Trait-based structure of invertebrates along a gradient of sediment colmation: Benthos versus hyporheos responses

Science of the Total Environment 466–467 (2014) 265–276

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Trait-based structure of invertebrates along a gradient of sediment colmation: Benthos versus hyporheos responses S. Descloux a,b,c,⁎, T. Datry b, P. Usseglio-Polatera d a

Université Lyon 1, UMR-CNRS 5023 Ecology of Natural and Anthropized Hydrosystem Laboratory, 43 Bd du 11 novembre 1918, 69622 Villeurbanne Cedex, France IRSTEA, UR MALY, F-69336 Lyon, France Electricité de France, Savoie Technolac, 73373 Le Bourget du Lac, France d Université de Lorraine, UMR-CNRS 7360, Interdisciplinary Laboratory of Continental Environments, Avenue du Général Delestraint, 57070 Metz, France b c

H I G H L I G H T S • • • •

We studied the effects of colmation on the biological attributes of invertebrates of three rivers. A higher number of traits were significantly modified with colmation in the benthic vs. hyporheic assemblages. Most of the biological attributes impaired were different in the two zones. A potential indicator of river colmation may be based on the functional traits of benthic communities.

a r t i c l e

i n f o

Article history: Received 25 February 2013 Received in revised form 7 June 2013 Accepted 19 June 2013 Available online 31 July 2013 Editor: Christian EW Steinberg Keywords: Biological attributes Fine sediment Functional groups Community Mixed-effects models

a b s t r a c t Streambed colmation by fine sediment, e.g. the deposition, accumulation and storage of fines in the substrate, is known to have severe effects on invertebrate assemblages in both the benthic and hyporheic zones but the changes in biological attributes of invertebrate assemblages related to colmation have never been considered simultaneously for these two zones. We studied the effects of colmation on the invertebrate assemblages of three rivers, testing a priori hypotheses on the biological attributes that should be more selected in communities subjected to different levels of colmation in both zones. Only the proportion of organisms with high fecundity increased and the proportion of small-sized organisms decreased along the colmation gradient in both zones simultaneously. As expected, a higher number of traits were significantly modified with colmation in the benthic vs. hyporheic assemblages. Most of the biological attributes impaired were different in the two zones. In the benthic zone, colmation mainly selected particular physiological or trophic characteristics of species and features related to their resistance or resilience capacities. In contrast, the morphological attributes of species were much more impaired by colmation in the hyporheic zone than in the benthic zone. In clogged benthic habitats, traits seemed to be more impaired by an increase in physico-chemical constraints (e.g. the reduction of oxygen availability) and a reduction of potential exchanges (including exchanges of food resources) due to a decline in stream bed conductivity. The morphological attributes of the hyporheic species were probably more influenced by changes in interstitial space characteristics. A potential indicator of the effects of colmation on river health may be based on the functional traits of benthic communities because they (i) satisfy the WFD recommendations, (ii) respond consistently along a colmation gradient and (iii) are comparable among assemblages even across ecoregions that differ in their taxonomic composition. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Human activities including agriculture and forestry accelerate land erosion and increase fine sediment inputs into aquatic ecosystems (Waters, 1995; Hancock, 2002). This problem is widespread worldwide

⁎ Corresponding author at: Electricité de France, Savoie Technolac 73373 Le Bourget du Lac, France. Tel.: +33 4 79 60 63 21; fax: +33 4 79 60 61 26. E-mail address: [email protected] (S. Descloux). 0048-9697/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.scitotenv.2013.06.082

and, in addition to flow perturbations (e.g. damming, withdrawal), often leads to the clogging of substrate interstitial spaces with fine sediments (i.e. colmation), reducing hydrological exchanges between surface and ground waters, then impairing the availability of dissolved oxygen, nutrients and organic matter within the streambed (Petts et al., 1989; Waters, 1995; Matthaei et al., 2006; Boulton, 2007). As a result, colmation could have severe negative consequences on the structure and function of aquatic assemblages which may be impaired in terms of both taxonomic and functional trait structure (Gayraud and Philippe, 2001).

266

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

Trait structure refers to the combination of traits present in a given community. These traits are considered as ‘functional’ because they can affect ecosystem functioning by influencing organismal performance (McGill et al., 2006). The combination of traits in a community results from adaptive pressures (Usseglio-Polatera et al., 2000a, 2001). Indeed, the ‘habitat filtering hypothesis’ considers that the characteristics of habitats, in particular those related to disturbance regime and habitat diversity, provide the evolutionary conditions from which life history traits and community properties are derived (Southwood, 1977, 1988; Townsend and Hildrew, 1994). Numerous workers have confirmed predictions from habitat templet theory (Menezes et al., 2010; Statzner and Bêche, 2010), giving evidence that the functional trait approach can provide a direct link between organisms and ecosystems (Petchey and Gaston, 2006). Therefore, using species traits as response variables should facilitate the successful explanation of the mechanisms structuring communities, including human-induced disturbances such as toxic contamination (Archaimbault et al., 2010), salinity (Piscart et al., 2006), waste water discharge (Charvet et al., 2000); agricultural pressure and/or hydrological disturbance (Lecerf et al., 2006; Townsend et al., 2008); land use in the watershed (Dolédec et al., 2006), hydrological connectivity alteration (Paillex et al., 2009) or multiple stressors (Gayraud et al., 2003; Dolédec and Statzner, 2008). Colmation impairs vertical connectivity and could have severe effects on river ecosystem function (Stanford and Ward, 2001; Boulton, 2007). In the absence of scouring spates, the main direct physical effect is a reduction in habitat availability (Lenat et al., 1979; Maridet and Philippe, 1995). Colmation changes also the biogeochemical conditions of habitats, shifting from oxic to anoxic conditions and increasing toxicant concentrations (Minshall, 1984; Gayraud and Philippe, 2003). By reducing (i) interstitial pore size, (ii) habitat heterogeneity and (iii) hydrological disturbance effects, then increasing habitat stability (Petts et al., 1989; Ryan and Boufadel, 2007), colmation has the potential to modify community structure and major functional traits within both the benthic and hyporheic invertebrate assemblages. These modifications have been mainly studied for the benthic zone. A decrease in benthic invertebrate density and taxonomic diversity has been widely demonstrated by Lenat et al. (1979) and Richards and Bacon (1994). The percentage of the benthic Ephemeroptera, Plecoptera and Trichoptera (EPT) assemblage has been negatively correlated with an increase in fine sediment deposition (Bjornn et al., 1977; Lenat et al., 1979; Matthaei et al., 2006; Larsen et al., 2009; Larsen and Ormerod, 2010), while the abundance of other taxa, such as Oligochaeta, often increases with colmation (Lenat et al., 1979; Angradi, 1999; Zweig and Rabeni, 2001). The increase in fine sediment disadvantages organisms that need large interstitial spaces and high oxygen availability (James et al., 2008), potentially resulting in the selection of specific co-adapted traits within faunal assemblages. Gayraud and Philippe (2001) showed that the proportions of small-sized (b 5 mm) and cylindrical or spherical organisms, were negatively correlated to interstitial space availability. The balance between the major feeding groups can also be modified, with a shift from filter-feeders and scrapers to deposit-feeders (Lemly, 1982; Waters, 1995; Relyea et al., 2000). Very few studies on the impact of sediment colmation on hyporheic invertebrate communities exist, even though the functional role of the hyporheic zone and the importance of faunal exchanges between the epibenthic and hyporheic areas are well known. In particular, the hyporheic zone acts as a refugium and nursery for many benthic organisms (Marmonier et al., 1993; Ward et al., 1998). The densities of hyporheic invertebrate assemblages generally decrease with increasing fine sediment deposition (Richards and Bacon, 1994; Weigelhofer and Waringer, 2003; Descloux et al., in press), by reducing invertebrate dispersal within the substrate. However, the effects of colmation on the biological trait composition of hyporheic communities are poorly known, despite the strong

effects of colmation on the availability of habitats, their abiotic characteristics and biotic interactions within the resident community. The objective of this study was to compare the effects of an increasing gradient of fine sediment proportion in mineral substrate on the trait profiles of both benthic and hyporheic invertebrate assemblages at the reach scale. We aimed to assess whether changes in community trait combinations could be used to evaluate the impact of colmation. First, we predicted that colmation should less drastically affect the traits of invertebrate communities of the hyporheic zone than of the benthic zone because, the hyporheic zone collecting fine particles that are less frequently washed away by floods than in surface sediments, might host an assemblage naturally more adapted to fine sediment (hypothesis H1). Second, for ten biological traits, we predicted differences in attributes selected by invertebrates along an increasing colmation gradient (Table 1, Appendix 1), taking into account four potential driving processes: (i) a size reduction of interstitial spaces (process 1 = ‘proc 1’ hereafter); (ii) a reduction of potential exchanges due to a decline in streambed hydraulic conductivity (proc 2); (iii) an increase in physico-chemical constraints (e.g. a reduction in oxygen availability; proc 3) and (iv) an increasing temporal stability of harsh habitat conditions generated by the three first processes (proc 4). We hypothesized an increasing proportion of organisms (i) small-sized (due to proc 1), (ii) burrowing (proc 1 and 2), (iii) monovoltine (proc 4) with (iv) asexual reproduction (proc 2), (v) medium to high fecundity (proc 3), (vi) no or few resistance stages (proc 4), (vii) mainly tegumental respiration (proc 2 and 3), (viii) cylindrical body (proc 1), (ix) being deposit-feeders (proc 2) and (x) consuming fine organic detritus and microorganisms (proc 2) along an increasing colmation gradient (Table 1, Appendix 1, hypothesis H2).

