Macrophytes respond to reach-scale river restorations



1. In recent years, river restoration science has been searching for biological indicators of improvement in the physical habitats of streams. To date, research has mainly focused on the use of fish and macroinvertebrates as indicators. Despite their importance in aquatic ecosystems, the response of macrophytes to habitat restoration has been rarely studied.

2. We investigated the macrophyte communities of 40 restored river reaches in the lowland and lower mountainous areas of Germany. Each restored reach was compared to an upstream, unrestored reach using a space-for-time-substitution approach. At each reach, a 100 m stretch was surveyed for submerged and emergent macrophytes, recording the quantity, abundance and growth form of each species. Additionally, microhabitat patterns (substrate, depth, current velocity) and channel parameters (mean and bankfull width, number of channel elements) were recorded.

3. Restored reaches had a significantly higher macrophyte cover, richness, diversity and number of growth forms. Macrophyte diversity and richness were both positively correlated with depth, current and substrate.

4. The analysis of growth forms showed that Lemnids, Helodids, Parvopotamids, Elodids, Peplids and Juncids are all significant indicators of restoration. These species all responded directly to the restoration measures either by highly increasing in abundance or by being present in the restored reaches and absent in the unrestored reaches. While the restored reaches of the lowland rivers were characterized by a high abundance of Peplids and Parvopotamids, the restored reaches of the mountain rivers showed a significantly higher presence and abundance of Lemnids and Helodids.

5. Three macrophyte species (Lemna minor, Persicaria hydropiper, Potamogeton crispus) were regarded as significant indicators of restoration. No species were found to be indicators of unrestored reaches.

6. Synthesis and applications. Macrophyte communities benefit from river restoration by showing increased cover, abundance and diversity. The main drivers of this enhancement are more natural and diverse substrates and an increased floodplain area in the restored reaches, as well as a greater variability of current and depth patterns. Monitoring of macrophytes could thus be an easy and cost-effective means to gauge the success of river restoration measures.


River restoration is a central focus in stream management (Bernhardt et al. 2005; Palmer et al. 2005). Reduced faunal and floral biodiversity has raised awareness on anthropogenic degradation of streams and rivers. Furthermore, legislative frameworks, such as the EU-Water Framework Directive (WFD) and the habitat directive, have instituted restoration and protection programmes to counteract ongoing depletion and degradation of flowing waters. As the majority of streams in Europe fail to reach ‘good ecological status’ as defined by the WFD, restoration is urgently needed. In response, the number of restoration measures in place has increased greatly in recent years. Large amounts of money have been and will be spent in upcoming years on restoring streams and rivers (Bernhardt et al. 2005; Miller, Budy & Schmidt 2010). For example, the (German) federal state of North Rhine-Westphalia will spend 80 million Euros every year until 2027 on river restoration measures (

In Central and Northern Europe, a key target of river restoration measures is hydromorphological improvement (BMU 2005; Spänhoff & Arle 2007), and the physical habitat improvements caused by the restoration activities are apparent (e.g. significant increase in lentic and shallow areas and channel features like islands, backwaters or bars according to Jähnig, Lorenz & Hering 2008; Roni, Hanson & Beechie 2008; Jähnig et al. 2011). Nonetheless, restoration measures often only cover short stretches of rivers, i.e., they are reach-scale restorations. A holistic remediation of the catchment is seldom realizable and, therefore, not targeted. However, reach-scale river restoration measures often result in an increase in the diversity of substrates and number of channel features, such as islands, backwaters or bars, providing more possible niches for stream organisms than were available prior to restoration. Different reach-scale restoration measures have been implemented in different geographical regions. For example, lowland rivers in central Europe have been re-meandered (Lorenz, Jähnig & Hering 2009), whereas mountainous rivers in central Europe have been subject to re-braiding measures (Jähnig, Lorenz & Hering 2008). Similarly, boulders that had been previously removed for timber floating have now been replaced to enhance habitat diversity in boreal rivers of northern Sweden (Nilsson et al. 2005).

