Accepted: 20 February 2018 DOI: 10.1111/fwb.13101 ORIGINAL ARTICLE Evaluating effects of weed cutting on water level and ecological status in Danish lowland streams Annette Baattrup-Pedersen1 | Niels B. Ovesen1 | Søren E. Larsen1 | Dagmar K. Andersen1 | Tenna Riis2 | Brian Kronvang1 | Jes J. Rasmussen1 1 Department of Bioscience, Aarhus University, Silkeborg, Denmark 2 Department of Bioscience, Aarhus University, Aarhus C, Denmark Correspondence Annette Baattrup-Pedersen, Department of Bioscience, Aarhus University, Silkeborg, Denmark. Email: abp@bios.au.dk Abstract 1. At present, scientific evidence documenting effects of weed cutting in streams as a measure to improve flood protection and run-off from agricultural land is scarce, which is surprising considering the huge effect that it has on stream ecology. Instead, weed cutting is performed under the assumption that removal of aquatic plant biomass improves runoff from agricultural land and prevents flooding of adjacent areas provided that it is performed regularly. 2. In this study, we examined linkages between weed cutting practice and water level reductions in 126 small- and medium-sized Danish streams (catchment size <100 km2) with continuously monitored discharge and water level data (from 1990 to 2012). Specifically, we hypothesised that (1) weed cutting reduces stream water levels more in late summer when the biomass of aquatic plants is higher than in early summer; (2) the efficiency of cutting declines with increasing cutting frequency as the aquatic plant community changes with increasing abundance of species able to regrow fast following a cutting event; (3) the high-frequency cutting in Danish streams lowers the ecological status of the streams as evaluated from aquatic plant assemblages. 3. The average effect of weed cutting on the water level was largest in July, August and September with an average reduction of 16 cm and lowest in early spring and late autumn with an average reduction of 11 cm. Regrowth was largest in June, with an increase in water level of 0.41 cm/day, whereas regrowth was absent in autumn. Regrowth also varied with the frequency of weed cutting, from an average of 0.04 cm/day in streams subjected to one annual cutting to an average of 0.6 cm/day in streams subjected to >6 annual cuttings. Furthermore, we found that the ecological status was either moderate or poor/bad in streams with more than one annual cutting. 4. Our findings highlight that it is by no means certain that the current weed cutting practice is efficient for flood control since (1) regrowth is stimulated by frequent cuttings and a positive feedback loop may develop, necessitating even more frequent cuttings to maintain the discharge capacity of the streams, and (2) many species stimulated by weed cutting, like for instance Sparganium emersum, form dense canopy beds across the entire stream profile and therefore reduce the discharge capacity of the stream more than species growing in confined Freshwater Biology. 2018;1–10. wileyonlinelibrary.com/journal/fwb © 2018 John Wiley & Sons Ltd | 1 2 | BAATTRUP-PEDERSEN ET AL. patches. We encourage more studies with the aim to identify how stream maintenance should be performed to optimise flood control without compromising the ability to reach good ecological stream quality. KEYWORDS aquatic plants, macrophytes, maintenance, management, river 1 | INTRODUCTION scarce, though (but see e.g. Old et al., 2014). This is surprising considering that maintenance has a tremendous effect on the ecology Aquatic plants are central elements of many small and shallow low- of the streams (Armitage, Blackburn, Winder, & Right, 1994; Baat- land streams today and may also in former times under conditions € the, Riis, & O’Hare, 2016; Baattrup-Pedersen trup-Pedersen, Go relatively undisturbed by human settlement have played a key role et al., 2003; Dawson, Clinton, & Ladle, 1991; Kaenel & Uehlinger, for these ecosystems (Svenning, 2002). Aquatic plants regulate and 1999). Instead, maintenance is performed under the assumption that modify important in-stream habitat characteristics (e.g. substrate removal of aquatic plant biomass improves flood protection and run- composition and retention of fine particulate organic matter) and off from agricultural land if applied regularly (Bach et al., 2016; Old hereby the structural and functional characteristics of the biological et al., 2014; Vogelsang, 2016). stream communities (Caffrey, 1991; Carpenter & Lodge, 1986; Gur- In Denmark, plant growth is highly seasonal and begins in early nell, 2014; Sand-Jensen, 1997). For example, aquatic plants usually spring when light and temperature conditions increase to levels suit- have a stimulating effects on macroinvertebrate abundance and spe- able for photosynthesis. Subsequently, the accumulation of plant cies richness (Ferreiro, Feijoo, Giorgi, & Leggieri, 2011; Iversen, biomass gradually declines as the availability of light becomes the 1988), which may further support dense and productive populations limiting resource (Sand-Jensen, 1997) and die back occurs for most of fish such as brown trout (Salmo trutta; Madsen, 1995; Mortensen, species during autumn. Maintenance is performed regularly in a 1977; Teixeira-de Mello et al., 2015). majority of streams, but over the last three decades, the frequency Aquatic plants also influence the hydraulic conditions in streams of weed cutting has increased, from typically 1–2 to 2–3 times per by increasing the resistance against flow and, hence, the water level year, and the frequency is even higher