1. (continued)
1292
M. BRAINWOOD ET AL.
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
GEOMORPHOLOGY ROLE IN SUBSTRATUM PATCH SELECTION BY FRESHWATER MUSSELS
(Brainwood et al., 2006; Brainwood, in press). Data were
characterized by high numbers of zeros recorded for mussel
densities at sites sampled. This ‘zero inflated’ data set is
characteristic of populations with low abundance, and uneven
distribution or overdispersion (Martin et al., 2005). Log (ln)
transformation of mussel density data resulted in
approximately normal distribution of data, but excluded sites
without mussels. As a result, a two-part approach to data
analysis was adopted for this stage (Martin et al., 2005).
A hurdle model was applied to the full data set, so that
mussel distribution was correlated with substratum patch
structure as well as granulometric and biotic characteristics
(Table 1). The first step in this process involved the use of a
forward stepwise logistic regression (SPSS, version 12). This
process produces a model that identifies factors that are highly
correlated with the dependent variable, mussel presence, and
will add or subtract these factors to produce the best model
based on a continuous reappraisal (McNally, 2000). These
models deal with problems associated with colinearity by
condensing the data set to a subset that best correlates with the
presence of mussels, but need to be interpreted cautiously
without attributing causality (McNally, 2000). These analyses
were conducted separately with each group of microhabitat
characteristics, and then in combination with several reach
scale variables (geomorphic reach type, land use, and level of
disturbance to riparian vegetation). This aimed to reduce the
potential for wrongly attributing mussel distribution patterns
with an inaccurate, colinear, variable (Martin et al., 2005). The
second step involved the use of binomial regression (SPSS) to
compare mussel density with environmental factors. Binomial
regression indicates the direction of relationship trends, and
their significance; however, this process may only suggest the
factors that explain mussel distribution patterns (McNally,
2000).
Previous research identified mussel distribution patterns that
showed a high level of correlation with geomorphic reach type
(Brainwood et al., 2006; Brainwood, in press). To explore this
relationship further, the second stage of analysis involved
examining data separately within geomorphic reach types.
BIOENV (PRIMER-E, version 5) was used to compare mussel
densities within faunal groups with microhabitat parameters at
each site. Designed to relate multivariate ecological species
abundance data with environmental variables, it provides an
appropriate nonparametric procedure (Clarke and Ainsworth,
1993; Clarke and Warwick, 2001).
Assessment of sites based on overall density provided an
indication of differences among sites where microhabitats were
under-utilized (low density sites) with sites in the same
geomorphic reach where microhabitats were more fully
utilized (high density sites). When the data set was limited to
quadrats in high density sites the data approximated
normality. Use of a truncated data set such as this also deals
Copyright # 2008 John Wiley & Sons, Ltd.
1293
with any uncertainty regarding the type of zeros in the zero
inflated data (Martin et al., 2005). Having met the underlying
test assumption of normality, ANOVA was used to compare
mussel distribution, for the whole assemblage and for each
species, with habitat utilization for each type of substratum
patch.
RESULTS
Mussel presence and microhabitat factors
A logistic regression was used to relate mussel presence with
results recorded from all quadrats sampled (Table 1). Analysis
with substratum patch structure gave a weak correlation
(R2=0.058) with three variables: shallow undercut, silt on clay,
and silt on cobbles. Including reach scale variables
(geomorphic reach type, land use, and level of riparian
disturbance) in the analysis gave a stronger correlation
(R2=0.275), but only with these three variables. Combining
granulometric/biotic characteristics with patch structure and
reach scale variables showed mussel presence to be linked with
geomorphic reach type, level of riparian disturbance, and three
granulometric/biotic characteristics: % algae, % macrophytes,
% FPOM (R2=0.288; Table 1).
Relationships with microhabitat characteristics were
explored for each species (Table 1). For V. ambiguus, a
relationship for mussel presence was observed with % algae
and % FPOM (R2=0.105). Comparison of microhabitat
descriptors and reach scale variables with mussel presence
showed a stronger relationship (R2=0.293) with the three
reach scale variables only. Hyridella australis also showed a
weak relationship with % algae and % FPOM (R2=0.094).
Inclusion of reach scale variables gave a relationship with
geomorphic reach type, land use, level of riparian disturbance,
and quadrats with steep clay slopes (R2=0.116). For H.
depressa the relationship with granulometry and biotic factors
was markedly different. Mussel presence was linked with %
boulder, % algae, % FPOM, % CPOM, and % woody debris
(R2=0.215). Adding reach scale variables gave a much
stronger correlation with level of riparian disturbance, land
use and geomorphic reach type (R2=0.424), but not with
microhabitat descriptors.
Size classes were compared separately with substrate
conditions (Table 1). For small mussels (540 mm valve
length), a relationship was identified with a group of
granulometric variables and a single patch structure: steep
clay slope, % cobbles, % algae, % FPOM, % woody debris
(R2=0.158). Inclusion of reach scale variables continued to
result in these factors dominating the observed patterns,
excluding microhabitat factors completely (R2=0.197).
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
1294
M. BRAINWOOD ET AL.
Medium sized mussels (40–60 mm) showed a response to less
of the granulometric and biotic conditions (Table 1), having a
relationship identified with % boulders, % algae, % FPOM
and % woody debris (R2=0.197). Inclusion of patch structures
gave a weaker relationship with four patch structures and five
granulometric components (R2=0.127). When reach scale
variables were included in the analysis, the observed
relationship (R2=0.293) was with level of riparian
disturbance, land use, geomorphic reach type, and patches
characterized by a rocky slope with sand, and a steep
clay slope.