Despite improved habitat conditions, the response of the aquatic fauna to the reach-scale restoration measures is often minimal or absent (Roni, Hanson & Beechie 2008; Palmer, Menninger & Bernhardt 2010) and differs from expectations (e.g. increased biodiversity). Hence, it is difficult to draw a general picture (Palmer, Ambrose & Poff 1997). Moreover, outcomes have been quite contradictory, with some studies indicating that the diversity and richness of fish or macroinvertebrates did not significantly improve with improvements to habitat (e.g. Hamilton 1989; Lepori et al. 2005; Jähnig, Lorenz & Hering 2009; Sundermann et al. 2011) and other studies suggesting improvements in macroinvertebrate densities (Friberg et al. 1998) or fish diversity (Shields, Knight & Cooper 2007).

In most recent studies investigating river restoration measures, aquatic macrophytes have been largely neglected, despite their importance in stream habitats. Furthermore, the majority of restoration studies focussed on plants have looked at riparian vegetation (Jansson et al. 2005; Engström, Nilsson & Jansson 2009) rather than (aquatic) macrophytes. However, the few studies conducted in this field have reported a positive response of aquatic macrophytes to river restoration (Clarke & Wharton 2000; Pedersen, Baattrup-Pedersen & Madsen 2006). For example, planting native macrophytes in a restored New Zealand urban stream was followed by enhanced growth and dominance of the native plants over alien plants (Larned et al. 2006), although this technique has not been without difficulty (Suren 2009). In Danish lowland streams, Pedersen, Baattrup-Pedersen & Madsen (2006) showed that restored reaches had macrophyte communities similar to those of naturally meandering reaches, but their communities were different from those of channelized reaches.

Nonetheless, a general evaluation of macrophytes in restored reaches and the pairwise comparison of different restoration measures have so far not been performed. Furthermore, no studies have compared possible differences in the macrophyte communities inhabiting restored reaches in mountain streams and lowland streams. The differing morphological river patterns (e.g. gradient, substrate composition) of these two geographical regions might cause their respective macrophyte communities to react quite differently to restoration.

Hence, we aimed to investigate the effects of reach-scale river restoration measures on the macrophyte communities of German lowland and mountain rivers. We compared the abundance and diversity of macrophytes in restored reaches to those of degraded reaches to investigate the benefit of hydromorphological improvement on the community. Furthermore, we tried to analyse which hydromorphological improvements had the greatest effect on the macrophyte communities.

Moreover, we wanted to identify growth forms and species that might serve as indicators for evaluating restoration. Using macrophyte growth forms as an identification and interpretation tool for aquatic plant communities was first introduced by Mäkirinta (1978) and Den Hartog & Van der Velde (1988). In this system, macrophytes are separated into groups according to the form and structure of leaves and roots, as well as their aquatic status (emergent or submerged). Although species-based assessments are more commonly used, the use of growth forms to summarize different species could be a better alternative to identify the habitat structure of river reaches, as they might provide added value for the analysis and interpretation of aquatic plant communities (Cadotte, Carscadden & Mirotchnick 2011).

Materials and methods

We investigated 40 restored stream reaches in Germany (see Appendix S1, Supporting Information). Stream sizes varied between 9 and 2530 km2 in catchment area, except for the Rhine, with more than 152 000 km2. Sixteen reaches are located in the German lowlands (altitudes below 200 m above sea level), and 24 reaches are located in the lower mountainous areas (altitudes between 200 and 400 m a.s.l.). The reaches were subject to morphological restoration (see Appendix S1, Supporting Information) which target the ‘(morphological) natural state before human pressure’. This reference status (Leitbild) is defined in LUA NRW (1999, 2001). Guided by this reference, the physical restoration works focused on an improvement of the in-stream habitats on the reach-scale. The restoration measures did not concern the catchment of the rivers nor water quality issues. The lengths of the restored reaches ranged between 100 and 8000 m, and restoration was conducted between 1 and 13 years ago (mean 5 years). Each restored reach was compared to an unrestored reach in the same river, a few 100 m upstream. The unrestored reach was selected under the prerequisite that it was morphologically similar to the restored reach prior to restoration. Assessed by the German standard river habitat survey (LUA NRW 1998), the restored reaches are generally more natural than the respective unrestored reaches (Appendix S1, Supporting Information). This sampling design allows for a space-for-time-substitution, because no before-restoration data are available for these areas. In the Hase, Niers, Inde and Rur rivers, two restoration measures were investigated and compared to one unrestored reach upstream. According to German standards, the water quality of all sites was in a good status and no point sources exist between unrestored and restored paired reaches.

