Freshwater Biology (2003) 48, 404–417 Modelling the impact of benthic filter-feeders on the composition and biomass of river plankton J . - P . D E S C Y * , E . E V E R B E C Q †, V . G O S S E L A I N * , L . V I R O U X * AND J . S . S M I T Z † *URBO, Laboratory of Freshwater Ecology, Department of Biology, University of Namur, Namur, Belgium †CEME, Institute of Physics, University of Liège, Liège, Belgium SU M M A R Y 1. The POTAMON model [Everbecq E. et al. (2001) Water Research, 35, 901] has been used to simulate the effect of benthic bivalves (mainly Dreissena polymorpha) on the phytoplankton and zooplankton in a lowland Western European river (the Moselle). Here we use a modified version of the POTAMON model with five categories of phytoplankton (Stephanodiscus, Cyclotella-like, large diatoms, Skeletonema and non-siliceous algae) to model filter-feeding effects of benthic bivalves in the Moselle. Zooplankton has been represented in the model by two categories, Brachionus-like and Keratella-like rotifers. 2. According to density estimates from field surveys (Bachmann V. et al. (1995) Hydroécologie Appliquée, 7, 185, Bachmann V. & Usseglio-Polatera P. (1999) Hydrobiologia, 410, 39), zebra mussel density varied among river stretches, and increased through the year to a maximum in summer. Dreissena filtration rates from the literature were used, and mussels have been assumed to feed on different phytoplankton categories (but less on large and filamentous diatoms) as well as on rotifers. 3. The simulations suggest a significant impact of benthic filter-feeders on potamoplankton and water quality in those stretches where the mussels are abundant, their impact being maximal in summer. Consequently, different plankton groups were not affected to the same extent, depending on their period of development and on indirect effects, such as predation by mussels on herbivorous zooplankton. 4. A daily carbon balance for a typical summer shows the effect of benthic filter-feeders on planktonic and benthic processes: the flux of organic matter to the bottom is greatly enhanced at high mussel density; conversely, production and breakdown of organic carbon in the water column are reduced. Mussel removal would drive the carbon balance of the river toward autotrophy only in the downstream stretches. Keywords: lowland river, modelling, mussels, phytoplankton, zooplankton Introduction Zebra mussels (Dreissena polymorpha Pallas) have drawn considerable attention to benthic filter feeders after their recent invasion of large American lakes, and several studies have reported ecological effects of benthic bivalves in river systems. The main direct effect of mussels in freshwater ecosystems is their Correspondence: J.-P. Descy, URBO, Laboratory of Freshwater Ecology, Department of Biology, University of Namur, Namur, Belgium. E-mail: [email protected] 404 predation on plankton. Other effects of dense populations of benthic filters-feeders in rivers are (i) an increase of oxygen consumption through respiration of the benthos (Effler et al., 1996, 1998; Bachmann et al., 1998; Bachmann & Usseglio-Polatera, 1999; Garnier, Billen & Palfner, 2000), (ii) competition of exotic species with native unionids for food and through fouling, leading to their decline, along with several other changes in the zoobenthos (reviewed for North American rivers in Strayer & Smith, 1996), and (iii) resulting from predation on plankton, a channelling of planktonic production to the benthos with 2003 Blackwell Publishing Ltd Modelling benthic filter-feeders and river plankton corresponding changes to the food web and carbon pathways (Welker & Walz, 1998; Jack & Thorp, 2000). The size range of prey of young and adult zebra mussels is 20–450 lm, hence they can feed upon most phytoplankton and small zooplankton (Sprung & Rose, 1988; Grigorovich & Shevtsova, 1995). Strayer & Smith (1996) have attributed losses of phytoplankton and zooplankton in several North American rivers to feeding by this exotic bivalve. Predation on small zooplankton (including rotifers and nauplii), with grazing rates similar to those on phytoplankton, has been demonstrated in the laboratory (MacIsaac, Lonnee & Leach, 1995) and field (Jack & Thorp, 2000). Accordingly, data from studies in rivers have shown a significant impact on phytoplankton and zooplankton. Pace, Findlay & Fischer (1998) reported a >70% decline in zooplankton biomass (especially of ciliates, rotifers and nauplii) following invasion by Dreissena of the Hudson River, along with a 80–90% reduction in phytoplankton biomass. Despite their capacity to feed on a wide range of plankton, zebra mussels filter most efficiently on particles of 5–35 lm, and filtration rate varies with body size (Walz, 1978; Young et al., 1996). Several authors have measured the filtration rate of Dreissena experimentally and have explored the effects of mussel size, temperature and food concentration. For instance, Sprung & Rose (1988) found a typical functional response with maximum filtration of 230 mL ind)1 h)1 at low food concentration, and a minimum of 50 mL ind)1 h)1 at high food. Other published values lie in this range: Reeders, Bij de Vaate & Slim (1989) measured 78–170 mL ind)1 h)1 by 20–22 mm mussels. According to Walz (1978), the mean maximum filtration rate is 80 mL ind)1 h)1 and the incipient limiting level (ILL) is 2 mg C L)1; optimal temperature is 15 C and mussels filter 70% of the time. Other workers found optimal ingestion and assimilation rates between 20 and 24 C (Aldridge, Payne & Miller, 1995). On the other hand, in situ investigations seem to show an effect of temperature in relation to seasonality (see Reeders et al., 1989; Fanslow, Nalepa & Lang, 1995; Sprung, 1995). Selectivity has been rarely reported, although Bastviken, Caraco & Cole (1998) showed net selective removal of algae, not related to algal size, in laboratory experiments. Most studies have dealt with the invading species D. polymorpha, but a high