in many streams due to within the channel (Curran & Hession, 2013; Sand-Jensen et al., extensive regrowth. At the same time, stream maintenance has been 1989). The spatial variability of aquatic plant growth within the increasingly confined to the central parts of the stream channel, channel makes it difficult to predict the resistance from coverage which is supposed to be more environmentally friendly (Baattrup- alone, however, as it may change with the distribution of patches Pedersen, Skriver, & Wiberg-Larsen, 2000). However, it has never and patch architecture (Folkard, 2011; Green, 2005). Additionally, at been systematically explored to what extent the higher cutting fre- high flow plant configuration may change radically (e.g. O’Hare, quency has improved flood protection or run-off from agricultural Hutchinson, & Clarke, 2007; Sand-Jensen, 2003), making it even land, nor has it been explored how it affects the ecological quality of more difficult to predict their resistance. Flow resistance associated the streams also in terms of downstream reaches that are likely to with aquatic plants is the main reason for active maintenance of receive increased loads of eroded sediments and associated chemi- aquatic plant populations in streams (Dawson & Robinson, 1984; cals (Old et al., 2014). It can be argued that an increase in weed cut- Naden, Rameshwaran, Mountford, & Robertson, 2006; Pitlo & Daw- ting frequency reduces the efficiency of the cutting because son, 1990), and maintenance is performed regularly in many lowland improved light conditions stimulate the regrowth of aquatic plants streams to increase the discharge capacity (e.g. Baattrup-Pedersen and, furthermore, that plant decay may be delayed due to renewal et al., 2009; Caffrey, 1993; Fox & Murphy, 1990; Kaenel & Uehlin- of the biomass (Dawson, 1978a, 1978b). Additionally, weed cutting ger, 1999; Vereecken, Baetens, Viaene, Mostaert, & Meire, 2006; may act as a strong species-specific environmental filter generating € ring, Filetti, Brux, & Herr, 2014). Maintenance methods Wiegleb, Bro cascading effects on the composition of aquatic plant communities include simple weed cutting involving removal of aquatic plant bio- (Baattrup-Pedersen & Riis, 1999, 2004; Baattrup-Pedersen et al., mass or removal of selected species using a scythe or boat mounted 2003; Pedersen, Baattrup-Pedersen, & Madsen, 2006; Sabbatini & with knives (Baattrup-Pedersen, Larsen, & Riis, 2003; Old et al., Murphy, 1996b) and potentially compromise the ability to reach 2014; Schwarz & Snelder, 1999) and dredging of the total macro- good ecological status as meeting this goal requires that all ecologi- phyte biomass and associated accumulated sediments using an exca- cal quality elements are in at least good ecological status (the one- vator (Caffrey, 1993; Sabbatini & Murphy, 1996a, 1996b). out-all-out principle; European Commission, 2000). Scientific evidence documenting the effects of stream mainte- In this study, we examined 3,000 weed cuttings performed in nance as a measure to maintain the discharge capacity (Dawson, small- and medium-sized Danish streams (catchment size <100 km2) 1978a, 1978b; Gurnell & Midgley, 1994; Pitlo & Dawson, 1990) is to explore to what extent the stream water level declines following BAATTRUP-PEDERSEN | ET AL. cutting and how this response is influenced by the timing and frequency of weed cutting. Specifically, we hypothesised that (1) weed 3 2.2 | Hydraulic calculations cutting reduces stream water levels more in late summer when the The fundamental problem encountered when attempting to estimate biomass of aquatic plants is higher than in early summer; (2) the effi- discharge from stage at a stream gauging station at a site with ciency of cutting declines with increasing cutting frequency as the downstream growth of macrophytes is that the resistance of the aquatic plant community changes with an increasing abundance of flow is subject to variations with increasing resistance during the species able to regrow fast following a cutting event; (3) the high- spring and summer season (Dawson, 1978a, 1978b). The increase in frequency cutting in Danish streams lowers the ecological status of resistance might be counteracted by stream maintenance involving the streams as evaluated from aquatic plant assemblages applying cutting of the macrophytes. We have utilised this normally seen pat- the Danish Stream Plant Index (DSPI; Baattrup-Pedersen, Larsen, & tern in Danish lowland streams to analyse each stream hydraulically Riis, 2013; Baattrup-Pedersen et al. 2017) that is currently used to to register: (1) time of cutting, (2) water level drop associated with assess the ecological status of the aquatic plant community in com- cutting and (3) water level increases resulting from vegetation pliance with the Water Framework Directive. regrowth following cutting (Figure 2). Base discharge (Q)–stage (H) relationships were established for 2 | METHODS 2.1 | Study sites each stream gauging station based on monthly measurements of discharge and water stage utilising only measurements of discharge and stage during the winter and early spring period where the hydraulic influence of growing macrophytes in the downstream channel is A total of 126 Danish stream reaches with continuously monitored absent or very small (Figure 2a; Herschy, 1995). Discharge was mea- discharge and water level data for at least 10 years within the period sured following the guidelines in ISO-standard ISO748 according to 1990–2012 were included in this study (Figure 1; Ovesen et al., number