Macrophyte survey

Macrophyte sampling was conducted according to the German standard method (Schaumburg et al. 2005a,b) in summer 2007 and in some rivers in the summers of 2005 and 2008. Here, a 100-m reach was surveyed for macrophytes by wading through the river in transects and walking along the riverbank. In non-wadeable areas, the river bottom was raked with a rake (on a long pole or at the end of a rope) to reach the macrophytes. All macrophyte species were recorded and identified to the species level, except for Callitriche stands without fruits, which were only identified to genus level. The surveys included all submerged, free-floating, amphibious and emergent angiosperms, liverworts and mosses. In addition, plants were recorded which were attached or rooted in parts of the river bank that are likely to be submerged for more than 85% of the year. The abundance of each species was recorded according to the 5-point scale devised by Kohler (1978): 1 = very rare, 2 = rare, 3 = common, 4 = frequent, 5 = abundant, predominant. Therefore, the Kohler scale estimates the frequency of each species and takes into account the three-dimensional development of the plant stands. According to Kohler & Janauer (1997) and Schaumburg et al. (2004), the relationship between the five degrees of estimation and the actual quantity of the macrophytes can be described best by the function y = x3; accordingly, the values of the 5-point Kohler scale were x3-transformed into quantitative values. Additionally, the growth form of each species was recorded according to Den Hartog & Van der Velde (1988) and Wiegleb (1991; see Appendix S2, Supporting Information). The growth forms comprise different plant species that realized the same or comparable phenotypical adaptations to the aquatic environment.

Hydromorphological survey

In each reach, a hydromorphological survey was conducted on 10 transects. The transects were 10 m apart in small streams (catchment size > 10 km2) and 20 m apart in medium (catchment size > 100 km2) and large (catchment size > 1000 km2) rivers. In each transect, the number of channel elements (side arms, main channel, bars, islands) was recorded. The width of the aquatic areas (main channel, side arms, backwaters) and the bankfull width, i.e., the width of the floodplain area, were also measured. In the aquatic area of each transect, the substrate, depth and current velocity (in six classes) were recorded at 10 equidistant points. In the river Rhine, the restoration measure was a rehabilitation of several groynes. Therefore, the measurements here were restricted to the groyne areas and compared to conventional groynes. In the Inde, Niers_2, Rur and Wurm rivers, only macrophyte surveys and the general measurements (width, bankfull width and number of stream elements) were conducted on the transects.

Data analysis

Richness (expressed as number of taxa), abundance (sum of the Kohler scale), area covered (m3), diversity (Shannon–Wiener Index) and the total number of growth forms were calculated for each reach and compared using the Wilcoxon test between the restored and unrestored reaches, first for the whole data set and then for the mountain and lowland sites separately. The second analysis was performed to exclude the possibility of contradictory responses (of macrophytes in restored reaches) in the two geographical regions. Furthermore, comparisons were restricted to: (i) only submerged macrophytes and (ii) to submerged and emergent macrophytes together. In addition, Spearman rank correlations were used to test for a possible influence of time on restoration response. Therefore, the number of years since the completion of the restoration measures was correlated with the effects of restoration and tested with the Wilcoxon test.

To detect growth forms and species significantly indicative for either restored or unrestored reaches, an indicator species analysis was conducted (Dufrene & Legendre 1997) using PC-ORD (McCune & Mefford 2006). The analyses on species level concentrated on species present in at least 25% of the reaches (n = 10).

The hydromorphological data were summarized for restored and unrestored reaches. For the river width, the ratio of the width of restored reaches to that of unrestored reaches was calculated. The coefficient of variation was calculated from the depths and current velocities of each transect and reach to describe the degree of naturalness of the reaches, with higher values representing a near-natural state. The Shannon–Wiener Index was calculated from the channel elements (bars, islands, main channel, side arms) to express feature diversity on the mesohabitat scale. Moreover, the Spatial Diversity Index (Fortin, Payette & Marineau 1999) was calculated from the substrate data to express the feature diversity on a microhabitat scale. These parameters were selected because they are major determinants of high macrophyte diversity (Baattrup-Pedersen & Riis 1999).