Table 1 Trait-based predictions in both benthic and hyporheic site assemblages along a colmation gradient described by three colmation classes. ‘Intermediate’ means a selection of adaptations ‘intermediate’ between the predictions of the heavily and lightly clogged reaches. Proc 1 = process 1, a size reduction of interstitial spaces; proc 2 = a reduction of potential exchanges due to a decline in streambed hydraulic conductivity; proc 3 = an increase in physico-chemical constraints and proc 4 = an increasing temporal stability of harsh habitat conditions. Colmation class No Trait

1

Process Lightly involved clogged (LC)

3

Maximal proc 1 potential size Number of proc 4 reproductive cycles per year Fecundity proc 3

4

Body form

proc 1

5

Reproduction Technique Resistance forms Respiration

proc 2

2

6 7 8

9

10

Locomotion and substrate relation Food

Feeding habits

proc 4 proc 2+3 proc 1+2

Large/small

Moderately Heavily clogged (HC) clogged (MC)

Large/small to small Low to high Medium

Small Medium

Low to high Medium to Medium to high high Varied Varied/ Mainly cylindrical mainly cylindrical Varied Intermediate Asexual; free eggs and clutches Present/ Present/ Absent absent absent Varied Intermediate Mainly tegumental Varied

Intermediate Crawlers; burrowers attached

proc 2

Varied

proc 2

Varied

Intermediate mainly microorganisms in fine sediment; fine detritus Intermediate Mainly deposit-feeders and filterers reduced proportion of scrapers

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

267

Les Usses River U -MC U -LC

U -HC

I-HC

I-MC I-LC

Isère River

Drôme River D -LC D -MC

D -HC

Fig. 1. Location of the sampling area. I = Isère River, D = Drôme River; U = Usses River, HC = heavily clogged reach, MC = moderately clogged reach and LC = lightly clogged reach (Descloux et al., 2010).

2. Material and methods 2.1. Study area The field study was conducted in October 2007 on three reaches located on each of three rivers belonging to the Rhône River catchment: the Drôme River ‘D’ (fourth-order, sensu Strahler, 1952), the Usses River ‘U’ (fourth-order) and the Isère River ‘I’ (sixth-order; Fig. 1 and Supplementary material 1). Sampling sites were located on the downwelling zone of riffles within reaches of approximately 100 m long. The geological substrate of the Drôme catchment is primarily limestone. The Usses River catchment is mainly composed of limestone and marl and the Isère River catchment consists of granite and marl. Streambeds were dominated by coarse mineral substrate (pebbles and cobbles) mixed with various amounts of sand, silt and clay. The Isère River submontane reaches, at a mean altitude of 416 m, were colder (mean temperature of 9.0 °C) than the Drôme (288 m; 13.6 °C) and Usses (348 m; 15.3 °C) River reaches located in a more lowland context. The water of the three rivers was well oxygenated (ranging from 9.2 to 11.3 mg L−1) and had low to moderate nutrient contents. The three rivers had a pH between 8.1 and 8.6 (Supplementary material 1). 2.2. Gradient of colmation and invertebrate sampling In each river, three reaches were selected according to i) the recommendations of local experts, ii) combined measures of colmation through visual estimates, hydraulic conductivity evaluation and application of the wooden stake technique and iii) measures of their percentage of fine sediment (PFS), to reflect a colmation gradient (Descloux et al., 2010). Sediments were of natural origin. The three replicate cores, using the freeze-coring technique (Stocker and Williams, 1972) were performed in each of the nine reaches to identify its grain size distribution and PFS down to a depth of

60 cm. Then, the three reaches of each river were respectively classified as ‘heavily clogged’ (noted HC), ‘moderately clogged’ (noted MC) and ‘lightly clogged’ (noted LC), using the combined results. The colmation gradient and percentage of fines particles are available in Table 2. A ‘Hess sampler’ was used to collect benthic invertebrates (sampling area: 0.125 m2). Triplicate samples were performed after randomly selecting sampling points spaced by a maximum of 2 m from PFS evaluation sampling points, to reduce effects of small-scale habitat variability. Hyporheic invertebrates were collected at the same location, using the Bou Rouch system (Bou and Rouch, 1967) at three depths (i.e. − 10, −30 and − 50 cm deep) for collecting the whole hyporheic fauna. Six liters of water were pumped at a constant rate of 4 L min−1. The three depth sample units were gathered in a single ‘sample’ to eliminate the depth effect and reduce the number of confounding factors. All the invertebrate samples from both zones were preserved in 70% isopropyl alcohol, identified to the lowest possible taxonomic level [generally at the species or genus level, based on Harding and Smith, 1974; Tachet et al., 2002 and Scarbrook et al., 2003] and enumerated. 2.3. Species traits Ten biological traits were used to examine the effects of colmation on the invertebrate assemblages (Appendix 1). Each trait was described by several categories. Autecological information gathered from the literature was translated into relative frequency distributions, each describing the affinity of a given taxon for the different categories of a given trait. This procedure is similar to the fuzzy coding procedure (Chevenet et al., 1994). The trait profiles of 124 taxa from the meio- and macro-fauna of the benthic and hyporheic zones were described mainly at species or genus level, except for some Diptera (described at family or sub-family level), Nematoda and Oligochaeta (described as mean

Table 2 Percentage of interstitial fine sediments (i.e. particle size b2 mm) in the heavily clogged (HC), moderately clogged (MC) and lightly clogged (LC) reaches of the three rivers. After Descloux et al. (2010). Rivers

DROME D-HC

% of fine sediments b 2 mm 0–20 cm 10.0 (3.7) 20–40 cm 14.4 (2.4) 40–60 cm 16.2 (3.8)

USSES

ISERE

D-MC

D-LC

U-HC

U-MC

U-LC

I-HC

I-MC

I-LC

5.9 (1.5) 18.1 (10.6) 13.6 (4.5)

7.9 (7.9) 11.2 (4.7) 11.2 (2.8)

23.5 (18.7) 50.2 (24.2) 69.3 (26.6)

15.7 (6.7) 20.6 (2.3) 33.8 (28.0)

6.5 (4.0) 10.2 (1.9) 7.5 (0.8)

29.3 (2.7) 27.6 (1.1) 27.5 (2.3)

19.8 (8.9) 28.5 (9.7) 32.8 (3.1)

11.9 (1.9) 22.8 (3.7) 28.0 (7.4)