of depth transects established and point measurements of 2000). The streams were all small- to medium-sized and distributed current velocity utilising OTT C31, OTT C2. However, Doppler throughout Denmark (Figure 1). A total of 3,086 weed cutting instruments (ADCP) like StreamPro from RD Instruments have been events were registered in the 126 streams following the procedure applied in most larger and deeper Danish streams following the ISO- described below. standard ISO/TS 24154 in recent years. The annual number of discharge measurements was 10–12. The influence of macrophyte growth during the summer period can be observed in the Q/H relationship as deviations of water stage from the base Q/H relationship (Figure 2a; Dawson, 1978a, 1978b; Gurnell & Midgley, 1994). The deviations in H from the base Q/H relationship are caused by a change in resistance of flow in the downstream channel from especially growing macrophytes are described as new Q/H variants (Figure 2a). In fact, for calculation of discharge, a new Q/H relationship is established for every time step in the water stage record between the dates with actual measured discharge and stage utilising an interpolation procedure to produce the suite of Q/H curves. Discharge for every time step is then calculated based on the water stage record (every 10 min) making use of the established corresponding Q/H relationship variants (Figure 2b). To analyse and extract the influence of macrophyte growth and weed cutting during the summer period independent of changes in discharge, a control discharge for the summer period (median summer discharge) was estimated for all streams (Figure 2a). The water stage record for every time step (10 min) is then referred back to the control discharge (in the example stream control discharge is 800 L/s) and calculation of the water stage at that control discharge is used for showing the changes (increase) in stage more or less solely ascribed to changes in resistance due to regrowth of macrophytes in the channel downstream of the gauging station (Figure 2c). Similarly, cutting of weed in the channel downstream of the gauging FIGURE 1 streams Map showing the geographical location of the study station will result in an abrupt reduction in resistance shown as a sudden decline in the control water stage (Figure 2c). 4 | BAATTRUP-PEDERSEN (a) 240 Control Q (800 l/s) 220 Water stage (cm) ET AL. 200 180 Base 160 Q/H-r elatio nship 140 120 100 Discharge measurement Q/H variants 80 0 1,000 2,000 3,000 4,000 Discharge (l/s) (b) Water stage (cm) 260 240 220 200 180 160 Discharge (l/s) 140 5,000 4,000 3,000 2,000 1,000 0 Jan (c) Water stage (cm) 260 240 Feb Mar Apr May Jun Jul Discharge measurement Control point Control water stage Aug Sep Oct Nov Dec Weed cutting 220 200 180 160 140 Apr May Jun Jul Aug Sep Oct F I G U R E 2 An example stage (H)–discharge (Q) relationship at a gauging station in the stream Trend  A with inserted control Q (a), observed water stage and calculated discharge (b) and the calculated control water stage during summer period (c) 2.3 | Hydrographs and weed cutting the result. Finally, the water level increase due to regrowth of the vegetation following cutting was estimated as the increase in con- A sudden drop in water level on the hydrographs over a couple trol water level following cutting. The period used for this estima- of hours was registered as a weed cutting event. The drop had to tion was restricted to 3 weeks to prevent underestimation of be a sudden and not a continuous drop in water level over days regrowth due to intensified internal shading within the plant beds since the latter may reflect more factors than just weed cutting upon regrowth. In streams that were left uncut, growth was calcu- such as removal of cut biomass with reduced stowing in conse- lated as the increase in control water level from 1 May and quence (Figure 2). Often, a sudden increase was registered just 3 weeks ahead. This approach was chosen to avoid that the bio- prior to the drop, reflecting that the water stowed just upstream mass reached levels promoting internal shading with negative of the maintenance machinery. The water level drop associated effects on growth estimations. with a cutting was estimated as the difference between the water level before and after cutting (denoted a in Figure 2). The water level prior to cutting was estimated from a regression line cover- 2.4 | Physical and chemical variables ing the increase in control water level over the last 3 weeks, Physical and chemical environmental variables were assessed accord- thereby avoiding that minor water level fluctuations would affect ing to the technical guidelines of the Danish monitoring programme BAATTRUP-PEDERSEN | ET AL. 5 (Pedersen & Baattrup-Pedersen, 2003). Water chemistry parameters of the number of weed cuttings. The effect of site was included in (NO23-N, PO4-P, total-P) were determined at least twice annually the ANCOVA model as a random effect since sites are regarded as a (spring and autumn) at base-flow and means were used in the analy- random subsample from a larger population of sites. sis. Stream width, depth and slope were measured, and the stream Since weed cutting practice may change over time, we only reach was categorised according to (1) channel shape (i.e. a natural included sites with similar cutting frequency in 2012 and 2013 when shape being meandering or sinuous and a channelised shape being analysing the influence of cutting frequency on the ecological status straight) and (2) stream profile (i.e. a natural profile having natural applying DSPI. variations in depth, hydrology and substrate and a channelised profile having a rectangular shape). The number of sites with information on each variable is given in Table 2. 