Finally, the change of the macrophyte metrics (restored minus unrestored) was correlated with the change in the hydromorphological parameter. With this analysis, the main parameters were identified, which produced a significant response by the macrophyte communities.


Richness, diversity and abundance

In total, 86 species (and genera) were identified belonging to 15 different growth forms. The most common species Phalaris arundinacea L. (a Helodid) was present in 35 reaches (46·1% of all reaches). The highest number of species (23) was found in the restored reach of Hase_2. These 23 species comprised a total of 10 different growth forms. The average number of taxa was 4·4 in the unrestored reaches and 9·1 in the restored reaches. In four unrestored reaches, no species could be recorded.

In general, the macrophyte quantity, number of taxa, number of growth forms, the sum of the Kohler scale and Shannon–Wiener diversity were all significantly higher in restored reaches than in unrestored reaches (not shown). This pattern holds true for the comparison of the combined group of submerged and emergent macrophytes, as well as for the submerged macrophytes alone and also if lowland and mountain river sites are considered separately (Fig. 1), except for the quantity of submerged macrophytes in lowland rivers.

Figure 1.

 Box–Whisker plots for differences in restored (re) and unrestored (un) mountain and lowland reaches in macrophyte quantity (a, e), richness (b, f), number of growth forms (c, g) and sum of Kohler scale (d, h). Left panel (a–d) submerged and emergent macrophytes, right panel (e–h) only submerged macrophytes. sm, submerged macrophytes; em, emergent macrophytes, *< 0·05, **< 0·01, ***< 0·001.

The Spearman rank order correlation analysis between the time since restoration and the increase in metric values (difference in results of the parameters between restored reach and unrestored reach of each site) showed no significant results in any parameter.

Indicators for restored reaches

The Lemnid, Helodid, Parvopotamid, Elodid, Peplid and Juncid growth forms were all significant indicators of restoration (Table 1). They all exhibited higher abundances or were present only in the restored reaches. Haptophytes and Nymphaeids were more uniformly continuous, albeit not significantly so, in unrestored reaches compared to restored reaches (= 0·63 and 0·58, respectively). Helodids and Lemnids were significant indicators of restoration in the mountain rivers, whereas Parvopotamids and Peplids were indicators in lowland rivers. In contrast, no growth form showed a significant increase in presence in the unrestored reaches.

Table 1.  Significant indicator growth forms and species of an indicator species analysis of restored vs. unrestored reaches from all reaches at mountain (n = 24) and lowland (n = 16) sites
Growth form or speciesAll sites (n = 40)Mountain sites (n = 24)Lowland sites (n = 16)
Indicator for group P Indicator for group P Indicator for group P
  1. NS, not significant; *< 0·05; **< 0·01.

 Elodea canadensis RestoredNSRestoredNSRestored*
 Persicaria hydropiper Restored**Restored**RestoredNS
 Lemna minor Restored*Restored*RestoredNS
 Potamogeton crispus Restored**RestoredNSRestored*

Compared to results from the consideration of growth forms, a lower number of species served as significant indicators of restoration. Only Lemna minor L., Persicaria hydropiper (L.) and Potamogeton crispus L. occurred in significantly higher abundances in restored reaches compared to unrestored reaches (Table 1). However, no species was significantly more abundant in unrestored reaches than in restored reaches. Whereas Lemna minor and Persicaria hydropiper were significantly more abundant in mountain sites, Elodea canadensis Michaux and Potamogeton crispus were significant indicators for restored reaches in lowland sites. All other species showed no significant value as indicators.

Hydromorphological determinants

The channel width of nearly all restored reaches was greater than that of the respective unrestored reaches (Fig. 2). In the Kimmer Brookbäke, the average stream width was more than three times the width of the unrestored reach. In five rivers, the average width did not increase, but individual transects of the restored reaches were much wider than the average of those in the unrestored reach.

Figure 2.

 Mean (bars) and maximum (whiskers) change in river width (expressed as the ratio of restored/unrestored). If width increased on average, values >1 were obtained; if value = 2, average stream width doubled. Asterisks indicate significantly different river width values between unrestored and restored reaches (Wilcoxon test, < 0·05).

