RESEARCH COMMUNICATIONS RESEARCH COMMUNICATIONS
Colonization and development of an
Alaskan stream community over 28 years
Alexander M Milner1,2*, Anne L Rober tson3, Kieran A Monaghan4, Amanda J Veal5, and Elizabeth A Flory6
The application of general successional theory to stream ecosystems has not been widely addressed due to a lack
of long-term studies on stream channels at sufficiently large spatial scales. Wolf Point Creek in Glacier Bay,
Alaska, a lake-fed stream that began to emerge from under glacial ice in the mid-1940s, offers an opportunity to
address this imbalance. We examine the stream’s development from 1977 to 2005, with reference to concepts of
succession and community assembly. Dispersal constraints have influenced the succession, as non-insect taxa
required at least 20 years to colonize. We suggest that tolerance is a major mechanism of macroinvertebrate
community assembly. Most taxa, with the exception of the cold-tolerant first colonizers, have persisted within
the community following colonization, although relative abundance has changed markedly with time.
However, biotic processes do influence colonization and succession. Redd (nest) digging by spawning salmon
creates disturbed patches that facilitate the persistence of some early colonizers, and riparian vegetation facilitates colonization by caddisflies and chironomids. We further suggest that both deterministic and stochastic
elements influence succession and community assembly in streams. Our study highlights the importance of reestablishing riparian vegetation during stream restoration programs and of increased in-stream habitat complexity from inputs of coarse woody debris to improve nutrient retention, particularly of salmon carcasses.
Front Ecol Environ 2008; 6(8): 413–419, doi:10.1890/060149
S
uccession is a key concept in terrestrial ecology
(Chapin et al. 1994) and its study has revealed
mechanisms by which communities are assembled,
including the relative roles of deterministic and stochastic pathways. Connell and Slatyer (1977) proposed
three models of succession: (1) facilitation, in which
early species modify the environment to make it more
suitable for later colonizers; (2) tolerance, in which, as
the environment changes, established species exhibit a
progressive tolerance of invading species; and (3) inhibition, in which early colonizers restrict the invasion of
later colonizers. To date, stream ecology has contributed
little to general ecological theory (Fisher 1997). Here,
we begin to address this imbalance by examining a longterm stream colonization record in the context of successional theory.
Glacier Bay National Park and Preserve in southeast
Alaska is a unique natural laboratory for the study of
aquatic ecosystems and the processes of primary succession. Rapid glacial recession has created a de-glaciated
landscape with a temporal scale of 220 years across a spatial scale of 11 000 km2, providing a “blank slate” to study
changes in species composition (Platt and Connell
2003). We have studied Wolf Point Creek (WPC) in
1
School of Geography, Earth and Environmental Sciences, University of Birmingham, Edgbaston, Birmingham, UK; 2Institution of
Arctic Biology, University of Alaska, Fairbanks, AK
*
([email protected]); 3School of Human and Life Sciences,
Roehampton University, London, UK; 4Departmento de Biologia,
Universidade de Aveiro, Aveiro, Portugal; 5The Environment Agency,
Exeter, UK; 6Juneau, AK
© The Ecological Society of America
Glacier Bay since 1977, providing the longest continuous
record of stream colonization within a primary successional framework (1977 to 2005) of which we are aware.
Primary succession can be defined as colonization and
subsequent change in dominance within a community at
a specific site where no remnants of a biotic community
initially exist (Fisher 1983). This differs from secondary
succession, in which elements of the original biotic community persist following disturbance. In this study, we
focus on macroinvertebrates, copepods (meiobenthic
crustaceans retained by a 63-µm sieve), and fish.
Colonization and primary succession studies of streams
and rivers are rare for these groups, particularly at the spatial scale of entirely new river channels (Fisher 1990;
Robertson and Milner 2006). Salmon colonization and
establishment within streams represents a major nutrient
influx that can influence stream and riparian food webs,
thereby potentially influencing succession and community structure (Naiman et al. 2002).
There is little understanding of the mechanisms driving
succession in stream ecosystems. Here, we address
whether the models of succession proposed by Connell
and Slatyer (1977) are appropriate to stream ecosystems
using invertebrate data collected in WPC over 28 years
and whether mechanisms of succession in this system differ from those of other systems. We also examine whether
colonization patterns in stream development are deterministic or stochastic. The integration of ecological theory into stream restoration approaches has generally been
lacking (Lake et al. 2007); we therefore also link our findings on stream succession to implications for stream
restoration.
www.frontiersinecology.or g
413
Stream community succession
414
(a)
(c)
AM Milner et al.