Comparison of large mussels (>60 mm) with microhabitat
descriptors revealed a relationship (R2=0.156) with patches
characterized by a deep undercut, silt on clay, silt on cobbles,
steep clay slope, % cobbles, % algae, and % FPOM (Table 1).
Reach scale variables continued to dominate distribution
patterns, but the correlation (R2=0.226) included one patch
structure, steep clay slope.
Mussel density and microhabitat factors
The highest mussel densities were recorded at Menangle
overpass (north side), a sandstone gorge site. Mean density
for 33 quadrats at this site was 50 mussels m 2 (S.D.=30
mussels), and quadrat densities ranged from 6 to 127 mussels
per quadrat. Anecdotal evidence suggests that numerous other
sites along the river historically had similar densities, while
recent surveys indicate that this is no longer the case
(Brainwood et al., 2006; Brainwood, 2007).
Densities for the whole mussel assemblage were positively
correlated with % boulders and % FPOM, and negatively
associated with % fine sand, % algae, and % macrophytes
(Table 2). The relationship was varied among the different
species in the assemblage. For V. ambiguus and H. australis,
increasing mussel densities were linked with increased %
FPOM and % CPOM and decreased % algae, and for H.
Table 2. Correlation
%
%
%
%
%
%
%
%
%
boulders
cobbles
coarse sand
fine sand
algae
macrophytes
FPOM
CPOM
woody debris
of
densities
of
mussels
with
australis, also with decreased % macrophytes. Densities of H.
depressa increased with increased % boulders and % FPOM,
and decreased % fine sand and % CPOM. Consistent
relationships were observed for small and medium mussels,
with increasing mussel densities associated with increased %
boulders and % FPOM, and less % fine sand, % algae, and %
macrophytes. For larger mussels the relationship for mussel
densities was reduced to increasing mussels with greater %
boulders and % FPOM, and decreased % algae (Table 2).
Distribution of mussels in geomorphic reaches in relation
to granulometric and biotic characteristics
For sandstone gorges, mussel density was well correlated with
presence of coarse and fine sand, cobbles and FPOM (range
for rho: 0.347 to 0.569 at low density sites, 0.242 to 0.348 at
high density sites; Table 3). In sandstone straights there was a
weak correlation between mussel density and the presence of
coarse sand and boulders. For shale reaches, mussel density
was well correlated with the presence of boulders, cobbles,
coarse sand and CPOM (r=0.251 to 0.428 at low density sites,
r=0.464 to 0.524 at high density sites; Table 3). For sites
above weirs in shale reaches, boulders, CPOM and woody
debris were most frequently correlated with mussel
distribution, while below weirs highly correlated factors were
coarse sand and cobbles.
The microhabitat characteristics identified as important for
the distribution of mussels differed among size classes
(Table 4). For small mussels the main factors were the
presence of coarse sand and boulders, while for medium
mussels the most important factors were coarse sand, cobbles
and fine sand. For the larger mussels the only frequently
correlated factor was coarse sand. These factors were broadly
consistent among size classes across geomorphic reaches
(Table 4).
substrate characteristics from
Hawkesbury–Nepean River
microhabitat
scale
for
sites
sampled
in
the
Spearmans Rho
all mussels
V. ambiguus
H. australis
H. depressa
all small
all medium
all large
correlation
correlation
correlation
correlation
correlation
correlation
correlation
correlation
correlation
0.217**
ns
ns
0.147**
0.100**
0.086*
0.262**
ns
ns
ns
ns
ns
ns
0.178**
ns
0.157**
0.099**
ns
ns
ns
ns
ns
0.143**
0.113**
0.163**
0.113**
ns
0.386**
ns
ns
0.306**
ns
ns
0.293**
0.196**
ns
0.216**
ns
ns
0.164**
0.113**
0.138**
0.169**
ns
ns
0.212**
ns
ns
0.144**
0.094*
0.124**
0.233**
ns
ns
0.112**
ns
ns
ns
0.089*
ns
0.232**
ns
ns
n=495 quadrats.
**P50.01, *P50.05.
ns=not significant.
Copyright # 2008 John Wiley & Sons, Ltd.
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
1295
GEOMORPHOLOGY ROLE IN SUBSTRATUM PATCH SELECTION BY FRESHWATER MUSSELS
Table 3. Results of BIOENV correlations (Spearmans rho) for mussel distribution and quadrat substrate data for various geomorphic reach types in
the lower Hawkesbury–Nepean River
Site (mussel density)
r (P50.05)
Impacting factors
Sandstone gorges
Bargo Junction (low)
Cataract Junction (low)
Menangle o’pass north (high)
Wilton Park (high)
0.569
0.347
0.348
0.242
%
%
%
%
boulders, % cobbles, % fine sand/clay, % CPOM
cobbles, % coarse sand, % FPOM, % woody debris
boulders, % fine sand/clay, % macrophytes, % CPOM
boulders, % FPOM, % coarse sand
Sandstone straights
Agnes Banks (low)
Leonay (low)
Theresa Park above (high)
Menangle o’pass south (high)
0.048
0.061
0.110
0.086
%
%
%
%
cobbles, % CPOM
boulders, % coarse sand, % woody debris
coarse sand, % algae
boulders, % coarse sand, % fine sand/clay, % algae
Shale sinuous/straight reaches
Thurns weir above (low)
Thurns weir below (low)
Camden weir above (low)
Camden weir below (low)
Sharpes weir above (low)
Sharpes weir below (high)
Cobbitty weir above (high)
0.270
0.251
0.428
0.329
0.342
0.464
0.524
%
%
%
%
%
%
%
boulders, % cobbles, % woody debris
cobbles, % coarse sand, % CPOM
boulders, % coarse sand, % CPOM, % woody debris
cobbles, % coarse sand
fine sand/clay, % CPOM
boulders, % coarse sand
cobbles, % CPOM
n=33 quadrats per site.