The restored reaches showed a clear tendency towards higher substrate diversity (SDI_substrate), depth diversity (cv_depth) and richness of stream elements (SWI_channel elements) (Fig. 3). The diversity in flow patterns was, in general, higher in the restored reaches, but this difference was only significant in some cases. The unrestored reaches were mainly associated with low diversities in substrate, depth, stream elements and current velocity. Specified by the two ecoregions, the differences between restored and unrestored reaches showed similar significant results for most of the hydromorphological parameters (Table 2). Total hardness expressed no significant differences between reaches.

Figure 3.

 Comparison of unrestored (X-axis) and restored metric values (Y-axis). Points above the dashed line have a higher value in the restored reach. Dark points indicate significantly different metric values between unrestored and restored reaches (Wilcoxon test, < 0·05).

Table 2.  Mean morphological parameters and ranges for the restored and the unrestored reaches of the lowland and mountain sites
ParameterMountain sitesLowland sites
Mean (range)cvMean (range)cvMean (range)cvMean (range)cv
  1. Asterisks indicate significant differences (Wilcoxon test, *< 0·05, **< 0·01).

Channel width13·0 (1·6–31·9)0·1017·5 (1·7–50·7)**0·24**10·7 (0·7–25·0)0·0414·5 (2·4–33·5)*0·20**
Bankfull width18·3 (4·7–37·0)0·1235·1 (4·4–89·7)**0·25**13·0 (2·1–26·8)0·0630·6 (3·2–63·8)**0·23**
Current velocity0·320·45**0·070·11NS
Substrate diversity0·29 (0·0–0·6)0·41 (0·1–0·7)*0·19 (0·0–0·3)0·33 (0·1–0·5)*
Channel element diversity0·61 (0·0–1·5)1·44 (0·5–2·0)**0·28 (0·0–1·2)1·09 (0·0–1·9)**
Total hardness2·3 (1·1–7·0)2·4 (1·2–6·8)NS1·9 (0·4–3·2)2·0 (0·8–3·1)NS

Correlations between the change of the hydromorphological determinants and the change of the macrophyte metrics showed that increased substrate diversity had highly significant positive effects on all macrophyte metrics (Table 3). This is also true if only the mountain sites are analysed. In the lowland sites, the number of growth forms and the Shannon–Wiener diversity responded significantly. Additionally, depth variability in the lowlands had a significant influence on the area covered by macrophytes. The second morphological parameter that led to significant responses in most of the macrophyte measures was the increase in floodplain area. The variability of current velocity was important for the number of growth forms, particularly in the mountain sites. Also, increased depth variability led to an increase in richness, abundance and area covered by submerged taxa in the lowland sites. Furthermore, increased channel width was followed by positive responses in submerged macrophyte abundance, area covered and diversity. The number of growth forms showed more significant responses than the number of taxa.

Table 3.  Correlation coefficients of change (restored minus unrestored) in macrophyte metrics to change (restored minus unrestored) in hydromorphological parameters
 Channel widthBankfull widthCv current velocityCv depthSubstrate diversityChannel element diversity
All (N=38)M (N=24)L (N=14)All (N=38)M (N=24)L (N=14)All (N=32)M (N=22)L (N=10)All (N=31)M (N=22)L (N=9)All (N=31)M (N=22)L (N=9)All (N=39)M (N=24)L (N=15)
  1. Bold figures significant with *< 0·05, **< 0·01.