density of native unionid mussels can have similar impacts, as described in a 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 405 river-lake system (River Spree, Germany) by Welker & Walz (1998). In most cases in Europe, however, the proliferating bivalves are exotic. The most famous is Corbicula sp., which lives on a wide range of substrata (Bachmann et al., 1995, 1997), and possibly competes with Dreissena for space and resources. The filtration rate of Corbicula on phytoplankton is about 50 mL ind)1 h)1 (Boltovskoy, Izaguirre & Correa, 1995), and seems non-selective, despite the fact that diatoms and Chrysophytes were found in greater abundance in the mussels’ digestive tracts. In North America, phytoplankton depletion following invasion by the exotic bivalve Corbicula fluminea (Müller) was observed in the Potomac River in the early 1980s (Cohen et al., 1984). Some simulation models have been developed to predict the effects of mussels on freshwater communities and ecosystems. For instance, Padilla et al., (1996) used a model for investigating zebra mussels impact in Green Bay, Lake Michigan. They predicted a greater impact on large phytoplankton than on small, as small algae may compensate for grazing losses by enhanced growth, owing to an increase in phosphorus cycling and in water clarity. Cladocerans feeding mainly on small phytoplankton were not greatly affected by zebra mussels. Overall, these authors held the view that the ecological effects of invasion in North American lakes may be less than expected. Caraco et al. (1997) developed a box-flow model of the Hudson River and demonstrated that grazing by dense zebra mussel population is likely to result in low algal biomass. Indeed, in the light-limited conditions prevailing in turbid, nutrient-rich rivers, zebra mussel filtration does not improve water clarity to the extent that phytoplankton could compensate for grazing losses. Furthermore, the concentration of suspended solids in rivers (in the 10 mg L)1 range) does not affect the filtration activity of the molluscs. For these reasons, it seems that rivers may be more sensitive than lakes to development of dense mussel populations and to dramatic plankton decline (Gosselain et al., 1998b). Garnier et al. (2000) applied the RIVERSTRAHLER model to the River Moselle and its tributaries, and showed the effect of filter feeders on total chlorophyll a and on the oxygen budget. They demonstrated that a low summer oxygen concentration in some river stretches results from high organic matter loading, reduced phytoplankton biomass and respiration by the macrozoobenthos. 406 J.-P. Descy et al. Here we apply a non-stationary river plankton model, POTAMON, which was developed for predicting potamoplankton composition and biomass (Everbecq et al., 2001). We use model simulations and data on the plankton of the River Moselle in 1993 to explore the effects of benthic grazers on plankton composition and biomass. In particular, we consider mussels as feeding on different categories of phytoplankton and on herbivorous rotifers. We also investigate some impacts on ecosystem function. Methods Site description and ecological characteristics A 200-km stretch of the river Moselle was studied (Fig. 1), from Frouard in France (river km 183) to Detzem in Germany (river km 376), from 1993 to 1995. Various environmental, water quality and plankton data were collected in each year at several sites over the growing season (see Gosselain et al., 1998b). Details of the potamoplankton (phyto- and zooplankton) composition and biomass can be found in Gosselain (1998) and Viroux (1997), along with methods of sampling and data acquisition. Briefly, 1993 was characterised by differences in the development of the potamoplankton at different points along the river. In the upstream reach (i.e. up- and downstream of the city of Nancy), phytoplankton biomass varied little over the growing period; zooplankton peaked in spring, and again more strongly in summer. In the downstream reach (around Metz and downstream), the plankton was abundant in spring, with the highest peaks of both phyto- and zooplankton observed in early June, but declined sharply in summer and autumn. Gosselain et al. (1998b) hypothesised that filtration by zebra mussels and other benthic filterfeeders abundant in this downstream reach could cause the summer decline of potamoplankton in the River Moselle. Composition, abundance and biomass estimates of the macrozoobenthos for the same period were published in Bachmann et al. (1995, 1997), Bachmann & Usseglio-Polatera (1999) and Bachmann (2000). In the period 1993–95, the filter-feeders were dominated by zebra mussels with some Asiatic clams (Corbicula spp.), together comprising 75% of macrozoobenthic biomass. Maximum recorded mussel abundance on substrata in the main channel was about 7000 m)2 (in Fig. 1 Location map of the studied stretch of the River Moselle. the vicinity of Metz and downstream, but was lower between Nancy and Metz; Fig. 1). Accordingly, average mussel biomass ranged between 5 and 20 g C m)2, for the whole river bed, depending on the stretch and the time of the year (see below). The same range in mussel biomass was used for modelling purposes by Garnier et al. (2000) in the River Moselle and by Schöl et al. (2002) in the River Rhine. Implementation of the model The POTAMON model has been described in detail in Everbecq et al. (2001). Except for the inclusion of the benthic filter-feeders and the different categories of phytoplankton, the model used here is essentially the same, using as input data the River Moselle’s 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 Modelling benthic filter-feeders and river plankton (a) At 0.5 mg C L–1 4000 3000 Filtration rate (mL g C–1 h–1) 2000 1000 0 5 10 15 20 25 30 Temperature (°C) At 20 °C (b) 4000 3000 2000 1000 0 1 2 3 4 5 6 7 –1 Algal biomass (mg C L ) Fig. 2 Dependence of zebra mussel filtration rate (mL g C)1 h)1) on algal biomass (a) and