(b)
Figur e 1. Wolf Point Creek at different stages of floodplain
succession: (a) bare gravel in 1981, (b) small clumps of alder
and willow in 1988, and (c) dominated by alder and willow
reaching over 3 m in 1997.
Physical variables
Methods
Site description
During the Little Ice Age, around 1700 AD, glacial
advance reached the mouth of Glacier Bay, followed by an
extensive retreat uncovering a fjord over 100 km long.
This recession left remnant ice sheets in valleys and the
resultant meltwaters fed new streams. One of these
streams, on the west side of Muir Inlet, was WPC, whose
mouth was uncovered in the mid-1940s. Lawrence Lake,
the lake that now feeds the stream, emerged in the early
1970s and is currently 1.45 km2 in size, with a maximum
depth of 35 m. WPC is approximately 2 km in length, 6 to
10 m wide, and flows over glacial moraine, till, and outwash deposits. In 1977, the lower floodplain was essentially
barren (Figure 1a). Isolated clumps of alder (Alnus crispa)
and willow (Salix spp) were evident on upper terraces,
where mats of mountain aven (Dryas spp) were nearly continuous. By 1988, the lower terraces supported a few
clumps of alder and willow (Figure 1b) but by 1997 these
plants were dominant, with riparian plants exceeding 3 m
in height (Figure 1c). In 2005, the upper terraces supported increasing numbers of cottonwood trees (Populus
trichocarpa), along with the occasional Sitka spruce (Picea
sitchensis). In 20 to 30 years, we predict that cottonwood
and Sitka spruce will dominate valley vegetation, with
alder and willow restricted to the riparian corridor.
www.fr ontiersinecology.or g
Initially, water temperature was recorded by spot measurements with hand-held thermometers, but since 1992,
data loggers have been used to continually record daily
means and maxima. Annual degree-days were determined, from 1992 onward, using mean water temperature
for any calendar day, summed over the entire year.
Habitat heterogeneity and the occurrence of pools and
backwaters was estimated by comparing data from Sidle
and Milner (1989) with that of Milner et al. (2000) and
by using available aerial imagery over the study period.
Invertebrates
From 1977, macroinvertebrates were collected yearly, in
August or early September (except 1984–85, 1987, 1995,
and 2003), from a representative sampling station located
0.75 km from the stream mouth, typically using a
Surber net (ten replicates with a 330-µm mesh). However,
between 1978 and 1983, qualitative samples were taken
with a pond net. Between 1994 and 2002, microcrustacea
were collected yearly from the same sampling station, during the same time period, with a 63-µm mesh Surber sampler (five replicates). Invertebrates were preserved in 70%
ethanol and, in the laboratory, sorted from detritus and
inorganic matter and identified. Chironomid larvae were
identified using methods outlined in Milner et al. (2000).
Fish
Numbers of spawning adult salmon were estimated by regular foot counts along the stream length, throughout the
study period. In southeast Alaska, the number of spawning
pink salmon (Oncorhynchus gorbuscha) is an order of magnitude higher in odd years than in even years. To investigate the potential effect of redd (nest) digging for egg depo© The Ecological Society of America
AM Milner et al.
Stream community succession
415
Coho salmon
Pink salmon
Dolly Varden
Chum
salmon
Coastrange
sculpin
Figure 2. Year of first colonization by fish and macroinvertebrate orders and families with corresponding maximum water
temperature. Total number of macroinvertebrate taxa in August for any particular year is given in purple.
sition by pink salmon, macroinvertebrates were collected
from the streambed prior to, during, and after peak redd
construction, using six replicate Surber samples in years
when salmon abundance was high (1997) and low (1996
and 1998). Duplicate drift samples were collected downstream from observed redds, over a 24-hr period, between
the end of July and early September in 1997 and 1998.
Marine-derived nutrient sources
In 1998, marked salmon carcasses were staked into the
streambed to investigate their role as a potential food source
for macroinvertebrates through direct scavenging. Tagged
carcasses were experimentally released to estimate carcass
retention. In 1998 and 2004, samples of algae, juvenile fish,
and macroinvertebrates were collected for 15N stable isotope analyses, following the approach of Milner et al. (2000).