  2. M, mountain; L, lowland; sm, submerged macrophytes; em, emergent macrophytes.

Area covered sm and em0·280·330·100·140·090·32−0·09−0·09−0·180·220·08 0·70* 0·44* 0·44* 0·43−0·11−0·04−0·31
Area covered sm 0·34* 0·50** −0·020·230·300·11−0·020·03−0·25 0·38* 0·21 0·72* 0·40* 0·43* 0·31−0·13−0·01−0·31
No. of taxa sm and em0·260·230·33 0·41* 0·350·520·290·370·130·330·240·49 0·55** 0·60** 0·45−0·030·21−0·36
No. of taxa sm0·290·220·46 0·50** 0·59** 0·36 0·36* 0·50* −0·30 0·39* 0·18 0·72* 0·40* 0·49* 0·090·140·21−0·07
No. of growth forms sm and em0·290·160·52 0·52** 0·49* 0·55* 0·45** 0·56** 0·200·240·300·16 0·54** 0·53* 0·77* 0·040·17−0·12
No. of growth forms sm0·320·22 0·54* 0·52** 0·49* 0·54* 0·39* 0·51* −0·230·220·210·24 0·55** 0·56** 0·620·120·190·03
Sum of Kohler scale sm and em0·300·300·32 0·35* 0·260·520·140·180·070·280·160·56 0·54** 0·57** 0·45−0·070·09−0·33
Sum of Kohler scale sm 0·36* 0·400·33 0·46** 0·56** 0·320·200·34−0·21 0·41* 0·18 0·74* 0·45* 0·56** 0·19−0·010·11−0·15
Shannon–Wiener sm and em 0·32* 0·280·45 0·59** 0·58** 0·62* 0·30 0·53* 0·070·270·290·27 0·52** 0·61** 0·73* −0·010·19−0·18
Shannon–Wiener sm0·280·190·50 0·46** 0·49* 0·440·220·33−0·080·160·120·22 0·56** 0·57** 0·76* 0·100·140·07


Richness, diversity and abundance

In all the biological parameters tested, the restored reaches showed clear improvements compared to the unrestored reaches, irrespective of stream type, size or time since restoration. Macrophyte quantity (i.e. the area covered by macrophytes) nearly doubled in the restored reaches because of the wider and shallower stream channels. These wide channels are less shaded, thus enhancing macrophyte growth. Macrophyte richness and diversity both nearly doubled in the restored reaches, probably due to the higher variety in niches provided by the more diverse flow and depth patterns, as opposed to the deep, uniformly flowing character of the unrestored reaches. The higher habitat heterogeneity of the restored reaches was also mirrored in the significant increase in growth forms, as each growth form requires different habitat combinations of substrate, current and depth (Den Hartog & Van der Velde 1988). Observed differences between restored and unrestored reaches were even larger when submerged and emergent macrophytes were combined. Emergent macrophytes (Helodids) are generally present in semi-aquatic areas, a habitat which is much more abundant in restored reaches, as highlighted in several studies (e.g. Jähnig, Lorenz & Hering 2008). Higher species richness and diversity were also found by Baattrup-Pedersen & Riis (1999) in unregulated Danish lowland streams. However, in their study, total macrophyte cover did not differ between sites, possibly because of the site selection and the fact that all streams were unshaded.

On the contrary, Baattrup-Pedersen et al. (2000) found no change in the species richness or recovery of submerged vegetation cover to pre-restoration levels in their study of the short-term effects (2 years) of a river restoration. In contrast, the authors noted a shift in plant dominance from non-riparian species on the riverbanks to riparian gramineous species, such as Juncus effusus L. This species requires higher soil moisture content and thus suggests an increase in semi-aquatic habitats, similar to that observed in our study.

Our analyses also showed significant differences in richness and diversity between restored and unrestored reaches. This result was surprising when considering the small distance between restored and unrestored reaches (mean distance = 500 m). Species easily colonized and became established and proliferated in these areas after restoration, despite being only a few metres downstream from where they were not found. Many of these new species are generally common in Germany [e.g. Elodea nuttallii (Planch.) or Potamogeton crispus], but nonetheless they increase key factors with which restoration success is evaluated: diversity and ecological functioning, in this case defined as the diversity of macrophyte growth forms. The significant increase in the diversity of growth forms indicates a higher ecological functioning, where each individual growth form represents a potential habitat for benthic invertebrates (e.g. Collier, Champion & Croker 1999; Shupryt & Stelzer 2009) and different life stages of fish, or a colonization platform for epiphytic algae or diatoms (Den Hartog & Van der Velde 1988; Cheruvelil et al. 2002). Thus, in-stream habitat restoration not only benefits macrophytes directly, but also indirectly aids in increasing site diversity and provides a variety of habitats for other taxonomic groups.

In our study, no discernable effect of time on the number of taxa, the area covered, the diversity or any other parameter tested was observed, despite the sometimes short time (1 year) since restoration. Thus, pioneer effects might be negligible. Furthermore, neither the number of taxa nor the diversity of macrophytes in the old restored reaches appeared to be decreasing; thus, any adverse effects of sedimentation or succession were not discernable.