temperature (b) as calculated by the model (equations and parameters in Tables 1 and 2, respectively), based on experimental data of Bachmann (2000) and other references (see text). characteristics. Morphometric data (slope and width of the river, presence and characteristics of weirs) were available at a scale of a few km. We used mean daily discharge and temperature just upstream the Cattenom Power plant, and hourly or semihourly solar radiation measured at Florange (by Meteo France) and at Trier (by the Deutscher Wetterdienst). A few changes were made in the POTAMON model in order to describe adequately the River Moselle phytoplankton, where the filamentous diatoms Skeletonema potamos (Weber) Hasle and S. subsalsum (A. Cleve) Bethge develop significant numbers. As Skeletonema spp. usually achieve maximum biomass in summer, they were described by physiological and growth parameters similar to those for green algae, but with Si required for their growth. The five algal categories considered in the model predictions are: (1) Stephanodiscus hantzschii, which is assumed to represent the following taxa: S. invisitatus Hohn & 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 407 Hellermann, S. hantzschii Grun., S. hantzschii fo. tenuis (Hust.) Håkansson & Stoermer and S. minutulus (Kütz.) Round. (2) Small centric diatoms, or Cyclotella-like, to represent Cyclotella meneghiniana Kütz., Cyclostephanos dubius (Fricke) Round, and other ‘small’ Thalassiosiraceae. (3) Large centric diatoms, corresponding to Aulacoseira spp. and large Thalassiosiraceae. (4) Non-siliceous algae, as a broad category for describing green algae (mostly Chlorococcales) and Cryptophytes. (5) Skeletonema, which includes essentially S. potamos and S. subsalsum. As for the zooplankton, the model describes two categories of rotifers, Brachionus-like and Keratellalike rotifers, differing mainly in size, filtration rate (Gosselain et al., 1996) and general food preferences (Pourriot, 1977). A similar distinction was successfully applied in previous analyses of riverine rotifer populations (Gosselain et al., 1996). Indeed, rotifers dominate the zooplankton in the River Moselle, although small cladocerans may occasionally be abundant (Viroux, 1997). In eutrophic lowland rivers, rotifers may reach a density of thousands per litre in late spring blooms, and exert a significant grazing pressure on small phytoplankton (Gosselain, 1998; Gosselain, Viroux & Descy, 1998a; Gosselain et al., 1998b): filtration rates measured at high rotifer density regularly reached 20–30% day)1, and small centric diatoms, that constituted the bulk of the phytoplankton biomass, were particularly affected. We added a macrozoobenthos compartment, which was essentially composed of Dreissena-like filter feeders. The biomass of the molluscs was set according to the available data for the river (see above). Unlike Schöl et al. (2002), we did not describe explicitly the population dynamics of the mussels, considering that the data available on both in situ growth and colonisation were insufficient. Rather, we used an average minimum value for winter biomass and allowed the mussel biomass to increase according to a simplified profile, involving a linear increase up to a summer maximum of 3.5 times the winter biomass; thereafter, Dreissena biomass was assumed to decline linearly back to winter values. This seasonal pattern was suggested by the detailed study in the River Rhine by Schöl et al. (2002). Although mussel biomass was not computed explicitly, oxygen consumption because of J.-P. Descy et al. Chl a (µg L–1) Frouard 180 160 140 120 100 80 60 40 20 0 J F M A M J J A S O N D A M J J A S O N D M J J A S O N D M J J A S O N D Chl a (µg L–1) Hauconcourt 180 160 140 120 100 80 60 40 20 0 J F M Koenigsmacher Chl a (µg L–1) their presence was split between true respiration, which depended on temperature and mussel biomass, and oxygen consumption from the degradation of mussels faeces, which was related to the amount of planktonic biomass ingested. We used a filtration rate of 3600 mL g C)1 h)1, which is derived from the literature (Walz, 1978), corresponding to 20 C for adult mussels of average size and biomass feeding 14–15 h day)1. Parameters describing the dependence of the filtration rate upon temperature and food concentration were chosen (see Table 2) in accordance with experimental data (Bachmann, 2000). We chose an optimal filtration rate at 25 C, although the range of optima found in the literature is quite large (15–27 C), suggesting possible acclimation. A rather high optimal temperature for ingestion and assimilation (24 C) was reported by Aldridge et al. (1995). ReedAndersen et al. (2000) used an upper thermal threshold of 27 C for computing Dreissena activity in Lake Mendota. We also described the filtration rate as depending on a saturation constant, according to the results of Bachmann (2000), who measured a mussel filtration rate 20% lower at high food than at low food ration. Figure 2 presents the dependence of the filtration rate of mussels upon temperature and algal biomass, as used in our model. As the River Moselle is regulated and rather intensely navigated, we assumed turbulent mixing over the whole water column. In other words, we assumed that the plankton was equally distributed over the whole river depth and that the zebra mussel filtration did not affect this distribution. Tables 1 and 2 summarise the main processes taken into account in the biological sub-model. Those are essentially the same as in Everbecq et al. (2001), but with the addition of the Skeletonema algal category and of the macrozoobenthos compartment. For comparison with calculated values, we used: • Phytoplankton biomass from measured chlorophyll a (Chl a), expressed as lg Chl a L)1 or as lg C L)1, based on a single C : Chl a ratio of 30 (Gosselain et al., 1998b); • Carbon biomass of phytoplankton categories, estimated from relative biomass from microscope counts and measured