Table 4. Comparison of mussel size class distributions with quadrat substrate data using BIOENV (Spearmans ranked correlation, r) for high
mussel density sites sampled in various geomorphic reach types in the Hawkesbury–Nepean River
Site (density)
Small
Medium
Large
r
Factor (%)
r
Factor (%)
r
Factor (%)
Sandstone gorge (high)
Sandstone gorge (high)
0.096
0.105
coarse sand, FPOM
boulders, coarse sand
0.065
0.078
coarse sand, cobbles
coarse sand, woody debris
0.136
0.095
coarse sand, FPOM
coarse sand, woody
debris
Sandstone straight (high)
0.210
0.178
fine sand, algae
0.031
Sandstone straight (high)
0.291
coarse sand, cobbles,
CPOM
boulders, coarse sand,
fine sand, woody debris
0.534
boulders, coarse sand, fine sand,
woody debris
0.138
coarse sand, algae,
CPOM
FPOM, macrophytes
Shale reach (high)
Shale reach (high)
0.075
0.384
coarse sand, boulders, algae
boulders, cobbles, fine sand,
FPOM
0.409
0.378
cobbles, FPOM
boulders, cobbles, fine sand
0.521
0.482
Summary
Frequency of factors more
commonly correlated:
Small
Medium
cobbles, CPOM
boulders, coarse
sand
Large
5
coarse sand
3
coarse sand
4
coarse sand
4
2
2
2
boulder
cobbles
fine sand
FPOM
3
3
2
2
cobbles
fine sand
boulder
woody debris
2
2
FPOM
CPOM
Distribution of mussels in geomorphic reaches in relation
to substrate patch structure
Analysis of the utilization of different types of substratum
patch involved a truncated data set comprising only sites with
Copyright # 2008 John Wiley & Sons, Ltd.
high overall mussel density. High density sites had mussel
populations that maximized the colonization of suitable
microhabitat types, giving a clearer picture of preferred
habitat, and generated a data set that met the test
assumptions for ANOVA. Much of the observed distribution
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
1296
M. BRAINWOOD ET AL.
Table 5. Comparison of mean mussel densities in quadrats with different benthic structure at sites with high density populations in different
geomorphic reach types
Species
F
P5F
Substrate description
Shale reaches (n=66 quadrats)
V. ambiguus
H. australis
H. depressa
All species
11.4200
0.5870
50.0001
0.7677
11.0769
50.0001
Silt cobbles>LWD>steep slope
No significant difference
Not recorded
Silt cobbles>LWD>steep slope>fine silt
Sandstone straights (n=66 quadrats)
V. ambiguus
H. australis
H. depressa
All species
3.4749
2.9007
22.6269
15.7555
0.0127
0.0290
50.0001
50.0001
Shallow undercut>deep undercut
Deep undercut>shallow undercut
Boulder>deep undercut>rocky slope>shallow undercut>fine sand
Deep undercut>shallow undercut>boulder
Sandstone gorge (n=99 quadrats)
V. ambiguus
H. australis
H. depressa
All species
11.6535
2.1487
10.2054
2.9104
50.0001
0.0665
50.0001
0.0174
Sand boulders>shallow undercut>silt boulders
No significant difference
Boulder>rocky slopecdeep undercut> shallow undercut
Sand boulders>boulder>shallow undercut
patterns in shale reaches were derived from habitat usage by V.
ambiguus. Hyridella australis showed no difference in habitat
utilization, and H. depressa was not recorded from these sites.
For V. ambiguus, more mussels were present in quadrats
characterized as silt on cobbles, followed by those with large
woody debris, and less were present in quadrats with a steep
clay slope (Table 5).
In sandstone straights, V. ambiguus and H. australis showed
similar distributions, with more of each of these species found
in deep or shallow undercuts than in other quadrats (Table 5).
For H. depressa, higher densities were recorded in quadrats
characterized by boulders, then deep undercuts and rocky
slopes. Overall, the whole assemblage reflected this habitat
utilization. A different pattern emerged for populations in
sandstone gorges (Table 5). Hyridella australis showed no
significant difference in colonization of types of substratum
patch. Highest densities of V. ambiguus were recorded from
quadrats described as sand in boulders, with less in quadrats
with shallow undercuts, and silt in boulders. For H. depressa,
highest densities were recorded from quadrats with boulders,
less in rocky slopes, and less again in deep undercuts (Table 5).
Recruitment and benthic structure
Small mussels were found in more quadrats in sandstone
gorges than sandstone straights, and relatively few were
recorded from quadrats in shale reaches. The presence of
mussels in medium and large size classes for assemblages in
quadrats with small mussels showed some partitioning based
on size. In shale reaches, higher densities of large mussels
occurred with higher densities of small and medium sized
Copyright # 2008 John Wiley & Sons, Ltd.
mussels, while in sandstone straights and gorges high densities
of large mussels were rarely accompanied by higher densities of
small or medium sized mussels.