3 | RESULTS 3.1 | Effects of weed cutting on water level 2.5 | Aquatic plant data The average discharge of the 126 streams was 1.71 m3/s, average Aquatic plant data for 2012 and 2013 were obtained from the width (66 streams) was 7.4 m, and average depth 42.7 cm (Table 1). Danish environment, Some of the streams were quite small, however, with a mean width NOVANA (Friberg, Baattrup-Pedersen, Pedersen, & Skriver, 2005). of only 1.0 m and a mean depth of 5.6 cm (Table 1). The plant cov- Aquatic plant data were collected following the protocol described erage of the streams was, on average, 59% but with a large in Pedersen and Baattrup-Pedersen (2003). In each stream reach, between-stream variability (Table 1). Of the total of 3,086 weed cut- plant recordings were made in July/August at maximum biomass. tings registered in the 126 streams, most were performed from June Recordings were made in approximately 150 plots (25 9 25 cm) to September and <15% outside this period (Figure 3). The earliest placed side by side in cross-sectional transects at a 100-m-long cutting was performed at the start of April and the latest at the end stream reach. Depending on the width of the stream, the number of November. monitoring programme on the aquatic of transects varied from a minimum of 10 to maximum 20 in small We found significant effects of weed cutting on the water level streams. A cover score was allocated to each species present in in the streams and that these varied over time (Figure 4; Mixed the plots using the following abundance scale: 1 = 1%–5%, model ANOVA; F = 5.73; p < .0001). Generally, the effect was 2 = 6%–25%, 3 = 26%–50%, 4 = 51%–75%, 5 = 76%–100%. Spe- strongest in July, August, and September, with an average reduc- cies abundance at each stream reach was then calculated as the tion of 16 cm, and lowest in early spring and late autumn, with an sum of cover scores to the maximum score sum (i.e. the number of average reduction of 11 cm. The smallest effect of weed cutting plots multiplied by the maximum score of five; Pedersen et al., was a few cm and the greatest was 73 cm, such large drops being rare, though, as the upper 75% quantile was a little <20 cm and 2006). the lower 25% quantile <5 cm (data not shown). The observed 2.6 | Ecological status assessment from aquatic plant data variability in water level drop associated with cutting may be partly explained by differences in the slope of the streams and stream size in terms of stream width and depth (data not shown). Thus, Ecological status was calculated from the Danish Stream Plant Index the effect of weed cutting declined with increasing stream slope in (DSPI; Baattrup-Pedersen et al., 2013). The index gives an Ecological the range of 0&–3& (r = Quality Ratio (DSPI_EQR) that expresses the deviation from the nat- increasing stream width in the range from 2 to 16 m (r = .38; .32; p < .0001) and increased with ural undisturbed conditions. The ratio is expressed as a numerical p = .016) and with stream depth within the range of 8–90 cm value between 0 and 1 and translates in to high, good, moderate, (r = .51; p < .0001). poor and bad using the flowing boundary between ecological status classes: high/good = 0.70; good/moderate = 0.50; moderate/ poor = 0.35; poor/bad = 0.20. DSPI complies with the requirements of the Water Framework Directive (WFD) and is currently used in the ecological assessment of aquatic plant communities in mediumsized and large streams in Denmark. 2.7 | Statistics T A B L E 1 Characteristics of the 126 streams used in the study showing mean, minimum and maximum discharge, depth, width and plant coverage. Values on depth, width and plant coverage are based on measurements in summer (July, August), while discharge measurements are yearly measurement means. Note that plant coverage can exceed 100% in multi-layer communities (see for instance Baattrup-Pedersen et al., 2002) No. of streams The effect of weed cutting measured in terms of a water level drop was tested in a linear covariance model (ANCOVA) with month as the categorical independent variable and the number of weed cuttings as the covariate. The covariate model evaluates the means of the effect of weed cutting across months controlling for the effect 3 Discharge (m /s) Depth (cm) Width (m) Plant coverage (%) 126 Mean Min 1.71 Max 0.05 26.10 42.7 5.6 89.9 66 7.4 1.0 34.8 126 59.5 4.6 97.9 66 6 | BAATTRUP-PEDERSEN T A B L E 2 Results of linear regression models between regrowth, expressed as the increase in water level following weed cutting and stream water chemistry (PO4-P, total-P, and NO23-N), stream channel morphology (shape and profile) and vegetation characteristics (coverage and community-weighted means [CWM] of growth from basal, multi-apical and single-apical meristems). All regression models were statistically significant (p < .05). NS denotes that the obtained model was without significance (p > .05) 800 Number of weed cuttings 700 600 500 400 300 200 Variable, category Variable Chemistry PO4-P 100 0 Apr May Jun Jul Aug Sep Oct Nov F I G U R E 3 The number of weed cuttings performed in the 126 investigated streams in months with weed cutting Morphology Vegetation 18 16 14 No. of streams 54 Fvalue Pr > F 3.35 NS total-P 54 0.47 NS NO23-N 54 0.04 NS Channel shape 20 1.04 NS Channel profile 20 0.18 NS Coverage 126 2.40 NS CWM basal growth 126 7.63 0.0076 CWM