Data analysis
The relationship between macroinvertebrate taxa richness, cumulative macroinvertebrate taxa added to the
assemblage, and mean harpacticoid copepod abundance
© The Ecological Society of America
and stream temperature were explored using linear regression analysis. As an estimate of year-to-year community
similarity, Jaccard’s similarity coefficients were calculated
for corresponding year pairs, where a value of 0 indicates
no similarity and a value of 1 indicates identical communities. Compositional stability of macroinvertebrate communities (similarity over time with respect to relative taxa
abundance) was investigated using Bray–Curtis distances.
Bray–Curtis (scale 0 to 1) is a measure of dissimilarity,
with high values indicating low compositional stability.
To estimate macroinvertebrate persistence, the number of
taxa was plotted against year of collection and compared
to the number of those taxa still present in 2005.
Results
Water temperature rose from a maximum of 2˚C in August
1977 to 18.6˚C in August 2003, as the upstream lake
increased in size, resulting in an increase in the number of
annual degree days for the stream from < 500 Centigrade
temperature units (CTU; estimated) to 1945 CTU. In
1977, habitats were initially dominated by riffles and runs,
but the frequency of pools and backwaters increased over
www.frontiersinecology.or g
Stream community succession
416
AM Milner et al.
time, particularly between 2000 and 2005, when they
formed approximately 20% of the available habitat.
The first macroinvertebrate colonizers were chironomids (genus Diamesa and Paratrichocladius), followed by
mayflies (Baetidae) and stoneflies (Capniidae) between
1984 and 1986, when maximum water temperature
reached 6.0 ˚C (Figure 2). Blackflies (Simuliidae) and
chloroperlid stoneflies were first collected in 1988. The
first macroinvertebrate non-insects to colonize WPC
were Oligochaeta in 1992, followed by snails (Planorbidae) and freshwater amphipods (Gammaridae) in 1998.
Water beetles (Dysticidae) colonized in 2000 and water
boatmen (Corixidae) in 2003. No previously unrecorded
taxa were found in 2004 and 2005 (Figure 2).
Unlike the other macroinvertebrate taxa, chironomids
(the dominant taxon) displayed a distinct successional
sequence over the study period, with a number of extinctions (Table 1). Some of the early colonizing Diamesa
species (notably Diamesa sommermanni) were not found
after 1988, when water temperature reached 7.5˚C.
Diamesa davisi decreased in abundance after 1978 and
were not collected in August after 1992, when water temperature reached 10˚C. Pagastia partica, which first colonized WPC in 1986, was numerically dominant from 1990
through 1998. Orthocladius mallochi and Eukiefferiella gracei, which colonized in 1988 and 1991, respectively, were
both relatively abundant in the mid-1990s, but numbers
subsequently declined. Eukiefferiella claripennis, first collected in 1986, was co-dominant in 2001 with Cricotopus
tremulus, which first colonized in 1996. Tanytarsus, a
taxon that typically inhabits stable habitats, was dominant in 2004. Paratrichocladius was the only chironomid
collected every sampling year from 1977 to 2005.
Eight copepod species were collected in WPC during
the period 1994–2002. Of the five Harpacticoida, two
species, Mesochra alaskana and Atheyella sp, were found
only in 1996. A third, Epactophanes richardi, was collected
in 1994 and in 1997. Moraria affinis and Maraenobiotus
brucei were more abundant, the former reaching densities
of 470 individuals m–2 in 1994. Harpacticoid copepods
were absent in WPC from 1999 to 2002. Of the three
cyclopoid copepods found, Acanthocyclops vernalis was the
only species found in all years and reached densities of
> 1100 m–2, Eucyclops agilis was found at low densities
during the first 2 years of the study, and Paracyclops poppei
was collected only in 1997. Ostracoda were also present
in WPC, but two species of chydorid cladocera (Chydorus
sp and Alona sp) were rare.
Macroinvertebrate taxon richness and the cumulative
number of taxa over the study period displayed a strong
positive relationship with water temperature (r2 = 0.90 and
0.97, respectively; P < 0.05). In contrast, both copepod
taxon richness (r2 = 0.66, P < 0.05) and cumulative taxon
richness (r2 = 0.89, P < 0.001) exhibited a significant negative relationship with temperature. Harpacticoid copepod
abundance (r2 = 0.83, P < 0.05) displayed a similar pattern.