chlorophyll a (according to Gosselain et al., 1998b and Gosselain, Hamilton & Descy, 2000); • Biomass of rotifers estimated from counts, using published values of specific dry weight and conversion to carbon biomass (Viroux, 1997, 2000); 180 160 140 120 100 80 60 40 20 0 J F M A Sierck Chl a (µg L–1) 408 180 160 140 120 100 80 60 40 20 0 J F M A Month Fig. 3 Simulated (lines) and observed (dots) chlorophyll a in the studied stretch of the River Moselle in 1993, from the upstream site (Frouard, top) to the downstream site (Sierck, bottom). • Oxygen concentration measured during routine surveys by the French authorities. Results In order to evaluate the effect of mussels on potamoplankton and on ecological processes, the model was run for the conditions of the year 1993, for the whole sector of the River Moselle between the source and the French–German border (km 310). Simulations of the 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 Modelling benthic filter-feeders and river plankton 409 Table 1 Internal fluxes of the plankton sub-model Description Units Equation Phytoplankton phyj Tpphy Kopt (T) T Red_P PPO4 Red_Si Phytoplankton concentration Production rate Light saturated rate of photosynthesis Temperature Phosphorus limitation factor Phosphates concentration Silica limitation factor mg C L)1 s)1 mg C mg C)1 s)1 C – mgP L)1 – f(I) Light limitation factor – I (z,t) I0 (t) Ik wmoy ke k1, k2, k3 POC Tmphy Trphy Tsphy Graz tfzoophy tfdressphy Available light Incident light energy Light saturation constant Average light over the preceding 7 days Extinction coefficient Extinction coefficient parameters Particular organic carbon Mortality rate Respiration rate Sedimentation rate Grazing rate Edibility coefficient by zooplankton Edibility coefficient by mussels lE m)2 lE m)2 lE m)2 lE m)2 m)1 Zooplankton Zooi tfzoom tfzoo Zooplankton concentration Maximum filtration rate Filtration rate mg C L)1 L mg C)1 day)1 L mg C)1 day)1 phyG Grazable phytoplankton mg C L)1 tfzoom20 Æ qgzoo e[(T)20) ⁄ 10] tfzoom tfzoom Æ Cphyzoo/phyG if phyG > Cphyzoo P tfzoophyjÆphyj trzoo tmzoo Respiration rate Mortality rate s)1 s)1 trzoo20 Æ qrzoo e[(T)20) ⁄ 10] tmzoo20 Æ qmzoo e[(T)20) ⁄ 10] Mussels Dress Tfiltdrm Tfiltdr phytotdr Kphydr Mussel concentration Maximum filtration rate Filtration rate Grazable phytoplankton Saturation constant mg C m)2 L mg C)1 day)1 L mg C)1 day)1 mg C L)1 mg C L)1 Tfiltdrm Æ (Topt) Æ e[–(T–Topt)2 ⁄ dtek2] Tfiltdrm Æ Kphydr ⁄ (phytotdr + Kphydr) P tfdressphyjÆphyj s)1 s)1 s)1 s)1 Kopt (T) Æ Red_P Æ Red_Si Æ f(I) Kopt (Topt) Æ e[–(T–Topt)2 ⁄ dtek2] PPO4 ⁄ (PPO4 + KPO4) SiO2 ⁄ (SiO2 + KSiO2) h i R Iðz;tÞ=ð2IkÞ I0 ðtÞ I0 ðtÞekeH 2 H 2 2Ik H 0 ½1þðIðz;tÞ=ð2IkÞÞ2 dz ¼ keH arctg 2Ik arctg I0(t) Æ e–ke Æ z b0 Æ wmoy ⁄ (b1 + wmoy) k1+ P k2j{{phy}}j+k3{{POC}} j mg C L)1 s)1 s)1 s)1 s)1 plankton were examined for four sites: Frouard and Hauconcourt, located at both ends of a stretch where the bivalve density was low, and Koenigsmacher and Sierck, in stretches with high mussel density. Simulations without zebra mussels were run, in order to reveal the changes because of filter-feeders. Simulations of variations through time at four sites and comparison with observations The model provided a satisfactory overall description of temporal and longitudinal variations in chlorophyll 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 tmphy20 Æ qmphy e[(T)20) ⁄ 10] trphy20 Æ qrphy e[(T)20) ⁄ 10] vsphy ⁄ H tfzoo Æ tfzoophy Æ zoo j j a (Fig. 3). However, it failed to simulate a phytoplankton peak at the beginning of June in a part of the study stretch. Small centric diatoms were particularly abundant during this bloom. The fact that this peak was not observed at the upstream site (Frouard) indicates that it could have resulted from an advective input of phytoplankton from the River Meurthe, a tributary of the River Moselle (Fig. 1). Temporal variation of the five phytoplankton categories is shown in Fig. 4 for an ‘upstream’ site (Frouard) and for the most downstream site studied (Sierck). Here, the typical sequence of phytoplankton in lowland Light saturated rate of photosynthesis Optimal temperature for photosynthesis Determines the shape of the Kopt: T curve Parameter of the Ik equation Parameter of the Ik equation Specific extinction coefficient Half saturation constant for P-limited growth Half saturation constant for Si-limited growth Silica: carbon ratio in diatoms Mortality rate (at 20 C) Respiration rate (at 20 C) Sedimentation velocity Edibility coefficient by Brachionus-like Edibility coefficient by Keratella-like Edibility coefficient by mussels Maximum filtration rate for Brachionus-like (at 20 C) Maximum filtration rate for Keratella-like (at 20 C) ILL for Brachionus-like ILL for Keratella-like Growth yield for Brachionus-like Growth yield for Keratella-like Mortality rate (at 20 C) of Brachionus-like Mortality rate of Keratella-like (at 20 C) Respiration rate of Brachionus-like (at 20 C) Respiration rate of Keratella-like (at 20 C) Edibility coefficient of Brachionus-like by mussels Edibility coefficient of Keratella-like by mussels Maximum filtration rate (at topt C) Optimal temperature for filtration Determines the shape of the Tfilt: T curve Saturation constant for mussels filtration Growth yield of mussels Respiration rate of mussels (at 20 C) Phytoplankton Kopt (at Topt) Topt dtek b0 b1 k2 KPO4 KSiO2 SiO2: C tmphy20 trphy20 vsphy tfzoo1phy tfzoo2phy tfdressphy Zooplankton tfzoo1m20 tfzoo2m20 cphyzoo1 cphyzoo2 yzoo1 yzoo2 tmzoo120 tmzoo220 trzoo120 trzoo220 tfdresszoo1 tfdresszoo2 Mussels tfiltdrm20 Toptfilt dtfilt Kphydr ydress trdress20 L mg C)1 day)1 C C mg C L)1 0.25 day)1 day)1 day)1 day)1 day)1 0.25 0.40 L mg C)1 day)1 L mg C)1 day)1 mg C L)1 mg C L)1 mg C mg C)1 day)1 C C lE m)2 s)1 lE m)2 s)1 m2 mg C)1 mg P L)1 mg SiO2 L)1 mg SiO2 mg C)1 day)1 day)1 m day)1 1.00 1.00 1.00 Units 0.03 0.086 25 15 20 1.70 1.45 2.0 1.5 0.25 0.22 0.13 0.11 0.13 0.11 6.00 14 8 200 20 0.50 0.01 0.02 0.50 0.15 0.15 0.70 0.90 0.90 1.00 Stephanodiscus hantzschii Table 2 Parameters of the plankton sub-model. All rates were temperature adjusted assuming a Q10 of 2.0 0.20 0.17 6.00 21 12 250 25 0.50 0.01 0.06 0.50 0.14 0.14 0.80 0.45 0.60 1.00 Small centric diatoms 0.15 0.13 6.75 24.5 14 350 30 0.50 0.01 0.00 0.00 0.16 0.16 0.50 0.10 0.10 0.50 Non-siliceous algae 0.12 0.10 3.75 28 10 90 10 0.50 0.01 0.10 0.75 0.06 0.06 0.90 0.25 0.35 1.00 Large centric diatoms 0.15 0.12 6.25 25 10 275 10 0.50 0.01 0.04 0.25 0.16 0.16 0.70 Skeletonema 410 J.