Indicators for restored reaches

Growth forms are likely to be a better indicator of restoration than taxon richness (i.e. species number). First, measures of growth forms represent a more stable element in the community (Wiegleb 1984, 1988). Our study reveals that individual species are not very common throughout the individual rivers and are replaced in the next river by different species of the same growth form. Thus, the number of species that significantly indicate a restored reach is low compared to the number of (significant) growth forms. Secondly, the identification of growth forms is significantly easier and more cost-effective than species identification and is, therefore, more practical for future routine monitoring.

In our study, Lemnids, Helodids, Parvopotamids, Elodeids, Peplids and Juncids directly responded to restoration measures by either increasing in abundance or simply by being present in the restored reaches and absent in the unrestored reaches. The increased diversity of growth forms results from: (i) increased areas of slower or no current velocities, (ii) increased areas of shallow water characterized by frequently altering water levels and sufficient exposure of the river bed and (iii) a higher number of valuable substrate types, e.g., sand and mud in combination with (i) and (ii) (Remy 1993). Helodids (e.g. Veronica beccabunga L., Alisma plantago-aquatica L., Sparganium erectum L.) increased in abundance and colonized the re-profiled, wider, smoother banks of the restored reaches. Lemnids (Spirodela polyrhiza (L.), Lemna spec.), Parvopotamids (Potamogeton berchtoldii Fieber) and Juncids (Juncus spec.) colonized the backwater areas created by the restoration efforts. They benefit from areas of low to no flow and nutrient enrichment (Willby & Eaton 1996). Furthermore, Elodeids (Elodea canadensis, E. nuttallii) benefit from areas of clear, shallow and slow flowing water in the restored reaches (Remy 1993). These colonization patterns were suggested by Nilsson et al. (2005), who predicted that ‘restoration may allow helophytes (i.e. Helodids) and some Elodeids to increase in abundance’. Our findings, as well as the increase in diversity and abundance of helophytes also found by others (e.g. Clarke & Wharton 2000; Pedersen, Baattrup-Pedersen & Madsen 2006), clearly support this hypothesis. In conclusion, the growth forms of lentic and shallow habitats benefit most from restoration, as these areas are significantly increased in restored reaches (Jähnig et al. 2009).

Baattrup-Pedersen & Riis (1999) identified Sparganium emersum Rehmann (a Vallisnerid) as the most abundant species in their regulated streams and Ranunculus peltatus Schrank (a Myriophyllid) as the most abundant species in unregulated streams. This contrasts with results from our study, where neither Sparganium emersum nor Ranunculus peltatus showed any preference for restored reaches. In fact, R. peltatus was not found at all. But, the Myriophyllids, the respective growth form of R. peltatus, are close to being a significant indicator growth form (P = 0·08) for restored reaches. Thus, species may change between regions, but growth forms are more consistent, and certain growth forms indicate physical habitat improvements, e.g., by restoration measures.

Hydromorphological determinants

Restored and unrestored (i.e. degraded) reaches are associated with different hydromorphological features. In unrestored reaches, habitat uniformity appears to directly influence the establishment and number of macrophyte growth forms. Our hydromorphological analysis indicates that depth, velocity and substrate variability, as well as stream element diversity, i.e., mesohabitat quality, are all higher in restored reaches. Increased channel width in the restored reaches results in higher substrate and stream element diversity, increasing the possibility for macrophytes to become established in a site by providing higher habitat heterogeneity. The importance of bank and river bed morphology as key factors for increased diversity and abundance was found by Pedersen, Baattrup-Pedersen & Madsen (2006), although mainly for terrestrial species that migrate through lower banks and shallow channels into the river channel. A similar result was found by Clarke & Wharton (2000), who showed that restored reaches had significantly higher wetland species diversity, bank width, and soil moisture, as well as significantly lower bank angles. Therefore, connectivity between the river and its banks seems to enhance macrophyte diversity. This was true in most of our restored reaches. Bank profiles were rebuilt and secondary channels were initiated in many cases. The creation of islands and side arms nearly doubles the possible area of connectivity between the rivers and their floodplains. The high correlations between the increase in bankfull width in our study and the significant increases in macrophyte richness (taxa and growth forms) as well as in abundance and diversity of macrophytes provide evidence for the success of the restoration measures. Additionally, substrate diversity is enhanced because different substrates aggregate according to flow parameters. Substrate diversity is shown to be the overarching factor, enhancing significantly all macrophyte metrics. Baattrup-Pedersen & Riis (1999) also found that substrate heterogeneity is positively correlated with macrophyte coverage and diversity. In the uniform, unrestored reaches, the habitat heterogeneity was low (Sear, Briggs & Brookes 1998). Furthermore, within in the lowland sites, the variability in depth had the highest influence on the quantity, richness and abundance of macrophytes, whereas in the restored mountain reaches, the increase in the variability of current velocity was more important. In degraded mountain rivers, the flow patterns are uniform which limit macrophyte richness and diversity, while lowland rivers naturally have a low current velocity diversity, and the limiting factor for macrophytes in degraded reaches are straightened and steep banks.