Macroinvertebrate year-to-year community similarity, as
estimated by Jaccard’s index, was typically above 0.6 after
1988, except in 1990 and 1991, when five previously
unrecorded chironomid species colonized the stream.
Bray–Curtis values, as an estimate of compositional stability, typically exceeded 0.5 (Figure 3a), indicating low stability. Of the 10 macroinvertebrate taxa collected in 1988,
eight were still found in 2005, and of the 19 found in 1996,
16 were present in 2005 (Figure 3b).
Dolly Varden (Salvelinus malma) were the first fish to colonize in 1987, followed by pink (Oncorhynchus gorbuscha) and
coho (Oncorhynchus kisutch) salmon in 1989. The initial run
of spawning pink salmon included approximately 100 to 200
fish, but increased to 1250 in 1991 and to 3600 in 1993. By
1997, pink salmon exceeded 12 000 individuals, with subsequent estimates in odd years above 10 000 and a peak of
15 700 in 2005 (Figure 3c). Even-year runs of pink salmon
have typically not exceeded 1000 fish. Chum salmon
(Oncorhynchus keta) colonized the stream in 1996, but numbers have typically been below 100. Coast Range sculpin
(Cottus aleuticus) were found from 1998 onward. A waterfall
below the outlet of Lawrence Lake acts as a barrier to
upstream fish migration, most notably to sockeye salmon
(Oncorhynchus nerka), which typically spawn in streams
above lakes. Nevertheless, more than 500 sockeye salmon
were observed in 2005, in a large pool at the base of the falls.
During peak spawning periods in late August 1997,
macroinvertebrate abundance in reaches with redds was significantly lower than in August 1996 or 1998, or in the
period prior to spawning, in 1997. Macroinvertebrate densities were reduced to < 100 individuals per 0.1 m2 from a
mean of 480 individuals per 0.1 m2, while taxon richness
declined from 18 to 10. MacroTable 1. Major chironomid taxa in WPC, showing successional changes over time
invertebrate drift densities
increased fourfold during peak
Year first Year peak Peak abundance Max water temp at Year last
spawning in 1997 compared to
–2
Species
colonized numbers
(#m )
peak abundance (˚C) found
drift densities in the low salmon
Diamesa davisi
na
1978
2956
2.0
1991
run year of 1998. The 15N values
indicate no evidence of the
Diamesa sommermanni
na
1978
1026
2.0
1986
incorporation of marine-derived
Pagastia partica
1986
1994
2715
12.7
nutrients into lotic food webs (ie
Eukiefferiella claripennis
1986
2001
907
16.0
in fast-moving water) or riparian
Eukiefferiella gracei
1991
1994
2555
12.7
vegetation in WPC in 1998 and
Cricotopus tremulus
1996
1997
1290
15.8
2004, despite over 9000 pink
Tanytarsus spp
1993
2004
1466
18.3
salmon spawning in the previous
www.fr ontiersinecology.or g
© The Ecological Society of America
AM Milner et al.
Bray-Curtis distance
417
Jaccard’s index
years. Direct consumption of
carcasses by scavenger macroinvertebrates was not observed. Experimental release of
tagged salmon carcasses
indicated less than 10%
retention at relatively low
flows, with most carcasses
retained within 200 m of
the release point, principally
(62%) in stream margins.
Stream community succession
Jaccard’s index
Bray-Curtis distance
Discussion
The colonization and succession of the macroinvertebrate community within
WPC is the longest documented such sequence in a
stream environment and
major taxonomic groups
were still colonizing at least
40 years after the stream
formed. In contrast, Minshall et al. (1983) estimated
that recolonization of the
original taxa in the Teton
River, Idaho, took 3 years,
following a major flood
that initiated primary succession and, in an Oregon
stream impacted by the
Mount St Helens eruption,
over 200 taxa were found
after 10 years (Anderson
1992). We suggest that the
relatively long time period
over which invertebrate
colonization occurs at WPC
is due to the dominant
roles of water temperature,
dispersal limitation, and
geomorphological habitat
in this recently deglaciated Figure 3. (a) Jaccard’s Similarity Index values for year pairs, with higher values indicating greater
and isolated environment. similarity of the macroinvertebrate communities from year to year, and Bray–Curtis distances for year
Water temperature was an pairs with higher values, indicating lower compositional stability. (b) A comparison of the number of
important determinant for taxa for any particular year with the number of those taxa still present in 2005. (c) The number of
the colonization of many spawning pink salmon in odd years from 1977 to 2005.