-P. Descy et al. 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 Modelling benthic filter-feeders and river plankton Frouard 411 Sierck –1 Biomass (mg C L ) Stephanodiscus 3 2 1 0 Biomass (mg C L ) 1.0 1.5 –1 Biomass (mg C L ) Biomass (mg C L ) 1.0 Biomass (mg C L ) Small centrics 2 1 0 –1 Large diatoms 0.5 0.0 –1 Skeletonema 0.0 Non-siliceous algae –1 Fig. 4 Simulated (lines) and observed (dots) carbon biomass of the five algal categories represented in the POTAMON model, at the most upstream sampling site (Frouard, left) to the most downstream studied site (Sierck, right) of the River Moselle in 1993. 0.5 1.0 0.5 0.0 J F M A M eutrophic rivers was observed in both simulations and observations. In particular, there was a large spring bloom, dominated by Stephanodiscus gr. hantzschii, followed by an increase of various other centric diatoms, green algae and cryptophytes (here included in the non-siliceous algae category). The lack of an autumn bloom was marked in the River Moselle in 1993, and all groups declined from September onwards. In the upstream stretch (Frouard, Hauconcourt, see Fig. 3), phytoplankton biomass remained relatively high throughout the growing season. In contrast, a summer phytoplankton decline occurred at the downstream sites, affecting mainly the small centric diatoms, Skeletonema spp. and the non-siliceous algae. Zooplankton showed a similar pattern (Fig. 6), with conspicuous spring maxima and lower biomass throughout summer. For these organisms, a sharp 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 J J A S O N D J F M A M J J A S O N D peak in abundance was also observed at the beginning of June, which was not replicated by the model, probably because it originated from a upstream tributary, the River Meurthe (see Fig. 1). The effect of benthic filter-feeders on the potamoplankton: simulations without mussels Comparison of simulations with and without mussels allows isolation of the effect from all the other factors that control potamoplankton (Figs 5 and 6). For all categories, the major impact was seen from June to September, which may be related to (i) an increase in filter-feeder biomass; (ii) an increase of their feeding rate as temperature rises; (iii) lower discharge which leads to increased predation impact on plankton. For rotifers, observed and simulated data (Fig. 6) show the expected downstream increase of biomass, with J.-P. Descy et al. 412 Frouard Sierck Biomass (mg C L–1) Stephanodiscus 3 2 1 0 Biomass (mg C L–1) Small centrics 2 1 0 Biomass (mg C L–1) Biomass (mg C L–1) Large diatoms 1.0 1.0 0.5 0.0 Skeletonema 0.5 0.0 Biomass (mg C L–1) Non-siliceous algae 1.5 1.0 0.5 0.0 J F M A M J J A S O N D J F M A M J J A S O N D Fig. 5 Model simulations of the carbon biomass of the five algal categories represented in the POTAMON model, at the most upstream sampling site (Frouard, left) to the most downstream studied site (Sierck, right) of the River Moselle in 1993. Black lines: actual situation, with mussels; grey lines: predicted situation, without mussels. Frouard Sierck 0.8 0.6 0.4 0.2 0.0 Keratella - like Biomass (mg C L 1) Fig. 6 Simulated (lines) and observed (dots) carbon biomass of the two metazooplankton categories represented in the POTAMON model, at the most upstream sampling site (Frouard, left) to the most downstream studied site (Sierck, right) of the River Moselle in 1993. Black lines: actual situation, with mussels; grey lines: predicted situation, without mussels. Biomass (mg C L 1) Brachionus - like 0.2 0.1 0.0 J F M A M longer retention time allowing more generations to be produced. The difference between simulations shows the very strong effect of predation on metazooplankton by mussels in the study stretch. It appears that J J A S O N D J F M A M J J A S O N D both rotifer categories were strongly reduced from June onwards, despite their high potential population growth. Note that larger zooplankton, although present, were usually scarce, with the small cladoceran 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 Modelling benthic filter-feeders and river plankton Bosmina longirostris (Müller) reaching biomass maxima at about 0.05 mg C L)1. Among phytoplankton (Fig. 5), two contrasting patterns can be identified. Two algal groups, Stephanodiscus and the ‘small centrics’, were only slightly affected by mussel predation, although all algae were described as 100% edible by the bivalves. The remaining algal groups were much more affected by mussel feeding activity (Fig. 5). The model simulations indicated that these algae would have maintained higher biomass without the high density of benthic filter-feeders. Simulations of heterotrophic activity, oxygen and extinction coefficient Built as an ecological model which describes all important processes in the river ecosystem, POTAMON enables us to explore the impact of benthic filter-feeders on ecosystem function and water quality. Figure 7 presents the simulations of chlorophyll a, heterotrophic bacteria, light attenuation in the water column and dissolved oxygen at both ends of the studied stretch of the River Moselle. The removal of plankton biomass from the water column by benthic Fig. 7 Model simulations of chlorophyll a, heterotrophic bacteria, vertical attenuation coefficient and dissolved oxygen, at the most upstream sampling site (Frouard, left) to the most downstream studied site (Sierck, right) of the River Moselle in 1993. Black lines: actual situation, with mussels; grey lines: predicted situation, without mussels. 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 413 grazers led to two contrasting effects (observed essentially during the growing season). On the one hand, the biomass of heterotrophic bacterioplankton was reduced, as less dissolved organic carbon (DOC) available to bacteria was released in the water by phytoplankton excretion and mortality (Descy et al., 2002). On the other hand, water transparency was somewhat improved by the decrease of algal biomass, particularly in the downstream part of the river, which supported a high density of filter-feeders. With regard to the oxygen budget, the simulated removal of benthic grazers increased dissolved oxygen concentration at the downstream site from June to September. These results clearly show that, although water transparency was increased by filter-feeder activity, light penetration remained in the range limiting to photosynthesis. This was attributable to the high inorganic turbidity of such large river systems. The relative improvement in light penetration did not compensate for the decrease in autotrophic biomass from mussel grazing. Also, the oxygen budget was affected by mussel metabolism and increased organic matter degradation in the sediment (see below). J.-P. Descy et al. 414 Effect of mussels on the organic carbon transfer between compartments in the river ecosystem We calculated carbon fluxes, with and without mussels, at the two ends of the study stretch of the River Moselle (Fig. 8) for a typical summer day. At the upstream site (Frouard, with low mussel density), mussel predation resulted in a slight reduction in production and respiration, as well as in organic matter breakdown by the heterotrophic bacteria in the water column. Conversely, more organic matter was broken down and consumed on the bottom as a result of mussel activity. These changes in carbon fluxes were much stronger in the downstream site, where there were abundant benthic filter-feeders. Note, however, that photosynthetic production was reduced proportionately less than algal respiration. With regard to the balance of autotrophy versus heterotrophy, all sites of the studied stretch were heterotrophic, depending on their physical characteristics (particularly from the low Frouard (low mussel density) 5 Without mussels With mussels 4 Organic carbon flux (g C m 2 day 1) 3 2 1 0 Sierck (high mussel density) 5 Without mussels With mussels 4 3 2 1 0 Phytoplankton production Phytoplankton respiration OM breakdown in OM breakdown / the water column consumption on the bottom Fig. 8 Organic carbon fluxes in the two extreme sites of the studied stretch of the River Moselle, for the Julian day 228, 1993, as simulated by the POTAMON model. White bars: simulated situation without mussels; grey bars: simulated situation with mussels. OM ¼ organic matter. euphotic depth: total depth ratio, which limits primary production). Removal of mussels would not greatly affect this balance in the upstream part of the river, while downstream the balance would be improved toward autotrophy. At Sierck, for instance, the carbon production: carbon consumption ratio would change from 0.42 to 0.69, and the organic carbon transfer to the benthic compartment would be reduced about threefold. Discussion Regarding changes in ecosystem functions in lowland rivers brought about by the development of large populations of zebra mussel and other benthic grazers, we have reached similar conclusions to other authors. Garnier et al. (2000) identified zebra mussel predation as the main cause for the decline of chlorophyll a in the R. Moselle where it is heavily colonised by D. polymorpha and other benthic grazers. They proposed that the balance between autotrophy and heterotrophy varied in successive river stretches. Schöl et al. (2002) refined the description of zebra mussel population dynamics in the River Rhine, showing that grazing by benthic filter-feeders exhibited strong seasonal variations and greatly exceeded grazing by metazooplankton in summer. However, total calculated grazing in the River Rhine reached a maximum of 0.2 days)1, which is much less than in other rivers, where both rotifer and benthic grazer abundance may be higher. For instance, Gosselain et al. (1998a) measured rotifer grazing rates up to 0.32 (1994) and 1.13 days)1 (1996) in the River Meuse, and Bachmann et al. (1998) estimated that 57% of the water volume may be filtered by the macrobenthos of the River Moselle. Caraco et al. (1997) explored the sensitivity of the River Hudson ecosystem to the zebra mussel invasion with a mechanistic model and were able to demonstrate that the dramatic phytoplankton decline observed over several years resulted entirely from zebra mussel grazing. They emphasised that river depth, complete mixing of the water column, high nutrient concentration and high inorganic (nonphytoplankton) turbidity are characteristics of lowland rivers that render them sensitive to invasion by zebra mussels. For instance, nutrients and turbidity cannot be influenced by the organisms to a point where there could be some compensation for zebra mussel predation, as is observed in lakes. 