Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Restoration ecology wikipedia , lookup
Occupancy–abundance relationship wikipedia , lookup
Reforestation wikipedia , lookup
Habitat conservation wikipedia , lookup
Ecological fitting wikipedia , lookup
Lake ecosystem wikipedia , lookup
Biological Dynamics of Forest Fragments Project wikipedia , lookup
Theoretical ecology wikipedia , lookup
Keystone Interactions: Salmon and Bear in Riparian Forests of Alaska Author(s): James M. Helfield and Robert J. Naiman Source: Ecosystems, Vol. 9, No. 2 (Mar., 2006), pp. 167-180 Published by: Springer Stable URL: http://www.jstor.org/stable/25470328 . Accessed: 05/11/2014 10:44 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . Springer is collaborating with JSTOR to digitize, preserve and extend access to Ecosystems. http://www.jstor.org This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions (2006) 9: 167-180 Ecosystems DOI: 10.1007/sl0021-004-0063-5 ES*S\C\/CTEAAC E VA/3 T O 1 EfVlD ? 2006Springer Science+Business Media,Inc. Keystone in Bear Interactions: Forests Riparian James M. Helfield,1,3*1 of Forest College Sciences, of Alaska and Robert Box 352100, Seattle, University of Washington, Box 355020, 98195, Seattle, Washington of Washington, Umed and Environmental Science, University, Resources, University and Salmon J.Naiman2 98195, Washington USA;2 School of Aquatic USA;3 Landscape Ecology Group, SE-901 Sweden 87, Umed, and Fishery Department of Ecology Abstract in the presence of increased forest is significantly not both salmon and bear, but by either species of salmon and bear The interactions individually. to of 24% may provide up riparian N budgets, but varies in time and space according this percentage is used to describe species" "keystone that exert a disproportionately organisms impor on the ecosystems tant influence in which they as "keystone such live. Analogous concepts The term in and "mobile links" illustrate how, of two or more cases, the interactions species an effect greater than that of any one produce Because in of their role species individually. ocean to river nutrients from the and transporting mutualism" to variations many ties, rather animals than are equal, but some animals Orwell are more (1945) Received 25 May 2004; accepted 16 December 2004; published online 15March 2006. * Corresponding author; e-mail: [email protected] ^Current address: Department of Environmental Sciences, Huxley College Western of the Enviroment, Bellinaham, University, Washington 98225-9181, mor than their specific component parts. that control the integrity describe those animals Since then, the and stability of their communities. used in ecology and con concept has been widely has been and the keystone servation, designation range of taxa at various trophic applied to a wide levels in various ecosystems (see Bond 1993; Mills and others 1993; Power and others 1996). there is no universally accepted oper Although a keystone of what constitutes ational definition Ecological theory holds that certain animals exert a disproportionately on the influence important in Paine which live. ecosystems (1966) first they a in reporting described this phenomenon how starfish influences the (Pisaster ochraceus) predatory and population of an species composition density Washington channel of barnacles, intertidal ecosystem. By eating masses Pisaster prevents exclusion competitive by domi nant organisms, thereby creating open space for a Paine of species. subse greater number (1969) the term "keystone introduced species" to quently others." George escapement, salmon; bear; riparian forest; marine Key words: derived nutrients; nitrogen; species. keystone Introduction equal salmon characteristics, vegetation phology and functional redun suggesting interdependence sources. These illustrate N findings dancy among of interspecific the the complexity interactions, across bound of ecosystem linkages importance of examining aries and the necessity the processes communi and interactions that shape ecological Pacific salmon riparian ecosystems, (Oncorhynchus and bear brown (Ursus arctos) have been de spp.) scribed as keystone links, al species and mobile are to few data the available though quantify to of this interaction relative other importance vectors. nutrient of a mass balance Application to data from a southwestern stream model Alaskan to that influx the suggests nitrogen (N) riparian "All in and watershed species, certain requisite traits have been identified. are generally so designated native Animals species their activities and abun that regulate, through USA 167 This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions 168 J. M. Helfield and R. J. Naiman or physical productivity, diversity their communities, with influences those organisms extending beyond directly affected through trophic interactions (Paine 1966, 1969). is that keystone Implicit in the concept species are in their importance relative to the rest exceptional dances, structure the of of the community (Mills and others 1993), that in their functioning within the they are unique and that their impacts community (Kotliar 2000), are disproportionately large relative to their abun dances (Power and others 1996). Loss of a keystone in the struc species results in significant changes ture or of organization a ecosystem, given pre with adverse for the consequences sumably survival of other native species or populations. Most of the keystone descriptions phenomenon focus on a single species, although it is understood cases keystone that in many effects arise through the interactions of two or more species. For exam the "key ple, studies of mutualism (for example, sensu Gilbert stone mutualist 1980; hypothesis" and Christian 2001) Bertness and Shumway facilitation (for example, 1993; Bertness and Leonard and others 2001) demonstrate how 1997; Mulder interactions among different positive species help maintain the structure and diversity of various plant and animal under ad communities, particularly verse environmental conditions. There has been some debate as to whether such positive interac tions exert amore widespread influence than do the in forces of competition and predation described Paine's definitive (Bruno and keystone example are dependent on others 2003), but both paradigms interspecific tion strength" interactions. as a measure Use of the term "interac of the importance of most of their life histories and (see Lundberg 2003). Moberg Another of this sort of interspecific, example interaction is the transfer of nutri trans-boundary ents from the Pacific Ocean to river and riparian ecosystems by sockeye salmon (Oncorhynchus nerka) and brown bear (Ursus arctos). Pacific salmon have as keystone been described species in coastal eco as a food re because of their importance for vertebrate and scavengers predators and others 1995; Willson (Willson and Halupka salmon and bear have been de 1998). Similarly, scribed as interacting mobile link organisms because of their role in transporting marine-derived nutri ents to forest ecosystems in Alaska (Lundberg and systems source marine-derived nutrients 2003). Although Moberg to influence have been shown riparian structure and dynamics 2001, 2002; (Helfield and Naiman Bartz and Naiman to few data are available 2005), of the salmon-bear the importance inter quantify to other nutrient action relative vectors. Here we for nitrogen present a mass balance model (N) flux in the riparian forest adjacent to a boreal Alaskan stream. The objectives of this study are to quantify in N contributions variations via spatiotemporal salmon and bear relative to other N sources, and to elucidate the potential effects of salmon long-term on the productivity bear interactions and species of riparian ecosystems. In so doing, we composition to build on the existing body of theoretical hope related to the keystone and species concept our understanding in which enhance of the ways are shaped by interspecific communities ecological across and interactions ecosystem linkages work boundaries. Macarthur 1972; (for example, keystone predators Paine 1992) implies that this importance is derived through the reactions of other species. Accordingly, some authors com have that ecological argued more and be community plexity stability may on synergistic from effects dependent resulting on or actions interactions the than interspecific of any one species (Naiman and Rogers and Jones and others 1993; Lawton 1995; Soule and others 2003). In some cases these keystone interactions involve across boundaries. that ecosystem migrate species abundances 1997; Mills