multi- apical growth 126 4.68 0.0345 CWM single-apical growth 126 5.91 0.0181 12 0.6 10 8 0.5 6 0.4 4 2 0 Apr May Jun Jul Aug Sep Oct Nov F I G U R E 4 The effect of weed cutting in terms of cm drop in water level in months with weed cutting Regrowth (cm/day) Cutting effect (cm water level) ET AL. 0.3 0.2 0.1 0 –0.1 –0.2 3.2 | Regrowth The increase in water level following weed cutting due to vegetation regrowth did not vary significantly with either stream water chem- Apr May Jun Jul Aug Sep Oct Nov F I G U R E 5 Regrowth in terms of increases in water level after weed cutting applying a period restricted to 3 weeks in months with weed cutting istry (PO4-P, total-P, NO23-N; Table 2; p > .05), morphometric properties of the stream (channel shape, channel profile; Table 2; Regrowth also varied significantly with cutting frequency as a p > .05), or total plant coverage in the streams (Table 2; p > .05). significant interaction was observed between the time and the fre- However, regrowth did vary significantly with growth characteristics quency of cutting (mixed model covariance analysis; F = 2.61; of the vegetation (Table 2). Thus, regrowth increased with an p < .0163). Thus, regrowth increased with enhanced cutting fre- increasing abundance of species growing from basal growth meris- quency (Figure 6), from an average of 0.04 cm/day in streams sub- tems (Table 2; p < .05) but decreased with increasing abundance of jected to one annual cutting to an average of 0.6 cm/day in streams species growing from single or multi-apical growth meristems (Fig- subjected to >6 annual cuttings. No significant differences in ure 5, Table 2; p < .05). regrowth were found in streams cut 6, 7 and 8 times annually (Fig- The increase in water level following weed cutting due to vegetation regrowth varied significantly with cutting time (Mixed model ure 7). Regrowth in June was 0.2 cm/day in streams cut once per year and 0.9 cm/day in streams cut eight times per year. covariance analysis; F = 4.04; p = .0005; Figure 5). Regrowth was largest in June, with an increase in water level of 0.41 cm/day, whereas no regrowth took place in autumn. In October, negative 3.3 | Ecological status regrowth was registered, indicating that the plant biomass was We found that ecological status assessed applying DSPI depended decaying. significantly on the annual number of weed cuttings (ANOVA; BAATTRUP-PEDERSEN | ET AL. 7 with those in comparable streams outside Denmark. However, in 1.0 support of our results, a study conducted in River Lambourn, UK, revealed responses to cutting comparable with those observed in Regrowth (cm/day) 0.8 our investigation (average 16 cm), varying between 16 and 29 cm in three successive years of cutting (Old et al., 2014), and another 0.6 study, although based on modelling, reported a change in water level of app. 50 cm (Hearne & Armitage, 1993). 0.4 We also observed that plant regrowth was highest in early summer, in June, and absent in autumn, and furthermore, and in accor- 0.2 dance with our second hypothesis, that regrowth occurred faster in streams that were cut frequently. Actually, we observed a fourfold higher regrowth in streams cut eight times during summer compared 0 0 1 2 3 4 5 6 8 7 Number of cuttings (per year) with that observed in streams cut once during summer. Fast regrowth reduces the time where the discharge capacity is main- F I G U R E 6 Regrowth in terms of increases in water level after weed cutting applying a period restricted to 3 weeks after weed cutting as a function of the number of weed cuttings performed in the streams tained or partly maintained in the channels. For example, if the increase in water level is 0.6 cm/day, then it takes 4 weeks to reach pre-cutting conditions (applying the average water level drop of 16 cm). In contrast, in June when regrowth is at its maximum (0.9 cm/day), pre-cutting conditions are obtained after only Weed cutting frequency (per year) 2.5 weeks. 2.0 A The finding that plant regrowth increased with increasing weed A cutting frequency is unlikely to reflect differences in growth conditions among the streams. Thus, regrowth did not vary following a 1.5 predictable pattern with either stream water chemistry or morpho- B 1.0 metric properties of the stream channel. Instead, we observed that B regrowth increased with the abundance of plant species growing from basal meristems and decreased with the abundance of species growing from single- and multi-apical meristems. In accordance with 0.5 this, previous studies have also shown that weed cutting prompts higher abundance of species growing from basal meristems while 0 being less abundant in streams without cutting (Baattrup-Pedersen 1–2 3 4 5 Ecological status (DSPI) F I G U R E 7 Relationships between weed cutting frequency (annual cuttings) and ecological status class based on aquatic plant assemblages in the streams applying DSPI (Danish Stream Plant Index). Different letters denote significant difference between mean values (ANOVA; p < .5) et al., 2016), whereas species growing from apical meristems show the opposite pattern (Baattrup-Pedersen et al., 2016). Hence, the position of the growth meristem likely determines the potential for plant regrowth following weed cutting, reflecting that species that have an intact growth point after weed cutting can start regrowth immediately after the intervention, whereas species with apical growth meristems exhibit delayed regrowth. Consequently, the increased weed cutting frequency in Danish streams may have p < .05). Stream sites that were cut more than once