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 Modelling benthic filter-feeders and river plankton Our application of the POTAMON model to the River Moselle in 1993 supports these conclusions. Moreover, we draw attention to the role of flow rate in a river system, which influences many processes, including the impact of benthic grazers on phytoplankton. Indeed, keys to understanding the strong seasonal differences in the impact of mussel grazing may be population dynamics (Schöl et al., 2002) and temperature dependence of the filtration rate, as well as the fact that, at lower discharge, residence time increases, allowing a greater volume of water to be filtered per unit time. Accordingly, ecological effects are major in summer as seen for instance on chlorophyll a and dissolved oxygen. As in the case of the River Hudson, improved water clarity has little effect on photosynthesis and phytoplankton growth in the River Moselle. This contrasts with lakes that have been invaded by the zebra mussel, where some of the grazing losses may be compensated by enhanced growth rate (Padilla et al., 1996; Reed-Andersen et al., 2000) from improved water transparency and nutrient cycling. In addition, our plankton sub-model has enabled us to describe with more detail the impact of filterfeeders on the whole potamoplankton community, and to explore their effects on the planktonic food web in a lowland river. A high density of mussels, filtering actively as temperature increased, suppressed the density of rotifers in summer, as in other rivers, e.g. the River Meuse (Viroux, 2000). As observed in the laboratory (MacIsaac et al., 1995) and in enclosure experiments in the Ohio River (Jack & Thorp, 2000), rotifers can be ingested by zebra mussels and mussel invasions may lead to a spectacular decline of small zooplankton in rivers (Pace, Findlay & Fischer, 1998). Rotifers usually constitute the bulk of metazooplankton in lowland rivers, where short residence time does not allow large crustaceans to develop. Predation on rotifers by benthic filter-feeders is likely to lead to alterations of food web interactions within the plankton, which may affect the whole river food web. As shown by field observations and by the results of our modelling study, rotifers are particularly sensitive to mussel predation. It is interesting to note that the zooplankton declined proportionately more than the phytoplankton, despite the fact that the predation coefficient of filter-feeders on rotifers in the model was lower. The reason for this is that zooplankton grows more slowly, and cannot compensate for predation losses. 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 415 Similar community changes brought about by grazing of the zoobenthos can be seen in the phytoplankton, although with strong differences among algal categories. Whereas all phytoplankton may be equally grazed by mussels, based upon the settings of the model, different algae were not affected in the same way. On one hand, Stephanodiscus escaped predation because it developed in spring, when mussel biomass and temperature were still low, so that total activity of filter-feeders was low and their overall impact on the water column reduced by high flow rate. Some algae like ‘small centrics’, on the other hand, although preyed upon by mussels at the same rate as other algae, may have benefited from the simultaneous reduction of herbivorous zooplankton. In other words, it is likely that compensatory effects occurred for these diatoms, and that their conspicuous downstream decrease (Fig. 4) was because of factors other than grazing and predation, i.e. other losses and ⁄ or competition with other algal groups. The remaining algal groups were much more affected by mussel feeding activity (Fig. 5). These algae, i.e. colonial green algae, large unicellular and filamentous diatoms, are usually less sensitive than small centric diatoms to rotifer grazing and thrive in lowland rivers such as the Meuse in summer (Gosselain et al., 1998a). In contrast, they have declined spectacularly in stretches of the River Moselle heavily colonised by benthic filter-feeders. To summarise the effect of high zebra mussel density on the river plankton, spring blooms of diatoms and rotifers persist, whereas part of the algae and virtually all small zooplankton are lost in summer. This is a dramatic situation, which may affect the entire food web in lowland river systems, in ways which have not yet been fully evaluated. Acknowledgments The field study on the River Moselle (1993–95) was funded by the CIPMS-IKSMS (International Commission for the Protection of the Moselle and Saar rivers). References Aldridge D.W., Payne B.S. & Miller A.C. (1995) Oxygen consumption, nitrogenous excretion, and filtration rates of Dreissena polymorpha at acclimation temperature between 20 and 32C. Canadian Journal of Fisheries and Aquatic Sciences, 52, 1761–1767. 416 J.-P. Descy et al. Bachmann V. (2000) Dynamique spatio-temporelle d’espèces invasives, particulièrement Corbicula spp. et Dreissena polymorpha (Mollusca: Bivalvia) en hydrosystème fluvial: évolution des populations et effets de l’artificialisation sur les peuplements macrobenthiques. PhD Thesis, Equipe de Démoécologie, Université de Metz, Metz. Bachmann V. & Usseglio-Polatera P. (1999) Contribution of the macrobenthic compartment to the oxygen budget of a large regulated river: the Mosel. Hydrobiologia, 410, 39–46. Bachmann V., Cegielka E., Wagner P., Usseglio-Polatera P. & Moreteau J.-C. (1995) Installation de l’amphipode Corophium curvispinum et de la palourde asiatique Corbicula sp. dans la partie française de la Moselle. Hydroe´cologie Appliquée, 7, 185–191. Bachmann V., Usseglio-Polatera P., Cegielka E., Wagner P., Poinsaint J.F. & Moreteau J.C. (1997) Premières observations sur la coexistence de Dreissena polymorpha, Corophium curvispinum et Corbicula spp. dans la rivière Moselle. Bulletin Français de la Pêche et de la Pisciculture, 344–345, 373–384. Bachmann V., Usseglio-Polatera U., Moreteau J.C. & Wagner P. (1998) A new method to assess the contribution of the macrobenthic compartment to the dissolved oxygen budget of a large river. In: Advances in River Bottom Ecology (Ed. B.G. Helesic), pp. 143–153. Backhuys Publishers, Leiden, The Netherlands. Bastviken D.T.E., Caraco N.F. & Cole J.J. (1998) Experimental measurements of zebra mussel (Dreissena polymorpha) impacts on phytoplankton community composition. Freshwater Biology, 39, 375–386. Boltovskoy D., Izaguirre I. & Correa N. (1995) Feeding selectivity of Corbicula fluminea (Bivalvia) on natural phytoplankton. Hydrobiologia, 312, 171–182. Caraco N.F., Cole J.J., Raymond P.A., Strayer D.L., Pace M.L., Findlay S.E.G. & Fischer D.T. (1997) Zebra mussel invasion in a large, turbid river: phytoplankton response to increased grazing. Ecology, 78, 588–602. Cohen R.R.H., Dresler P.V., Phillips E.J.P. & Cory R.L. (1984) The effect of the Asiatic clam, Corbicula fluminea, on phytoplankton of the Potomac river, Maryland. Limnology and Oceanography, 29, 170–180. Descy J.