Distribution patterns for small mussels within reach types
suggests some association with microhabitat factors. In
sandstone gorges and straights high numbers of small H.
depressa were recorded from quadrats characterized by deep
undercuts and rocky slopes. Low densities of small V.
ambiguus were also recorded in most of the patch structures
that supported small mussels in these reaches, and showed a
similar use of a range of structural habitats in shale reaches. In
contrast, small H. australis were reported from quadrats with a
limited range of patch structures, including deep or shallow
undercut banks and rocky slopes in sandstone reaches, and
deep undercut banks in shale reaches.
DISCUSSION
The results of this study indicate that, overall, there are
consistencies in habitat preferences of mussels across all reach
types. Partitioning of species was clearly evident at a reach
level, but less apparent at a microhabitat scale, where
partitioning appeared to be based as much on mussel size
as species. At this scale, mussel distribution was related to
aspects of patch structure, and granulometry and biotic
characteristics.
The geomorphic reaches present in the Hawkesbury–
Nepean River are similar to those that constitute the main
channel bed forms for many coastal rivers in south-eastern
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
GEOMORPHOLOGY ROLE IN SUBSTRATUM PATCH SELECTION BY FRESHWATER MUSSELS
Australia (Erskine, 1998). Within each geomorphic reach,
mussels showed strong associations with some substrate
characteristics. For sandstone straights, mussel populations
were associated with boulders and coarse sand. In sandstone
gorges the relationship between mussel distribution and
granulometry was stronger, with mussels mainly associated
with coarse and fine sands, and boulders. Boulders create flow
refuges in a sandstone straight, but dominate the gorge benthic
habitat, leaving less room for mussel assemblages. Mussels in
gorges tend to cluster in pockets of sand among the boulders
where the substrate is more stable despite the higher velocity
flow conditions. In shale reaches, mussels were most strongly
linked with boulders, cobbles, coarse sand and CPOM. All of
these factors constitute roughness elements in the finer
sediments suitable for mussel burrowing (Box et al., 2002).
The presence of smaller roughness elements, such as cobbles,
creates flow refuges at the sediment–water interface, increasing
the suitability of the microhabitat for mussels.
Among the mussel species investigated here, each size class
showed some degree of habitat preference. While mussels of all
sizes were most consistently associated with coarse sand,
smaller mussels were also associated with boulders, medium
sized mussels with cobbles, and larger mussels with little else.
Thus, as mussel size increased there was less apparent need for
an association with roughness elements. This suggests that the
dependence on the availability of flow refuges decreases with
increasing mussel size. The relationship between mussel size
and different habitat preference has been reported in other
studies where older (and larger) mussels from several unionid
species were recorded in a wider range of habitats, including
those considered to be less favourable, such as silt and mud
(Downing et al., 2000; Hastie et al., 2000). Strayer et al. (2004)
refer to the possibility of ‘organized’ beds, where larger, firmly
buried adult mussels stabilize sediments within the mussel bed.
Mussel populations in sandstone gorges and shale reaches
showed a stronger association with some types of patch
structure. In sandstone gorges not all parts of the reach
provided suitable habitat, principally due to variations in flow
conditions, but also due to the preponderance of boulders.
Along with bedrock and deep silt beds, boulders have been
long recognized as unsuitable habitat for mussels as these
substrates lack anchorage (Sietman et al., 1999). For shale
reaches some microhabitats also appeared unsuitable, namely
the fine sediments associated with riparian and channel bed
disturbances (Brainwood et al., 2006). In sandstone straights
no type of substratum patch structure was considered
unsuitable for freshwater mussels, and there was no apparent
relationship between mussel distribution and microhabitat
characteristics for this reach type.
Of the three main species in this river, V. ambiguus, known
as the billabong mussel (Walker, 1981), has been recorded
from several coastal rivers, but is more widely distributed
Copyright # 2008 John Wiley & Sons, Ltd.
1297
throughout the Murray–Darling River in central Australia,
where it shows strong habitat preference for slow flowing
shallow reaches often associated with floodplains (McMichael
and Hiscock, 1958; Sheldon and Walker, 1989). In the
Hawkesbury–Nepean, microhabitat preferences for this
species parallel this, with a marked preference for patch
structures dominated by the presence of silt. Higher densities
of V. ambiguus were recorded in shale reaches and other
reaches where flow regimes have been moderated by weirs,
creating conditions where fine sediment particles are not
carried in the water column. In many of these reaches V.
ambiguus dominated the observed mussel assemblage. This
development of distinct assemblages in low-flowing fine
substrate habitats, such as those associated with
impoundments, has been reported elsewhere (Haag and
Warren, 2007).
Hyridella australis and H. depressa have been observed to
occur commonly in the Hawkesbury–Nepean, and are broadly
distributed in eastern coastal rivers of Australia (McMichael
and Hiscock, 1958). Hyridella depressa has been noted to cooccur with H. drapeta, a species that is abundant in other
eastern coastal rivers but sparse in the Hawkesbury–Nepean
(Brainwood, 2007), and not recorded in this study. McMichael
(1955) reported that Hyridella species were predominantly
found in ‘medium flow’ reaches in fine sand. Of these, H.
depressa also occurs in root-stabilized banks, while H. australis
was reported from slow-flowing muddy reaches, generally in
flow refuges associated with shallow banks (Jones, 1983). In
this study, H. australis was recorded from the fine sediment
shale reaches, predominantly in undercut banks. However, it
was reported in higher densities from sandstone gorges and
straights, and was broadly distributed throughout the
available habitat types from both sandstone reaches. This
suggests it is a habitat generalist, and concurs with previous
observations in relation to water quality that the species is
physiologically robust, and able to survive in a range of
conditions (Brainwood, 2007). In contrast, H. depressa was
absent from the shale reaches in this study, and in general
(unpublished data). Instead, higher densities of H. depressa
were observed in sandstone gorges in association with patch
structures dominated by a roughness element that provided a
flow refuge.