per year were caused a shift in the composition of plant species towards increasing either classified as poor/bad or moderate ecological status sites, dominance of species possessing traits that enable them to cope whereas sites cut once per year or less were classified as either with frequent cutting (Abernethy, Sabbatini, & Murphy, 1996; Baat- good or high ecological status sites (Figure 7). trup-Pedersen et al., 2016; Sabbatini & Murphy, 1996b). Consequently, our findings strongly question the underlying assumption that the change in cutting practice in Danish streams around 1990 4 | DISCUSSION from less comprehensive to more frequent weed cutting (BaattrupPedersen et al., 2000) has maintained the discharge capacity of the In accordance with our first hypothesis, we found that weed cutting streams. In fact, the observed increase in plant regrowth with lowered the water levels in the investigated streams and that the enhanced frequency of weed cutting may have intensified the need effects of cutting were strongest in late summer when the biomass for regular cutting to maintain the same discharge capacity. of aquatic plants was highest. As stated already in the introduction, Other studies have shown that species with high tolerance quantifications of the hydraulic impacts of weed cutting are rare, towards weed cutting often exhibit a high overwintering capacity as making it difficult to compare the here observed water level drops well by producing vegetative propagules that remain dormant during 8 | BAATTRUP-PEDERSEN ET AL. the coldest seasons (Baattrup-Pedersen et al., 2003, 2016; Wiegleb combination of increased average flow and higher plant biomass can et al., 2014). Consequently, the resilience of these species may be intensify the need for management targeting flood prevention. Our huge once they become dominant. This may also explain why Spar- findings highlight that it is questionable if the current weed cutting ganium spp. and other species that both have basal growth and a practice is an efficient flood protection measure since (1) regrowth high overwintering capacity are highly tolerant of weed cutting, as seems to be stimulated by frequent weed cutting and a positive observed previously (Baattrup-Pedersen et al., 2003, 2016; Wiegleb feedback loop may develop, requiring even more frequent cuttings et al., 2014). For example, Wiegleb et al. (2014) noticed an increase to maintain the discharge capacity, and (2) many species that are in S. emersum in response to increasing anthropogenic disturbance in stimulated by weed cutting, such as for instance S. emersum, form German rivers when combining hydromorphological stressors (e.g. dense canopy beds across the entire stream profile and may there- weed cutting, dredging, and construction work) and stressors associ- fore cause a greater reduction in the discharge capacity than species ated with water quality such as malfunctioning sewage plants and growing in confined patches (Sand-Jensen & Mebus, 1998). At pre- conventional agriculture. Interestingly, S. emersum is also widely dis- sent, however, very limited information exists on how different plant tributed in less disturbed lowland streams in Europe (Baattrup-Ped- morphologies impact flow (Naden et al., 2006) and how a shift in ersen et al., 2008) but less abundant than in anthropogenically species composition can influence flow resistance (e.g. Pitlo & Daw- disturbed streams (Baattrup-Pedersen, Larsen, & Riis, 2002; Birk & son, 1990).As an alternative to weed cutting, the establishment of Willby, 2010; Pedersen et al., 2006; Riis, Sand-Jensen, & Vester- buffers along streams that can act as measures against flooding of gaard, 2000; Steffen, Becker, Herr, & Leuschner, 2013). It has been downstream areas under high flow events may be more efficient. suggested that the success of this species in less impacted streams is linked to a low light compensation point and a low photosaturation level (Sand-Jensen et al., 1989). Hence, different traits can ACKNOWLEDGEMENT contribute to its success in different types of streams (Baattrup- This study was supported by the European Union Seventh Frame- Pedersen et al., 2002). work Project MARS under contract no. 603378. We thank Anne In accordance with our third hypothesis, we found that weed cutting clearly conflicts with the probability of reaching good ecolog- Mette Poulsen for manuscript editing and Tinna Christensen for figure layout. ical status in the investigated streams as meeting this goal requires that all ecological quality elements are in at least good ecological status (the one-out-all-out principle; European Commission, 2000). We ORCID found that the ecological status assessed from species composition Annette Baattrup-Pedersen applying DSPI (Baattrup-Pedersen et al., 2013) was either moderate 344X http://orcid.org/0000-0002-3118- or poor/bad in streams with more than one annual weed cutting. This finding highlights the necessity of developing new management practices for lowland streams to secure both the discharge capacity and the ecological quality. At present, little research has been conducted into developing alternative management methods, but other measures should be considered. Based on our findings, we suggest a reduction in weed cutting frequency and its postponement to late summer to avoid fast regrowth. Additionally, we suggest that weed cutting should be risk-based so that cutting is only performed when there is a risk of flooding of adjacent areas as opposed to cuttings performed at fixed time periods. As long as the frequency is reduced, such a practice may hinder cuttings that only lower the water table marginally and counteract the development of aquatic plant communities with fast regrowth. 