link" pollinators and include the "mobile Examples and the seed dispersers described by Gilbert (1980), and Polis others subsidies described by spatial or that carry nutrients, energy (1997). Animals food material otherwise among separate genetic on the exert a significant influence webs may structure and dynamics of receiving communities, even if they are extrinsic to those communities for of Salmon, Interactions Vegetation Riparian Bear and Having spent most of their lives feeding and grow at Pacific salmon returning to spawn and die sea, ing in their natal streams carry marine-derived nutri ents in their body tissues. Returning salmon are eaten mammal and bird species by numerous and others and others 1989; Willson (Cederholm car from nutrients salmon and 1998), decaying casses support the production of periphyton, aqua resident freshwater fishes tic macroinvertebrates, streams (Mathis salmon in spawning and juvenile en and others 1990, 1993; 1988; Kline and others Schuldt and Hershey 1995; Michael 1995; Bilby and and others 1998). These others 1996, 1998; Wipfli are also delivered to terrestrial vegetation. nutrients stable isotopes indi Studies of naturally-occurring This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions Keystone Interactions 169 to spawning cate that riparian plants adjacent as as streams may much derive 18-26% of their N from salmon and others foliar 1996; Ben (Bilby and Hilderbrand others David and others 1998; and midge (Diptera: Chironomidae) feed directly on salmon carcasses in streams, species whereas blackflies mayflies (Ephemeroptera), on and fine par feed (Diptera: Simuliidae) midges 2001, 2002). 1999a; Helfield and Naiman from nutrients Transfer of marine-derived streams to is medi ecosystems spawning riparian are an important ated largely by bears. Salmon downstream from decomposing ticles immediately carcasses and others 1998; 1997; Wipfli (Minakawa nutri and Gara 1999). Salmon-derived Minakawa ents are also incorporated into the body tissues of stoneflies and dragonflies (Odonata) predatory interactions (Kline and others trophic through are 1990; Bilby and others 1996). These materials to transferred terrestrial subsequently adjacent source for coastal bear populations and others 1996, 1999b), and bears (Hilderbrand a consume of total large proportion frequently or either biomass, spawner through predation seasonal food carcasses of post-spawn (Quinn and scavenging Kinnison and Buck 1999; Quinn 2000; Reimchen and others and 2000; Gende 2000; Ruggerone others 2001; Quinn and others 2001). Marine-de are then made to riparian rived nutrients available vegetation through carcasses salmon dissemination and of partially-eaten salmon-enriched wastes. Hilderbrand and others (1999a) report that brown bears on the Kenai Peninsula, deliver 83 Alaska, in white N detected 84% of marine-derived spruce 500 m of spawning (Picea glauca) foliage within streams. These findings are consistent with other studies demonstrating between spatial correlations N in riparian fo bear activity and marine-derived and others and 1998; Helfield liage (Ben-David Naiman 2002). bears are the most visible consumers of Although are use not to the make salmon, they only species of this anadromous food source. River otter (Lontra canadensis) and mink (Mustela vision) feed regularly on salmon and others 1998; Ben-David (Blundell marten and others whereas 1997a), (Martes ameri into salmon their diets during cana) incorporate of low abundance of years prey species principal and others Bald (Ben-David 1997b). eagle (Haliaeetus leucocephalus) and gulls (Larus spp.) also and occasionally kill salmon in Alaskan scavenge streams and others and 1998; Quinn (Ben-David Buck 2000). On the Olympic Peninsula of Wash bear are absent and black brown ington, where bear (U. Cederholm mammals carcasses. americanus) are relatively scarce, and others list 22 species of (1989) to consume and birds known salmon These animals all transport marine-de rived to terrestrial nutrients and organic matter with amounts and spatial distributions ecosystems, to the abundance, varying according longevity, and per capita salmon of mobility consumption each species. At finer spatial scales, insects play an important role in mediating (Jackson numerous nutrient transfer stream-riparian and Fisher larvae of 1986). Aquatic stonefly (Plecoptera), (Tri caddisfly choptera) of the adult insects the emergence or death subsequent consumption by terrestrial insectivores. Salmon carcasses deposited on exposed gravel bars and in the riparian zone are and consumed quickly colonized by larval forms of flies terrestrial and (Diptera: Calliphoridae carrion and beetles Scathophagidae) (Coleoptera: which be an important factor may Silphidae), ecosystems and their with and distribution of sal decomposition controlling mon-derived nutrients (Meehan 2000). are also transferred to nutrients Marine-derived In habitats abiotic lar processes. riparian through carcasses on salmon ger rivers, flooding deposits stream banks and others 1989; Ben (Cederholm David and others 1998). In streams with hydrauli conductive cally from decomposing riparian ecosystems dissolved nutrients substrates, carcasses may be transferred to via shallow subsurface (that is, Studies of non-salmon flowpaths. hyporheic) streams have demonstrated that the hyp bearing orheic zone can as serve as a transient solutes storage area for from surface nutrients, downwelling zone can be rapidly atten water to the hyporheic uated and uptake through sorption physical by communities microbial and Walters 1983; (Bencala zones Triska and others 1994). Where hyporheic are shallow and extend the active laterally beyond the roots of riparian plants may extend floodplain, zone and take up transiently into this saturated stored nutrients. and Edwards O'Keefe (2003) concentrations that of ammonium reported (NH4) and soluble reactive phosphorus increased signifi stream following Alaskan cantly in a southwestern of spawning and that salmon, entry sockeye water nutrient-enriched surface subsequently zone beneath entered the hyporheic the riparian forest. The importance as a of hyporheic exchange vector for marine nutrients is controlled by the physical hyporheic salmon extent zone, carcasses and hydraulic of the conductivity as well as by the abundance of within the stream. This marine nutrient subsidy has a potentially important influence on riparian plant communities. This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions 170 and R. J. Naiman J. M. Helfield Recent found that foliar N content, studies have and stem densities of riparian basal area growth streams trees are enhanced at sites near spawning and Bartz and Naiman 2002; 2001, (Helfield and richness whereas 2005), species are relative of decreased, understory plants density to comparable sites without salmon (Bartz and It marine N influx is that Naiman 2005). possible of N also decrease the may advantage competitive as in such alder (Alnus spp.), resulting fixing plants of these species near spawn decreased abundance 2002). Because ing streams (Helfield and Naiman are generally contents foliar N with plants higher more nutritious to browsers such as and palatable Naiman moose hare snowshoe (Lepus (Alces dices) and and others Pastor 1987; americanus; Bryant 1988), of marine also influence N inputs may patterns turn nutrient in affects which cycling, browsing, and plant species composi successional processes and and Bryant tion (Kielland 1998; Suominen to -6?C. Annual -15 ranges from 250 precipitation cm falls as snow to 340 cm, of which 200-250 and Johnson Creek is 1984). Lynx (Hartman at 2.3 km Lynx long, originating approximately into Lake Nerka. Mean Lake and discharging summer 500 is approximately 1 s-1 discharge Including Lynx Lake (O'Keefe and Edwards 2003). cov the Lynx Creek watershed and its tributaries, ers approximately 2760 ha The (USGS 1979). a is association of boreal forest riparian vegetation and paper birch (Betula papyrifera) stands of balsam poplar with (Populus interspersed willow (Salix spp.) and cottongrass balsamifera), are dominated (Eriophorum spp.). Upland hillslopes alder stands of dense green (Alnus crispa). by of sockeye sal Since 1947, annual escapement mon to Lynx Creek has ranged from 464 to 17,023 = from with spawning occurring 3,084), (mean and late July (Rogers Rogers August through that the average adult sockeye con 1998). Given white spruce others 1999). The consequences 82 g of N in its body tissues runs 1997), annual spawning = 38 N -1,397 (mean 253) kg bring approximately to the Lynx Creek watershed during the growing of