macroinvertebrate taxa in
WPC, particularly temperature limitations imposed when al. 2001), that a deterministic element to community
remnant ice still influenced the stream during deglaciation assembly and succession exists in the initial stages of develand maximum temperature did not exceed 10 ˚C. In con- opment in post-glacial streams, when water temperature is
trast, water temperature was not a major variable in the low. However, it is unclear whether the macroinvertebrate
Teton River and in streams around Mount St Helens. assemblages of other streams, formed due to ice recession,
Unlike Fisher (1983), who considered that succession and would follow the same trajectory as that at WPC, as water
community assemblage in streams is entirely stochastic, we temperature increases and more colonizers arrive. In other
believe, based on studies in WPC and elsewhere (Milner et words, is the successional pathway in these new streams
© The Ecological Society of America
www.frontiersinecology.or g
Stream community succession
418
deterministic as water temperature increases? Data that we
have accumulated from another new watershed in Glacier
Bay indicate that this is not the case, as the stream lacks
three orders/families and chironomid diversity is considerably greater when compared to WPC at the same temperature. Thus, stream colonization appears to be initially
deterministic and then stochastic, based on life history
attributes and dispersal. This finding differs from that of
terrestrial invertebrate succession on an alpine glacier foreland. Hodkinson et al. (2004) considered this succession to
be deterministic and linked to plant community and soil
development, with little further change in invertebrate
communities after 50 years (Kaufmann 2001).
Non-insect macroinvertebrates lack the obvious interstream dispersal mechanisms of winged insects and the
delay in colonization of these groups appears to result from
difficulties in crossing land barriers from proximal streams,
rather than from water temperature limitations. The
Copepoda also lack the winged dispersal stages typical of
insects, and the copepod assemblage of WPC contained
species with good dispersal abilities. Acanthocyclops,
Eucyclops, and Epactophanes richardi produce dessicationresistant eggs or resting stages capable of dispersal between
systems by wind, waterfowl, and other animals (eg Caceres
and Soluk 2002). The feces of Canada geese, collected in
wetlands adjacent to a nearby new stream and subsequently wetted and incubated, yielded viable copepod
nauplii (M McDermott pers comm), suggesting that
waterfowl may be a key dispersal route. The appearance of
water mites (Hydracarina) and springtails (Collembola)
in 1999, water beetles (Dysticidae) in 2000, and water
boatmen (Corixidae) in 2003, follows an increase in habitat heterogeneity in WPC with age and the creation of
slower flowing habitats, such as pools and backwaters.
This 28-year record of site-specific temporal succession
in a riverine system indicates that the abiotic variables of
water temperature and geomorphological habitat dominate
colonization and succession. Tolerance would appear to be
a major mechanism of macroinvertebrate community
assembly as, following colonization, most taxa have persisted from year to year within the community (supported
by the relatively high similarity between years and the
number of taxa still present in 2005 compared with earlier
years). However, relative abundance within the community has changed markedly with time (as demonstrated by
high Bray–Curtis distance values). If inhibition were a
major mechanism of community assembly in WPC, we
would expect macroinvertebrate diversity to remain stable
or to increase slowly with stream development. If facilitation were a major mechanism, marked extinctions would
occur throughout the succession. In fact, we found a
marked increase in diversity, with few extinctions, apart
from the cold stenothermic early chironomid colonizers.
Our results therefore do not support inhibition or facilitation as major mechanisms in this stream succession. There
may, however, be facilitation elements that influence successional patterns in WPC. Redd digging by salmon crewww.fr ontiersinecology.or g
AM Milner et al.
ated disturbed patches, where the relative abundance of
blackflies (Simuliidae) and the chironomid Cricotopus
intersectus, potential fugitive taxa (good dispersers but poor
competitors), increased. Increases in fugitive taxa following redd digging were also documented by Minakawa and
Gara (2003), and potentially provide a recurrent disturbance that can increase the persistence of early species
(Platt and Connell 2003). However, this role may not constitute facilitation, as envisaged by Connell and Slatyer
(1977), because redd digging facilitates the persistence of
an earlier colonizer, not a later one. In addition, colonization by some taxa, notably some caddisflies and chironomids, was facilitated by the growth of riparian vegetation
along the stream as terrestrial succession progressed, and by
the provision of willow catkins as a food source or alder
roots as a substrate (Flory and Milner 1999). However, this
facilitation is from an adjacent ecosystem.