4.1 | Perspectives under climate change The ongoing climate change will increase the average flow in lowland streams, mainly due to higher precipitation (and less snow accumulation) during winter months with an increase in the frequency and magnitude of extreme flow events (Karlsson, Sonnenborg, Seaby, Jensen, & Refsgaard, 2015; van Roosmalen, Sonnenborg, & Jensen, 2009; Thodsen et al., 2014). In addition, increasing temperatures may prolong the plant growing season in the streams. The REFERENCES Abernethy, V. J., Sabbatini, M. R., & Murphy, K. J. (1996). Response of Elodea Canadensis Michx. and Myriophyllum spicatum L. to shade, cutting and competition in experimental culture. Hydrobiologia, 340, 219–224. https://doi.org/10.1007/978-94-011-5782-7_34 Armitage, P. D., Blackburn, J. H., Winder, M., & Right, J. F. (1994). Impact of vegetation management on macroinvertebrates in chalk streams. Aquatic Conservation: Marine and Freshwater Ecosystems, 4, 95–104. https://doi.org/10.1002/aqc.3270040202 €the, E., Riis, T., Andersen, D. K., & Larsen, S. E. Baattrup-Pedersen, A., Go (2017). A new paradigm for biomonitoring: An example building on the Danish stream plant index. Methods in Ecology and Evolution, 8, 297–307. https://doi.org/10.1111/2041-210X.12676 € the, E., Riis, T., & O’Hare, M. (2016). FuncBaattrup-Pedersen, A., Go tional trait composition of aquatic plants can serve to disentangle multiple interacting stressors in lowland streams. Science of the Total Environment, 543, 230–238. https://doi.org/10.1016/j.scitote nv.2015.11.027 Baattrup-Pedersen, A., Kristensen, E. A., Jorgensen, J., Skriver, J., Kronvang, B., Andersen, H. E., . . . Kjellerup Larsen, L. M. (2009). Can a priori defined reference criteria be used to select reference sites in Danish streams? Implications for implementing the water framework directive. Journal of Environmental Monitoring, 11, 344–352. https:// doi.org/10.1039/b808779b Baattrup-Pedersen, A., Larsen, S. E., & Riis, T. (2002). Long-term effects of stream management on plant communities in two Danish lowland BAATTRUP-PEDERSEN ET AL. streams. Hydrobiologia, 48, 33–45. https://doi.org/10.1023/A: 1021296519187 Baattrup-Pedersen, A., Larsen, S. E., & Riis, T. (2003). Composition and richness of plant communities in small Danish streams – influence of environmental factors and weed cutting. Hydrobiologia, 495, 171– 179. https://doi.org/10.1023/A:1025442017837 Baattrup-Pedersen, A., Larsen, S. E., & Riis, T. (2013). From expert judgment to supervised classification: A new approach to assess ecological status in lowland streams. Science of the Total Environment, 447, 116– 122. https://doi.org/10.1016/j.scitotenv.2012.12.062 Baattrup-Pedersen, A., & Riis, T. (1999). Macrophyte diversity and composition in relation to substratum characteristics in regulated and unregulated Danish streams. Freshwater Biology, 42, 375–385. https://doi.org/10.1046/j.1365-2427.1999.444487.x Baattrup-Pedersen, A., & Riis, T. (2004). Impacts of different weed cutting practices on plant species diversity and composition in a Danish stream. River Research Applications, 20, 103–114. https://doi.org/10. 1002/rra.722 Baattrup-Pedersen, A., Skriver, J., & Wiberg-Larsen, P. (2000). Weed-cutting practice and impact on trout density in Danish lowland streams. Verhandlungen der Internationale Vereinigung der Limnologie, 27, 674– 677. https://doi.org/10.1080/03680770.1998.11901319 Baattrup-Pedersen, A., Springe, G., Riis, T., Larsen, S. E., Sand-Jensen, K., & Kjellerup, L. M. (2008). The search for reference conditions for stream vegetation in northern Europe. Freshwater Biology, 53, 1890– 1901. https://doi.org/10.1111/j.1365-2427.2008.02003.x Bach, H. (red.), Baattrup-Pedersen, A., Holm, P. E., Jensen, P. N., Larsen, T., Ovesen, N. B., . . . Styczen, M. (2016). Faglig udredning om grødeskæring i vandløb. Aarhus Universitet, DCE – Nationalt Center for Miljø og Energi, 106 s. - Videnskabelig rapport fra DCE - Nationalt Center for Miljø og Energi nr. 188. Retrieved from http://dce2.au.dk/ pub/SR188.pdf. Birk, S., & Willby, N. (2010). Towards harmonization of ecological quality classification: Establishing common grounds in European plant assessment for rivers. Hydrobiologia, 652, 149–263. https://doi.org/10. 1007/s10750-010-0327-3 Caffrey, J. M. (1991). Aquatic plant management in Irish rivers. In M. W. Steer (Ed.), Irish rivers: Biology and management (pp. 85–98). Dublin, Ireland: Royal Irish Academy. Caffrey, J. M. (1993). Plant management as an integrated part of Ireland’s aquatic resources. Hydroecologie Appliquee, 5, 77–96. https://doi.org/ 10.1051/hydro:1993104 Carpenter, S. R., & Lodge, D. M. (1986). Effects of submersed macrophytes on ecosystem processes. Aquatic Botany, 26, 341–370. https://doi.org/10.1016/0304-3770(86)90031-8 Curran, J. C., & Hession, W. C. (2013). Vegetative impacts on hydraulics and sediment processes across the fluvial system. Journal of Hydrology, 505, 364–376. https://doi.org/10.1016/j.jhydrol.2013.10. 