Providing evidence of recruitment, small mussel distribution
showed strong differentiation across microhabitat types.
Described as a ‘boom-bust’ species (Byrne, 1998), H.
depressa is smaller overall and shows less resistance to
adverse conditions (Brainwood et al., 2006, in press;
Brainwood, 2007). Generally, small H. depressa were present
in high numbers in stable flow refuges created by deep
undercut banks, rocky slopes and boulders. In contrast, low
numbers of small V. ambiguus and H. australis were present
across the range of microhabitats in sandstone reaches. In high
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
1298
M. BRAINWOOD ET AL.
density shale reaches, however, small H. depressa were absent
and small H. australis were present only in deep undercuts, a
comparatively stable flow refuge. Small V. ambiguus were
present across a wide range of microhabitats with higher
densities in stabilized habitats. Higher numbers of small H.
depressa implies recruitment in recent times and suggests that
an increased rate of reproduction, perhaps coupled with a
behaviourally derived juvenile resilience, provides an
important strategy for survival of this species. This appears
in direct contrast with V. ambiguus, where adult resilience is
characteristic (McMichael and Hiscock, 1958; Sheldon and
Walker, 1989).
In addition to the variations observed in species distribution
among benthos types associated with different geomorphic
reach types, mussel species showed varied relationships with
aspects of granulometry and biotic characteristics. While the
whole mussel assemblage was positively associated with the
presence of boulders, and negatively associated with fine sand,
only H. depressa reflected this relationship as a single species.
The absence of a strong link for V. ambiguus and H. australis
with any aspect of granulometry supports their status as
habitat generalists. The consistent relationship observed
between mussel density and increased levels of FPOM and
decreased levels of algae raises several questions. Increasing
mussel density associated with decreased algal presence implies
consumption of a food resource (Nichols and Garling, 2000;
Christian et al., 2004), but does not identify the density at
which mussel populations are able to affect their ecosystem
significantly (see Brainwood, 2007). The observed relationship
between mussel density and FPOM may be an artefact of the
keystone role that mussels play in freshwater ecosystems
(Vaughn and Hakenkamp, 2001; Gutierrez et al., 2003).
Recent research in UK lowland rivers suggested that
macroinvertebrate taxon richness is strongly associated with
increased mussel density for these streams (Aldridge et al.,
2007). The observed increase in diversity comprised taxa across
a range of feeding guilds, including those involved in the
breakdown of coarse organic matter to fine organic matter.
Interestingly, in the Hawkesbury–Nepean, mussel species
showed different responses to increased levels of CPOM.
Both V. ambiguus and H. australis were present in higher
numbers with more CPOM, while H. depressa demonstrated
the opposite relationship. It is unclear whether this is in
response to variations in flow (H. depressa densities were
greater in gorges), or whether this species can become
smothered by a layer of coarse organic debris.
Natural distribution of mussels is undoubtedly complicated
by a number of factors that influence their growth and
survival. Thus, within a hierarchy of potential impacting
factors, substrate conditions may not be the dominant
determinant of mussel population distribution (Box et al.,
2002). This was observed in the Hawkesbury–Nepean River
Copyright # 2008 John Wiley & Sons, Ltd.
where variables, such as geomorphic reach type and level of
riparian disturbance, that operate at a reach scale were
consistently more strongly associated with mussel
distribution than most microhabitat variables. As well, there
was little consistency in the microhabitat variables that
dominated within this hierarchy, or even among groups of
variables. As a result, neither patch structure nor
granulometric and biotic characteristics dominated the
relationships observed for mussel distribution with
microhabitat. Both aspects of geomorphology, however, were
relevant in the determination of species and size class
distribution.
Hastie et al. (2004) observed that the abundance and
distribution of freshwater mussels was not limited by the
availability of suitable habitat. Lack of limitation by habitat
availability was offered as an explanation for lower than
expected mussel densities at optimal locations. The implication
is that population density is not necessarily defined by the
availability of suitable habitat, but may also be influenced by
other potentially limiting factors such as the absence of host
fish. While the results of this study indicated some habitat
preference for local mussel species, identification of suitable
habitat is difficult owing to the influence of other unknown
factors including host availability.
Partitioning of habitat by different unionid species is
generally minimal, with multiple species assemblages the
most common condition (Strayer and Ralley, 1993). The
absence of partitioning, at a microhabitat scale, in favour of
multi-species mussel assemblages was observed among the
hyriids examined in this study. Non-partitioning of habitat by
mussels in the Hawkesbury–Nepean may be due to the general
abundance of food resources, typical of most rivers (Strayer,
1981). Since juveniles have little control over the process of
dispersal there may be selective pressure for generalized
habitat use, exhibited by two of the species present.
Moreover, with little competition for food there is little
benefit to be derived from habitat specialization. Alternatively,
multispecies assemblages may gain some advantage over single
species assemblages, with benefits such as sediment stability
derived from distribution of differently sized mussels.
Results of this study indicate that microhabitat conditions
may affect the size class representation of local mussel
populations. However, neither granulometry and biotic
characteristics nor patch structures are dominant in defining
this population structure. With the exception of the absence of
mussels from clearly unfavourable microhabitats, mussel
presence (or absence) does not appear to be strongly linked
to any particular habitat characteristic. Access to stable
substrates such as coarse sand is important, and the need for
physical flow refuges apparently decreases with increasing
mussel size. Selection of different habitat conditions in
different geomorphic reaches indicates that within a
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
GEOMORPHOLOGY ROLE IN SUBSTRATUM PATCH SELECTION BY FRESHWATER MUSSELS
hierarchy of impacts for freshwater mussels, microhabitat
conditions are not dominant, although this should not be
interpreted as unimportant.