268

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

trait profiles of their potential families in the corresponding biogeographic area). Traits reflected the life history of taxa (e.g. ‘maximal potential size’, ‘number of cycles per year’, ‘fecundity’), their mobility (‘locomotion and substrate relation’), general morphological (‘body form’) or physiological (‘respiration’, ‘reproduction technique’, ‘feeding habits’, ‘food’, ‘resistance forms’) features of organisms. Trait category description, traditionally used for benthic invertebrates (e.g. Statzner et al., 1994; Usseglio-Polatera et al., 2000a), was modified to take into account the particular features of the invertebrates from the hyporheic zone (see Appendix 1 and Supplementary material 2). The mean trait profile of each reach sample assemblage was obtained by weighing the individual trait profiles of taxa by their log-transformed abundances in the sample. Then, the sums of the weighted scores (one per trait category) were rescaled to sum to one for each trait and each river reach, following Dolédec et al. (2000), Statzner et al. (2001), Gayraud et al. (2003), Archaimbault et al. (2010) and Demars et al. (2012). 2.4. Data analyses 2.4.1. “Within river” Fuzzy Correspondence Analysis Decomposing the total variance in the data, a widely used method for identifying the major explanatory variables of differences within a given array (e.g. Dolédec and Chessel, 1987, 1989; Bournaud et al., 1998; Dangles et al., 2001), was applied to the variance in the mean trait profiles of reach sample assemblages, broken down into ‘river’ (i.e. D, U, and I) ‘compartment’ (i.e. benthos vs. hyporheos) and ‘colmation’ (i.e. LC, MC, and HC) variation. Because our major objective was to focus only on both the ‘colmation’ and ‘compartment’ effects, we removed the ‘between-river’ variation (considered as a confounding effect) in reach assemblage trait profiles by centering the data ‘by river’ before applying a multivariate approach to the data array. As a result, a ‘within-river’ Fuzzy Correspondence Analysis (FCA; Chevenet et al., 1994) was performed to assess how substrate colmation can modify combinations of trait categories within the reach invertebrate assemblages of both the benthic (B) and hyporheic (H) zones. 2.4.2. Linear mixed-effects models To take into account the inner variation (i.e. rivers) and the variation due to the experimental design (colmation), linear mixed-effects models with Gaussian error distribution were fitted to the observed data for each compartment (benthos vs. hyporheos) independently. They allowed examining the relationships between colmation and the relative abundance of invertebrate trait categories. For each dependent variable (trait categories), we tested for the effects of colmation, and then tested if these effects differed among rivers. Hence, we compared nested mixed-effects models with increasing complexity. The first model (intercept only model) was a null model with a random intercept. The second model (colmation fixed model) was an average model with a fixed effect of colmation across rivers. The third model (colmation random model) was a model with a random effect of colmation, which was allowed to vary among rivers (Bolker et al., 2009; Öckinger et al., 2010). An effect of colmation across the three rivers was observed when the difference between the ‘intercept only’ model versus the ‘colmation fixed’ model was significant. This effect was river-specific when the difference between the ‘colmation fixed’ model versus the ‘colmation random’ model was significant. The statistical significance levels for the fixed and random effects in the best-fitting models were determined using likelihood-ratio tests (chi-squared distribution, p b 0.05 before Bonferroni correction for multiple comparisons of categories among traits) on models with and without each effect (Bolker et al., 2009; Öckinger et al., 2010). To select the most parsimonious model, we used the minimum Akaike's Information

Criterion (AIC) (Bolker et al., 2009). Residual plots did not reveal any substantial departures from the assumption of normality for any ‘trait category × compartment’. Statistics and graphical outputs were computed with ade4 library (Thioulouse et al., 1997) and the package ‘nlme’ (Pinheiro et al., 2009) implemented in the R software (R Development Core Team, 2009). 3. Results 3.1. Mean biological trait profiles of each assemblage among the three colmation classes The decomposition of the total variance in site assemblage trait profiles revealed that the effect of ‘colmation’ accounted for only 8.5% of the total variance, whereas ‘river’ and ‘compartment’ effects respectively accounted for 33.9 and 15.9% of the total variance in the biological trait profiles of reach sample assemblages. The ‘river’ effect importance justified the ‘within river’ FCA approach selected to optimally evidence the ‘colmation’ and ‘compartment’ effects by removing ‘between-river’ variation from the FCA analysis (Dolédec and Chessel, 1989). The plane defined by both the first (F1; 42.9% of explained variance) and third (F3; 10.4%) factorial axes in the ‘within-river’ FCA best described the colmation effect (Fig. 2); the second axis (11.4%) captured more ‘between-replicates’ variability within each ‘reach type’ (not presented). Considering their mean trait profiles, most of benthic assemblages (F1 N 0) were clearly opposed to hyporheic assemblages (F1 b 0) along the F1 axis. River reach assemblages were mainly located along the F3 axis according to the colmation level of the corresponding habitats from lightly (F3 b 0; opened circles in Fig. 2) to moderately/heavily (F3 N 0; solid circles) clogged reaches, except ‘I3-B’. Along the F1 axis, that reflects differences in benthos vs. hyporheos, the traits ‘respiration’ [tegument (F1 b 0) vs. gill or aerial (F1 N 0)], ‘reproduction technique’ [free eggs or clutches and parthenogenesis vs. fixed eggs and/or clutches or ovoviviparity], ‘body form’ [cylindrical or spherical vs. flattened and/or streamlined], ‘feeding habits’ [predator or deposit-feeder vs. scraper, shredder and piercer] and ‘maximal size’ [very small vs. other sizes] best discriminated õinvertebrate assemblages of the different river reaches (Fig. 3). These traits exhibited the strongest differences in trait category õselection when comparing benthic (F1 N 0) to hyporheic (F1 b 0) assemblages. The categories of the traits ‘fecundity’ [intermediate (F3 b 0) vs. high (F3 N 0)], ‘reproduction technique’ [ovoviviparity vs. õparthenogenesis] and ‘food’ [macrophytes and invertebrates vs. microorganisms and fine detritus] were well separated along the F3 axis. Such differences in trait category selection were mainly related to differences in colmation classes. 3.2. Identity of changing traits and magnitude of change with colmation for the two zones Colmation significantly modified eight trait profiles (19 trait category frequencies modified; Table 3) of the invertebrate assemblages in the benthic zone, including (i) maximal size, (ii) fecundity, (iii) reproduction technique, (iv) resistance forms, (v) respiration, (vi) locomotion and substrate relation; (vii) food and (viii) feeding habits. Across the three rivers, crawlers (Loco2 in Table 3) and attached organisms (Loco5), species with high fecundity (Fec3), branchial respiration (Resp2) and/or consuming dead plants or living microphytes (Food3 and 4), especially scrapers (Fhab3), were increasingly selected in benthic assemblages along an increasing colmation gradient. In contrast, colmation has a negative effect on organisms with tegumental respiration (Resp1 in Table 3), shredding (Fhab2) and/or eating

(10.4%)

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

269

0.24 -0.32 0.28 -0.16

F3 I1-B U2-H

D2-H

D1-H U1-B D2-B

I3-H

U2-B

F1

U1-H D1-B D3-H

I1-H

D3-B

I2-H

Lightly Moderately Heavily

U3-B

I2-B

I3-B U3-H

(42.9%)

Fig. 2. Ordination of 54 site samples (small black squares) according to the mean biological trait profiles of their macroinvertebrate assemblages by “within-river” Fuzzy Correspondence Analysis. Ten traits including 48 categories (cf. Appendix 1) were simultaneously taken into account. Solid circles correspond to the mean location of the different [River × Compartment × Colmation class] combinations. They were positioned at the weighted average of the corresponding faunal replicate locations (small squares). Lines link replicates to their mean location. The percentages of explained variance were indicated on each axis. H = hyporheos samples; B = benthos samples, I = Isère River, D = Drôme River; U = Usses River; Reaches from a given river are labeled from ‘1’ (upstream) to ‘3’ (downstream) along the longitudinal gradient.

microinvertebrates (Food7), using ovoviviparity as reproduction strategy (Repr1) and/or living in substrate interstices (Loco4) in the benthic zone. The slope of the relationship between the relative frequency of each of these biological attributes and the colmation level did not significantly differ among rivers (the difference between the ‘intercept only’ model versus the ‘colmation fixed’ model was significant and the ‘colmation fixed’ model was the best model; cf. Appendix 2), indicating the similarity of the relationship ‘trait category frequency/colmation level’ in the three rivers. The relationship between the benthic proportion of (i) deposit-feeders (Fhab 2 in Table 3) consuming microorganisms in fine sediment (Food1) or (ii) individuals using asexual reproduction (including parthenogenesis, Repr7) and the colmation gradient varied according to rivers (i.e. positive in D and U, but negative in I). Similarly, the proportion of invertebrates using cocoons as a resistance stage (Resi2) in I or exhibiting medium fecundity (Fec2) in U, significantly decreased with colmation, but these proportions increased in the two other rivers (Table 3). Only five traits (seven trait categories in Table 3) of the hyporheic assemblages were significantly modified by colmation: (i) maximal size, (ii) fecundity, (iii) body form, (iv) resistance forms and (v) food. In the three rivers, colmation both (i) enhanced the proportion of organisms with large size (Size4 in Table 3) and/or high fecundity (Fec3) and (ii) decreased the proportion of organisms exhibiting opposite attributes (Size1, Fec1), spherical form (For4), diapause/dormancy as resistance strategy (Resi3) or living microinvertebrates as food (Food7). For these attributes, the ‘trait category frequency’/‘colmation level’ linear models exhibited a similar slope in the three rivers, except for For4 (Table 3). 4. Discussion 4.1. Were traits of the hyporheos less drastically affected by colmation than those of the benthos? (Hypothesis H1) Fine sediments filling interstices in clogged substrates induce changes in habitat conditions and then in local fauna. The first evidence of community impairment is (i) a drastic decrease in