approx season. The area supports a population stream, potentially increasing aquatic productivity. trees en Increased growth and density of riparian fil sediment bank hances stabilization, shading, debris of large woody tration and production enhance spawning and rearing (LWD), all of which and others fishes for salmonid habitat (Meehan and Naiman Harmon and others 1986; 1977; imately 2.8 bears per 100 km2 (Van Daele 1998), all the spawning feed on salmon throughout of whom season. Recent studies have characterized patterns on sockeye of bear predation (Quinn and Buck and and others 2000; Gende 2000; Ruggerone as well as and others 2001), others 2001; Quinn tains nutrient subsidies to of marine also affect ecosystems aquatic ecosystems. riparian in in riparian Increased N content foliage entails to the of litter delivered creased nutritional quality 1997; Bilby Decamps others 1998). Because element retaining and Bisson LWD and 1998; Naiman is a key structural carcasses salmon in streams and Peterson 1985), increased riparian (Cederholm of the availability enhances further production effects in nutritive salmon carcasses and associated As a result of these lotic ecosystems. linkages, nutrients carried upstream by adult salmon help to of subsequent the productivity enhance genera fishes. tions of salmon and other stream-dwelling A Mass Balance at Lynx Creek Model for Riparian N estimates N flux at Lynx The mass balance model River Lakes system Creek, a tributary of the Wood in the Bristol Bay region of southwestern Alaska, a tran area in is The 158?55' USA N, W). (59?29' as as well climatic sitional zone, with maritime influences continental patterns. affecting weather summer range from 6 to temperatures Average 20?C, and average winter temperatures range from approximately (Larkin and Slaney 2003) (O'Keefe and Edwards of riparian veg N enrichment and and Naiman etation 2002; Bartz (Helfield in River the Wood Naiman system. 2005) calculates N influx to a The mass balance model corridor (45.2 ha) of riparian forest on 200 m-wide processes hyporheic and marine-derived the model either side of Lynx Creek. Specifically, N of marine-derived estimates annual contributions for com via bear activity and hyporheic exchange sources from non-marine parison with contributions soils and leaching from upland (that is, precipitation, defines For each N source, the model N fixation). and identifies variables environmental parameters then be ad N fluxes. Parameters may controlling justed to estimate N fluxes under different scenarios. from empirical observa Input data were derived not data from Lynx Creek were tions. Where sites we from comparable data used available, or River the Wood within ranges of pub system Mini for boreal Alaskan lished values ecosystems. were mean identified values and mum, maximum calculations were then for all input variables. Model combina using all possible repeatedly, performed as so to full the tions of input data range of generate This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions Keystone values for possible results for each N source. Mean values for all in results were obtained using mean of salmon put variables. To assess the importance total annual N in bear interactions, we calculated all possible flux under 1,296 scenarios representing maximum and mean combinations of minimum, the from each N source, assuming contributions of both salmon and bear (729 scenarios), presence salmon but not bear but not salmon (243 scenarios), salmon nor bear bear (243 scenarios), and neither scenarios were Differences among (81 scenarios). evaluated with Kruskal-Wallis analyses of variance contrast using ranks, followed by Dunn's multiple test. Calculations, and in assumptions hypothesis put data are described below. Sample calculations A in Appendix using mean input data are provided Interactions 171 to export N to the riparian zone via potential (USGS 1979). In a study of downslope leaching at much forest sites encompassing nutrient cycling in boreal encountered of the range of conditions Alaska, Van Cleve and others (1983) report mean to 1.9 kg N ha"1 0.4 from annual fluxes ranging mean via This range =1.2) (all-site leaching. with values reported in subsequent encompasses (Kaye and others 2003). Using these data, at Lynx Creek are calculated contributions NL = JL ' AL studies leaching as (2) via NL is the mass of N delivered annually from upland soils, JL is the per unit area leaching annual flux of N via leaching, and AL is the upland where leaching area. (http://www.springerlink.com). N Fixation Precipitation source of N is a variable but continuous data are available describing of N within the Wood River atmospheric deposition system, but the US National Atmospheric Deposition (NADP 2003) reports amean annual flux of Program 0.19 kg ha"1 (range = 0.02-0.63) for the years 1980 at each of its two closest monitoring sites -2001 AK03 Poker Denali National Creek; (AK01 Park). these sites are in the interior of Alaska and Although Precipitation in boreal forests. No receive less annual rainfall than does the Bristol Bay and Johnson 1984), region (NADP 2003; Hartman studies of global patterns of N cycling indicate that or receive greater, they comparable, slightly amounts of reactive N via atmospheric deposition (Galloway and Cowling 2002). Using these data, the as model calculates N influx via precipitation NF=JF'AR (1) via NF is the mass of N delivered annually area is the unit flux of annual JP per precipitation, N via precipitation, and AR is the area of the where riparian Leaching zone. from Upland Soils accumu forests are areas of net nutrient Riparian lation within the watershed (Van Cleve and Yarie 1986) and typically derive proportions significant of their total nutrient from upland soils. budgets Whereas in the surface runoff is likely negligible forested Creek undisturbed, watershed, Lynx of soil solution leaching and downslope movement the lake may be important vectors for N. Excluding its tributaries, and the Lynx Creek watershed area 273 ha of upland comprises approximately Biological N fixation may be an important source of N in northern forests. Through the symbiosis with Frankia actinomycete, alders fix atmospheric N2, to ammonium is converted and transferred which to surrounding secretions soils via root and nodule of N-rich leaf litter (Binkley and production 1986; in Wurtz This results accelerated N process 2000). as in N and increased forest soils, availability cycling as in increased growth and foliar N content in and conifers sympatric understory plants (Binkley 1983; Binkley and others 1985, 1992; Wurtz 1995; and Gower Rhoades and others 1998; 2001). Vogel Shrub alders of boreal ecosystems (for example, A. well forests crispa;A. tenuifolia) inhabit early-successional Van and others Cleve and Viereck Cleve 1971; (Van stable 1981; Wurtz 1995), but also form long-term, Wurtz communities and others 1985; (Wilson influence soil N dynamics which may 2000), the later of succession. stages throughout at No data are available N fixation describing Lynx Creek, but other studies have characterized rates of N fixation in boreal Alaskan forests. Van Cleve and others annual N (1971) report a mean fixation rate of 156 kg ha-1 (range = 72 -362) for alder stands of various the Tanana ages within River floodplain. These values are similar to those area in subsequent in the studies reported and Van Cleve (Klingensmith 1993) and at other boreal and Alaskan sites (Lawrence 1958; Daly that N fixation rates are compa 1966). Assuming rable at Lynx Creek and proportionate with alder as N fixation may be calculated abundance, h where fixation, = a (Jaref / tfref) (3) Ja is the per unit area annual flux of N via N a is alder abundance of (that is, proportion This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions 172 J. M. Helfield and R. J. Naiman cover or aboveground forest biomass), Jaref is the per unit area annual flux of N via N fixation at the site (for example, reference Van Cleve and others at the reference 1971), and aTci is alder abundance site. Alder comprises 83 - 89% (mean = 86) of total at the Tanana River aboveground plant biomass sites (Van Cleve and others 1971). As alder abun dance data for Lynx Creek are reported in terms of uses percent cover as percent cover, this calculation a surrogate for percent biomass and (see Mynemi others 2001; Fensham and others 2002). covers approximately Alder 44% of the Lynx Creek watershed (O'Keefe and Edwards 2003), but zone is negligible alder abundance in the riparian and Naiman The of alder (Helfield 2002). majority fixed N affecting the Lynx Creek riparian forest is an therefore leached from upland stands. Mean nual losses described leaching by Van Cleve and others 0.2% of (1983) represent approximately a similar rate of forest floor soil N pools. Assuming leaching at Lynx Creek, total riparian N influx via N as fixation is calculated Na = ' ' (JaU PL AL) -r (JaR AR) (4) via N Na is the mass of N delivered annually fixation, JaU is the per unit area annual flux of N via as in Eq. (3), pL is the upland N fixation, calculated to of soil N lost proportion leaching, AL is the up where land leaching area, JaR is the per unit area annual as in Eq. flux of N via riparian N fixation, calculated area zone. is and the of the This AR (3), riparian assumes upland alder distribution to be calculation relatively homogenous (that is, non-patchy). alder is the primary N-fixing Although species in Alaska's boreal forests (Wurtz 1995), there are other the capable of fixing N. For example, organisms Schreber's feather moss (Pleurozium ubiquitous (Nos schreberi) forms symbioses with cyanobacteria as 1.5 - 2 kg N ha-1 toc spp.) and may fix as much in mid- to late-successional boreal forests, annually on abundance and the length of the depending season Simi (DeLuca and others 2002). growing form symbioses with larly, some lichen mycobionts fix N and (Rai 1988; Honegger cyanobacteria rates of N fixation by lichens and 1991). However, are 1-2 lower orders of magnitude bryophytes Alex for alder (for example, than those reported and Billington 1986; Van Cleve and others may 1971), and N fixed by lichens and bryophytes to soil N pools. Previous be less likely to be exported fix barely enough N studies suggest that bryophytes own metabolic to satisfy their requirements and that N and Billington 1981), (Alexander trees or transfer from feather moss to neighbouring ander tortuous and Alexander (Van Cleve a of N 1981). Similarly, large proportion relatively fixed by lichens is retained for cyanobacterial metabolism it (Rai and others 1983). Nonetheless, should be recognized that the mass balance model underestimates total biological N fixation by con shrubs is sidering only Hyporheic alder-mediated N fixation. Exchange O'Keefe and Edwards the exis (2003) confirmed tence of an extensive zone parafluvial hyporheic to Lynx Creek and demonstrated that adjacent is a feasible mechanism for hyporheic exchange to riparian nutrients marine-derived transporting a spawning run of 9460 vegetation. Following sockeye in 2000 (T. P. Quinn, unpublished data), a flux of 4.3 kg N ha-1 yr"1 was observed moving through a 100-cm deep column of hyporheic water to Lynx Creek and T. C. adjacent (R. T. Edwards a propor O'Keefe, unpublished data). Assuming tional relationship between escapement sockeye of N downwelling and the mass into hyporheic as annual flux may be calculated flowpaths, JH = E where JH is the per unit 100 cm deep column JHref is the escapement, 100 cm deep column of (JHref/ ?ref) (5) area annual flux of N per of hyporheic E is water, per unit area flux of N per water observed hyporheic during the reference year (for example, 2000), and the reference ETei is escapement year. during zone to the parafluvial The depth hyporheic cm to to 30 150 Creek from ranges Lynx adjacent below the soil surface (O'Keefe and Edwards 2003). Because of white the maximum rooting depth 120 cm (Nienstaedt and spruce is approximately Zasada 1990), only the upper portion of the hyp is accessible orheic water column for uptake, and in some cases none of it is accessible. Accessible as flux is therefore calculated hyporheic JHacc = JH [(Zrt - ZH) / 100] (6) N where Jnacc is the per unit area flux of hyporheic to riparian plants, JH is the flux of hypor accessible heic N as in Eq. (5), Zrt is the rooting depth of riparian zone. On plants, and ZH is the depth to the hyporheic zone 35 extends the parafluvial average, hyporheic m (range = 0 -100) at the active channel beyond and T. C. O'Keefe, (R. T. Edwards Lynx Creek data). As the spatial extent and con unpublished zone varies according to figuration of the hyporheic in and variations channel morphology reach-scale substrate permeability (Edwards 1998), the riparian This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions Interactions Keystone area underlaid flows at Lynx Creek by hyporheic to 45.2 N contri 0 ha. from may range Hyporheic butions are therefore calculated as Nh = * ^Hacc ^H (7) to is the mass of N delivered annually is via the JHaCc exchange, hyporheic riparian plants flux of accessible hyporheic N as in Eq. (6), and AH is the area of the parafluvial zone, cal hyporheic as stream twice and the product of culated length where the NH extent lateral active of hyporheic flow beyond 173 to salmon and abundance also varies according a greater proportion ease of capture, with eaten or harder to capture when fish are less abundant and others (2001) and others Gende 2001). (Gende that per fish killed percent report consumption = mean 4 to 80 from the ranges 25) within (all-site the uneaten Wood River system. Assuming portion is left on the riparian of each bear-killed salmon forest floor, N influx via discarded as may be calculated the Ncarc channel. = E PBK ' (1 ~ Peat) salmon carcasses ' ^Nsock (10) via ATcarc is the mass of N delivered annually E is salmon of carcasses, deposition partially-eaten is the of salmon killed pBK escapement, proportion of by bears as in Eq. (9), peai is the proportion biomass eaten per fish, and mNsock is the mass of N in the body tissues of each adult sockeye contained where Bear Activity in the northern Bristol Bay region is density 2.8 individuals per 100 km2 (Van approximately from 0.53 to Daele with densities 1998), ranging 100 interior 30.4 individuals km2 per throughout Local densities Alaska Daele and others (Van 2001). near streams are typically greater during salmon Bear For example, and others Hilderbrand spawning. on 43 of -57% bear observations (1999a) reported as being within 500 m of Peninsula the Kenai a similar degree of streams. Assuming spawning at be bear Creek, Lynx aggregation density may as calculated ?agg = B (pbf / Paf) (8) spawning Bagg is local bear density during season, B is regional bear density, pB? is the pro 500 m of spawn portion of bears observed within is and of the total the proportion pAF ing streams, m area observed is 500 of that within spawning where streams In the and others 1999a). (Hilderbrand no assumes absence of salmon, model bear the = (that is, Bagg B). aggregation On average, bears kill approximately 37% of in the Wood River system spawning sockeye rates vary (Quinn and others 2001), but predation to bear density from year to year according and salmon abundance Kinnison and 1999; (Quinn and others and others 2000; Quinn Ruggerone Creek, another 2001). At Hansen tributary of the River system that is comparable Wood in length to Lynx Creek (Rogers and and average escapement and others Rogers 1998), Ruggerone (2000) report an inverse relationship between and escapement the proportion of salmon killed by bears. This as is quantified relationship pBK= where bears (2493 E-0531) j 100 salmon 1997). (Larkin and Slaney The other means bear contribute N to by which waste is via forests excretion. riparian Among Kenai 500 m of spawning Peninsula bears observed within and others (1999a) report that streams, Hilderbrand each adult female passes 0.046 -0.051 kg ha"1 of N annually via urine and, to a lesser salmon-derived feces and decomposition of body tissues extent, after death. On average, salmon represent 62% of carbon and N for Kenai Peninsula bears, dietary with values for coastal, salmon-eating populations in Alaska ranging from 33 to 79% (Hilderbrand and others 1999b). Using these excretion may be calculated ^excr ^Mexcr / influx /?Ndiet via bear / V11 of N per JexCr is the per unit area excretion area is unit excretion the of salmon bear, JMexcr per derived N per bear, and /?Ndiet is the proportion of where bear of dietary salmon, N from derived the model between tributions via bear excretion Nexcr Jexcr = salmon. calculates difference and (#agg ' ^F JMexcr- In this the flux Annual absence as N are then calculated * )* ^excr ^R the con as (12) where via iVexcr is the mass of N delivered annually as bear excretion, is local bear in density Eq. Bagg 500 m of the stream, Jexcr (8), AF is the area within is the per unit area excretion of N per bear as in Eq. and is the this AR (11), riparian area over which flux is deposited. Total bear-mediated N contribu as tions are calculated (9) of salmon killed by pBK is the proportion and E is escapement and others (Ruggerone of biomass eaten per fish 2000). The proportion = data, N as NB = NCarc + NeXcr (13) where NB is the mass of N delivered annually bear activity, NcaTC is the mass of N delivered This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions via via J. M. Helfield 174 and R. J. Naiman D escapement * 464 escapement* annual nitrogen contributions (kg N yr"1) and percentage of total annual nitrogen influx (%) to the riparian zone of Lynx Creek via precipitation (P), leaching from upland soils (L), alder-mediated nitrogen fixation (A), hyporheic exchange (H), 0 escapement*17023 1200 1. Estimated Figure 3084 | and bear 800*4^-1 (B), as a function are mean values, activity Data of salmon calculated escapement. balance model values using mean by the mass within of for all input variables the parameters Error scenario. bars represent each escapement the full range of results obtained z 400-?-fm*mm P possible L A H B discarded salmon carcasses as in Eq. (10), and Nexcr is as in Eq. (12). via excretion the mass of N delivered Model Results and Discussion is the Model calculations indicate that leaching source of N in the riparian zone at Lynx dominant from contributions Creek (Figure 1). Annual sources and other non-marine (that is, leaching are not influenced and N-fixation) by precipitation in salmon abundance, but these decrease changes sources as contributions in importance from marine in and bear activity) (that is, hyporheic exchange crease with At levels of escapement. increasing minimum accounts marine-derived N levels of escapement, for approximately 7% of total annual levels of escape riparian N influx. At maximum to 39%. increases this percentage ment, bear activity When salmon are abundant, rep source resents of the second most important for an average of 15% of riparian N, accounting and total N influx at mean levels of escapement at maximum levels of escapement (Figure 1). 24% of combinations input using all values. Total Model spawning calculations season indicate there are, that on average, during 8.4 the bears the vicinity of Lynx Creek, and per 100 km2 within that bears kill approximately 35% of returning in years of mean This cor spawners escapement. mean roborates the value (37%) reported forWood streams by Quinn River and others (2001). The via par N is distributed of bear-mediated majority salmon carcasses (that is, ort). In years tially-eaten of mean bears discard an average of 66 escapement, as in 0.7 kg N yr"1 ort, compared with kg N yr-1 a in urine and feces. Although greater mass passed via ort, N distributed via excre of N is distributed more tion is Carcass scraps are found widespread. the immediate within vicinity of exclusively streams and Naiman 2002), spawning (Helfield whereas bear excretion may be detected more than 500 m from the stream and others (Hilderbrand because the majority of excreted 1999a). Moreover, converts to in is which N delivered urine, rapidly ex ammonium and others 1999a), (Hilderbrand to riparian creted N may be more readily available almost This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions Interactions Keystone trees than organic forms of N bound in carcass tis sues (Schulze and others 1994). It should be recognized is not that bear activity zone. in the riparian distributed homogenously streams are Certain locations adjacent to spawning for fishing and feeding on salmon, and preferred these are characterized elevated by significantly rates of excretion and carcass deposition (Hilder 600 500 zone riparian interactions salmon-bear source of marine-derived at Lynx Creek, there are generally N in the are ?'? X a * 400 a _L -T-3 ' I ci cp 300J_ B N 2. Estimated Figure to yr"1) the presence of salmon (B), and bear intervals, scenarios imum, riparian neither S total annual nitrogen zone SB influx of Lynx Creek, assuming nor bear salmon but (N), bear (kg N the not salmon bear but not salmon (S), and both are mean ? 95% values confidence (SB). Data on mass based balance for 1,296 calculations all possible combinations of min representing maximum and mean from each N N contributions source. letters denote subsets. homogenous Superscript ranks indicates Kruskal-Wallis of variance analysis by = means P < differences 89.86, among significant (x23 tests indicate contrast Dunn's 0.001). multiple significant differences between SB and N (Q4 = 5.42, P < 0.001), but not between B and N (Q4 = 0.20, P > 0.5) or S and N (Q4 = 2.26, 0.2 > P> 0.1). circum stances under which contributions may hyporheic be equally if not more of important. The magnitude N flux via hyporheic is controlled exchange largely to and lateral extent of hyporheic by the depth are in turn controlled flowpaths, which by topog soil texture and raphy, channel geomorphology, all of hydraulic 1998), conductivity (Edwards which within the At the watershed. vary spatially maximum extents and minimum lateral depths observed at Lynx Creek, hyporheic flows are below to the area the reach of riparian roots or confined so beneath active the that N in channel, directly is nonexistent. In exchange hyporheic as as annual contributions be N contrast, may high 316 kg, or 65% of total influx in areas where are at minimum hyporheic flowpaths depth and ex maximum lateral extent (Figure 1). Hyporheic a source of therefore be dominant N change might some reaches of Lynx in other systems or within flux ' *L and others and Naiman 1999a; Helfield for localized effects in 2002). By not accounting the mass balance model may these bear middens, underestimate the importance of bear-mediated N at smaller (for example, patch-level) spatial scales. The importance of bear activity as a source of N is individually. Although the greatest "j _b brand Con derived interactions with salmon. through the importance of salmon-derived N to versely, on facilitation by bear. riparian forests is dependent Model simulations indicate that total N influx to zone at Lynx Creek is significantly the riparian in the presence increased of both salmon and bear, but not in the presence of either salmon or bear alone suggest that the (Figure 2). These findings interactions of salmon and bear are more important to riparian N budgets, to riparian and by extension than are the actions of either species ecosystems, 175 via Creek. in hyporheic Temporal variability exchange also be than indicated by the mass greater might balance model. The mass of dissolved nutrients to increases with downwelling hyporheic flowpaths not but this be escapement, might relationship a is When escapement proportionate. high, large of spawner biomass will proportion decompose instream and potentially into the hyp down well is low, however, orheic zone. When escapement a consume bears and other terrestrial piscivores of biomass spawner greater proportion (Ruggerone and others 2000; Gende and others 2001), leaving a small amount of carcass material disproportionately to decompose instream. Accordingly, the impor tance of hyporheic be overestimated exchange may and underestimated for years of low escapement for years of high escapement. alder-mediated N fixation contributes Although a relatively of total riparian N small proportion influx at Lynx Creek (Figure 1), this process may source of N at other sites. Among be the dominant other sub-basins within the Wood River system, alder coverage ranges from 5 to 82% in the uplands and R. T. Edwards, (T. C. O'Keefe unpublished 0 to 19% in riparian zones (Helfield and from data) are and Naiman these parameters 2002). When results indicate applied to the Lynx Creek model, that annual N contributions from alder could the be as high as 3,800 kg, accounting for oretically 97% of total influx, or as little as 2 kg, accounting for less than 1% of total influx (Figure 3). Riparian This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions J. M. Helfield 176 and R. J. Naiman tance of marine-derived N and the relative impor tance of different N vectors vary from year to year to variations in escapement and other according climatic and parameters (for example, ecological bear abundance, Seasonal patterns precipitation). might also be important. For example, precipitation and leaching occur throughout the year, and a large OtcgNyr1 0% 10000i 1000-1 o.i I | in winter may be lost to of N deposited proportion In contrast, the system during the spring thaw. N is deposited marine-derived primarily during the a and of annual season, greater proportion growing H? J?I-Hi?,?I-HH?,?ILynx Creek Wood Minimum Wood Rtver Maximum River influx may (kg N Figure 3. Estimated annual nitrogen contributions of influx and total annual percentage (%) nitrogen y"1) zones riparian are presented tershed area, 0% to via Data alder-mediated for Lynx Creek of riparian area) fixation. nitrogen = 50% of wa (alder and for theoretical sites (5% watershed, 0% riparian) representing the minimum and maximum 19% riparian) alder (82% watershed, abundances observed within the Wood River system (Helfield and Naiman 2002; R. T. Edwards and T. C. O'Keefe, culated by unpublished the mass all variables input Error nario. bars tained using are mean Data data). balance model within the parameters the full represent all possible combinations values, mean using range of values for each sce results ob of of cal input values. is the most abundance factor important N contributions, alder-related with each affecting 1% increase in riparian alder abundance resulting in an increase of approximately 80 kg in annual N contributions. alder Keystone Interactions: versus Parts Processes and bear meet the basic criteria of keystone in that their loss entails significant species changes in the N budget and, by extension, the productivity and structure of the riparian forest. Organisms af fected include terrestrial scavengers, plants, as well as stream and browsers, decomposers fishes (Figure 4). To the extent that bear dwelling food source, and to rely on salmon as an essential the extent that bear-mediated fertilization of Salmon habitat for juve helps to enhance are salmon and bear populations not only in their role as mutually dependent nutrient vectors, but also for their long-term per sistence. These findings illustrate the complexity of interspecific linkages and interactions structuring river and riparian ecosystems. riparian forests nile salmonids, Yet varies of this keystone the significance both temporally and spatially. interaction The impor therefore be retained and assimilated by bear excretion and dis riparian plants. Spatially, carcasses occur at varying semination of salmon rates of whereas of hyporheic degrees aggregation, to reach and vary according exchange leaching in topography, soil texture scale differences and In general, N processing. microbial the influence of to the N increases with proximity marine-derived stream channel and others 1998; (Ben David and others 1999a; Helfield and Naiman Hilderbrand differences 2001, 2002), but at broader resolutions, in climatic and biogeographic factors (for example, alder abundance, bear density) entail significant in the relative importance differences of N sources. their the Moreover, despite interdependence, are some in to involved this interaction species extent Marine nutrients may be interchangeable. various carried upstream of salmon, species by those with low spawning den including typically sities Schuldt and 1995; (for example, Hershey 1996), or by other anadromous Bilby and others such as char fishes, (Salvelinus spp.) or smelt Willson and others spp.; 1998). Simi (Thaleichthys vectors be bears the dominant for may larly, of marine in transfer nutrients stream-riparian but this function may many Alaskan watersheds, be supplemented and others 1997a, b; (Ben-David and Buck 2000; Blundell and others 2003) Quinn bears are scarce, replaced by or, in regions where and others other species 1989; Bilby (Cederholm in and others nutrient 1996). Long-term cycling watersheds therefore de salmon-bearing might of different pend on a combination species and ex abiotic processes (for example, hyporheic alternate over time in their relative change), which none of these is functional importance. Although in its function, loss of any one will entirely unique if not immediate consequences. long-term reason why is not easily Another this interaction in terms of the keystone characterized species is its reliance on linkages across ecosystem concept boundaries. the impact of a given species Assessing or guild relative to its abundance within the com munity implies an ability to discern the boundaries have This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions Keystone 4. Figure of marine-derived Cycling MDN transport salmon-enriched upstream; wastes (B) Bears and nitrogen and other partially-eaten (MDN) and on effects consume piscivores carcasses salmon in river and salmon; the riparian salmon carcasses, enhancing decomposition and MDN diffusion; beneath is the forest and taken up by tree roots; (F)MDN flowpaths riparian colonize rates of riparian trees; (G) Riparian trees provide shade, bank riparian (C) Bears forest; ecosystems; and plants, potentially altering patterns of browsing, which in turn affects cause of salmon (A) Spawning other into hyporheic (E) Dissolved N downwells content enhance foliar N and growth inputs organic matter debris (LWD), enhancing the quality of instream habitat for salmonid fishes; (H) LWD retains post-spawn in streams, further enhancing MDN availability; (/) Increased foliar N content enhances palatability riparian 177 disseminate piscivores and insects (D) Terrestrial aquatic allochthonous stabilization, Interactions patterns of riparian and large woody salmon carcasses and nutrition of productivity and species composition. of that community Paine 1992; (for example, Power and others is This 1996; Hurlbert 1997). easier to accomplish in an for sessile organisms intertidal Paine community (for example, 1966) than for animals like salmon and bear that migrate over long distances freshwater spanning marine, and terrestrial habitats. As with functional impor to of a community is sensitive tance, the definition the spatial and temporal scales of observation. are especially Riverine characterized ecosystems by connectivity and others with (Naiman surrounding landscapes even 1987; Ward 1989), although are discrete af ecosystems seemingly invariably fected to some degree by external influences. or possibly because of, its importance as Despite, a theoretical con the keystone construct, species cept has Whereas been the subject of considerable debate. some authors have recommended that the focus of environ species be made keystone mental management and conservation efforts (for 1989; Woodruff 1989; Rohlf Conway example, the 1991; Carroll 1992), others have questioned of the keystone be species concept applicability in quantifying inherent the difficulties and effects which keystone distinguishing species merit such distinction 1993; (Mills and others Hurlbert Our illus 1997; Kotliar 2000). findings trate the extent to which keystone species rely on interactions and linkages across eco interspecific extent to which and the their boundaries, system in varies time functional and space. importance Losses of species identified as keystones may have adverse for other members of their consequences but protection of these species does communities, not guarantee the avoidance of such consequences. more to con it be constructive may Accordingly, the interactions and processes that structure communities rather than their specific com for and protection of ponent parts. The search a to tacit species keystone implies permission more obscure but other, ignore impor potentially tant taxa, along with and interactions interspecific abiotic processes any ecosystems, linking multiple of which may be essential combination for main the taining long-term productivity, diversity, or viability structure of a given commu physical sider those This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions J. M. Helfield 178 and R. J. Naiman is to nity. If the first rule of intelligent tinkering save all the parts (Leopold 1953), the second rule how they fit together. should be to understand ACKNOWLEDGEMENTS thank M. We C. R. T. Ed R. T. D. E. T. C. O'Keefe, Liermann, T. P. Quinn, L. Peterson, D. Paine, R. E. Bilby, C. A. two anonymous reviewers for their to this paper. We also thank contributions valuable L. Strom for illustrating Figure 4 and the University for logistic Alaska Salmon Program of Washington was in Financial Alaska. support support provided and Schindler Science Foundation National (DEB by Service's Pa States Forest the United 98-06575), Station and the University Research cific Northwest the of Washington. Ecol Bond WJ. Bruno V, Billington taiga. In: Van Alaskan Nitrogen FS K, Chapin Cleve in the fixation CT, Eds. Forest ecosystems LA, Dyrness p 112-20. taiga. New York: Springer, Viereck PW, III, Flanagan in the Alaskan M, flooding Ben-David M, in diets of salmon. Pacific and Can J Zool 1997a. DM. Schell mink: riverine 75:803-11. and sea 1997b. Annual Flynn RW, Schell DM. from stable isotope in diets of martens: evidence sonal changes 111:280-91. Oecologia analysis. Ben-David Bertness tions M, GH. Leonard MD, in communities: 1997. The from lessons role intertidal interac of positive habitats. 78:1976-39. Bertness MD, in marsh Bilby RE, Shumway plants. Am Bisson debris. woody and management. Bilby RE, Fransen PA. SW. Nat Competition 142:718-24. 1998. Function In: Naiman New RJ, York: BR, Bisson and 1993. and Bilby RE, Springer, PA, Walter Eds. River Bilby gen JK. of Response steelhead and 1998. coho RE, Fransen and carbon evidence from 5:215-7. stable Can to J Fish of nitro the isotopes. trophic Can J in Douglas-fir site fertility. planta For Ecol in river stategies? Behav MD. 2003. Inclusion Ecol Evol of 18:119 hare interactions: and floodplain succession. of management Ecological PL, Jain SK, Eds. natural sensitive Conservation nature biology: conservation, York: preserva & Hall, p 347 Chapman 72. CJ, Peterson Cederholm Can Christian the importance 413:635-9. the 1998. RJ, Quantitative from aerial JN, Galloway world: Gende 200 University fixation York: Oxford by nodulated Alnus M-C, in feather zone. hyporheic and management. The Fairfax RJ, Holman assessment photography. moss In: Naiman vegetation Int J Remote feeding MF. TP, Willson on salmon. 2001. Oecologia Franklin JD, Anderson JF, Swanson NH, Cline GW, Cromack Linkaemper debris ogy of coarse woody Res 15:133-302. attributes 23:2293-317. and the choice of the conservation FJ, Sollins SP, Aumen BA, Eds. Conser Sun perspective. P, Gregory SV, Sedell JR, NG, KW. 1986. Ecol K Jr., Cummins in temperate ecosystems. This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions 2002. PJ. Consumption 127:372-82. and LE. 1980. Food web organization In: Soule ME, Wilcox diversity. neotropical an evolutionary-ecological vation biology: Sinauer. derland, MA: p 11-33. ME, Springer, Reactive EB. 2002. nitrogen Cowling 31:64-71. years of change. Ambio Gilbert Harmon RJ, Bilby structural Sens of bor York: JE, Whitehead of 2002. A. carpets New Can rugosa. Sellstedt Nilsson O, ecology SM, Quinn by bears reveal Nature New century. nitrogen-fixation 419:917-20. Nature RT. of and species Conservation Zackrisson TH, Fate for sustaining The prospects Eds. In: Western D, Pearl MC, Daly GT. 1966. Nitrogen J Bot 44:1607-21. Quantifying eal forests. small spawning invasion of a biological for plant communities. Consequences of mutualism twenty-first 199-209. p in 1989. WG. evolution. salmon in 1989. DB, Cole DL, Scarlett WJ. carcasses kisutch) {Oncorhynchus Sci 46:1347-1355. J Fish Aquat CE. 2001. Deluca of coho debris CJ, Houston streams. Conway their retention 1985. The NP. carcasses kisutch) by organic J Fish Aquat Sci 42:1222-5. salmon Lattin D. 1983. Ecosystem production Binkley and of red alder interactions tions: Manage carcasses PA. 1996. Incorporation BR, Bisson into coho salmon from spawning system of small streams: Sci 53:164-73. Fish Aquat ecology p 324-46. salmon kisutch) (Oncorhynchus to the addition of salmon mykiss) (Oncorhynchus U.S.A. in southwestern two streams Washington, Sci 55:1909-18. Aquat juvenile large Sociality willow-snowshoe balance of and practice theory New and management. tion Fensham of distribution 1992. In: Fiedler RE, Eds. River p 399^129. facilitation Biogeo stands. Ecol 68:1319-27. CR. Edwards Ecology Feltleaf carbon/nutrient Press, spawning 1992. E-D, Mooney HA, York: Springer, Trends theory. of Sci Soil New function. Bertness JJ, ecological 1987. JP. plant for changes of role role of the Seasonal the of Fertilization 1998. salmon: Pacific spawning 83:47-55. Oikos activity. by TA, Hanley coastal DM. Schell TA, Hanley vegetation and predator abundance 25. Bryant coho in of solute transport RW. 1983. Simulation KE, Walters a mountain stream: a transient storage model. pool-and-riffle Water Resour Res 20:1797-803. terrestrial into Cederholm Bencala Ben-David Stachowicz JF, (Oncorhynchus streams. Can nutrients RJ. 2005. KK, Naiman Impacts of salmon-borne in southeast on riparian Alaska. soils and vegetation Ecosys tems (in press). Bartz York: Wi nitrogen. In: Schulze 1993. Keystone species. and ecosystem facilitation Natural RT. 2003. Bowyer or reproductive Eds. Biodiversity p 237-253. the 1986. M, foraging cooperative 13:13441. areas. MM. Ben-David GM, otters: 1985. alder-fixed D. R, Sachs D, Myrold conifer and alder-conifer P, Bell Sollins D, of adjacent chemistry ogy 73:2022-33. Carroll Alexander New management. P, McGill WB. as a tool for tracing Nitrogen-15 Soc Am J 49:444-7. Ecology REFERENCES nutrition Forest Sollins D, Binkley Blundell Reidy, 1986. ley, p 304. Binkley Ben-David, M. J. Elser, wards, D. Binkley Adv Ecol Interactions Keystone Hartman JM, Naiman on riparian gen productivity. Helfield sources 2001. RJ. forest Ecology to riparian 133:573-82. Hilderbrand atlas of Alaska. 2002. Salmon into Hilderbrand as nitrogen watershed. Alaskan Robbins Schwartz CT, in bears arctos) a ecosystem. (Ursus terrestrial the of flow CC, Robbins C. SM, particularly conservation to body size, population salmon, American of North brown Servheen 1999b. Meehan and productivity, bears. Can J Zool 77:132-*. Titus TA, SD, Robbins CT, Hanley Farley C. 1996. Use of stable isotopes to determine Servheen bears. Can J Zool 74:2080-88. living and extinct GV, R. 1991. Rev SH. Hurlbert Functional Plant Physiol 1997. Functional some reformulating J Ecol 22:369-82. Jackson of the lichen aspects Plant Mol Biol 42:553-78. of 1986. emergence Secondary production, insects of a Sonoran stream. Desert and of aquatic export 67:629-38. Ecology C. 2003. Stable soil nitrogen JP, Binkley D, Rhoades Kaye accumulation and flexible organic matter stochiometry during succession. 63:1-22. primary floodplain Biogeochemistry K, Bryant JP. 1998. Moose and vegetation biogeochemistry sion. Oikos 82:377-83. Kielland in taiga: effects on succes in primary herbivory dynamics TC Jr., Goering JJ, Mathisen OA, Poe PH, Parker PL. 1990. of elements upstream Recycling transported by runs of Pacific I. 815N and 513C evidence salmon: in Sashin south Creek, Kline eastern Kline Can Alaska. TC J Fish Aquat Scanlan Sci 47:136-44. JJ, Mathisen Jr., Goering RS. 1993. Recycling by runs of Pacific salmon: Kvichak river watershed, stream the Can Alaska. Van KM, fixation Klingensmith nitrogen Tanana Kotliar J Fish Aquat Larkin GA, derived Slaney nutrient salmonid Lawrence Alaska. Lawton PA. Am JH, Bristol K. Cleve 1997. influx 1958. southwestern Bay, 1993. of to Glaciers the well Denitrification keystone-species it work? Conserv Biol in marine Columbia 22:16-24. and in southeastern vegetation Sci 46:89-122. Jones CG. 1995. species In: Jones as ecosystem organisms engineers. Eds. Linking species and ecosystems. Hall, p 141-50. LB, Ed. 1953. Round Leopold New York: Oxford Leopold. river: New and ecosystems: JH, CG, Lawton York: Chapman and University the journals Press, p 286. of Aldo of spawning pink salmon: managing Sci 69:228-33. sal one con 1997. The dynamics of aquatic insect communities salmon PhD Dissertation, Seattle spawning. of Washington. (WA): University N. Minakawa associated with effects of a chum N, Gara RL 1999. Ecological run in a small stream (Oncorhynchus keta) spawning Pacific Northwest. J Freshw Ecol 14:327-35. Minakawa Mulder Uliassi CPH, DF. Doak DD, 2001. States of America Physical role the diversity-productivity relationships: Proc National interactions. Academy of Sciences salmon of stress the and of positive of the United 98:6704-6708. RB, Dong J, Tucker CJ, Kaufmann RK, Kauppi PE, Liski J, L, Alexeyev V, Hughes MK. 2001. A large carbon sink in the woody biomass of northern forests. Proc Nat Acad Sci USA Mynemi Zhou 98:14784-9. Naiman zones. H. 1997. The ecology RJ, Decamps Ann Rev Ecol Syst 28:621-58. Naiman Naiman RJ, New York: Eds. RE, Bilby Springer, Lock MA, riparian - and system 47:521-529. J. SJ, Chen KL, McKay In: Naiman RJ, Melillo of interfaces: 1997. Large animals corridors. BioScience Fetherston RJ, forests. 1998. River level Riparian and ecology p 289-323. Ford SR. 1987. TE, Reice of ecosystem and community patterns Longitudinal processes structure in a subarctic river continuum. 68:1139-56. Ecology National JM, Trends Deposition Program/National 2003. Annual data, sites AK01 (NADP). deposition and AK03. NADP Program Office, Illinois State Water Survey, 2204 Griffith Dr., Champaign, IL 61820. Atmospheric Network Honkala H, BH, 1. Conifers. USDA O'Keefe from effects coho rearing juvenile to benefit another. Northwest LS, Soule ME, Doak DF. 1993. The keystone-species in ecology BioScience and conservation. 43:219-24. Nienstaedt Linking Enhancement stream management. new does of Washington. RJ, Rogers KH. in river characteristic the along 23:956-63. flies (Diptera: Calliphoridae) in movement for nutrient implications levels. Senior Thesis. Seattle trophic Naiman and soils of trends Implications coastal south British Fisheries production. DB. PL, Sci 50:2350-65. Application dogs: how to prairie concept 14:1715-21. Parker up transported II. 515N and 513C evidence in in 2000. PH, elements successional floodplain interior Alaska. Can J For Res River, NB. Poe OA, of 1995. JH. on resource Mills role of The and management of Importance, preservation, a symposium. Fort Collins habitat: (CO): USDA Forest General Technical RM-43, Report Rocky Mountain and Range Experiment Station, p 137-45. Michael mon keystoneness: Aust biocenology. und Angewandte Eds. Forest symbiosis. Conn of riparian FJ, Sedell JR. 1977. Influences on aquatic with particular to reference ecosystems In: Johnson fishes and their food supply. RR, Jones WR, riparian Service, cept SG. JK, Fisher diets vs. importance in theoretical questions K, Trans Swanson vegetation salmonid DA, Hilderbrand 2000. decomposition: intermediate (WA): University CT, Jacoby ME, Hanley of meat, The importance TA, Arthur Honeggar Annu EP. Meehan interactions? fur Theoretische Vereinigung 23:2249-58. Limnologie salmon Oecologia and Parker PL, Goering JJ, Kline TC, Poe PH, Scalan of marine into elements transported Recycling salmon. der In by anadromous Verhandlungen OA, freshwater CC. or weak, 1972. Strong, Sci 44:177-88. 1988. RS. through Schwartz GV, Arts Acad Mathisen TA, nitrogen 121:546-50. marine R. Macarthur ternationalen Hanley brown Role nitro stream for alder and Mobile link organisms and for ecosystem resilience implications 6:87-98. Ecosystems functioning: management. implications in a boreal forests system eco F. 2003. J, Moberg Lundberg /Engineering of salmon-derived and GV, of 1999a. Effects growth 82:2403-9. RJ. JM, Naiman Oecologia 1984. Environmental of Water Resources (AK): Institute of Alaska. Station, University Experiment Helfield PR. Johnson CW, Fairbanks 179 and Forest Zasada Service, in southwest 1990. Handbook Silvics Burns RM, spruce. of North America: 654, Washington (DC): p 204-26. RT. 2003. of marine-derived Alaska. White coordinators. Agriculture TC, Edwards removal JC. technical Am Fish This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions Evidence nutrients Soc Symp for hyporheic in a sockeye 34:99-107. transfer stream J. M. 180 1945. G. Orwell RT. Paine of per analysis strength. through Nature on A note Am Nat RT. Nat Food-web interaction 1969. stability. Paine York: 1992. capita RT. Paine farm. New Animal Brace Harcourt, and 118. p Company, and R. J. Naiman Helfield 1966. field measurement 355:73-5. trophic complexity complexity and and U.S. community Van species Am diversity. Pastor JP, Naiman crobes RJ, Dewey the boreal forest. and P. mi 1988. Moose, Van 38:770-7. BioScience GA, Anderson BA, Bond WJ, Mills LS, JA, Menge J, Paine RT. 1996. Challenges Daily G, Castilla JC, Lubchenco BioScience 46:609-20. in the quest for keystones. Tilman ME, D, Estes GB. TP, Buck Quinn 2000. bears, Ursus by brown on adult gulls, Larus glaucescens, 114:217-23. nerka. Can Field-Nat Scavenging and arctos, glaucous-winged salmon, Oncorhynchus sockeye Size-selective salmon. predation 121:273-82. Oecologia by Can Rai AN. J Zool 79:1782-93. In: Galun metabolism. Nitrogen volume of Lichenology, 1988. Handbook I. Boca Eds. M, Raton CRC (FL): CRC. p 201-37. Rai Rowell P, WDP. Stewart 1983. Interactions between 137:241-51. and evolutionary aspects of ecological Can J in coastal British Columbia. Some interactions bear-salmon 78:448-57. B. 2001. Alder D, Stottlemyer H, Binkley C, Oskarsson on in of the effects soils ecosystems Agashashok (Alnus crispa) 8:89-95. northwest Alaska. Ecoscience River valley, Rhoades BJ. 1998. DE, Rogers Rogers River Lakes. Seattle Wood Salmon Alaska DJ. Rohlf 1991. ground Spawning (WA): University Technical Report No. Program reasons biological - and what work Six Species Act doesn't Biol 5:273-32. in the surveys of Washington FRI-UW-9803. the Endangered why to do about it. Conserv Schuldt JA, Hershey on Lake position nitrogen Arct Alp Res Van AE. 1995. Superior Effect tributary of salmon streams. carcass decom J N Am Benthol and concepts and and Game, Daele Game bear aid in report. report of survey-inventory (AK): Alaska Department management bear. Juneau of 147-57. p JR, Hinkes LJ, Morgan 2001. Kay con cultural of Alaska. Fairbanks, 17 brown unit management In: Hicks MV, Eds. Federal brown CT. eco 1971. Accumulation near 1998. restoration, activities: RL. ecosystems inventory RH. Grizzlies, SD, Denton Kovach MT, JW, cross and eskimos, biologists: Alaska. Ursus in southwest bear management 12:141 of ST. 1998. Carbon and nitrogen JG, Gower dynamics Vogel a green alder and without stands with boreal jack pine 1:386-400. Ecosystems understory. Ward Willson species Willson benthic browsers in an Alaskan K, JP. Bryant on vegetation floodplain. J N Am 1999. and Ecoscience Indirect effects ground-dwelling 6:505-10. nature dimensional 1995. of in lotic ecosys Conserv SM, Marston J, Caouette fish as keystone 9:489-97. Anadromous communities. Hudson on of Soc 8:2-8. KC. Gende MF, four The Benthol MF, Halupka in vertebrate BH. on J. Biol 1998. Fishes JF. the 1985. line on 1998. and interactions. fish-wildlife and Longevity the Seward Influence of Pen salmon and of biofilm response productivity: in southeastern macroinvertebrates Alaska, USA. stream Can Sci 55:1503-11. J Fish Aquat genes and spe for the twenty Press, p 76-88. of conserving 1989. The problems cies. In:Western D, Pearl MC, Eds. Conservation Woodruff DS. New York: Oxford University in three boreal alder Understory and effects on soil fertility. Alaska: local distribution Res 25:987^96. Wurtz Danell 1989. JV. tems. first century. Estes O, mammalian DC, (Alnus) 3:101-14. LJ. Daele survey wildlife Wipfli MS, carcasses Eco Rio C. 2003. J, Del Martinez JA, Berger conservation species. goals for interactive logical effectiveness: Biol 17:1238-50. Conserv Suominen In: West 13:747-66. LA, Schlentner K, Viereck in alder Cleve III, O'Keefe BF, Patterson WA of the tree of alder west persistence Can J Bot 63:1870-5. insula, Alaska. 14:259-68. ME, to in relation succession R, Viereck LA, Dyrness in taiga forest cycling Wilson nutrition FS III, Gebauer G. 1994. Nitrogen ED, Chapin tre life forms at the northern and isotope differences among 100:406-12. eline of Alaska. Oecologia sects and Productivity Can J For Res systems. perspectives expanding 48:455-62. BioScience Schulze Soule LK, Schlentner nutrient K, Oliver Cleve forest: Selective R, Rogers DE. 2000. GT, Hanson predation Ruggerone on spawning bears sockeye (Ursus arctos) foraging by brown salmon nerka). Can J Zool 78:974-81. (Oncorhynchus Soc Forest 52. TE. 2000. Reimchen Zool Van Van and 15N2-incorporation canina. Arch Microbiol and fungus during cyanobacterium in the lichen Peltigerna metabolism 1981. of temperature, Cleve K, Yarie J. 1986. Interaction moisture, in controlling nutrient and soil chemistry cycling and ecosystem in the taiga of Alaska. In: Van Cleve K, Chapin FS development LA, Dyrness CT, Eds. Forest ecosys III, Flanagan PW, Viereck tems in the Alaskan New York: p 160-89. taiga. Springer, Fish AN, LA. K, Viereck Van Van DE. 2001. K, Rogers S, Overberg TP, Wetzel L, Bishop Quinn on bear predation and age at habitat of breeding Influence of sockeye salmon and sexual dimorphism popula maturity tions. in tundra and 1981. Nitrogen cycling T, Eds. Terrestrial FE, Rosswall 33. vol. Stockholm: Bulletin, Ecological Research Science Council, p 375-404. cycles, Natural Cleve topo Alaska. In: Clarke ecosystems. 1983. 1999. MT. TP, Kinnison Quinn bears on sockeye brown V. K, Alexander series quadrangle, in the boreal forest of Alaska. cycling succession: Shugart HH, Botkin DB, Eds. Forest New York: Springer, p 179-211. application. of Power Dillingham 360 1:63 (B-8 and C-8) nutrient an integration WB, Holt RD. 1997. Toward of spatially the dynamics and food web ecology: landscape Annu Rev Ecol Syst 28:289-316. food webs. subsidized Polis Cleve boreal nitrogen Swedish 100:65-75. B, Mclnnes map, 1979. (USGS). Survey Geological graphic 103:91-3. Food web RJ. 1994. Ammonium FJ, Jackman AP, Duff JH, Avanzino a transient to channel and riparian sediments: storage sorption 26:67-83. inorganic nitrogen. pool for dissolved Biogeochemistry Triska TL. 1995. forests Can of J For and white between Interactions TL. 2000. spruce Portland forest sites in Alaska. alders at three boreal shrubby Sta Research Service Pacific Northwest Forest (OR): USDA Wurtz tion, General Technical Report This content downloaded from 128.163.8.62 on Wed, 5 Nov 2014 10:44:57 AM All use subject to JSTOR Terms and Conditions PNW-GTR-481.