The risk of extinction is not imminent for many Australian
species, except H. glenelgensis (Playford and Walker, 2007).
However, the loss of local populations has serious implications
for riverine health and management, both in Australia and
overseas. Catchment scale changes such as channel
modification, flow alteration, clearing of riparian vegetation,
and changes in land use, are all reflected in changes in habitat
conditions at the patch scale. These changes are occurring with
increasing frequency, and each may seriously impede the
normal functioning of mussel populations. Regardless of
causal factors for declines in mussel populations, successful
conservation remains strongly dependent on a better
understanding of the factors that enhance their growth and
reproduction (Warren and Haag, 2005). The comparatively
long lifespan of freshwater mussels means that deleterious
mechanisms that occur infrequently, including burial, crushing
or dislodgement, can have a significant impact on a mussel
population (Strayer et al., 2004). Development of management
actions that increase stream stability has the potential to
facilitate population growth rather than the declines currently
being documented (Haag and Warren, 2007). This study
demonstrates that these management processes need to include
the maintenance of favourable habitat structure and
conditions, even at a patch scale. While the results of this
study are broadly similar to those from many northern
hemisphere studies, this research fills a key knowledge gap
for members of this family.
ACKNOWLEDGEMENTS
The authors would like to thank Penrith City Council and
NSW Fisheries for financial support. Thanks also to the divers
for their assistance, especially Caroline Forest, Geoff Hunter,
Becca Saunders, and Mark Spencer.
REFERENCES
Aldridge DC, Fayle TM, Jackson N. 2007. Freshwater mussel
abundance predicts biodiversity in UK lowland rivers.
Aquatic Conservation: Marine and Freshwater Ecosystems
17: 554–564.
Arbuckle KE, Downing JA. 2002. Freshwater mussel
abundance and species richness: GIS relationships with
watershed land use and geology. Canadian Journal of
Fisheries and Aquatic Sciences 59: 310–316.
Bauer G. 2001. Factors affecting naiad occurrence and
abundance. Ecology and Evolution of the Freshwater
Copyright # 2008 John Wiley & Sons, Ltd.
1299
Mussels Unionoida, Bauer G, Wachtler K (eds). Springer:
Berlin; 155–162.
Blalock HN, Sickel JB. 1996. Changes in mussel (Bivalvia:
Unionidae) fauna within the Kentucky portion of Lake
Barkley since impoundment of the lower Cumberland River.
American Malacological Bulletin 13: 111–116.
Box JB, Mossa J. 1999. Sediment, land use, and freshwater
mussels: prospects and problems. Journal of North American
Benthological Society 18: 99–117.
Box JB, Dorazio RM, Liddell WD. 2002. Relationships between
streambed substrate characteristics and freshwater mussels
(Bivalvia: Unionidae) in Coastal Plain streams. Journal of the
North American Benthological Society 22: 253–260.
Brainwood M. 2007. The hierarchical effects of anthropogenic
influences and natural filters on the distribution of
freshwater mussels in the Hawkesbury–Nepean River,
NSW, Australia. PhD thesis, University of Western
Sydney, Richmond, Australia.
Brainwood M, Burgin S, Byrne M. 2006. Is the decline of
freshwater mussel populations in a regulated coastal river in
south-eastern Australia linked with human modification of
habitat? Aquatic Conservation: Marine and Freshwater
Ecosystems 16: 501–516.
Brainwood M, Byrne M, Burgin S. in press. The impact of
small and large impoundments on freshwater mussel
distribution
in
the
Hawkesbury–Nepean
River,
southeastern Australia. River Research and Applications.
Bronmark C, Malmqvist B. 1982. Resource partitioning
between unionid mussels in a Swedish lake outlet.
Holarctic Ecology 5: 389–395.
Brown KM, Banks PD. 2001. The conservation of unionid
mussels in Louisiana rivers: diversity, assemblage
composition and substrate use. Aquatic Conservation:
Marine and Freshwater Ecosystems 11: 189–198.
Byrne M. 1998. Reproduction of river and lake populations of
Hyridella depressa (Unionacea: Hyriidae) in New South
Wales: implications for their conservation. Hydrobiologia
389: 29–43.
Christian AD, Smith BN, Berg DJ, Smoot JC, Findlay RH.
2004. Trophic position and potential food sources of 2
species of unionid bivalves (Mollusca: Unionidae) in 2 small
Ohio streams. Journal of North American Benthological
Society 23: 101–113.
Clarke KA, Ainsworth M. 1993. A method of linking
multivariate community structure to environmental
variables. Marine Ecology Progress Series 92: 205–219.
Clarke KR, Warwick RM. 2001. Change in marine
communities: an approach to statistical analysis and
interpretation. PRIMER-E, Plymouth, UK.
Cope WG, Hove MC, Waller DL, Hornbach DJ, Bartsch MR,
Cunningham LA, Dunn HL, Kapuscinski AR. 2003.
Evaluation of relocation of unionid mussels to in situ
refugia. Journal of Molluscan Studies 69: 27–34.
Di Maio J, Corkum LD. 1995. Relationship between the
spatial distribution of freshwater mussels (Bivalvia:
Unionidae) and the hydrological variability of rivers.
Canadian Journal of Zoology 73: 663–671.
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
1300
M. BRAINWOOD ET AL.
Di Maio J, Corkum LD. 1997. Patterns of orientation in
Unionids as a function of rivers with differing hydrological
variability. Journal of Molluscan Studies 63: 531–539.
Downing JA, Van Leeuwen H, Di Paolo LA. 2000.
Substratum patch selection in the lacustrine mussels
Elliptio complanata and Pyganodon grandis grandis.
Freshwater Biology 44: 641–648.
EPA. 1993. An inventory of pollutant sources in the
Hawkesbury–Nepean River Catchment. Technical Report
No. 94/103, Environmental Protection Authority of New
South Wales, Chatswood, Australia.
Erskine W. 1998. Channel morphology. Geomorphology of the
Hawkesbury–Nepean River System. Hawkesbury–Nepean
Catchment Management Trust: Windsor, Australia; 27–48.
Gangloff MM, Feminella JW. 2007. Stream channel
geomorphology influences mussel abundance in southern
Appalachian streams, USA. Freshwater Biology 52: 64–74.
Gordon ND, McMahon TA, Finlayson BL, Gippel CJ,
Nathan RJ. 2004. Stream Hydrology: An Introduction for
Ecologists. John Wiley: Chichester.
Gutierrez JL, Jones CG, Strayer DL, Iribarne OO. 2003.
Mollusks as ecosystem engineers: the role of shell production
in aquatic habitats. Oikos 101: 79–90.
Haag WR, Warren ML. 1998. Role of ecological factors and
reproductive strategies in structuring freshwater mussel
communities. Canadian Journal of Fisheries and Aquatic
Sciences 55: 297–306.
Haag WR, Warren ML. 2007. Freshwater mussel assemblage
structure in a regulated river in the Lower Mississippi River
Alluvial Basin, USA. Aquatic Conservation: Marine and
Freshwater Ecosystems 17: 25–36.
Hardison BS, Layzer JB. 2001. Relations between complex
hydraulics and the localised distribution of mussels in three
regulated rivers. Regulated Rivers: Research and
Management 17: 77–84.
Hastie LC, Boon PJ, Young MR. 2000. Physical microhabitat
requirements of freshwater pearl mussels, Margaritifera
margaritifera (L.). Hydrobiologia 429: 59–71.
Hastie LC, Cooksley SL, Scougall F, Young MR, Boon PJ,
Gaywood MJ. 2003. Characterization of freshwater pearl
mussel (Margaritifera margaritifera) riverine habitat using
River Habitat Survey data. Aquatic Conservation: Marine
and Freshwater Ecosystems 13: 213–224.
Hastie LC, Cooksley SL, Scougall F, Young MR, Boon PJ,
Gaywood MJ. 2004. Applications of extensive survey
techniques to describe freshwater pearl mussel distribution
and macrohabitat in the River Spey, Scotland. River
Research and Applications 20: 1001–1013.
Hastie LC, Cosgrove RJ. 2002. Intensive searching for mussels
in a fast flowing river: an estimation of sampling bias.
Journal of Conchology 38: 309–316.
Hiscock KM. 2005. Hydrogeology: Principles and Practice.
Blackwell Science: Carlton, Australia.
HNCMT. 1998. Geomorphology of the Hawkesbury–Nepean
River System: a review of landforms, processes and management. Hawkesbury–Nepean Catchment Management Trust,
Windsor, Australia.
Copyright # 2008 John Wiley & Sons, Ltd.
Howard JK, Cuffey KM. 2003. Freshwater mussels in a
California North Coast Range river: occurrence,
distribution, and controls. Journal of the North American
Benthological Society 22: 63–77.
Howell J, Benson D. 2000. Predicting potential impacts of
environmental flows on weedy riparian vegetation of the
Hawkesbury–Nepean River, southeastern Australia. Austral
Ecology 25: 463–475.
Johnson PD, Brown KM. 2000. The importance of
microhabitat factors and habitat stability to the threatened
Louisiana pearl shell, Margaritifera hembeli (Conrad).
Canadian Journal of Zoology 78: 271–277.
Jones HA. 1983. The reproductive biology of the freshwater
mussels (Pelecypoda: Hyriidae) of the Macleay River,
Northern N.S.W., with special reference to the reproductive
cycle in Cucumerunio novaehollandiae (Gray). BSc(Hons),
University of New England, Armidale, Australia.
Laut P, Marston R, Cuddy S. 1995. Land cover and
infrastructure in the Hawkesbury–Nepean Basin. Division
of Water Resources, Report 95/3, CSIRO, Canberra.
Layzer JB, Madison LM. 1995. Microhabitat use by
freshwater mussels and recommendations for determining
their instream flow needs. Regulated Rivers: Research and
Management 10: 329–345.
Layzer JB, Scott EMJ. 2006. Restoration and colonisation of
freshwater mussels and fish in a southeastern United States
tailwater. River Research and Applications 22: 475–491.
McMichael DF. 1955. A revision of the freshwater mussels of
the Australasian region. PhD, Harvard University,
Cambridge, Massachusetts, USA.
McMichael DF, Hiscock ID. 1958 A monograph of
the freshwater mussels (Mollusca: Pelecypoda) of the
Australian region. Australian Journal of Marine and
Freshwater Research 9: 372–508.
McNally R. 2000. Regression and model-building in
conservation biology, biogeography and ecology: the
distinction between - and reconciliation of – ‘predictive’ and
‘explanatory’ models. Biodiversity and Conservation 9: 655–671.
McRae SE, Allan JD, Burch JB. 2004. Reach- and catchmentscale determinants of the distribution of freshwater mussels
(Bivalvia: Unionidae) in south-eastern Michigan, U.S.A.
Freshwater Biology 49: 127–142.
Martin TG, Wintle BA, Rhodes JR, Kuhnert PM, Field SA,
Low-Choy SJ, Tyre AJ, Possingham HP. 2005. Zero
tolerance ecology: improving ecological inference by
modelling the source of zero observations. Ecology Letters
8: 1235–1246.
Morales Y, Weber LJ, Mynett AE, Newton TJ. 2006. Effects
of substrate and hydrodynamic conditions on the formation
of mussel beds in a large river. Journal of the North American
Benthological Society 25: 664–676.
Morris TJ, Corkum LD. 1999. Unionid growth patterns in
rivers of differing riparian vegetation. Freshwater Biology 42:
59–68.
Neves RJ, Widluk JC. 1987. Habitat ecology of juvenile
freshwater mussels (Bivalvia: Unionidae) in a headwater
stream in Virginia. American Malacological Bulletin 5: 1–7.
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
GEOMORPHOLOGY ROLE IN SUBSTRATUM PATCH SELECTION BY FRESHWATER MUSSELS
Nichols SJ, Garling D. 2000. Food-web dynamics and trophiclevel interactions in a multispecies community of freshwater
unionids. Canadian Journal of Zoology 78: 871–882.
Pickup G, Warner RF. 1986. Effects of hydrologic regime on
magnitude and frequency of dominant discharge. Journal of
Hydrology 29: 51–75.
Playford TJ, Walker KF. 2007. Status of the endangered
Glenelg River mussel. Hyridella glenelgensis (Unionidae:
Hydriidae) in Australia. Aquatic Conservation: Marine and
Freshwater Ecosystems DOI: 10.1002/aqc.877.
Poole KE, Downing JA. 2004. Relationship of declining
mussel biodiversity to stream-reach and watershed
characteristics in an agricultural landscape. Journal of the
North American Benthological Society 23: 114–125.
Recher HF, Hutchings PA, Rosen S. 1993. The biota of the
Hawkesbury–Nepean catchment: reconstruction and
restoration. Australian Zoologist 29: 3–42.
Robson BJ, Barmuta LA. 1998. The effects of two scales of
habitat architecture on benthic grazing in a river. Freshwater
Biology 39: 207–220.
SCA. 2005. Annual Water Quality Monitoring Report, 2004–
2005. Sydney Catchment Authority, Penrith, Australia.
Sheldon F, Walker KF. 1989. Effects of hypoxia on oxygen
consumption by two species of freshwater mussel
(Unionacea: Hyriidae) from the River Murray. Australian
Journal of Marine and Freshwater Research 40: 491–499.
Sietman BE, Furman MA, Pursell FA. 1999. Colonisation of
bedrock by freshwater mussels (Bivalvia: Unionidae).
American Midland Naturalist 141: 209–211.
Strayer DL. 1981. Notes on the microhabitats of unionid
mussels in some Michigan streams. American Midland
Naturalist 106: 411–415.
Strayer DL. 1993. Macrohabitats of freshwater mussels (Bivalvia:
Unionacea) in streams of the northern Atlantic Slope. Journal
of North American Benthological Society 12: 236–246.
Copyright # 2008 John Wiley & Sons, Ltd.
1301
Strayer DL. 1999. Use of flow refuges by unionid mussels in
rivers. Journal of North American Benthological Society 18:
468–476.
Strayer DL, Ralley J. 1993. Microhabitat use by an assemblage
of stream dwelling unionaceans (Bivalvia), including two
rare species of Alasmidonta. Journal of North American
Benthological Society 12: 247–258.
Strayer DL, Hunter DC, Smith LC, Borg CK. 1994.
Distribution, abundance, and roles of freshwater clams
(Bivalvia: Unionidae) in the freshwater tidal Hudson
estuary. Freshwater Biology 31: 239–248.
Strayer DL, Downing JA, Haag WR, King TL, Layzer JB,
Newton TJ, Nichols SJ. 2004. Changing perspectives on
pearly mussels, North America’s most imperiled animals.
Bioscience 54: 429–440.
Vaughn CC, Hakenkamp CC. 2001. The functional role of
burrowing bivalves in freshwater ecosystems. Freshwater
Biology 46: 1431–1446.
Walker KF. 1981. Ecology of freshwater mussels in the River
Murray. Australian Water Resources Council Technical
Paper 63. Australian Water Resources Council Technical
Paper, Canberra, Australia.
Walker KF, Byrne M, Hickey CW, Roper DS.
2001. Freshwater mussels (Hyriidae) of Australasia.
Ecology and Evolution of the Freshwater Mussels
Unionoida. Bauer G, Wachtler K (eds). Springer: Berlin;
145, 5–31.
Warner RF, Erskine WD. 1998. Channel changes.
Geomorphology of the Hawkesbury–Nepean River System:
A review of landforms, processes and management.
Hawkesbury–Nepean Catchment Management Trust:
Windsor, Australia; 49–68.
Warren MC, Haag WR. 2005. Spatio-temporal patterns of the
decline of freshwater mussels in the Little South Fork
Cumberland River, USA. Biodiversity and Conservation 14:
1383–1400.
Aquatic Conserv: Mar. Freshw. Ecosyst. 18: 1285–1301 (2008)
DOI: 10.1002/aqc
Документ
Категория
Без категории
Просмотров
5
Размер файла
190 Кб
Теги
patch, rivers, selection, geomorphology, mussel, south, substrate, уnepean, new, wales, freshwater, hawkesbury, role, australia
1/--страниц
Пожаловаться на содержимое документа