total abundance and taxonomic richness for both the benthic and hyporheic zones (Bjornn et al., 1977; Lenat et al., 1979) and (ii) a shift from EPT to Oligochaeta and Diptera dominance (Lenat et al., 1979; Waters, 1995). A major result of our study is that colmation selects also combinations of traits in invertebrate assemblages. Our first hypothesis (H1) predicting a lower filtering effect of colmation on the traits of invertebrate assemblages in the hyporheic zone than in the benthic zone has been supported: both the numbers of impaired traits (8 vs. 5 from 10) and modified trait category frequencies (19 vs. 7 from 48) were lower in the hyporheos, probably due to a local invertebrate assemblage more adapted to fine sediment. In the studied river reaches, the hyporheos was mainly composed of epigean species (see Supplementary material 3) that need effective vertical exchanges between the two zones via sediment interstices. However the species inhabiting the hyporheos area are probably more adapted to fine particles naturally collected by this zone (e.g. Baetis sp.), so that further increase in fine particles due to colmation may result in lower community changes than in the benthos (Robertson and Wood, 2010). Moreover, clogged hyporheic habitats are temporally more stable than clogged benthic habitats, because they are less physically disturbed during high flow conditions. As a result, hyporheic assemblages should be less impaired by colmation, losing probably fewer animals and less varying in both their taxonomic and trait composition. Second, the trait categories enhanced in the invertebrate assemblages differed between the two zones. In the benthic zone, colmation selects biological attributes that can be related to the resistance [e.g. substrate relation (four categories in Table 3)] or resilience [e.g. fecundity or reproduction technique (four categories)] abilities of species and their physiological [respiration (two categories)] or trophic [food or feeding habits (seven categories)] functions. In contrast, the selection frequencies of specific morphological characteristics [maximal size or body form (three categories)] were more modified in the hyporheic zone. Therefore, it seems that traits of clogged habitat assemblages are more impaired by an increase in physico-chemical constraints (e.g. the reduction of oxygen availability) and a reduction of potential exchanges (including exchanges of

270

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

Maximal potential size > 2 cm > 1-2 cm

Nb cycles/year

Fecundity

F1 1

> 100 -1000

streamlined flattened

parthenogenesis free eggs free clutches terrestrial clutches

sphaerical

cocons cemented eggs

Resistance forms Food

gill

burrower swimmer crawler attached

aerial

microorganisms

fine detritus

microphytes dead plant

microinvertebrates dead animal

macrophytes macroinvertebrates

Locomotion & substrate relation deposit feeder scraper

none

diapause

ovoviviparity

Reproduction Technique

interstitial

Feeding habits

eggs

cemented clutches

Respiration tegument

<= 100

<1

> 0.25-0.5 cm

Body form cylindrical

> 1000

>1

> 0.5-1 cm <= 0.25 cm

F3

0.2 - 0.55 0.4 - 0.4

shredder predator filter -feeder piercer

Fig. 3. Ordination of 48 categories (small black squares) from 10 biological traits by “within-river” Fuzzy Correspondence Analysis according to their selection within 54 benthic and hyporheic assemblages. Locations of trait categories in the F1–F3 factorial plane [cf. Fig. 2 for the corresponding locations of reach samples].

food resources) in the benthic than in the hyporheic zone. In the hyporheos, morphological attributes are probably selected more by changes in interstitial space characteristics. Third, colmation simultaneously modified only three trait categories in the hyporheic and benthic assemblages (i.e. Size1; Fec3; and Food7). For each of these trait categories, the pattern of the relationship between the relative frequency of the trait category and the colmation level was similar in both zones (less small-sized organisms and microinvertebrate consumers, more highly fecund individuals with increasing colmation), indicating the similarity of the underlying selective processes in the two zones for these attributes. The reduction of pore spaces within the substrate limits the development and dispersal of small-sized and slow-moving organisms, e.g. the meiofauna (i.e. Copepoda, Ostracoda and Cladocerans), by reducing the possibility of interstitial movements and between-zone exchanges. In both zones, the higher investment of organisms in egg production along the colmation gradient could be also related to the more severe habitat conditions in clogged habitats.

Fourth, the filtering effect of colmation on the biological attributes of invertebrate assemblages was significantly more homogeneous among rivers in the hyporheic zone than in the benthic zone, indicating a higher homogeneity of invertebrate assemblage composition and/or colmation effects in the hyporheos than in the benthos. The filtering effect of colmation statistically differed among rivers only for For4, but the observed trends in ‘trait category frequency’ versus ‘colmation level’ relationship remained similar in the hyporheic assemblages of the three rivers (slopes with the same sign). 4.2. Were the predictions on community traits matching the observations along the gradient of colmation? (Hypothesis H2) Predicting combinations of traits (Hypotheses H2, Table 1) that may be appropriate to a wide-range of invertebrates in the clogged habitats of the two zones remains difficult, because colmation reduces spatial heterogeneity [i.e. the availability and diversity of refugia (Townsend and Hildrew, 1994)], but simultaneously

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276 Table 3 Results of the linear mixed-effects models analyzing the relationship between the relative occurrence of trait categories and colmation within the benthic and hyporheic assemblages. *** = p-value b0.001; ** = 0.001 ≤ p-value b0.01; * = 0.01 ≤p-value b0.05; n.s = non significant. Trends give the information on the slope of the model: ‘+’ = trait category more selected with increasing colmation and ‘−’ = trait category less selected with increasing colmation. ‘Colm.’ = colmation, ‘D’ = Drôme River, ‘U’ = Usses River and ‘I’ = Isère River. See text for further details. Benthos

Hyporheos

Colm. Colm. Trend Trend fixed random colm. colm. fixed random

Colm. Colm. Trend Trend fixed random colm. colm. fixed random

D Size1 Size4 Fec1 Fec2 Fec3 For4 Repr1 Repr7 Resi2 Resi3 Resp1 Resp2 Loco1 Loco2 Loco4 Loco5 Food1 Food3 Food4 Food7 Fhab1 Fhab2 Fhab3

* n.s n.s ** *** n.s ** * * n.s * ** * * ** * n.s * * *** * *** ***

n.s n.s n.s * ** n.s ** ** ** n.s n.s * n.s n.s n.s n.s ** n.s n.s n.s ** n.s n.s

U

I

−

+ − + + − + + − − + − − + − + − + + + − + + − + + − − +

D * ** * n.s *** ** n.s n.s n.s *** n.s n.s n.s n.s n.s n.s n.s n.s n.s * n.s n.s n.s

n.s n.s * n.s n.s * n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s n.s

U

I

− + − + − − −

−

enhances temporal stability (a function of disturbance frequency, magnitude and predictability) of habitats, especially in the hyporheic zone. Furthermore, the spatial and temporal heterogeneity of habitats can act as an environmental filter with different weights according to traits and taxa. In other words, it is not clear if clogged habitats could act as a templet in a uniform way for all the traits of the species of an assemblage, because selected combinations of attributes are the specific results of potentially synergistic and/or antagonistic drivers, and depend on the adaptation capacities of the different taxonomic lineages. However, the results of the linear mixed-effects model approach seem to indicate a more homogeneous filtering effect of colmation on the traits of the hyporheic assemblages than the benthic assemblages in the three rivers. The colmation gradient seemed to strongly select morphological features (e.g. size, body shape) in the hyporheic zone. Some of the trait categories selected in the assemblages in clogged habitats were in accordance with our predictions (Table 1). The gradual shift from spherical (For4; i.e. Gastropoda and Microcrustacea; see Supplementary material 3) to cylindrical (mainly worms and Diptera) organisms in the hyporheos, might be linked to the reduction of interstitial pore size with colmation. Gayraud and Philippe (2001) already showed that cylindrical organisms were preferentially selected in habitats with interstitial spaces clogged by fine sediments, and suggested a body shape adaptation more related to bed sediment characteristics than to hydraulic conditions (because a cylindrical form does not significantly minimize the flow constraints on organisms; Statzner and Holm, 1982). The increasing fecundity observed in both areas may be considered as the result of an increasing species investment in reproduction in adverse conditions (e.g. Stearns, 1976; Verberk et al., 2008a). Reproduction can be also facilitated by ovoviviparity (Repr1) and asexual strategies (Repr7) in benthic habitats when colmation

271

impairs both egg development and invertebrate dispersal (see, for example, strategy T2, in Verberk et al., 2008b; p. 1744). The reduction of tegumental respiration (Resp1) and the increase in branchial respiration (Resp2) in the benthic zone may appear counter intuitive because (i) organisms with tegumental respiration are considered as more tolerant to the predicted dissolved oxygen depletion in interstitial water with colmation (Tomanova et al., 2008; Larsen et al., 2011), and (ii) organisms relying on gill respiration are supposed to be especially sensitive to fine particles that can impair their respiratory structures (Lemly, 1982). Dolédec et al. (2006) did not notice significant variation in the proportion of invertebrates with branchial respiration in benthic assemblages along an increasing gradient of fine sediment cover. In contrast, a significant decline of organisms with gills was already observed (i) along an increasing gradient of streambed fine sediment cover in New Zealand grassland streams (Townsend et al., 2008) and (ii) in artificial substrates filled with different amounts of sand in Welsh streams (Larsen et al., 2011). The results observed on the three rivers could be related to (i) hypoxic conditions not severe enough to limit the vertical distribution of animals with branchial respiration, and/or (ii) the presence of young insect instars that ‘potentially’ have gills (‘potential’ trait profiles being described over the whole biological cycle of taxa) but use mainly direct oxygen exchanges across tegument due to the low efficiency of their small-sized gills. Several trends related to ‘substrate relation’ also supported our predictions; e.g. fewer endobenthic organisms and swimmers, more crawlers and firmly attached individuals were observed in the epibenthic area with colmation. In addition, even if fine sedimentation has been known to constrain invertebrate dispersal within the substrate, e.g. by altering the vertical connectivity (Boulton, 2007), locomotion-related adaptations seemed to be less constrained by colmation in the hyporheos than in the benthos, maybe because the hyporheic fauna is already adapted to high level of fine sediments (Table 3). In contrast, the observed trends in maximum potential size did not match our predictions. A significant reduction of small-sized (Size1) individuals was observed in both zones, even if mean invertebrate size has been usually predicted to decrease in temporally unstable habitats (Townsend and Hildrew, 1994) or in harsh environments because of an increasing energy allocation of organisms in reproduction and/or specific physiological and morphological adaptations (Verberk et al., 2008b). We can hypothesize a major influence of the increasing colmation-related stability of hyporheic habitats in the selection of large-sized organisms. Boulton et al. (1998) have already observed the functional importance of the larger groundwater invertebrates (e.g. amphipods, isopods and syncarids), because their movements within the substrate and their ability to create and maintain voids by pelletizing fine interstitial material and biofilm, help to prevent substrate porosity loss and severe clogging effects. Lenat et al. (1981) and Broekenhuizen et al. (2001) demonstrated a decline in the feeding efficiency of filter-feeders and grazers with fine sediment deposition. However, substrate colmation has had little effect on the trophic resource utilization (i.e. ‘food’) and on the functional feeding group (FFG) structure (i.e. ‘feeding habits’) of the hyporheic assemblages in the three rivers (Table 3; Supplementary material 4). These assemblages seemed already adapted to the available food resources in fine sediment enriched habitats. However, a significant increase in deposit-feeders (except in the Isère River; Fhab1) and a significant decrease in shredders (Fhab2) were observed in the benthic assemblages, matching our a priori hypotheses and previous observations (e.g. Richards et al., 1997; Wantzen, 2006; Townsend et al., 2008). In contrast, benthic scrapers (Fhab3) seemed to be favored by colmation, despite adverse biofilm development conditions. Non embedded coarse particles at the streambed surface probably (still) allow the local development of perilithon eaten by scarpers (e.g. Heptageniidae). The higher number of food categories

272

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

exhibiting significant variation – with colmation – in their utilization by macroinvertebrate assemblages in the benthic (Food1, 3, 4 and 7) versus hyporheic (only Food7) zones may be related to a more flexible strategy of hyporheic species in food harvesting, leading to a more stable FFG structure of hyporheic assemblages despite potential variations in trophic resources of clogged habitats.

river functionality and can be viewed as the first step towards an in situ diagnostic tool of the effects of river bottom colmation at the community level. Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.scitotenv.2013.06.082.

4.3. Perspectives

Acknowledgments

Many invertebrate-based strategies have been developed (i) to evaluate the actual ecological status of water bodies and (ii) to assess water body impairment risk individually considering different anthropogenic pressure categories (e.g. Böhmer et al., 2004; Buffagni et al., 2004; Lorenz et al., 2004; Gabriels et al., 2010; Mondy et al., 2012) in the European Water Framework Directive (WFD) context. If traditional bio-assessment, including colmation detection, has been mainly based on structural attributes (Hering et al., 2006), a growing number of studies have used species traits as a proxy for function (Usseglio-Polatera et al., 2000b; Dolédec and Statzner, 2008; Archaimbault et al., 2010) to target anthropogenic stress. Indeed, by mechanistically inferring potential cause–effect relationships between stressors and biological impairments (Archaimbault et al., 2010), species traits can link community organization to ecosystem goods and services (Devin et al., 2005; Lecerf et al., 2006). However, none of these studies have simultaneously considered the benthic and hyporheic communities, even if many benthic invertebrate species rely so closely on the hyporheic zone that restoration of such zones may be needed to allow their successful recolonization of rivers (Collier et al., 2004; Boulton, 2007). Streambed colmation has a mechanistic action on invertebrate trait selection (Statzner and Bêche, 2010), increasing with colmation level. This action is biologically more selective on benthic than hyporheic assemblages, because benthic assemblages are naturally less adapted to fine particulate matter sedimentation. As a result, a potential indicator of the effects of colmation on river health may be based on the functional traits of benthic macroinvertebrates because such traits (i) satisfy the WFD recommendations (i.e. implicitly evaluating the ecological status of water bodies by comparing BQEs – e.g. macroinvertebrates – between an observed vs. a reference situation, regarding abundance, diversity and pollution sensitivity of taxa; Furse et al., 2006; Mondy et al., 2012), (ii) respond consistently along a colmation gradient and (iii) are comparable among benthic communities even across ecoregions that differ in their taxonomic composition (Statzner et al., 2001; Horrigan and Baird, 2008). Because the present study was restricted to streams with gravel bed, we have also to test the consistency of observed response patterns on a more diverse set of streams in terms of (hydro)morphological and geological characteristics.

We are grateful to the Zone Atelier Bassin du Rhône (ZABR) for information and contacts. The lead author (SD) was funded by a grant from Electricité De France, the second author (TD) by the IRSTEA. We are grateful to Professor Pierre Marmonier for his help in coding Ostracoda and Cladocera trait profiles.

Appendix 1. Biological traits and categories (= modalities) of aquatic macroinvertebrates used in this study. Trait codes into brackets were used as labels

N° Traits

N° Modalities

1

Maximal potential size (Size)

2

Number of cycles/year (Cycl)

3

Fecundity (Fec)

4

Body form (For)

5

Reproduction technique (Repr)

1 2 3 4 5 1 2 3 1 2 3 1 2 3 4 1 2 3 4 5 6 7

6

Resistance forms (Resi)

7

Respiration (Resp)

8

Locomotion and substrate relation (Loco)

9

Food (Food)

5. Conclusion Defining tools for evaluating the hyporheic zone integrity is of high interest for river biomonitoring because this zone (i) plays an important role in the life cycle of many epibenthic invertebrates (Orghidan, 1959; Puig et al., 1990) and many of their potential consumers (e.g. fish), and (ii) is involved in major ecosystem services such as water purification, water infiltration and bioremediation (Boulton et al., 2008). As a result, the restoration of clogged hyporheic habitats, e.g. by flushing clear waters in regulated rivers or streambed mechanical scraping, should be a priority of decision makers to enhance global river health and resilience. Our study not only confirmed the significant effect of substrate colmation on benthic invertebrate traits, but also highlighted a significant effect of colmation on the hyporheic community of rivers, even if obviously less severe than on the benthic community. Moreover, our results have underlined the potential of biological life history traits for assessing

10 Feeding habits (Fhab)

1 2 3 4 1 2 3 1 2 3 4 5 1 2 3 4 5 6 7 8 1 2 3 4 5 6

≤0.25 cm N0.25–0.5 cm N0.5–1.0 cm N1.0–2.0 cm N2.0 cm b1 1 N1 ≤100 N100–1000 N1000 Streamlined Flattened Cylindrical (+ geometric) Spherical Ovoviviparity Isolated eggs, free Isolated eggs, cemented Clutches, cemented or fixed Clutches, free Clutches, terrestrial (+ in vegetation) Parthenogenesis (+ asexual reproduction) Eggs, statoblasts Cocoons (+ housings against desiccation) Diapause or dormancy None Tegument Gill Aerial (plastron + spiracle) Swimmer Crawler Burrower (epibenthic) Interstitial (endobenthic) Attached (temporarily or permanently) Fine sediment + microorganisms fine detritus (b1 mm) Dead plant (N1 mm) Living microphytes Living macrophytes Dead animal (N1 mm) Living microinvertebrates Living macroinvertebrates (+ vertebrates) Deposit-feeder (+ absorber) Shredder Scraper Filter-feeder Piercer (plants or animals) Predator (carver/engulfer/ swallower) + parasite

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

273

Appendix 2. Linear mixed-effect model selection procedure with dF (number of parameters), AIC (Akaike's Information Criterion), BIC (Bayesian Information Criterion), Log. Lik (Log-Likelihood), Test (comparison between models) and L ratio (Likelihood ratio)

Trait Model categories Benthos Size1

1. Intercept only 2. Colmation fixed

Fec2

Fec3

Repr1

Repr7

Resi2

Resp1

Resp2

Loco1

3. Colmation random 1. Intercept only 2. Colmation fixed

m

Loco5

Log Lik Test

− 111.97 5 − 117.60 10 − 109.22 3 −57.93 5 −66.31

− 108.09 − 111.12 −96.26

58.98 63.80

1 vs. 2 9.62

64.61

−54.05 −59.83

3

L ratio P-value

Interpretation of model comparison

Conclusion

0.0081

Models sign. diff. — lower AIC (Colmation fixed model)

Colmation fixed

2 vs. 3 1.61

0.89

Models not sign. diff.

31.96 38.15

1 vs. 2 12.37

0.0021

Models sign. diff.

Colmation random

10 −72.56

−59.60

46.28

2 vs. 3 16.24

0.0062

Models sign. diff. — lower AIC (Colmation random model)

3 −75.93 5 −91.35 10 −82.62

−72.04 −84.87 −69.67

40.96 50.67 51.31

1 vs. 2 19.41 2 vs. 3 1.27

0.0001 0.93

Models sign. diff. — lower AIC (Colmation fixed model) Models not sign. diff.

Colmation fixed

5

68.25

1 vs. 2 14.02

0.0009

Models sign. diff. — lower AIC (Colmation fixed model)

Colmation fixed

3. Colmation random 1. Intercept only

10

− 112.58 − 120.02 − 106.79 −96.37

61.23

2. Colmation fixed

69.87

2 vs. 3 3.24

0.06

Models not sign. diff.

2. Colmation fixed

5

1 vs. 2 10.40

0.0053

Models sign. diff.

10

68.94

2 vs. 3 21.14

0.0008

Models sign. diff. — lower AIC (Colmation random model)

2. Colmation fixed

5

70.72

1 vs. 2 9.56

0.0084

Models sign. diff.

3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random 1. Intercept only

10

− 100.26 − 104.92 − 122.00 − 124.97 − 131.72 −69.71 −74.08 −59.79

58.36

3. Colmation random 1. Intercept only

− 116.47 − 126.50 − 119.75 − 100.26 − 106.73 − 117.88 − 125.88 − 131.45 − 144.68 −73.60 −80.56 −72.74

82.34

2 vs. 3 23.22

0.0003

Models sign. diff. — lower AIC (Colmation random model)

39.80 45.28 46.37

1 vs. 2 10.96 2 vs. 3 2.18

0.0042 0.82

Models sign. diff. — lower AIC (Colmation fixed model) Models not sign. diff.

Colmation fixed

3 −82.39 5 −91.38 10 −82.06

−78.50 −84.90 −69.10

44.19 50.69 51.03

1 vs. 2 12.98 2 vs. 3 0.68

0.0015 0.98

Models sign. diff. — lower AIC (Colmation fixed model) Models not sign. diff.

Colmation fixed

− 111.78 5 − 117.85 10 − 112.47 3 −78.86 5 −87.25 10 −87.08

− 107.89 − 111.37 −99.51

58.89 63.92

1 vs. 2 10.07

0.0065

Models sign. diff. — lower AIC (Colmation fixed model)

Colmation fixed

66.23

2 vs. 3 4.61

0.46

Models not sign. diff.

−74.97 −79.77 −74.13

42.93 48.12 53.54

1 vs. 2 11.39 2 vs. 3 10.83

0.0034 0.056

Models sign. diff. — lower AIC (Colmation fixed model) Models not sign. diff.

Colmation fixed

−88.34 −94.70

49.11 55.59

1 vs. 2 12.94

0.0015

Models sign. diff. — lower AIC (Colmation fixed model)

Colmation fixed

−91.61

62.28

2 vs. 3 13.38

0.02

Models not sign. diff. after Bonferroni correction

− 105.74 − 111.09 − 103.47 − 154.88 − 156.46 − 159.74

57.81 63.78

1 vs. 2 11.43

0.0026

Models sign. diff. — lower AIC (Colmation fixed model)

68.21

2 vs. 3 8.85

0.11

Models not sign. diff.

86.47

1 vs. 2 8.16

0.016

Models sign. diff.

96.34

2 vs. 3 19.75

0.0014

Models sign. diff. — lower AIC (Colmation random model)

3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random 1. Intercept only 2. Colmation fixed

Food1

BIC

3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random 1. Intercept only

2. Colmation fixed

Loco2

dF AIC

3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random

3

3

3

3 5 10

3

−92.23 − 101.18 10 − 104.57 3 − 109.63 5 − 117.57 10 − 116.43 3 − 158.77 5 − 162.94 10 − 172.69 3 5

53.13 Colmation random

65.94 Colmation random

Colmation fixed

82.38 Colmation random

(continued on next page)

274

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

Appendix 2 (continued) Trait Model categories

dF AIC

BIC

Log Lik Test

Food3

− 151.78 5 − 160.82 10 − 158.57 3 − 111.84 5 − 118.50 10 − 109.35 3 − 146.28 5 − 161.94 10 − 153.24 3 −74.81 5 −81.79

− 147.89 − 154.34 − 145.61 − 107.96 − 112.02 −96.40

78.89

− 142.39 − 155.46 − 140.28 −70.93 −75.31

10 −92.08 −99.00 − 113.00

1. Intercept only 2. Colmation fixed

Food4

3. Colmation random 1. Intercept only 2. Colmation fixed

Food7

3. Colmation random 1. Intercept only 2. Colmation fixed

Fhab1

Fhab2

Fhab3

3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random 1. Intercept only 2. Colmation fixed

3

3 5

L ratio P-value

Resi3

Food7

3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random 1. Intercept only 2. Colmation fixed 3. Colmation random

Conclusion

Colmation fixed

85.41

1 vs. 2 13.04

0.0015

Models sign. diff. — lower AIC (Colmation fixed model)

89.28

2 vs. 3 7.74

0.17

Models not sign. diff.

64.25

1 vs. 2 10.65

0.0048

Models sign. diff. — lower AIC (Colmation fixed model)

64.67

2 vs. 3 0.85

0.97

Models not sign. diff.

85.97

1 vs. 2 19.66

0.0001

Models sign. diff. — lower AIC (Colmation fixed model)

86.62

2 vs. 3 1.29

0.93

Models not sign. diff.

40.40 45.89

1 vs. 2 10.97

0.0041

Models sign. diff.

−79.13

56.04

2 vs. 3 20.29

0.0011

Models sign. diff. — lower AIC (Colmation random model)

−95.11 − 106.52

52.50 61.50

1 vs. 2 18.00

0.0004

Models sign. diff. — lower AIC (Colmation fixed model)

58.92 Colmation fixed

76.14 Colmation fixed

Colmation random

Colmation fixed

No convergence of the model 3 5

−87.68 − 103.34

−83.79 −96.86

46.84 56.67

1 vs. 2 19.00

0.0004

3. Colmation random Hyporheos Size1 1. Intercept only 2. Colmation fixed 3. Colmation random Size4 1. Intercept only 2. Colmation fixed 3. Colmation random Fec1 1. Intercept only 2. Colmation fixed 3. Colmation random Fec3 1. Intercept only 2. Colmation fixed 3. Colmation random For4 1. Intercept only 2. Colmation fixed

Interpretation of model comparison

Models sign. diff. — lower AIC (Colmation fixed model)

Colmation fixed

No convergence of the model

3 −54.37 5 −61.94 10 −58.01

−50.49 −55.46 −45.05

30.18 35.97 39.60

1 vs. 2 11.56 2 vs. 3 6.06

0.0031 0.29

Models sign. diff. — lower AIC (Colmation fixed model) Models not sign. diff.

Colmation fixed

−77.43 −89.52

−73.54 −83.04

41.71 49.76

1 vs. 2 16.08

0.0003

Models sign. diff. — lower AIC (Colmation fixed model) No convergence of the model

Colmation fixed

3 −45.42 5 −52.27 10 −47.84

−41.53 −45.80 −34.88

25.71 31.13 33.92

1 vs. 2 10.85 2 vs. 3 5.56

0.0044 0.35

Models sign. diff. — lower AIC (Colmation fixed model) Models not sign. diff.

Colmation fixed

3 5

−43.83 −57.28

−39.94 −50.80

24.91 33.64

1 vs. 2 17.45

0.0002

Models sign. diff. — lower AIC (Colmation fixed model) No convergence of the model

Colmation fixed

3 5

−73.66 −82.27

−69.78 −75.79

39.83 46.13

1 vs. 2 12.60

0.0018

Models sign. diff. — lower AIC (Colmation fixed model)

Colmation random

10 −89.78

−76.82

54.89

2 vs. 3 17.51

0.0036

Models sign. diff. — lower AIC (Colmation random model)

−87.71 − 105.21 10 −99.11

−83.82 −98.73

46.85 57.60

1 vs. 2 21.50

b0.0001 Models sign. diff. –—AIC (Colmation fixed model)

−86.15

59.55

2 vs. 3 3.89

0.56

Models not sign. diff.