013 Dawson, F. H. (1978a). The seasonal effects of aquatic plant growth on the flow of water in a stream. Proceedings of the European Weed Research Society 5th Symposium on Aquatic Weeds, 71–78. Dawson, F. H. (1978b). Aquatic plant management in semi natural streams: The role of marginal vegetation. Journal of Environmental Management, 6, 213–221. Dawson, F. H., Clinton, E. M. F., & Ladle, M. (1991). Invertebrates on cut weed removed during weed-cutting operations along an English river, the River Frome, Dorset. Aquatic Fisheries Management, 22, 113–121. https://doi.org/10.1111/j.1365-2109.1991.tb00500.x Dawson, F. H., & Robinson, W. N. (1984). Submerged macrophytes and the hydraulic roughness of a lowland chalk stream. Verhandlungen der Internationale Vereinigung der Limnologie, 22, 1944–1948. European Commission (2000). Directive 2000/60/EC of the European Parliament and of the Council – establishing a framework for community action in the field of water policy. | 9 Ferreiro, N., Feijoo, C., Giorgi, A., & Leggieri, L. (2011). Effects of macrophyte heterogeneity and food availability on structural parameters of the macroinvertebrate community in a Pampean stream. Hydrobiologia, 664, 199–211. https://doi.org/10.1007/ s10750-010-0599-7 Folkard, A. M. (2011). Vegetated flows in their environmental context: A review. Engineering and Computational Mechanics, 164, 3–24. https://doi.org/10.1680/eacm.8.00006 Fox, A. M., & Murphy, K. J. (1990). The efficacy and ecological impacts of herbicide and cutting regimes on the submerged plant communities of four British rivers. I. A comparison of management efficacies. Journal of Applied Ecology, 27, 520–540. https://doi.org/10.2307/ 2404299 Friberg, N., Baattrup-Pedersen, A., Pedersen, M. L., & Skriver, J. (2005). The new Danish stream monitoring programme (NOVANA) – preparing monitoring activities for the water framework directive era. Environmental Monitoring Assessment, 111, 27–42. https://doi.org/10. 1007/s10661-005-8038-3 Green, J. C. (2005). Comparison of blockage factors in modelling the resistance of channels containing submerged macrophytes. River Research and Applications, 21, 671–686. https://doi.org/10.1002/ (ISSN)1535-1467 Gurnell, A. (2014). Plants as river system engineers. Earth Surface Processes and Landforms, 39, 4–25. https://doi.org/10.1002/esp.3397 Gurnell, A. M., & Midgley, P. (1994). Aquatic weed growth and flow resistance – influence on the relationship between discharge and stage over a 25 year river gauging station record. Hydrological Processes, 8, 63–73. https://doi.org/10.1002/(ISSN)1099-1085 Hearne, J. W., & Armitage, P. D. (1993). Implications of the annual macrophyte growth cycle on habitat in rivers. Regulated Rivers Research and Management, 8, 313–322. https://doi.org/10.1002/rrr. 3450080402 Herschy, R. W. (1995). Streamflow measurements (2nd ed., p. 525). London, UK: Published by E & FN Spoon, Chapman & Hall. ISO 748 (2007). Hydrometry – Measurement of liquid flow in open channels using current-meters or floats. ISO/TS 24154 (2005). Hydrometry – Measuring river velocity and discharge with acoustic Doppler profilers. Iversen, T. M. (1988). Secondary production and trophic relationships in a spring invertebrate community. Limnology and Oceanography, 33, 582–592. https://doi.org/10.4319/lo.1988.33.4.0582 Kaenel, B. R., & Uehlinger, U. (1999). Aquatic plant management: Ecological effects in two streams of the Swiss Plateau. Hydrobiologia, 415, 257–263. https://doi.org/10.1023/A:1003860831654 Karlsson, I. B., Sonnenborg, T. O., Seaby, L. P., Jensen, K. H., & Refsgaard, J. C. (2015). Effect of a high-end CO2-emission scenario on hydrology. Climate Research, 64, 39–54. https://doi.org/10.3354/cr01265 Madsen, B. L. (1995). Danish watercourses. Miljønyt nr. 11. Ministry of Environmental Protection. Copenhagen, Denmark. Mortensen, E. (1977). Density-dependent mortality of trout fry (Salmo trutta) and its relationship to the management of small streams. Journal of Fish Biology, 11, 613–617. https://doi.org/10.1111/j.10958649.1977.tb05719.x Naden, P., Rameshwaran, P., Mountford, O., & Robertson, C. (2006). The influence of macrophyte growth, typical of eutrophic conditions, on river flow velocities and turbulence production. Hydrological Processes, 20, 3915–3938. https://doi.org/10.1002/hyp.6165 O’Hare, M. T., Hutchinson, K. A., & Clarke, R. T. (2007). The drag and reconfiguration experienced by five macrophytes from a lowland river. Aquatic Botany, 86, 253–259. https://doi.org/10.1016/j.aquab ot.2006.11.004 Old, G. H., Naden, P. S., Rameshwaran, P., Acreman, M. C., Baker, S., Edwards, F. K., . . . Neal, M. (2014). Instream and riparian implications of weed cutting in a chalk river. Ecological Engineering, 71, 290–300. https://doi.org/10.1016/j.ecoleng.2014.07.006 10 | €ller-Wohlfeil, D.-I., SvendOvesen, N. B., Iversen, H. L., Larsen, S. E., Mu sen, L. M., Blicher, A. S., & Jensen, P. M. (2000). Afstrømningsforhold i danske vandløb. Danmarks Miljøundersøgelser. 238 s. Faglig rapport fra DMU nr. 340. In Danish. Pedersen, M. L., & Baattrup-Pedersen, A. (2003). Ecological monitoring in streams and riparian areas under NOVANA 2004–2009 (in Danish). Aarhus, Denmark: National Environmental Research Institute, Aarhus University. Pedersen, T. C. M., Baattrup-Pedersen, A., & Madsen, T. V. (2006). Effects of stream restoration and management on plant communities in lowland streams. Freshwater Biology, 51, 161–179. https://doi.org/ 10.1111/j.1365-2427.2005.01467.x Pitlo, R. H., & Dawson, F. H. (1990). Flow resistance of aquatic weeds. In A. H. Pieterse, & K. J. Murphy (Eds.), Aquatic weeds: The ecology and management of nuisance aquatic vegetation (pp. 74–94). Oxford, UK: Oxford University Press. Riis, T., Sand-Jensen, K., & Vestergaard, O. (2000). Plant communities in lowland streams: Species composition and environmental factor. Aquatic Botany, 66, 255–272. https://doi.org/10.1016/S0304-3770 (99)00079-0 van Roosmalen, L., Sonnenborg, T. O., & Jensen, K. H. (2009). