Physicochemical stream bed characteristics and recruitment of the
Transcription
Physicochemical stream bed characteristics and recruitment of the
Freshwater Biology (2007) 52, 2299–2316 doi:10.1111/j.1365-2427.2007.01812.x Physicochemical stream bed characteristics and recruitment of the freshwater pearl mussel (Margaritifera margaritifera) JUERGEN GEIST*,† AND KARL AUERSWALD‡ *Fish Biology Unit, Department of Animal Science, Technische Universität München, Freising, Germany † Aquatic Toxicology Laboratory, Department of Anatomy, Physiology and Cell Biology, University of California, Davis, CA, U.S.A. ‡ Lehrstuhl für Grünlandlehre, Technische Universität München, Freising, Germany SUMMARY 1. The freshwater pearl mussel (Margaritifera margaritifera) is endangered and of conservation importance. We used its survival/mortality during the critical post-parasitic phase as a biological indicator for the habitat quality of the stream substratum. 2. We established and tested biological, physical and chemical methods of assessing the stream bed in 26 streams from seven European countries. We analysed penetration resistance, texture, the concentrations and ratios of C, N, S, P, Fe, Mn in fine material <100 lm, and redox, pH and electric conductivity at the surface and at 5 and 10 cm into the substratum. 3. Sites with high stream bed quality (promoting pearl mussel populations with good juvenile recruitment) had coarser and better sorted substrata with significantly lower quantities of fines, and a higher Mn concentration in the fines, than poor quality sites. Redox potential (Eh) at sites without recruitment differed markedly between the freeflowing water at the surface and at 5 and 10 cm in the bed, whereas no differences were detectable at good quality sites. This was also true of electric conductivity and, to a lesser extent, pH. The stream bed at sites lacking pearl mussel recruitment had a more variable and higher penetration resistance, indicating clogging of the interstitial macropore system by the deposition of mud and compaction of the stream bed. 4. Our results show that habitat quality for pearl mussels depends strongly on the exchange between the surface and the interstices, which is governed by physicochemical characteristics of the stream substratum. Combined measurements of penetration resistance, depth gradients of Eh and texture were most suitable for assessing stream bed quality, while water chemistry was insufficient because of the decoupling of interstitial and free-flowing water at poor quality sites. Keywords: ecosystem health, redox, sediment, stream substrate, unionid conservation Introduction Increasing evidence suggests that freshwater organisms are among the most threatened biota (Stein & Correspondence: Juergen Geist, Fish Biology Unit, Department of Animal Science, Technische Universität München, D-85350 Freising, Germany. E-mail: [email protected] Flack, 1997; Ricciardi & Rasmussen, 1999) and that degradation of running waters is globally widespread (Gleick, 2003). Generally, overexploitation, water pollution, flow modification, destruction or degradation of habitat, and invasion by exotic species are the five major threats to global freshwater biodiversity (Dudgeon et al., 2006). Although water quality of European rivers has steadily improved over recent decades 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd 2299 2300 J. Geist and K. Auerswald (Aarts, Van den Brink & Nienhuis, 2004), the recovery of many endangered aquatic species, mostly habitat specialists, is not proceeding accordingly. Most of these species depend on a high-quality stream substratum as a permanent or temporary key habitat for completion of their life-cycles. Aquatic ecologists have only recently become increasingly aware that biodiversity, ecosystem processes and ecosystem health in streams often depend on interactions between surface and groundwater (Palmer et al., 1997; Boulton et al., 1998; Altmüller & Dettmer, 2000; Battin et al., 2003), in the hyporheic zone (Orghidan, 1959). Stream bed conditions and their dynamics are governed by the flow regime that determines erosion and sediment deposition and remobilization. Changes in climate and land-use, as well as flow regulation and in-stream habitat alteration, can easily change these sensitive processes, have severe impacts on stream bed and, in turn, on overall ecosystem functioning. Major shifts in stream bed composition and processes can alter species distributions, productivity and even change the production of greenhouse gases (Palmer et al., 1997). Particular interest in the substratum quality of streams has mostly been focused on the impacts of the hyporheic environment to the survival of salmonid eggs in gravel bed streams (e.g. Acornley & Sear, 1999; Soulsby et al., 2001a,b; Malcolm, Youngson & Soulsby, 2003; Greig, Sear & Carling, 2005). For a much greater fraction of their long lives, freshwater mussels depend on the stream substratum quality and are perhaps the most endangered group of all freshwater animals (e.g. Bogan, 1993, 1998; Williams et al., 1993; Neves et al., 1997; Strayer et al., 2004). Given their high original numbers and biomass, and the role of bivalve molluscs in particle processing, nutrient release and sediment mixing (reviewed by Vaughn & Hakenkamp, 2001), the decline of freshwater mussels could affect profoundly ecosystem processes in streams. The characteristics of freshwater pearl mussels (Margaritifera margaritifera Linnaeus) match those of ’indicator’, ’flagship’, ’umbrella’ and ’keystone’ species, making them the focus of conservation efforts (Geist, 2005). Margaritifera margaritifera was formerly a widespread and abundant species, distributed from the Arctic and temperate regions of Western Russia through Europe to the north-eastern seaboard of northern America (Jungbluth, Coomans & Groh, 1985). Several studies have revealed dramatic declines throughout its range (e.g. Bauer, 1988), and the species is at present under a serious threat of extinction in Europe with only a small number of successfully recruiting populations remaining (Ziuganov et al., 1994; Young, Cosgrove & Hastie, 2001; Geist, 2005). As with all unionid mussels, freshwater pearl mussels have a complex life-cycle, comprising a parasitic phase on a host fish and a post-parasitic phase, in which the juvenile mussels live buried within the stream substratum. Single females can produce several million larvae (glochidia) per year (Young & Williams, 1984). As the proportion of adults producing glochidia is relatively high, even in sparse populations (Young & Williams, 1983; Schmidt & Wenz, 2000, 2001; Hastie & Young, 2003a), reduced fecundity does not seem to be the limiting factor preventing juvenile recruitment in most pearl mussel populations. Similarly, a recent study into the availability of host fishes in European freshwater pearl mussel streams revealed that a sparse host fish population seems to limit the recruitment in only a fraction of populations (Geist, Porkka & Kuehn, 2006). As juvenile pearl mussels, in their post-parasitic phase, depend on a continuously well-aerated and partly stable substratum for a period of at least 5 years, this phase in their development in which they live buried in the bed is considered to be the most vulnerable and limiting for juvenile recruitment (e.g. Bauer, 1988; Buddensiek et al., 1993; Geist, 1999a,b). River bed substratum characteristics appear to be the best physicochemical parameters for describing M. margaritifera habitat and for explaining their highly aggregated, non-random spatial dispersion (Hastie, Boon & Young, 2000). Survival of the juvenile, postparasitic stage is also therefore a potentially useful measure of stream-bed quality. Integrative methods are needed to assess and monitor the substratum quality of streams and rivers. Such methods are not only necessary with respect to the conservation of freshwater invertebrates (Strayer, 2006), but are applicable in a much broader context. Here, we suggest and test an integrative assessment of stream substratum quality, based partly on a transfer of methods from soil science to stream ecology. The recruitment of M. margaritifera was used as a biological indicator for habitat quality to assess the discrimination of sites based on the various physicochemical habitat parameters. 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Stream bed characteristics and mussel recruitment Methods Study area and biological categorization In 2002–05, 26 European streams with extant freshwater pearl mussel populations from the drainage systems of the rivers Elbe, Danube, Weser, Rhine, Aulne, Lorient, Kemijoki, Tuloma and Owenriff, in Germany, the Czech Republic, Belgium, Finland, France, Luxembourg and Ireland were investigated (Table 1; Fig. 1). Criteria for the selection of the streams were the occurrence of significant numbers Fig. 1 Map of sampling sites. 2301 or densities of pearl mussels (at present or in the past), the requirement for a wide and representative geographical range of sites, and the availability of actual information about the distribution and recruitment status of pearl mussels and their host fishes (for details see Geist et al., 2006). Additionally, we aimed to include as many populations with juvenile recruitment as possible, and matched such ’functional’ (F) streams with ’non-functional’ (NF) streams (without recent recruitment) of similar geomorphological and physicochemical properties. Prior to investigation of the stream bed, the substratum was checked for the presence of mussels with glass bottom viewing baskets or by snorkelling. Within each stream, the present range of mussels was subdivided into stretches of equal length. Depending on the total length of river with mussels, their distribution pattern, and the structural variability of the stream, five to 17 stretches along each stream were defined (mostly 5–50 m long). Within each stretch, a typical transect across the stream was analysed, including at least the middle of the channel and two spots on either side near the bank. In ’F’ populations, additional spots were analysed directly in areas where high densities of living juvenile post-parasitic pearl mussels were found, to include a sufficient number of ’F’ areas. These additional areas and the transects are referred to as ’sites’ throughout the text. In each spot, all measurements were made directly adjacent to each other, in order to be able to link the different parameters. Individual ‘spots’ were 1 m2 in size, as defined by a metal frame placed on the stream bottom. First, substratum penetration resistance was measured, followed by redox measurements and subsequent water sampling at various depths in the substratum. Finally, substratum samples for texture analysis were collected (see below). In 15 of the streams investigated, there has been no recruitment of juvenile pearl mussels for at least 30 years. These streams are referred to as ’NF’. Recent recruitment (more than two juvenile mussels younger than 5 years per m2 at sampling sites) was observed in seven ’F’ streams. A third category ’potentially functional’ (PF) was established for streams and sites where there was either (i) only very sparse juvenile recruitment at the sites investigated (i.e. only one or two mussels younger than 5 years per m2) or (ii) no or sparse juvenile recruitment in the stream for reasons clearly other than substratum quality, e.g. lack/local 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Sächsische Saale Sächsische Saale Sächsische Saale Sächsische Saale Sächsische Saale Weiße Elster Weiße Elster Weiße Elster Moldau Naab Naab Regen Gaißa Fränkische Saale fi Main Sauer fi Mosel Aller Elbe Danube Rhine Lutto Tuloma Owenriff OU LU EL LF RE Our Lutter Elez Le Fao Ruisseaux de l’ Etang du Loc’h Pikku-Luiro Jojojoki Kolmiloukkonen Ruohojärvenoja Owenriff PI JO KK FR OW ZI SR WB HB MB TB RB HaB BL CC WN BI WR KO RA SC Code Zinnbach Südliche Regnitz Wolfsbach Höllbach Mähringsbach Triebelbach Raunerbach Haarbach Blanice Hruška channel Waldnaab Biberbach Wolfertsrieder Bach Kleine Ohe Ranna Schondra Population FIN FIN FIN FIN IRE L, D, B D F F F D D D D D D D D CZ CZ D D D D D D Country F/NF F/NF F/NF F/NF F NF/PF F/NF PF PF PF NF NF NF NF NF NF NF NF F PF PF NF NF PF/NF NF NF + + + + + + + + + + + + + + + + + + + + + + + + Penetrometer + + + + + + + + + + + + + + + + + + + + + + + + + pH + + + + + + + + + + + + + + + + + + + + + Redox + + + + + + + + + + + + + + + + + + + + + + Conductivity + + + + + + + + + + + + + + + + + + + + + + + + Texture analyses + + + + + + + + + + + + + + + + + + + + + + + + C, N, P, Fe, S, Mn F, functional, i.e. recruitment takes place; NF, non-functional, i.e. without juvenile pearl mussels; PF, potentially functional, i.e. either indication but no sufficient proof for sufficient juvenile recruitment, or sites with lack of juvenile mussels because of other reasons, such as lack of host fish or recent construction of culturing channel; Country codes: B, Belgium; CZ, Czech Republic; D, Germany; F, France; FIN, Finland; IRE, Ireland; L, Luxembourg. Luiro Kemijoki Lorient Weser Aulne Subdrainage Drainage Functionality categories included (F/NF/PF) Table 1 Functionality classification for individual streams and applied analyses for bed quality assessment; ’functionality’ indicates if the stream supports recruitment of postparasitic freshwater pearl mussels (Margaritifera margaritifera) 2302 J. Geist and K. Auerswald 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Stream bed characteristics and mussel recruitment extinction of host fish. An artificial channel for the rearing of pearl mussels (the so-called ‘Hruška-channel’), with an artificial high-quality substratum, which will probably allow successful recruitment of pearl mussels, was also included in this category. Generally, all the pearl mussel streams included were oligotrophic, low in calcium, and were located in areas of similar geology with granite being the most common primary rock. In addition to the initial screening of the stream bed for mussels, grab samples for textural analyses were also screened for juvenile mussels in the field, and again when sieves were checked during textural analysis in the laboratory. Finally, after taking texture samples, the substratum at each spot was gradually removed by hand to a depth of 10–15 cm, to search for buried juvenile mussels. According to these screenings, each site was assigned to F, PF or NF. Areas without juveniles within F streams were also classified as NF. A total of 275 sites (about 1000 individual spots) comprising 46 F sites, 39 PF sites and 190 NF sites were analysed with a differing number of replicates depending on the parameter analysed (see below). In some cases, however, not all parameters could be measured because of technical (e.g. restricted substratum depth) or conservation limitations (e.g. restricted texture sampling at sites with many juvenile mussels). An unpublished pilot study showed that bed quality was poorest at the time of lowest discharge and highest water temperature during late summer, when depositions of fine sediments were greatest and oxygen concentration within the stream substratum was lowest. Thus, this period of worst-case conditions was selected for sampling, as it is limiting for the performance and survival of sediment-dwelling organisms. Texture analyses Stream substratum was sampled from 104 sites (each about 2–3 kg dry weight, each combined from all ’spots’ at one site). Sampling followed a slight modification of the cylinder method described previously (Geist, 1999a,b). Briefly, a plastic tube (total diameter of 7.4 cm, inner diameter 7.2 cm, total length 60 cm) was pushed into the substratum to a depth of 10 cm and closed with a trowel from below. The cylinder was then lifted and the material transferred into 2303 buckets. In accordance with the other analyses applied in this study, only the upper 10 cm layer was considered, as biologically the most important zone. Several studies state that species richness and abundance of the fauna in the substratum is highest in the upper layers (e.g. Palmer et al., 1997) and that this zone is also most relevant for spawning salmonids (e.g. Grost, Hubert & Wesche, 1991) and juvenile pearl mussels. Although the use of a corer is prone to the loss of fines, with greatest losses at sites with most fines and/or where large stones prevent easy sealing of the corer, the close correlation between undisturbed penetration resistance and texture shows that such losses are small compared with differences in textures within and between streams. Additionally, other sampling methods (e.g. freeze-core) would not have complied with regulations for sample collection in pearl mussel streams. Samples were cooled immediately to reduce degradation of organic matter, transported to the laboratory and stored for a maximum of 3 days at 4 C before analysis. Grain sizes were fractioned with a wet-sieving tower (Fritsch, IdarOberstein, Germany) of decreasing mesh sizes (20, 6.3, 3.15, 2.0, 1.0, 0.63, 0.2 and 0.1 mm). The fractions retained on each sieve were dried at 70 C and weighed (to nearest g). The proportions of the different grain fractions, the geometric mean particle diameter (dg) and the geometric standard deviation (SDg) were calculated according to Sinowski & Auerswald (1999). Considering the restricted grab sample volumes, the largest fraction >20 mm was generally excluded from further analyses. Chemical analyses The biologically most active fraction (<100 lm) retained from the texture analyses was homogenized with a mortar and pestle and 3 g were analysed for total C, N, S, P, Fe and Mn. Concentrations of C, N and S were measured with a Vario Max CNS analyser (Elementar, Hanau, Germany) according to German standard norm (DIN standard) protocols. The P, Fe and Mn were analysed by ICP-OES with an Optima 3000 (Perkin-Elmer, Norwalk, CT, U.S.A.) after acidic sample preparations according to DIN 38414 and DIN EN ISO/IEC 17025 (DIN, Deutsches Institut für Normung, 2001a,b). Pearl mussel streams are extremely low in carbonate and measurements of C thus reflect organic C. 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 2304 J. Geist and K. Auerswald Gradients in redox potential In order to measure redox potential (Eh) in stream water and at various substratum depths, a Platinum electrode was constructed, consisting of a Pt tip (2 · 5 mm free Pt) embedded and fixed into a 1 cm · 0.7 m plastic tube filled with epoxid resin. This resembles a basic design used previously with multiple and fixed probes for measuring Eh in microsites of the soil (Fischer, Flessa & Schaller, 1989). A standard copper electrode cable was linked to the Pt tip and also embedded into the epoxid resin inside the plastic tube. The electric potential between the Pt tip and an Ag/AgCl2 reference electrode (SCHOTT, Mainz, Germany) was measured with a handheld voltmeter (315i pH-meter; WTW, Weilheim, Germany) and Eh values referring to the standard hydrogen potential were obtained by correcting for Ag/AgCl2 and temperature. The low electric conductivity in most of the pearl mussel streams (often below 150 lS cm)1) was compensated by inserting the Ag/ AgCl2 reference electrode into a 3 M KCl agarose-gel filled salt bridge, a 2-cm wide and 0.7-m long PVC tube with a permeable ceramic membrane on the bottom end. The Eh and reference electrodes were calibrated with a standardized redox buffer solution of Eh ¼ 220 mV and pH 7.0 (Mettler Toledo Process Analytics, Greifensee, Switzerland). Redox potential was measured at 5–17 transects (45–153 spots) per stream. At each single spot, Eh was first measured in the free-flowing water, and then at depths into the substratum of 5 and 10 cm. Values below Eh ¼ 300 mV indicate anoxia and values above 300 mV oxic conditions (Schlesinger, 1991). Gradients in conductivity and pH Differences in electrical conductivity and pH between interstitial and free-flowing water indicate low hydrological exchange between the water column and interstices. Conductivity (corrected to 20 C) and pH were measured using handheld 315i conductivity and 315i pH-meters (WTW, Weilheim, Germany) in freeflowing water and in interstitial water from 5 and 10 cm depth. Samples of interstitial water were taken using a 30-cm long fixed PVC tube (outside diameter 5 mm, inner diameter 3.5 mm), attached to a 1-m long flexible plastic hose and a 50-mL syringe (Braun, Melsungen, Germany), which was used to create a vacuum. The tube was marked with coloured lines for sampling pore water from 5 and 10-cm depths. For each sampling spot, 15 mL of water were extracted from each depth and from the free-flowing water. Samples were transferred to 50-mL Falcon tubes (ROTH, Karlsruhe, Germany) for immediate measurement. Although tube sampling in coarse substrata may be prone to water being drawn down the sides of the tube from the surface, our experiments in various pearl mussel streams revealed significant physicochemical differences between interstitial and surface water. Moreover, the properties of water from different depths did not suggest mixing. Both observations indicate that the method is reliable. No measurements of dissolved gases were carried out in tube water samples, as we were cautious about changes in dissolved gases because of vacuum sampling. Penetration resistance of the substratum Resistance of the substratum surface was measured with a hand-held pocket penetrometer (Eijkelkamp Agrisearch Equipment, Giesbeek, the Netherlands) to assess the interface between free-flowing water and the interstitial zone. Low resistance indicates unconsolidated fine sediment, whereas high values can either indicate consolidation, e.g. by colmation, or an extremely coarse substratum. Although penetrometer analyses are frequently used in soil science (e.g. Nearing & West, 1988; Bradford, Truman & Huang, 1992; Becher, 1994), only a few authors (e.g. Johnson & Brown, 2000) have used them for assessment of stream beds. Normally, penetration resistance was measured nine times at each site (three in each spot on either side near the bank and in the middle of the stream). In F areas, measurements were carried out directly where juvenile pearl mussels occurred. At NF and PF areas, measurements were made where the stream bed seemed to be a potential habitat for sediment-dwelling organisms. This included gaps between larger stones and boulders, but excluded outcrops of pure bedrock, boulders or stones. The blunt ended tip of the penetrometer was pushed into the stream bed to a depth of 6 mm and the resulting resistance was read in kg cm)2. In order to ensure penetrometer readings over a wide variety of stream bed types, four metal adapter discs with diameters of 15, 18, 20 and 25 mm were used and the resulting 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Stream bed characteristics and mussel recruitment readings corrected according to the adapter area. A minimum value of 0.001 kg cm)2 was assigned to areas with extremely soft mud, where even the use of the largest adapter disc did not produce any detectable reading. 2305 (a) Statistical analyses The F, NF and PF sites were mainly compared by oneway A N O V A , ordinary linear regressions and t-tests implemented in S A S (version 9.1, SAS Institute Inc., Cary, NC, U.S.A.). In some cases, in which the averages between groups were similar but the variances were different, Cochran’s test (Cochran, 1941) was performed to test for significant (P < 0.05) differences in variances. To account for the different sample sizes between functionality groups, the harmonic mean of the sample size was used. To characterize the distribution of values within the different functionality groups, Box–Whisker plots were drawn, and additionally, kernel density distributions were estimated. Kernel density distribution is advantageous where the distribution is multi-modal or where differences between distributions are small. For the theory of kernel density estimation see Silverman (1986). Stepwise logistic regression was used to identify the most relevant parameters for discriminating between F and NF sites. Their interdependence was then tested further by multiple regression analyses containing a dummy variable, where the dummy variable was 1 for NF and 0 for F (Fox, 1997). The regression coefficient of the dummy variable then denotes the trued difference between F and NF. Results Single factor analysis Water chemistry Although F sites had significantly lower conductivity and pH, and higher Eh than NF sites, the quantitative differences were modest, the variation was wide and the groups overlapped (Fig. 2a–c). Generally, all streams included in this study had continuously high oxygen saturation in the water column and no fish or mussel mortality because of oxygen depletions had been observed in any of these streams. The free-flowing water conditions per se seem not likely to be crucial for the survival of juvenile freshwater pearl mussels. (b) (c) Fig. 2 Box–Whisker plots (crosses: minimum, maximum; Whisker: 0.05 and 0.95 percentiles; Box: 0.25 quartile, median and 0.75 quartile) of depth profiles for redox potentials (a) electrical conductivity (b) and pH (c) at functional (n ¼ 109, 46, 44, respectively), potentially functional (n ¼ 57, 42, 42, respectively) and non-functional sites (n ¼ 254, 244, 210, respectively), in the free-flowing water and at 5 and 10-cm depth. The right axis indicates experimentally determined Eh values at pH 7 where Mn2+ and Fe2+ formation start and below which O2 and NO 3 are no longer detectable (after Brümmer, 2002). Penetration resistance Log-transformed penetrometer readings differed highly significantly (P < 0.001) in SD between F, NF and PF according to Cochran’s test, but there was no significant difference between means. Geometric mean penetration resistance at F sites averaged 0.16 kg cm)2 and varied over one order of magnitude ranging from 0.04 to 0.39 kg cm)2 (Fig. 3). In contrast, NF sites had similar resistance (geometric mean, 0.18 kg cm)2), but the variability ranged over more than three orders of magnitude from <0.001 kg cm)2 at sites with accumulation of soft 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 J. Geist and K. Auerswald – 2306 Fig. 3 Box–Whisker plots (crosses: minimum, maximum; Whisker: 0.05 and 0.95 percentiles; Box: 0.25 quartile, median and 0.75 quartile) for penetration resistance at functional (n ¼ 182), potentially functional (n ¼ 160) and non-functional sites (n ¼ 659). mud up to 4.00 kg cm)2. PF sites were intermediate with resistances ranging from 0.03 to 0.80 kg cm)2. For a number of NF sites, where penetration resistance was monitored over several months, resistances were found to be smallest in October and highest in April, apparently caused by the accumulation of soft fine material over the summer and its remobilization during spring floods. Substratum texture The F and NF streams differed markedly in texture, with a higher percentage of fines at NF sites (Fig. 4). Grain size distribution generally followed a log-distribution and can thus be described by the mean geometric particle diameter dg and its SDg. On average, dg was 7 mm at F sites and 3 mm at NF sites, while average SDg was significantly larger (P < 0.001, Cochran’s test) for the NF sites. On average, F sites contained <3% (max. 7%) of particles <200 lm and <2% (max. 5%) of particles <100 lm. In contrast, NF sites had on average 13% (max. 78%) of material <200 lm and 9% (max. 56%) particles <100 lm. A similar relation was also evident for fines <1 mm, with NF sites having on average 35% (max. 96%), and F sites only containing 18% (max. 38%). Opposite patterns were observed for the larger grain sizes, with F sites generally having a coarser substratum. Chemical composition of fines The concentrations and ratios of N, C, S, P and Fe varied largely and overlapped widely between the different functionality categories (Table 2). Although the mean concentrations of S and P were significantly greater at F than at NF sites, concentrations overlapped between the two. Mn discriminated strongly between the two groups of streams, with an average 10-fold higher concentration at F than at NF sites. This difference is mainly attributable to one-third of the F samples, which had Mn concentrations up to about 150 g kg)1 in the material <100 lm. The highest concentrations of Mn were found at well-sorted sites with a low percentage of fines. The precipitation of the sensitive redoxindicator Mn (reduced and mobile when oxygen is just depleted, and reoxidized and immobile at Eh ¼ 0.35; Brümmer, 2002) was evident only at sites with SDg <10 lm, i.e. presumably where reduced Mn from deeper, anaerobic layers was oxidized and immobilized in the top layer because of a good exchange between free-flowing water and interstitial water. At NF sites with a high percentage of fines and anoxic conditions in the surface layer, Mn was also mobilized from these zones, resulting in Mn values below the minimum values for F sites (Table 2). For Fe, which only becomes mobile at Eh 0.15 V, no difference between F and NF sites was evident, presumably because such low Eh values did not occur. Spatial interactions Fig. 4 Grain size distribution of the stream bed at functional (n ¼ 14), potentially functional (n ¼ 21) and non-functional (n ¼ 69) sites; shaded areas indicate the range between minimum and maximum values. Depth profiles in redox potential Redox potential depth profiles differed markedly between F and NF sites (Fig. 2a). At F sites, Eh in the free-flowing water averaged 0.53 V with only marginal differences in Eh 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Stream bed characteristics and mussel recruitment Table 2 Comparison of the chemical composition of the biologically most active fraction <100 lm between non-functional (n ¼ 76), potentially functional (n ¼ 17) and functional streams (n ¼ 13) where the fraction <100 lm on average contributes 7.8%, 1.4% and 1.5%, respectively NonPotentially functional functional Functional N (%) Minimum 0.14 0.22 5 percentile 0.18 0.26 Arithmetic 0.32 A 0.42 B mean 95 percentile 0.53 0.59 Maximum 0.61 0.82 C (%) Minimum 1.30 1.80 5 percentile 1.70 2.90 Arithmetic 3.69 A 4.65 A mean 95 percentile 6.50 7.75 Maximum 9.60 9.85 C/N Minimum 8.08 8.18 5 percentile 8.76 9.64 Arithmetic 11.24 A 11.02 A mean 95 percentile 14.00 12.32 Maximum 15.74 13.08 S (%) Minimum 0.03 0.05 5 percentile 0.04 0.07 Arithmetic 0.12 A 0.16 B mean 95 percentile 0.35 0.35 Maximum 0.54 0.38 P (mg kg)1) Minimum 323 215 5 percentile 383 256 Arithmetic 765 A 1801 B mean 95 percentile 1306 3336 Maximum 2406 3556 Fe (g kg)1) Minimum 14.5 18.6 5 percentile 20.8 23.8 Arithmetic 40.0 A 37.4 A mean 95 percentile 53.9 60.0 Maximum 63.2 67.1 Mn (g kg)1) Minimum 0.3 0.4 5 percentile 0.6 0.6 Arithmetic 2.8 A 3.1 A mean 95 percentile 7.0 7.8 Maximum 10.3 9.2 All 0.14 0.18 0.38 AB 0.14 0.19 0.35 0.60 0.72 0.58 0.82 1.41 2.22 4.60 A 1.30 1.73 4.01 7.89 12.99 7.25 12.99 9.97 9.99 11.82 A 8.08 8.79 11.28 15.90 17.99 14.06 17.99 0.07 0.07 0.25 C 0.03 0.04 0.14 0.53 0.64 0.43 0.64 280 197 1451 B 215 338 1064 3034 3042 3039 3556 17.1 25.3 51.7 A 109.2 116.8 1.4 1.7 28.6 B 115.1 148.3 14.5 20.1 41.1 61.6 116.8 0.3 0.6 6.5 10.0 148.3 Means within a row followed by different letters differ significantly at P < 0.05. 2307 at 5 cm (on average 0.51 V) and 10-cm depth (0.47 V), indicating either an intense exchange between the water column and interstitial water or low levels of respiration. At all F sites, Eh at 5 and 10-cm depth (with a few exceptions at 10 cm) indicated oxic conditions even during summer (min. Eh at 5 and 10-cm depth at 0.33 V and 0.24 V, respectively). In contrast, average Eh at NF sites decreased from 0.47 V in the free-flowing water to 0.33 V at 5 cm and 0.27 V at 10 cm depth. PF sites were intermediate. Long-term monitoring series of Eh at specific sites (data not shown) showed that Eh was generally lowest during summer, coinciding with high water temperature, low water flow, accumulation of fines, low oxygen concentration and rapid decomposition. During winter, at low water temperatures and high flow rates, interstitial Eh was much higher and oxygen was often available at sites which were clearly identified as NF and oxygen-depleted during the summer. For instance, the average difference in Eh at 10 cm between F and NF sites for one continuously monitored stream was 0.24 V during ‘worst-case’ conditions in summer, but only 0.18 V at other times of the year. Depth profiles in conductivity and pH Conductivity in the free-flowing water ranged from 11 lS cm)1 in northern Lapland to 154 lS cm)1 in Germany. The mean electric conductivity differed significantly among stream groups. This was evident in the freeflowing water and at 5 and 10-cm depth (Fig. 2b). Mean electric conductivity was lowest at F sites. The mean electric conductivity of the free-flowing water was similar to that at 5 and 10-cm depth in all three functionality groups. This was explained by the fact that a separation between the water column and interstitial water resulted in either higher or lower conductivity in the interstitial, but without changing the average substantially. The absolute difference (positive or negative) between the free-flowing and interstitial water at 5 and 10-cm depth was about 20 lS cm)1 greater at NF sites than at F or PF sites, where only minor deviations of a few lS cm)1 were found (Fig. 5). The mean absolute difference from the free-flowing water was only slightly larger for 10-cm depth than for 5-cm depth (for NF sites on average 21 versus 18 lS cm)1), indicating a barrier between freeflowing and interstitial water within the first 5 cm. In particular in NF sites, apparently restricted exchange between free-flowing and interstitial water occasion- 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 2308 J. Geist and K. Auerswald Table 3 Final coefficients of a stepwise logistic regression with SE and error P of the Wald chi-square test for the odds of functional versus non-functional streams depending on geometric mean diameter dg and penetration resistance (n ¼ 78) Intercept dg (mm) Penetration resistance (kg cm)2) ally caused differences in conductivity of 100 lS cm)1 or more, with a maximum deviation of 300 lS cm)1, whereas F sites only had a maximum difference of 18 lS cm)1. Only four out of 138 F sites (<3%) had a deviation >10 lS cm)1 between surface and interstitial water. The pH differed less clearly between F and NF sites than did the other parameters. While on average the pH in the free flowing water was higher at PF and NF sites than at F sites (Table 2), this difference decreased with increasing depth into the substratum (Fig. 2c). The higher mean in the free-flowing water at NF and PF sites was mainly caused by some very high values (up to pH 8.8) indicating intense CO2 removal during the day. Because of a restricted hydrological exchange with interstitial water, especially at NF sites, these peaks were either only partly transferred to the interstitial water or compensated for by respiration. Depth profiles in conductivity and pH identified a number of NF sites with limited hydrological exchange between the surface and the interstitial zone. SE P-value )1.5 1.3 )31.1 1.0 0.4 11.9 0.131 0.001 0.009 classified 95% of the sites (whereas the best combination of all water chemistry parameters correctly classified only 77%). The mechanistic interdependence between both parameters, however, was better reflected in a multiple regression, using dg as independent and penetration resistance as dependent variable, and a dummy variable to separate sites according to the presence or absence of juvenile mussels (’F’ versus ’NF’) (Fig. 6): lgðrÞ ¼ 0:9ð0:3Þ þ 0:71ð0:09ÞlgðdÞ 0:5ð0:1Þf r2 ¼ 0:463; P < 0:001; n ¼ 78 ðSE’s in parenthesesÞ With r ¼ resistance in kg cm)2, d ¼ median grain diameter in mm and f ¼ functionality (1 ¼ F, 0 ¼ NF). The regression indicates that penetration resistance increases highly significantly with increasing d. Coarse substrata have a higher penetration resistance than fine textured substrata. Variation, however, is high and can be more than one order of magnitude, especially in coarse stream beds. The highly significant (P < 0.001) coefficient of parameter f – Fig. 5 Frequency of the absolute difference in electrical conductivity between the free-flowing water and the interstitial water in either 5 or 10-cm depth for functional (n ¼ 92), potentially functional (n ¼ 84) and non-functional sites (n ¼ 480). Coefficient Interactions between parameters Multivariate analyses were found to discriminate most effectively between F and NF sites. Stepwise logistic regression retained only substratum parameters, but no physicochemical parameters of the freeflowing water. In particular, Eh (which is, in turn, influenced most strongly by dg and penetration resistance), explained most of the differentiation between F and NF sites. The best logistic regression used only dg and penetration resistance to distinguish between F and NF sites (Table 3) and correctly Fig. 6 Dummy regression of penetration resistance depending on geometric mean diameter and functionality for functional (d) and non-functional (+) sites (r2 ¼ 0.463, P < 0.001, for equation see text). Potentially functional sites (s) are shown but not included in the regression. For comparison, the logistic regression function is displayed as shades of grey. 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Stream bed characteristics and mussel recruitment indicates that NF sites had a threefold higher penetration resistance than F sites at an identical d. The interaction of penetration resistance and grain size determines stream bed quality for pearl mussels, indicating that juveniles need a coarse but loose substratum where neither grain size nor compaction restrict the exchange of free-flowing water with the interstitial. In addition, there was a weaker though significant correlation between penetration resistance and SDg (r2 ¼ 0.046, P < 0.05). Heterogeneous substrata, where pores created by large particles are more prone to be filled or clogged by fine particles, have more inter-particle bonds than better sorted sediments. Penetration resistance was also found to decrease with increasing carbon content (r2 ¼ 0.117, P < 0.001) because of the lesser compaction caused by the approximately eight times lower density of organic particles in water (c. 1.2–1.0 ¼ 0.2 g cm)3) when compared with inorganics (c. 2.6–1.0 ¼ 1.6 g cm)3). However, as dg and organic matter content were linked (r2 ¼ 0.086, P < 0.01), both regressions were partly biased by this collinearity. Redox potential within the stream bed decreased strongly with decreasing dg (Fig. 7). Eh at 10-cm depth was about 100 mV lower than at 5-cm depth, although texture had identical effects at both depths. This relation was not caused by the organic matter content, which increased with decreasing dg, but presumably by the reduced exchange of water between the interstitial and the stream in finer sediments. Eh was independent of the organic matter content (r2 ¼ 0.007, NS at 5-cm depth and r2 ¼ 0.001, NS below), indicating that oxygen concentration and redox in the hyporheic zone were primarily governed by the 2309 supply of oxygen from the free-flowing water into the interstitial zone, and not primarily by the amount of labile organic material in the stream bed. The effect of dg on Eh in the free-flowing water was less pronounced (r2 ¼0.092, P < 0.01) than the effect within the stream bed (r2 ¼ 0.284, P < 0.001 at 5-cm depth and r2 ¼ 0.246, P < 0.001 below). Eh also decreased with SDg of the texture (r2 ¼ 0.171, P < 0.001, r2 ¼ 0.249, P < 0.01 and r2 ¼ 0.164, P < 0.001 in free-flowing water and at 5 and 10-cm depth, respectively), indicating that poorly sorted stream beds were more prone to oxygen depletion. An inverse correlation between the C/N ratio in fine material and electric conductivity was observed (a high C/N is expected when organic matter is derived from forested catchments and a low C/N indicates eutrophication or agricultural-derived organic matter) (r2 ¼ 0.042, P < 0.5, r2 ¼ 0.040, P < 0.01 and r2 ¼ 0.024 NS, in free-flowing water, and at 5 and 10-cm depth, respectively). Highest electric conductivity was found at sites with a C/N ratio of 8–10. Similarly, the C/N ratio decreased slightly with pH, presumably because of the stronger acidification of forested catchments as compared with agricultural, limed catchments (r2 ¼ 0.023 NS, r2 ¼ 0.091, P < 0.01 and r2 ¼ 0.067, P < 0.01 in free-flowing water, and at 5 and 10-cm depth, respectively). The pH decreased slightly with decreasing dg, indicating a higher probability for acidification in catchments with coarse textured stream beds, but the degree of acidification was low and thus not limiting (r2 ¼ 0.097, P < 0.01, r2 ¼ 0.065, P < 0.01 and r2 ¼ 0.061, P < 0.05 in free-flowing water, and at 5 and 10-cm depth, respectively). Discussion The discrepancy between free-flowing and interstitial water Fig. 7 Influence of geometric mean substratum particle diameter on redox potentials for the free-flowing stream water (x), and for 5 cm (s) and 10 cm (d) stream-bed depths (n = 90). The characteristics of the stream substratum, in particular the Eh depth profile and a combination of penetration and textural analyses, were powerful indicators of pearl mussel recruitment and discriminated well between ’F’ and ’NF’ sites. Physicochemical parameters of the free-flowing water overlapped between F and NF streams and were less useful for assessing the suitability of individual sites for pearl mussels. This concurs with observations that the survival of juvenile pearl mussels raised in cages in 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 2310 J. Geist and K. Auerswald the water column can strongly differ from that within the natural stream substratum in the same streams (Strecker, Bauer & Wächtler, 1990). Our findings do not necessarily contradict previous studies, which showed that long-term data sets for several water parameters can be useful in defining the water quality requirements of pearl mussels (e.g. Bauer, 1988; Sachteleben et al., 2004). The higher predictive power of the substratum parameters can be explained by the fact that: (i) the habitat requirements of adult and juvenile pearl mussels differ; (ii) only substratum parameters can accurately describe the habitat of juveniles and (iii) substratum parameters are less prone to short-term fluctuations (which limit the use of free-flowing water parameters if long-term data sets are not available). We observed that adult and juvenile pearl mussels used different microhabitats, although both life stages occurred simultaneously in the same habitat patches. During summer, adult mussels were usually only partly buried and filtered free-flowing stream water, whereas juveniles (younger than 5 years) were completely buried and thus exposed to interstitial water. Our findings that physicochemical conditions between the surface and interstices were very similar in F streams but could differ strongly in NF streams, shows that conditions in the free-flowing water are weak predictors of interstitial water quality and thus of habitat quality for juvenile pearl mussels and other interstitial organisms. Based on our results, habitat quality seems to be governed by the physical connectivity of free-flowing water and the interstitial zone. The hyporheic zone can account for a high proportion of the total respiration of the stream system (e.g. Fellows, Valett & Dahm, 2001; Battin et al., 2003) and, at sites with insufficient exchange between both compartments, anoxic conditions within the stream substratum can occur. Although pearl mussels can withstand shortterm oxygen deficits, the bivalve body plan imposes limitations to oxygen regulation (Wilson & Moorkens, 1992). A low concentration of dissolved oxygen can slow growth in clams and mussels (e.g. Belanger, 1991), and may result in reduced fitness. The pronounced Eh depth gradient, the discrepancy in electrical conductivity between free-flowing and interstitial water, and the threefold higher penetration resistance at NF sites, all suggest a separation of interstitial and free-flowing water. Factors controlling stream bed quality Previous studies suggested a link between the species composition and abundance of unionids and substratum particle size distribution (e.g. Neves & Widlak, 1987; Leff, Burch & McArthur, 1990; Hastie et al., 2000), although single parameters often failed to explain fully mussel distributions and recruitment. Brim Box & Mossa (1999) suggested that the lack of correlation between particle size and mussel distribution in some studies may result from inadequate sampling and insufficient substratum analyses, but it seems more likely that this was caused by the complexity of stream bed processes, which are determined by several factors at the same time. In our study, multivariate models discriminated best between F and NF pearl mussel habitats. Two different multivariate analyses were most successful; (i) the measurement of Eh at different depths and (ii) the combination of textural analysis and penetration resistance. The measurement of Eh depth gradients quantifies the effects of the hydrological barrier on juvenile mussel survival, while the combined measurement of texture and penetration resistance assesses the causes of hydrological segregation of surface and subsurface water, namely the consolidation of the stream bed surface. NF stream beds often suffered from a lack of sorting and a deposition of fines, which were most probably caused by anthropogenic habitat modifications, including changes of the natural flow regime, constructions of dams and weirs, the removal of wood and boulders from streams, abstraction and increased soil erosion. Both the smaller grain sizes and the larger variation in grain size at NF sites decrease the inter-particle voids, because smaller grains can occupy voids between larger grains (Hwang & Powers, 2003). Hence, smaller interstitial pore volumes and, particularly, lower interstitial water flow can be expected at NF sites. Furthermore, only a relatively small range of penetration resistance values seems to allow for successful mussel recruitment and a number of NF sites may be either too soft or too hard. Because of the large overlap of penetration resistance or texture between NF and F sites, however, only the combination of both is sufficient to characterize stream bed quality. Physicochemical gradients in the interstices result from several processes, including (i) the hyporheic flow pattern and the different properties of surface 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Stream bed characteristics and mussel recruitment and groundwater; (ii) retention, caused by the filtering effect of pore size and lithologic sorption as well as the transient storage of solutes caused by reduced water velocity and (iii) biogeochemical transformations in conjunction with local residence time (Brunke & Gonser, 1997). In our data set, these gradients seem to be mainly influenced by the physical state of the stream bed. F sites had a non-compacted, well-sorted, coarse substratum with a significantly lower fraction of fines than NF sites. Interstitial Eh, pH and electrical conductivity at F sites showed no or only weak deviations from those in the free-flowing water. Substratum quality and pearl mussel recruitment thus seem to be closely linked to the exchange rate between free-flowing and interstitial water. The intensity of this exchange largely depends on the stream-bed permeability, which is influenced by the available pore volume, pore size and the boundary layer at the substratum surface. Our results for depth profiles in Eh and the water chemical parameters suggest that the streambed surface is most crucial for the rate of exchange between free-flowing and interstitial water. In poorly sorted or fine-grained stream beds, the macropore system can clog and restrict these exchange rates (e.g. Schälchli, 1992; Richards & Bacon, 1994). Fine sediments may be fatal where they limit oxygen supply of the interstitial zone and where eutrophication effects foster high biological oxygen demands in both the free-flowing water and the interstitial zone. Hence, the lowest Eh values were observed during periods of high water temperature and at sites with a high percentage of fines <100 lm. Fine sediment loads thus can substantially impact the hyporheic exchange and associated ecological processes depending on the stream flow conditions, the rate and frequency of bed remobilization, and the extent of interaction of the introduced fines with bed sediments (Regh, Packman & Ren, 2005). Previous studies have reported negative effects of increased fine sediment deposition on pearl mussel habitats (Bauer, 1988; Buddensiek et al., 1993; Geist, 1999a,b), and for other invertebrate taxa (Strayer et al., 1997). The effect of fines on unionid growth under artificial culturing conditions remains controversial (Jones, Mair & Neves, 2005; Barnhart, 2006), and it seems they can promote microbial growth and have nutritional effects when oxygen is not limiting. Under natural conditions, only a few authors observed successful pearl 2311 mussel reproduction in fine sediment or peat-dominated areas (Cosgrove & Harvey, 2003; Altmüller, unpubl. data). Salmonid egg development is also negatively affected by fine sediment deposition (e.g. Magee, McMahon & Thurow, 1996; Malcolm et al., 2003; Curry & McNeill, 2004). However, poor quality (NF) streams in terms of Eh and texture (e.g. ZI, SR, BI, KO, WR, WB, SC) were nevertheless found to have high densities and biomass of sediment spawning fish, such as brown trout (Salmo trutta) and bullhead (Cottus gobio) (Geist et al., 2006). Gravel-spawning fish seem to be more tolerant of unfavourable substratum conditions than pearl mussels, as (i) spawning usually takes place in winter when flow and oxygen supply are high and respiration low; (ii) reproductive success depends on a much shorter time interval in the stream bed; (iii) fish can actively search for the most favourable sites and (iv) fish are able to actively engineer the substratum in redds and spawning pockets, i.e. by cleaning spawning sites from fines (e.g. Grost et al., 1991). Indeed, great differences in interstitial physicochemical conditions between summer and winter were evident for most NF sites (data not shown) and have also been described in other studies (e.g. Olsen & Townsend, 2003). Surprisingly, C, N, the C/N ratio, S and Fe and P in fines did not separate F and NF sites, although a link between eutrophication indicators and stream habitat quality for pearl mussels was expected from previous studies (Bauer, 1988; Sachteleben et al., 2004). Even though small significant differences between F and NF streams could be detected for these parameters, they resulted from collinearity with the causal factors, but could neither clearly indicate nor explain functionality. Slight eutrophication in itself does not seem to be harmful for pearl mussels at sites where the stream substratum still has a good hydrological exchange between surface and interstitial water. An ideal substratum for pearl mussels combines at the same time attributes of substratum quality and, to a certain extent, structural stream bed stability, preventing the destruction of the pearl mussel microhabitats and the scouring and drift of juvenile pearl mussels to less favourable sites. Howard & Cuffey (2006a) found highest recruitment and lowest mortality of a North American pearl mussel, Margaritifera falcata (Gould), at low shear stress in low-discharge years, while large floods can even produce severe 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 2312 J. Geist and K. Auerswald short-term population depletions of the adults (Hastie et al., 2001). Further, Strayer (1999), Johnson & Brown (2000) and Gangloff & Feminella (2007) report that mussel aggregations occurred primarily within hydraulic refugia, i.e. areas of the stream bed that remain stable at high flow. Our observations indicate that the best pearl mussel microhabitats in larger rivers are mostly at sites where coarse, well-sorted sediment is stabilized by boulders or stones and where both requirements of substratum quality and stability are fulfilled. Unstable substratum conditions could cause the lack of juvenile mussels in some of the PF sites. For instance, the transport of sand was associated with a lack of pearl mussel reproduction in northern Germany (Altmüller & Dettmer, 1996). On the other hand, substratum disturbance, mobility and rearrangement during floods may remove the fines and thus be important in preventing compaction of sediment. The requirements for structural substratum stability over the 5-year period of larval development and for clean, oxic substrata may be fulfilled under two principal conditions: (i) overall stable stream beds, with a low rate of fine sediment deposition and eutrophication, plus very scarce (every 5–10 years) flood events that clean the stream bed and (ii) dynamic co-occurrence and succession of unstable and stable habitat patches in one stream, resulting in different recruitment patterns. Implications for conservation and stream habitat restoration Stream bed quality appears to be the most important habitat factor limiting the recruitment of the endangered freshwater pearl mussel in many European rivers. Unionid mussels, in particular juvenile freshwater pearl mussels with their stringent habitat requirements, are useful indicators of stream substratum quality and processes, themselves playing an important role in aquatic ecosystems with impacts on other benthic organisms (Howard & Cuffey, 2006b; Spooner & Vaughn, 2006). Our results underscore the importance of considering the interstitial habitat for the conservation and restoration of functional stream ecosystems. The strong substratum-dependence of pearl mussel recruitment, and the small number of remnant pearl mussel populations indicate that the balance of erosion, deposition, sorting and flow in lotic ecosys- tems is a subtle and fragile system which can easily be altered. The role of natural disturbances (e.g. flood events) and human impacts (e.g. catchment land-use, flow regulation) governs the function of the ‘hyporheic zone’ (Boulton et al., 1998), and should be an important focus of long-term conservation and restoration strategies. The ecological integrity of river ecosystems depends on their dynamic character (Poff et al., 1997) and requires the restoration of natural flow dynamics in whole river catchments, rather than artificial and short-term measures, such as the local addition of gravel. Efforts to bridge substratum-dependent stages in the life cycles of endangered organisms, such as breeding and culturing of freshwater pearl mussels in cages (Buddensiek, 1995; Hastie & Young, 2003b) or the breeding of other endangered unionids in captivity (e.g. Jones et al., 2005; Barnhart, 2006 and references therein), can only be seen as emergency measures to retain the evolutionary and genetic potential of such species (Geist & Kuehn, 2005; Geist, 2005). They cannot replace measures to restore the natural habitat. Both bioindicators, such as the recruitment of unionid mussels, and technical approaches can be used to assess habitat quality, e.g. to evaluate habitat restoration measures. Biological indicators have high ecological relevance but usually involve long lag-times (about 5 years for pearl mussels). Technical measures, such as the measurement of Eh depth profiles suggested here, are a suitable and convenient indicator for stream-bed quality, and can be derived from in situ measurements instantaneously. The combination of penetration measurements and textural analyses (or their replacement by in situ field ratings, e.g. McDonald et al., 1990; Oberthür, Goovaerts & Dobermann, 1999) provides a powerful indicator of a decoupling between interstitial and free-flowing water and may also be useful in assessing the effects of changes in the flow-regimes, which seem to govern stream-bed processes. The methodology applied here may provide a tool for assessing and improving stream substratum quality for freshwater pearl mussels and other substratum-dependent species. The successful restoration of stream substrata is costly and time intense but is probably the most essential aspect of ecosystem health in rivers. 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Stream bed characteristics and mussel recruitment Acknowledgments We are grateful to A. Wurzinger and W. Sitte from ‘Bayerische Landesanstalt für Landwirtschaft’ for their support with chemical analyses, to F. Krüger for technical assistance with Pt electrode construction and to I. Kögel-Knabner and H. H. Becher for providing laboratory space for texture analyses. J. Geist acknowledges the financial support provided by ‘Bayerischer Naturschutzfonds’, ‘Landesfischereiverband Bayern’ and ‘Bayerische Forschungsstiftung’. The authors would like to thank Markku Porkka (Finland), S. Terren (Belgium), Dr E. Moorkens and I. Killeen (Ireland), E. Holder (France), J. Hruška (the Czech Republic), Dr R. Altmüller, R. Dettmer, M. Lange, F. Elender, W. Silkenat, K. Dietl (Germany) for their hospitality, their support and assistance during sample collections and field investigations. The assistance of A. Beck, C. Bottlender, M. Formánková, F. and Chr. Geist and B. Reindl during field sampling was highly appreciated. We also acknowledge the help of all government and park authorities involved for providing permits for investigating protected pearl mussel streams and habitats in various regions of Europe. References Aarts B.G.W., Van den Brink F.W.B. & Nienhius P.H. (2004) Habitat loss as the main cause of the slow recovery of fish faunas of regulated large rivers in Europe: the transversal floodplain gradient. River Research and Applications, 20, 3–23. Acornley R.M. & Sear D.A. (1999) Sediment transport and siltation of brown trout (Salmo trutta) gravels in chalk streams. Hydrological Processes, 13, 447–458. Altmüller R. & Dettmer R. (1996) Unnatürliche Sandifracht in Geestbächen – Ursachen, Probleme und Ansätze für Problemlösungen – am Beispiel der Lutter. Informationen des Naturschutz Niedersachsen, 16, 222– 237. Altmüller R. & Dettmer R. (2000) Erste Erfolge beim Arten- und Biotopschutz für die Flussperlmuschel (Margaritifera margaritifera L.) in Niedersachsen. Natur und Landschaft, 75, 384–388. Barnhart M.C. (2006) Buckets for muckets: a compact system for rearing juvenile freshwater mussels. Aquaculture, 254, 227–233. Battin T.J., Kaplan L.A., Newbold J.D. & Hendricks S.P. (2003) A mixing model analysis of stream solute dynamics and the contribution of a hyporheic zone to ecosystem function. Freshwater Biology, 48, 995–1014. 2313 Bauer G. (1988) Threats to the freshwater pearl mussel Margaritifera margaritifera in central Europe. Biological Conservation, 45, 239–253. Becher H.H. (1994) Evaluating the resistance to penetration within the surface layer of soil aggregates. Soil Technology, 7, 105–111. Belanger S.E. (1991) The effect of dissolved-oxygen, sediment, and sewage-treatment plant discharges upon growth, survival and density of Asiatic clams. Hydrobiologia, 218, 113–126. Bogan A.E. (1993) Freshwater bivalve extinctions (Mollusca: Unionoida): a search for causes. American Zoologist, 33, 599–609. Bogan A.E. (1998) Freshwater molluscan conservation in North America: problems and practises. Journal of Conchology Special Publication, 2, 223–230. Boulton A.J., Findlay S., Marmonier P., Stanley E.H. & Valett H.M. (1998) The functional significance of the hyporheic zone in streams and rivers. Annual Review of Ecology and Systematics, 29, 59–81. Bradford J.M., Truman C.C. & Huang C. (1992) Comparison of three measures of resistance of soil surface seals to raindrop splash. Soil Technology, 5, 47–56. Brim Box J. & Mossa J. (1999) Sediment, land use, and freshwater mussels: prospects and problems. Journal of the North American Benthological Society, 18, 99–117. Brümmer G. (2002) Redoxreaktionen. In: Lehrbuch der Bodenkunde (Eds H.-P. Blume et al.), pp. 136–143. Spektrum Akademischer Verlag, Heidelberg, Germany. Brunke M. & Gonser T. (1997) The ecolocial significance of exchange processes between rivers and groundwater. Freshwater Biology, 37, 1–33. Buddensiek V. (1995) The culture of juvenile pearl mussels Margaritifera margaritifera in cages: a contribution to conservation programmes and the knowledge of habitat requirements. Biological Conservation, 99, 183–190. Buddensiek V., Engel H., Fleischauer-Rossing S. & Wächtler K. (1993) Studies on the chemistry of interstitial water taken from defined horizons in the fine sediments of bivalve habitats in several North German lowland waters II: microhabitats of Margaritifera margaritifera L. Archiv für Hydrobiologie, 127, 151–166. Cochran W.G. (1941) The distribution of the largest of a set of estimated variances as a fraction of their total. Annals of Eugenics (London), 11, 47–61. Cosgrove P.J. & Harvey P.V. (2003) An unusual freshwater pearl mussel Margaritifera margaritifera (L.) population in Scotland. Journal of Conchology, 38, 139–146. Curry R.A. & McNeill W.S. (2004) Population-level responses to sediment during early life in brook trout. Journal of the North American Benthological Society, 23, 140–150. 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 2314 J. Geist and K. Auerswald DIN, Deutsches Institut für Normung (2001a) DIN 38 414 Aufschlüsse zur Bestimmung von Gesamtgehalten – Schlamm und Sedimente (Gruppe S). Beuth Verlag, Berlin, Germany. DIN, Deutsches Institut für Normung (2001b) DIN EN ISO/IEC 17025 Bestimmung von Elementen mittels ICP. Beuth Verlag, Berlin, Germany. Dudgeon D., Arthington A.H., Gessner M.O. et al. (2006) Freshwater biodiversity: importance, threats, status and conservation challenges. Biological Reviews, 81, 163–182. Fellows C.S., Valett H.M. & Dahm C.N. (2001) Wholestream metabolism in two montane streams: contribution of the hyporheic zone. Limnology and Oceanography, 46, 523–531. Fischer W.R., Flessa H. & Schaller G. (1989) pH values and redox potentials in microsites of the rhizosphere. Zeitschrift für Pflanzenernährung und Bodenkunde, 152, 191–195. Fox J. (1997) Applied Regression Analysis, Linear Models, and Related Methods. Sage Publications, Thousand Oaks, CA. Gangloff M.M. & Feminella J.W. (2007) Stream channel geomorphology influences mussel abundance in southern Appalachian streams, U.S.A.. Freshwater Biology, 52, 64–74. Geist J. (1999a) Schadwirkungen von Feinsedimenten in Flussperlmuschelgewässern, Die Flussmeister. Zeitschrift für Wasserwirtschaft, 1999/2000, 43–46. Geist J. (1999b) Ist die Flussperlmuschel noch zu retten? Geoökologische Aspekte im Gewässerschutz. Junge Wissenschaft, 55, 18–24. Geist J. (2005) Conservation Genetics and Ecology of European Freshwater Pearl Mussels (Margaritifera margaritifera L.). PhD Thesis, Technische Universität München, München, Germany. Available at: http:// www.weihenstephan.de/zpf/fisch/Mitarbeiter/Geist. htm [last accessed on 27 May 2007]. Geist J. & Kuehn R. (2005) Genetic diversity and differentiation of central European freshwater pearl mussel (Margaritifera margaritifera L.) populations: implications for conservation and management. Molecular Ecology, 14, 425–439. Geist J., Porkka M. & Kuehn R. (2006) The status of host fish populations and fish species richness in European freshwater pearl mussel (Margaritifera margaritifera) streams. Aquatic Conservation: Marine and Freshwater Ecosystems, 16, 251–266. Gleick P.H. (2003) Global freshwater resources: soft-path solutions for the 21st century. Science, 302, 1524–1528. Greig S.M., Sear D.A. & Carling P.A. (2005) The impact of fine sediment accumulation on the survival of incubating salon progeny: implications for sediment management. Science of the Total Environment, 344, 241–258. Grost R.T., Hubert W.A. & Wesche T.A. (1991) Descriptors of brown trout redds in a mountain stream. Transactions of the American Fisheries Society, 120, 582– 588. Hastie L.C. & Young M.R. (2003a) Conservation of the Freshwater Pearl Mussel 2. Relationship with Salmonids; Conserving Natura 2000 Rivers Conservation Techniques Series No.3. English Nature, Peterborough, 44 pp. Hastie L.C. & Young M.R. (2003b) Conservation of the Freshwater Pearl Mussel 1. Captive Breeding Techniques; Conserving Natura 2000 Rivers Conservation Techniques Series No.2. English Nature, Peterborough. Hastie L.C., Boon P.J. & Young M.R. (2000) Physical microhabitat requirements of freshwater pearl mussels, Margaritifera margaritifera (L.). Hydrobiologia, 429, 59–71. Hastie L.C., Boon P.J., Young M.R. & Way S. (2001) The effects of a major flood on an endangered freshwater mussel population. Biological Conservation, 98, 107–115. Howard J.K. & Cuffey K.M. (2006a) Factors controlling the age structure of Margaritifera falcata in 2 northern California streams. Journal of the North American Benthological Society, 25, 677–690. Howard J.K. & Cuffey K.M. (2006b) The functional role of native freshwater mussels in the fluvial benthic environment. Freshwater Biology, 51, 460–474. Hwang S.I. & Powers S.E. (2003) Using particle-size distribution models to estimate soil hydraulic properties. Soil Science Society of America Journal, 67, 1103–1112. Johnson P.D. & Brown K.M. (2000) The importance of microhabitat factors and habitat stability to the threatened Louisiana pearl shell, Margaritifera hembeli (Conrad). Canadian Journal of Zoology, 78, 271–277. Jones J.W., Mair R.A. & Neves R.J. (2005) Factors affecting survival and growth of juvenile freshwater mussels cultured in recirculating aquaculture systems. North American Journal of Aquaculture, 67, 210– 220. Jungbluth J.R., Coomans H.E. & Groh H. (1985) Bibliographie der Flussperlmuschel Margaritifera margaritifera (Linn. 1785). Verslagenen Technische Gegevens No 41; Institut voor Taxonomische Zoologie, Universiteit Amsterdam, Amsterdam, The Netherlands. Leff L.G., Burch J.L. & McArthur J.V. (1990) Spatial distribution, seston removal, and potential competitive interactions of the bivalves Corbicula fluminea and Elliptio complanata in a coastal plain stream. Freshwater Biology, 24, 409–416. Magee J.P., McMahon T.E. & Thurow R.F. (1996) Spatial variation in spawning habitat of cutthroat trout in a 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 Stream bed characteristics and mussel recruitment sediment-rich stream basin. Transactions of the American Fisheries Society, 125, 768–779. Malcolm I.A., Youngson A.F. & Soulsby C. (2003) Survival of salmonid eggs in a degraded gravel-bed stream: effects of groundwater-surface water interactions. River Research and Applications, 19, 303–316. McDonald R.C., Isbell R.F., Speight J.G., Walker J. & Hopkins M.S. (1990) Australian Soil and Land Survey Field Handbook, 2nd edn. Inkata Press, Melbourne, Vic. Nearing M.A. & West L.T. (1988) Soil strength indices as indicators of consolidation. Transactions of the American Society of Agricultural Engineers, 31, 471–476. Neves R.J. & Widlak J.C. (1987) Habitat ecology of juvenile freshwater mussels (Bivalvia: Unionidae) in a headwater stream in Virginia. American Malacological Bulletin, 5, 1–7. Neves R.J., Bogan A.E., Williams J.D., Ahlstedt S.A. & Hartfield P.W. (1997) Status of aquatic molluscs in the southeastern United States: a downward spiral of diversity. In: Aquatic Fauna in Peril: the Southeastern Perspective (Eds G.W. Benz & D.E. Collins). Special Publication 1, Southeast Aquatic Research Institute, Lenz Design and Communications, Decatur, GA, 43–85. Oberthür T., Goovaerts P. & Dobermann A. (1999) Mapping soil texture classes using field texturing, particle size distribution and local knowledge by both conventional and geostatistical methods. European Journal of Soil Science, 50, 457–479. Olsen D.A. & Townsend C.R. (2003) Hyporheic community composition in a gravel-bed stream: influence of vertical hydrological exchange, sediment structure and physiochemistry. Freshwater Biology, 48, 1363–1378. Orghidan T. (1959) Ein neuer Lebensraum des unterirdischen Wassers: der hyporheische Biotop. Archiv für Hydrobiologie, 55, 392–414. Palmer M.A., Covich A.P., Finlay B.J. et al. (1997) Biodiversity and ecosystem processes in freshwater sediments. Ambio, 26, 571–577. Poff N.L.R., Allan J.D., Bain M.B., Karr J.R., Prestegaard K.L., Richter B.D., Sparks R.E. & Stromberg J.C. (1997) The natural flow regime: a paradigm for river conservation and restoration. Bioscience, 47, 769–784. Regh K.J., Packman A.I. & Ren J. (2005) Effects of suspended sediment characteristics and bed sediment transport on streambed clogging. Hydrological Processes, 19, 413–427. Ricciardi A. & Rasmussen J.B. (1999) Extinction rates of North American fauna. Conservation Biology, 13, 1220– 1222. Richards C. & Bacon K.L. (1994) Influence of fine sediment on macroinvertebrate colonization of surface and hyporheic stream substrates. Great Basin Naturalist, 54, 106–113. 2315 Sachteleben J., Schmidt C., Vandré R. & Wenz G. (2004) Leitfaden Flussperlmuschelschutz. Bayerisches Landesamt für Umweltschutz, Augsburg, Germany. Schälchli U. (1992) The clogging of coarse river beds by fine sediment. Hydrobiologia, 235/236, 189–197. Schlesinger W.H. (1991) Biogeochemistry. Academic Press, San Diego, CA. Schmidt C. & Wenz G. (2000) Kontinuierliche Überwachung der Flussperlmuschel in Bayern und Maßnahmen zur Bestandsstützung. In: The Freshwater Pearl Mussel in Europe: Population Status and Conservation Strategies (Eds W.W.A. Hof, Albert-Ludwigs-Universität Freiburg), pp. 92–101. Wasserwirtschaftsamt Hof, Germany. Schmidt C. & Wenz G. (2001) Monitoring-Programm für ausgewählte Bestände der Flussperlmuschel (Margaritifera margaritifera L. 1758) als Datengrundlage für die Erfolgskontrolle von Schutzprojekten im Rahmen des Artenhilfsprogramms. Schriftenreihe des Bayerischen Landesamts fuer Umweltschutz, 156, 373–393. Silverman B. (1986) Density Estimation for Statistics and Data Analysis. Chapman and Hall, London. Sinowski W. & Auerswald K. (1999) Using relief parameters in a discriminant analysis to stratify geological areas of different spatial variability of soil properties. Geoderma, 89, 113–128. Soulsby C., Malcolm I.A. & Youngson A.F. (2001b) Hydrochemistry of the hyporheic zone in salmon spawning gravels: a preliminary assessment in a degraded agricultural stream. Regulated Rivers: Research & Management, 17, 651–665. Soulsby C., Youngson A.F., Moir H.J. & Malcolm I.A. (2001a) Fine sediment influence on salmonid spawning habitat in a lowland agricultural stream: a preliminary assessment. The Science of the Total Environment, 265, 295–307. Spooner D.E. & Vaughn C.C. (2006) Context-dependent effects of freshwater mussels on stream benthic communities. Freshwater Biology, 51, 1016–1024. Stein B.A. & Flack S.R. (1997) Species Report Card: The State of U.S. Plants and Animals. The Nature Conservancy, Arlington, VA, 26 pp. Strayer D.L. (1999) Use of flow refuges by unionid mussels in rivers. Journal of the North American Benthological Society, 18, 468–476. Strayer D.L. (2006) Challenges for freshwater invertebrate conservation. Journal of the North American Benthological Society, 25, 271–287. Strayer D.L., May S.E., Nielsen P., Wollheim W. & Hausam S. (1997) Oxygen, organic matter, and sediment granulometry as controls of hyporheic animal communities. Archiv für Hydrobiologie, 140, 131–144. 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316 2316 J. Geist and K. Auerswald Strayer D.L., Downing J.A., Haag W.R., King T.L., Layzer J.B., Newton T.J. & Nichols S.J. (2004) Changing perspectives on pearly mussels, North America’s most imperilled animals. BioScience, 54, 429–439. Strecker U., Bauer G. & Wächtler K. (1990) Untersuchungen über die Entwicklungsbedingungen junger Flussperlmuscheln. Schriftenreihe Bayerisches Landesamt für Umweltschutz, 97, 25–30. Vaughn C.C. & Hakenkamp C.C. (2001) The functional role of burrowing bivalves in freshwater ecosystems. Freshwater Biology, 46, 1431–1446. Williams J.D., Warren M.L., Cummings K.S., Harris J.L. & Neves R.J. (1993) Conservation status of the freshwater mussels of the United States and Canada. Fisheries, 18, 6–22. Wilson J.G. & Moorkens E. (1992) Oxygen consumption by Margaritifera margaritifera (Linnaeus) and Sphaerium corneum (Linnaeus) in relation to temperature and low oxygen concentration. In: Abstr. 11th Intern. Malacol. Congr. Siena 1992 (Eds F. Giusti & G. Manganelli), 11, 42–45. Unitas Malacologia, Siena. Young M.R. & Williams J.C. (1983) The status and conservation of freshwater pearl mussel Margaritifera margaritifera L. in Great Britain. Biological Conservation, 25, 35–52. Young M.R. & Williams J.C. (1984) The reproductive biology of the freshwater pearl mussel Margaritifera margaritifera (Linn.) in Scotland; I. field studies. Archiv für Hydrobiologie, 99, 405–422. Young M.R., Cosgrove P.J. & Hastie L.C. (2001) In: The Extent of, and Causes for, the Decline of a Highly Threatened Naiad: Margaritifera margaritifera; Ecology and Evolution of the Freshwater Mussels Unionoidea (Eds G. Bauer & K. Wächtler), Ecological Studies, 145, 337– 357. Springer Verlag, Heidelberg, Germany. Ziuganov V., Zotin A., Nezlin L. & Tretiakov V. (1994) The Freshwater Pearl Mussels and their Relationships with Salmonid Fish. VNIRO, Russian Federal Institute of Fisheries and Oceanography, Moscow, Russia, 104 pp. (Manuscript accepted 3 May 2007) 2007 The Authors, Journal compilation 2007 Blackwell Publishing Ltd, Freshwater Biology, 52, 2299–2316
Documents pareils
The functional role of burrowing bivalves in freshwater
experiments performed under unnatural conditions
(i.e. where bivalves were not allowed to burrow in the
substratum, where there was mechanical or chemical
disturbance, and/or where arti®cially high...
Previous Projects: Observing Systems and Monitoring in Lake Erie
with faculty and students from Central Michigan University, Cleveland State University and Kent State
University, we surveyed Lake Ontario bays, river mouths, and coastal wetlands to locate the pos...