− 117.98 5 − 125.11 10 − 126.63

− 114.09 − 118.63 − 113.67

61.99 67.55

1 vs. 2 11.13

0.0038

Models sign. diff. — lower AIC (Colmation fixed model)

73.31

2 vs. 3 11.51

0.042

Models not sign. diff. after Bonferroni correction

3 5

3 5

3

Colmation fixed

Colmation fixed

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

References Angradi TR. Fine sediment and macroinvertebrate assemblages in Appalachian streams: a field experiment with biomonitoring applications. J N Am Benthol Soc 1999;18:49–66. Archaimbault V, Usseglio-Polatera P, Garric J, Wasson JG, Babut M. Assessing pollution of toxic sediment in streams using bio-ecological traits of benthic macroinvertebrates. Freshw Biol 2010;55:1430–46. Bjornn TC, Brusven MA, Molnau MP, Milligan JH, Chacho E, Schaye C. Transport of granitic sediment in streams and its effects on insects and fish. Rapport B-036-IDA, College of orestry, wildlife and range sciences; 1977. Böhmer J, Rawer-Jost C, Zenker A. Multimetric assessment of data provided by water managers from Germany: assessment of several different types of stressors with macrozoobenthos communities. Hydrobiologia 2004;516:215–28. Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH, et al. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol Evol 2009;24:127–35. Bou C, Rouch R. Un nouveau champ de recherches sur la faune aquatique souterraine. C R Acad Sci Paris 1967;265:369–70. Boulton AJ, Findlay S, Marmonier P, Stanley EH, Valett HM. The functional significance of the hyporheic zone in streams and rivers. Annu Rev Ecol Evol S 1998;29:59–81. Boulton AJ. Hyporheic rehabilitation in rivers: restoring vertical connectivity. Freshw Biol 2007;52:632–50. Boulton AJ, Fenwick GD, Hancock PJ, Harvey MS. Biodiversity, functional roles and ecosystem services of groundwater invertebrates. Invertebr Syst 2008;22:103–16. Bournaud M, Tachet H, Berly A, Cellot B. Importance of microhabitat characteristics in the macrobenthos microdistribution of a large river reach. Ann Limnol 1998;34: 83–98. Broekenhuizen N, Parkyn S, Miller D. Fine sediment effects on feeding and growth in the invertebrate grazers Potamopyrgus antipodarum (Gastropoda, Hydrobiidae) and Deleatidium sp. (Ephemeroptera, Leptophlebiidae). Hydrobiologia 2001;457: 125–32. Buffagni A, Erba S, Cazzola M, Kemp JL. The AQEM multimetric system for the southern Italian Apennines: assessing the impact of water quality and habitat degradation on pool macroinvertebrates in Mediterranean rivers. Hydrobiologia 2004;516: 313–29. Charvet S, Statzner B, Usseglio-Polatera P, Dumont B. Traits of benthic macroinvertebrates in semi-natural French streams: an initial application to biomonitoring in Europe. Freshw Biol 2000;43:277–96. Chevenet F, Dolédec S, Chessel D. A fuzzy coding approach for the analysis of long-term ecological data. Freshw Biol 1994;31:277–94. Collier KJ, Wright-Stow AE, Smith BJ. Trophic basis of production for a mayfly in a North Island, New Zealand, forest stream: contributions of benthic versus hyporheic habitats and implications for restoration. N Z J Mar Freshw 2004;38: 301–14. Dangles O, Guérold F, Usseglio-Polatera P. Role of transported particulate organic matter in the macroinvertebrate colonization of litter bags in streams. Freshw Biol 2001;46:575–86. Demars BO, Kemp JL, Friberg N, Usseglio-Polatera P, Harper DM. Linking biotopes to invertebrates in rivers: biological traits, taxonomic composition and diversity. Ecol Indic 2012;23:301–11. Descloux S, Datry T, Philippe M, Marmonier P. Comparison of different techniques to assess surface and subsurface streambed colmation with fine sediments. Int Rev Hydrobiol 2010;95:520–40. Descloux S, Datry T, Marmonier P. Benthic and hyporheic invertebrate assemblages along a gradient of increasing streambed colmation by fine sediment. Aquat Sci 2013. http://dx.doi.org/10.1007/s00027-013-0295-6. [in press]. Devin S, Beisel JN, Usseglio-Polatera P, Moreteau JC. Changes in functional biodiversity in an invaded freshwater ecosystem: the Moselle River. Hydrobiologia 2005;542: 113–20. Dolédec S, Chessel D. Rythmes saisonniers et composantes stationnelles en milieu aquatique. I. Description d'un plan d'observation complet par projection de variables. I. Acta Oecol-Oec Gen 1987;8:403–26. Dolédec S, Chessel D. Rythmes saisonniers et composantes stationnelles en milieu aquatique. II. Prise en compte et élimination d'effets dans un tableau faunistique. Acta Oecol-Oec Gen 1989;10:207–32. Dolédec S, Statzner B. Invertebrate traits for the biomonitoring of large European rivers: an assessment of specific types of human impact. Freshw Biol 2008;53: 617–34. Dolédec S, Olivier JM, Statzner B. Accurate description of the abundance of taxa and their biological traits in stream invertebrate communities: effects of taxonomic and spatial resolution. Arch Hydrobiol 2000;148:25–43. Dolédec S, Phillips N, Scarsbrook M, Riley RH, Townsend CR. Comparison of structural and functional approaches to determining landuse effects on grassland stream invertebrate communities. J N Am Benthol Soc 2006;25:44–60. Furse MT, Hering D, Moog O, Verdonschot PFM, Sandin L, Brabec K, et al. The STAR project: context, objectives and approaches. Hydrobiologia 2006;566:3–29. Gabriels W, Lock K, Pauw ND, Goethals PL. Multimetric Macroinvertebrate Index Flanders (MMIF) for biological assessment of rivers and lakes in Flanders (Belgium). Limnologica — Ecology and Management of Inland Waters 2010;40:199–207. Gayraud S, Philippe M. Does subsurface interstitial space influence general features and morphological traits of the benthic macroinvertebrate community in streams? Arch Hydrobiol 2001;151:667–86. Gayraud S, Philippe M. Influence of bed-sediment features on the interstitial habitat available for macroinvertebrates in 15 French streams. Int Rev Hydrobiol 2003;88:77–93.