Our findings in WPC contrast with primary succession in
the terrestrial environment in Glacier Bay, which typically
involves dominance by a single species that is replaced as
succession proceeds (Chapin et al. 1994). Support for the
tolerance model in terrestrial succession has not been widespread, due to the extinction of colonizers throughout the
sequence, but we suggest that macroinvertebrate succession
in Glacier Bay streams provides some evidence for the tolerance model. Nevertheless, as is the case for terrestrial
ecosystems (Chapin et al. 1994), no single mechanism fully
accounts for primary succession of macroinvertebrate communities in streams. Dispersal dynamics, life history traits,
and elements of facilitation also play a role.
Unlike the macroinvertebrates, the copepod assemblage underwent substantial extinctions from 1994 to
2002 and generally exhibited low persistence. The relatively low habitat complexity and consequent lack of
refuge habitat may account for their low abundance and
poor persistence (Robertson 2000). As the colonization
record available for Copepoda is relatively limited
(1994–2002), the model of succession most relevant to
meiofaunal community assembly cannot be determined.
One key aspect of community assembly and succession in
WPC is that channel stability is relatively high, due to the
large headwater lake that buffers flow variations and
thereby limits habitat disturbance. Floods are typical recurrent disturbances in stream environments, influencing
community assembly and perhaps holding it in an earlier
successional stage (Fisher 1990). We know from studies of
older streams without lakes in Glacier Bay (Milner et al.
2000) that flood disturbance reduces taxon richness below
that of WPC, with several taxon groups not represented.
Implications for river management and
biodiversity
Our study demonstrates that the biodiversity of invertebrate assemblages in new stream channels may develop
slowly at first, as low water temperature may limit colonization and, in the absence of drift, water and mountain
© The Ecological Society of America
AM Milner et al.
barriers may make dispersal difficult. In such cases, assistance may be necessary to enhance dispersal. We have
shown that development of riparian vegetation is essential
for the establishment of some riverine taxa. Increases in
habitat heterogeneity are also necessary for the establishment of macroinvertebrate taxa associated with slower
flowing habitats and for meiofauna. The stochastic nature
of later successional stages indicates that the stream
macroinvertebrate community will not necessarily return
to the pre-disturbance community following restoration.
Our study also illustrates that, where a new stream channel with suitable habitat is formed, or access to previously
inaccessible habitat is created (eg above a dam), pink
salmon establish substantial populations from a small colonizing population of spawners within a few generations. The
recent creation of new stream habitat, linked to climate
change and local conditions, has been extensive along the
coast of southeastern and south–central Alaska. Although
many of these streams are relatively short in length, potential recruitment of new salmon stocks is substantial, at a time
when other stocks in the Pacific Northwest are threatened
by human activity (Nelson et al. 1991).
Although salmon colonization should enhance benthic
food webs through incorporation of marine-derived nutrients, WPC does not retain carcasses, which are flushed
back to the estuary by heavy rains in September and
October, as demonstrated by retention experiments. Hence,
recommendations for salmon carcass additions in the
Pacific Northwest to mitigate losses of salmon-derived
nutrients in streams (Compton et al. 2006) must consider
the varying ability of watersheds to retain these additions.
Retention ability in older Glacier Bay streams arises from
marginal habitats, pools, and, in particular, coarse woody
debris (CWD; Milner and Gloyne-Phillips 2005). Negligible amounts of CWD are present in WPC, as terrestrial succession has not yet progressed to a stage where recruitment
of larger trees (cottonwoods and spruce) into the stream
allows debris to accumulate. We estimate that this may take
a further 60 to 80 years. Mechanisms to improve retention
may have to be considered in streams where floodplain vegetation is regenerating or being planted in restoration
efforts (eg the introduction of large pieces of CWD). Our
long-term study has increased our understanding of the
processes by which stream ecosystems develop and thus can
improve the effectiveness and predictability of stream
ecosystem restoration as suggested by Lake et al. (2007).
Acknowledgements
We thank the many people who have assisted our research at
Wolf Point Creek over the past 28 years, including S Conn, L
Brown, M Dauber, S Mackinson, M McDermott, C McNeill,
I Phillips, J Smith, and C Soiseth. We thank M Winterbourn
and G Streveler for discussion of earlier versions of this manuscript. This paper would not have been possible without J
Luthy, skipper of the Park vessel MV Nunatak, who supported
our remote field camps in Glacier Bay for many years.
© The Ecological Society of America
Stream community succession
References
Anderson NH. 1992. Influence of disturbance on insect communities
in Pacific Northwest streams. Hydrobiologia 248: 79–92.
Caceres C and Soluk D. 2002. Blowing in the wind: a field test of overland dispersal and colonisation by aquatic invertebrates. Oecologia
131: 402–08.
Chapin III FS, Walker LR, Fastie CL, and Sharman LC. 1994.
Mechanisms of primary succession following deglaciation at
Glacier Bay, Alaska. Ecol Monogr 64: 149–75.
Compton JE, Andersen CP, Phillips DL, et al. 2006. Ecological and
water quality consequences of nutrient addition for salmon
restoration in the Pacific Northwest. Front Ecol Environ 4: 18–26.
Connell JH and Slatyer RO. 1977. Mechanisms of succession in natural communities and their role in community stability and organization. Am Nat 111: 1119–44.
Fisher SG. 1983. Succession in streams. In: Barnes JR and Minshall
GW (Eds). Stream ecology: application and testing of general ecological theory. New York, NY: Plenum Press.
Fisher SG. 1990. Recovery processes in lotic ecosystems. Environ
Manage 14: 725–36.
Fisher SG. 1997. Creativity, idea generation, and the functional morphology of streams. J N Am Benthol Soc 16: 305–18.
Flory EA and Milner AM. 1999. Influence of riparian vegetation on
invertebrate assemblages in a recently formed stream in Glacier
Bay National Park, Alaska. J N Am Benthol Soc 18: 261–73.
Kaufmann R. 2001. Invertebrate succession on an alpine glacier foreland. Ecology 82: 2261–78.
Hodkinson ID, Coulson SJ, and Webb NR. 2004. Invertebrate community assembly along proglacial chronosequences in the high
Arctic. J Anim Ecol 73: 556–68.
Lake PS, Bond N, and Reich P. 2007. Linking ecological theory with
stream restoration. Freshwater Biol 52: 597–615.
Milner AM, Knudsen EE, Soiseth C, et al. 2000. Colonization and
development of stream communities across a 220-year gradient in
Glacier Bay National Park, Alaska, USA. Can J Fish Aquat Sci 57:
2319–35.
Milner AM, Brittain JE, Castella E, and Petts GE. 2001. Trends of
macroinvertebrate community structure in glacier-fed streams in
relation to environmental conditions: a synthesis. Freshwater Biol
46: 1833–48.
Milner AM and Gloyne-Phillips IT. 2005. The role of riparian vegetation and woody debris in streams of Glacier Bay National Park,
Alaska. River Res Appl 21: 403–20.
Minakawa N and Gara RI. 2003. Effects of chum salmon redd excavation on benthic communities in a stream in the Pacific Northwest.
T Am Fish Soc 132: 598–604.
Minshall GW, Andrews DA, and Manuel-Faler CY. 1983. Application
of island biogeographic theory to streams: macroinvertebrate
recolonization of the Teton River, Idaho. In: Barnes JR and
Minshall GW (Eds). Stream ecology: application and testing of
general ecological theory. New York, NY: Plenum Press.
Naiman RJ, Bilby RE, Schindler DE, and Helfield JM. 2002. Pacific
salmon, nutrients, and the dynamics of freshwater and riparian
ecosystems. Ecosystems 5: 399–417.
Nehlson W, Williams JE, and Lichatowich JA. 1991. Pacific salmon at
the crossroads: stocks at risk from California, Oregon, Idaho and
Washington. Fisheries 16: 4–21.
Platt WJ and Connell JH. 2003. Natural disturbances and the directional replacement of species. Ecology 73: 507–22.
Robertson AL. 2000. Lotic meiofaunal community dynamics: colonization, resilience and persistence in a spatially and temporally
heterogeneous environment. Freshwater Biol 44: 135–47.
Robertson AL and Milner AM. 2006. The influence of stream age and
environmental variables in structuring meiofaunal assemblages in
recently deglaciated streams. Limnol Oceanogr 51: 1454–65.
Sidle RC and Milner AM. 1989. Stream development in Glacier Bay
National Park, Alaska, USA. Arctic Alpine Res 21: 350–63.
www.frontiersinecology.or g
419