-P., Leporcq B., Viroux L., François C. & Servais P. (2002) Phytoplankton production, exudation and bacterial reassimilation in the river Meuse (Belgium). Journal of Plankton Research, 24, 161–166. Effler S.W., Boone S.R., Siegfried C. & Ashby S.L. (1998) Dynamics of zebra mussel oxygen-demand in Seneca River, New-York. Environmental Science and Technology, 32, 807–812. Effler S.W., Brooks C.M., Whitehead K., Wagner B., Doerr S.M., Perkins M., Siegfried C.A., Walrath L. & Canale R.P. (1996) Impact of zebra mussel invasion on river water quality. Water Environment Research, 68, 205–214. Everbecq E., Gosselain V., Viroux L. & Descy J.-P. (2001) POTAMON: a dynamic model for predicting phytoplankton composition and biomass in lowland rivers. Water Research, 35, 901–912. Fanslow D.L., Nalepa T.F. & Lang G.A. (1995) Filtration rates of the zebra mussel (Dreissena polymorpha) on natural seston from Saginaw Bay, Lake Huron. Journal of Great Lakes Research, 21, 489–500. Garnier J., Billen G. & Palfner L. (2000) Understanding the oxygen budget and related ecological processes in the river Mosel: the RIVERSTRAHLER approach. Hydrobiologia, 410, 151–166. Gosselain V. (1998) Phytoplancton de la Meuse et de la Moselle et impact du broutage par le zooplancton. PhD Thesis, Department of Biology, University of Namur, Namur, Belgium. Gosselain V., Descy J.-P., Viroux L., Joaquim-Justo C., Hammer A., Métens A. & Schweitzer S. (1998b) Grazing by large river zooplankton: a key to summer potamoplankton decline? The case of the Meuse and Moselle rivers in 1994 and 1995. Hydrobiologia, 369, 199–216. Gosselain V., Hamilton P. & Descy J.P. (2000) Estimating phytoplankton carbon from microscope counts: an application for riverine systems. Hydrobiologia, 438, 75–90. Gosselain V., Joaquim-Justo C., Viroux L., Mena M., Métens A., Descy J.-P. & Thomé J.-P. (1996) Laboratory and in situ grazing rates of freshwater rotifers and their contribution to community grazing rates. Archiv für Hydrobiologie Suppl. 113. Large Rivers, 10, 353–361. Gosselain V., Viroux L. & Descy J.-P. (1998a) Can a community of small-bodied grazers control phytoplankton in rivers? Freshwater Biology, 39, 9–24. Grigorovich I.A. & Shevtsova L.V. (1995) Effect of Dreissena mussels on the distribution of zooplankton as exemplified by the Main Kakhovka Canal. In: The Fifth International. Zebra Mussel and Other Aquatic Nuisance Organisms, Pp. 183–193. EPRI Proceedings, Toronto, Canada. Jack J.D. & Thorp J.H. (2000) Effects of the benthic suspension feeder Dreissena polymorpha on zooplankton in a large river. Freshwater Biology, 44, 569–579. MacIsaac H.J., Lonnee C.J. & Leach J.H. (1995) Suppression of microzooplankton by zebra mussels: importance of mussel size. Freshwater Biology, 34, 379–387. Pace M.L., Findlay S.E.G. & Fischer D. (1998) Effects of an invasive bivalve on the zooplankton community of the Hudson River. Freshwater Biology, 39, 103–116. Padilla D.K., Adolph S.C., Cottingham K.L. & Schneider D.W. (1996) Predicting the consequences of dreissenid 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 Modelling benthic filter-feeders and river plankton mussels on a pelagic food web. Ecological Modelling, 85, 129–144. Pourriot R. (1977) Food and feeding habits of Rotifera. Archiv für Hydrobiologie Beiheft Ergebnisse der Limnologie, 8, 243–260. Reed-Andersen T., Carpenter S.R., Padilla D.K. & Lathrop R.C. (2000) Predicted impact of zebra mussel (Dreissena polymorpha) invasion on water clarity in Lake Mendota. Canadian Journal of Fisheries and Aquatic Sciences, 57, 1617–1626. Reeders H.H., Bij de Vaate A. & Slim F.J. (1989) The filtration rate of Dreissena polymorpha (Bivalvia) in three Dutch lakes with reference to biological water quality management. Freshwater Biology, 22, 133–141. Schöl A., Kirchesch V., Bergfeld T., Schöll F., Borcherding J. & Müller D. (2002) Modelling the Chlorophyll a content of the River Rhine – Interrelation between riverine algal production and population biomass of grazers, rotifers and the zebra mussel, Dreissena polymorpha. International Review of Hydrobiology, 87, 295–317. Sprung M. (1995) Physiological energetics of the zebra mussel Dreissena polymorpha in lakes II. Food uptake and gross growth efficiency. Hydrobiologia, 304, 133–146. Sprung M. & Rose U. (1988) Influence of food size and food quantity on the feeding of the mussel Dreissena polymorpha. Oecologia, 77, 526–532. 2003 Blackwell Publishing Ltd, Freshwater Biology, 48, 404–417 417 Strayer D.L. & Smith L.C. (1996) Relationships between zebra mussels (Dreissena polymorpha) and unionid clams during the early stages of the zebra mussel invasion of the Hudson river. Freshwater Biology, 36, 771–779. Viroux L. (1997) Zooplankton development in two large lowland rivers, the Moselle (France) and the Meuse (Belgium), in 1993. Journal of Plankton Research, 19, 1743–1762. Viroux L. (2000) Dynamique du métazooplancton en milieu fluvial. PhD Thesis, Department of Biology, University of Namur, Namur, Belgium. Walz N. (1978) The energy balance of the freshwater mussel Dreissena polymorpha Pallas in laboratory experiments and in Lake Constance. I. Pattern of activity, feeding and assimilation efficiency. Archiv für Hydrobiologie Suppl., 55, 83–105. Welker M. & Walz N. (1998) Can mussels control the plankton in rivers? – a planktological study applying a Lagrangian sampling strategy. Limnology and Oceanography, 43, 753–762. Young B.L., Padilla D.K., Schneider D.W. & Hewett S.W. (1996) The importance of size-frequency relationships for predicting ecological impact of zebra mussel populations. Hydrobiologia, 332, 151–158. (Manuscript accepted 15 September 2002)
© Copyright 2026 Paperzz