275

Gayraud S, Statzner B, Bady P, Haybach A, Schöll F, Usseglio-Polatera P, et al. Invertebrate traits for the biomonitoring of European large rivers: an initial assessment of alternative metrics. Freshw Biol 2003;48:2045–64. Hancock P. Human impacts on the stream-groundwater exchange zone. Environ Manage 2002;29:761–81. Harding JP, Smith WA. A key to the British freshwater cyclopids and calanoids copepods. Freshwater Biological Association; 1974 [Scientific publication]. Hering D, Johnson RK, Kramm S, Schmutz S, Szoszkiewicz K, Verdonschot PFM. Assessment of European streams with diatoms, macrophytes, macroinvertebrates and fish: a comparative metric-based analysis of organism response to stress. Freshw Biol 2006;51:1757–85. Horrigan N, Baird DJ. Trait patterns of aquatic insects across gradients of flow-related factors: a multivariate analysis of Canadian national data. Can J Fish Aquat Sci 2008;65:670–80. James ABW, Dewson ZS, Death RG. Do stream macroinvertebrates use instream refugia in response to severe short-term flow reduction in New Zealand streams? Freshw Biol 2008;53:1316–34. Larsen S, Ormerod SJ. Low-level effects of inert sediments on temperate stream invertebrates. Freshw Biol 2010;55:476–86. Larsen S, Vaughan IP, Ormerod SJ. Scale-dependent effects of fine sediments on temperate headwater invertebrates. Freshw Biol 2009;54:203–19. Larsen S, Pace G, Ormerod SJ. Experimental effects of sediment deposition on the structure and function of macroinvertebrate assemblages in temperate streams. River Res Appl 2011;27:257–67. Lecerf A, Usseglio-Polatera P, Charcosset J-Y, Lambrigot D, Bracht B, Chauvet E. Assessing functional integrity of eutrophicated streams using direct and indirect approaches: litter breakdown versus macrobenthic assemblage taxonomic and trait structure. Arch Hydrobiol 2006;165:105–26. Lemly AD. Modification of benthic insect communities in streams: combined effects of sedimentation and nutrient enrichment. Hydrobiologia 1982;87:229–45. Lenat DR, Penrose DL, Eagleson KW. Biological evaluation of non-point source pollutants in North Carolina streams and rivers. Biological series n° 102. Raleigh, North Carolina: North Carolina Department of Natural Resources and Community Development, Division of Environmental Management; 1979. Lenat DR, Penrose DL, Eagleson KW. Variable effects of sediment addition on stream benthos. Hydrobiologia 1981;79:187–94. Lorenz A, Hering D, Feld CK, Rolauffs P. A new method for assessing the impact of hydromorphological degradation on the macroinvertebrate fauna of five German stream types. Hydrobiologia 2004;516:107–27. Maridet L, Philippe M. Influence of substrate characteristics on the vertical distribution of stream macroinvertebrates in the hyporheic zone, 91. Folia Faculty of Science Natural University of Masarykianae Brunensis; 1995101–5. Marmonier P, Vervier P, Gibert J, Dole-Olivier M-J. Biodiversity in ground waters. Trends Ecol Evol 1993;8:392–5. Matthaei CD, Weller F, Kelly DW, Townsend CR. Impacts of fine sediment addition to tussock, pasture, dairy and deer farming streams in New Zealand. Freshw Biol 2006;51:2154–72. McGill BJ, Enquist BJ, Weiher E, Westoby M. Rebuilding community ecology from functional traits. Trends Ecol Evol 2006;21:178–85. Menezes S, Baird DJ, Soares A. Beyond taxonomy: a review of macroinvertebrate trait-based community descriptors as tools for freshwater biomonitoring. J Appl Ecol 2010;47:711–9. Minshall GW. Aquatic insect-substratum relationships. In: Resh VM, Rosenberg DM, editors. Ecology of aquatic insects; 1984. p. 358–400. Mondy CP, Villeneuve B, Archaimbault V, Usseglio-Polatera P. A new macroinvertebratebased multimetric index (I2M2) to evaluate ecological quality of French wadeable streams fulfilling the WFD demands: A taxonomical and trait approach. Ecol Indic 2012;18:452–67. Öckinger E, Schweiger O, Crist TO, Debinski DM, Krauss J, Kuussaari M, et al. Life-history traits predict species responses to habitat area and isolation: a cross-continental synthesis. Ecol Lett 2010;13:969–79. Orghidan T. Ein neuer Lebensraum des unterirdischen Wassers, der hyporheische Biotop. Arch Hydrobiol 1959;55:392–414. Paillex A, Dolédec S, Castella E, Mérigoux S. Large river floodplain restoration: predicting species richness and trait responses to the restoration of hydrological connectivity. J Appl Ecol 2009;46:250–8. Petchey OL, Gaston KJ. Functional diversity: back to basics and looking forward. Ecol Lett 2006;9:741–58. Petts GE, Thoms MC, Brittan K, Atkin B. A freeze-coring technique applied to pollution by fine-grained sediments in gravel-bed rivers. Sci Total Environ 1989;84:259–72. Pinheiro J, Bates D, DebRoy S, Sarkar D. nlme: linear and nonlinear mixed effects models. Vienna, Austria: The R Foundation for Statistical Computing; 2009. Piscart C, Usseglio-Polatera P, Moreteau J-C, Beisel J-N. The role of salinity in the selection of biological traits of freshwater invertebrates. Arch Hydrobiol 2006;166: 185–98. Puig MA, Sabater F, Malo J. Benthic and hyporheic fauna of mayflies and stoneflies in the Ter River basin (NE-Spain). In: Campbell IC, editor. Mayflies and stoneflies. Kluwer Academic Publishers; 1990. p. 255–8. R Development Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2009. Relyea CD, Minshall GW, Danehy RJ. Stream insects as bioindicators of fine sediment, 24. Proceedings of the Water Environment Federation, Watershed; 2000. p. 663–86. Richards C, Bacon KL. Influence of fine sediment on macroinvertebrate colonization of surface and hyporheic stream substrates. Great Basin Nat 1994;54:106–13. Richards C, Haro RJ, Johnson LB, Host GE. Catchment and reach-scale properties as indicators of macroinvertebrate species traits. Freshw Biol 1997;37:219–30.

276

S. Descloux et al. / Science of the Total Environment 466–467 (2014) 265–276

Robertson AL, Wood PJ. Ecology of the hyporheic zone: origins, current knowledge and future directions. Fundam Appl Limnol 2010;176:279–89. Ryan RJ, Boufadel MC. Evaluation of streambed hydraulic conductivity heterogeneity in an urban watershed. Stoch Env Res Risk A 2007;21:309–16. Scarbrook MR, Fenwick GD, Duggan IC, Haase M. A guide to the groundwater invertebrates of New Zealand. NIWA Science and Technology Series; 2003. Southwood TRE. Habitat, the templet for ecological strategies. J Anim Ecol 1977;46: 337–65. Southwood TRE. Tactics, strategies and templets. Oikos 1988;52:3–18. Stanford JA, Ward JV. Revisiting the serial discontinuity concept. Regul River 2001;17: 303–10. Statzner B, Bêche LA. Can biological invertebrate traits resolve effects of multiple stressors on running water ecosystems? Freshw Biol 2010;55:80–119. Statzner B, Holm TS. Morphological adaptations of benthic invertebrates to stream flow: an old question studied by means of a new technique (Laser Doppler Anemometry). Oecologia 1982;53:290–2. Statzner B, Resh VH, Roux AL. The synthesis of long-term ecological research in the context of concurrently developed ecological theory: design of a research strategy for the Upper Rhône River and its floodplain. Freshw Biol 1994;31:253–63. Statzner B, Bis B, Dolédec S, Usseglio-Polatera P. Perspectives for biomonitoring at large spatial scales: a unified measure for the functional composition of invertebrate communities in European running waters. Basic Appl Ecol 2001;2: 73–85. Stearns SC. Life-history tactics: a review of the ideas. Q Rev Biol 1976;51:3–47. Stocker ZSJ, Williams DD. A freezing core method for describing the vertical distribution of sediments in a streambed. Limnol Oceanogr 1972;17:136–8. Strahler AN. Hypsometric (area-altitude) analysis of erosional topography. Bull Geol Soc Am 1952;63:1117–42. Tachet H, Richoux P, Bournaud M, Usseglio-Polatera P. Invertébrés d'Eau Douce, 2nd corrected impression. Paris: CNRS Editions; 2002. Thioulouse J, Dolédec S, Chessel D, Olivier JM. ADE-4: a multivariate analysis and graphical display software. Stat Comput 1997;7:75–83.

Tomanova S, Moya N, Oberdorff T. Unisg macroinvertebrate biological traits for assessing biotic integrity of neotropical streams. River Res Appl 2008;24: 1230–9. Townsend CR, Hildrew AG. Species traits in relation to a habitat template for river systems. Freshw Biol 1994;31:265–75. Townsend CR, Uhlmann SS, Matthaei CD. Individual and combined responses of stream ecosystems to multiple stressors. J Appl Ecol 2008;45:1810–9. Usseglio-Polatera P, Bournaud M, Richoux P, Tachet H. Biological and ecological traits of benthic freshwater macroinvertebrates: relationships and definition of groups with similar traits. Freshw Biol 2000a;43:175–205. Usseglio-Polatera P, Bournaud M, Richoux P, Tachet H. Biomonitoring through biological traits of benthic macroinvertebrates: how to use species traits database? Hydrobiologia 2000b;442/423:153–62. Usseglio-Polatera P, Richoux P, Bournaud M, Tachet H. A functional classification of benthic macroinvertebrates based on biological and ecological traits: application to river condition assessment and stream management. Arch Hydrobiol Suppl Monogr Stud 2001;139/1:53–83. Verberk WCEP, Siepel H, Esselink H. Life-history strategies in freshwater macroinvertebrates. Freshw Biol 2008a;53:1722–38. Verberk WCEP, Siepel H, Esselink H. Applying life-history strategies for freshwater macroinvertebrates to lentic waters. Freshw Biol 2008b;53:1739–53. Wantzen KM. Physical pollution: effects of gully erosion on benthic invertebrates in a tropical clear-water stream. Aquat Conserv Mar Freshwat Ecosyst 2006;16:733–49. Ward JV, Bretschko G, Brunke M, Danielopol D, Gibert J, Gonser T, et al. The boundaries of river systems: the metazoan perspective. Freshw Biol 1998;40:531–69. Waters TF. Sediment in streams: sources, biological effects and control. Betheda: American Fisheries Society; 1995. Weigelhofer G, Waringer J. Vertical distribution of benthic macroinvertebrates in riffles versus deep runs with differing contents of fine sediments. Int Rev Hydrobiol 2003;88:304–13. Zweig LD, Rabeni CF. Biomonitoring for deposited sediment using benthic invertebrates: a test on 4 Missouri streams. J N Am Benthol Soc 2001;20:643–57.