Colonizing potential

The (re-) colonizing potential of macrophytes is greatly underestimated in comparison with the much more mobile benthic invertebrates and fish, which are known to be good colonizers. Macrophytes can display three contrasting modes of colonization (catchment delivery, seed bank and migratory animals/birds), whereas only catchment delivery is relevant for fish and macroinvertebrates (Figuerola & Green 2002; Green, Figuerola & Sánchez 2002; Jensen, Walker & Paton 2008).

The drift of propagules from the catchment is the main mode of colonization (Riis 2008). Drifting propagules are more likely to be retained in shallow and slow flowing river reaches than in uniformly fast flowing reaches (Riis & Sand-Jensen 2006). Therefore, restoration of channel complexity can enhance the retention of plant propagules (Engström, Nilsson & Jansson 2009). This was probably the case in most of the restored reaches, where obstacles in the larger and wider channels retained propagules. The high abundance and indicator function of Helodids and the backwater species (Lemnids) are good examples of species drifting there and being retained in these restored areas. Thus, in terms of the macrophytes, indicators for successful restoration had been recorded in the majority of the studied restoration measures.

Conclusions and implications

River restoration supports and enhances the diversity and abundance of macrophytes in lowland and mountain rivers even over short distances (a few 100 m). Therefore, macrophytes are potentially well suited for the evaluation of reach-scale river restoration measures, particularly when artificially narrowed river sections are later widened and more natural channel and floodplain patterns are established. Nonetheless, local physical restoration measures are subject to catchment influences like nutrient loading, which enhance macrophyte growth. Even if more natural substrate and flow pattern in restored reaches provide habitats for a more abundant and diverse macrophyte community, catchment remediation is additionally necessary to establish reference communities. Increased macrophyte cover and diversity subsequently increase the diversity of food sources, shelter or habitats in general for other organisms, including fishes and invertebrates. Furthermore, increased macrophyte growth engineers the restored reaches even further by accumulating sediment and organic matter or by re-directing flow.

Macrophytes respond relatively quickly to habitat improvement because of the increase in diversity of hydromorphological niches. Diversity is enhanced and abundance increases, but in some sites, the low number of species could be misleading. However, the use of growth forms to evaluate the effects of restoration emphasises the differences between sites and improves data interpretation possibilities. In this context, growth form communities have been proposed as tools to assess habitat degradation in different river types in Europe (LANUV NRW 2008; Gurnell et al. 2010).

Finally, from a practical perspective, it is important to note that standardized sampling and the identification of macrophytes is more easily performed and is more cost- and time-effective than fish or macroinvertebrate monitoring. According to Birk (2003), who reviewed European assessment methods, the costs of surveying macrophytes are, on average, half those of surveying fish and macroinvertebrates and can be carried out in a quarter of the time.


We thank three anonymous reviewers for their comments and suggestions, which improved the manuscript. This work was financially supported by Deutsche Bundesstiftung Umwelt (FK 25032-33/2), Hesse’s Ministry of Environment (FK III 2-79i 02) and by the research funding programme ‘LOEWE – Landes-Offensive zur Entwicklung Wissenschaftlich-ökonomischer Exzellenz’ of Hesse’s Ministry of Higher Education, Research and the Arts.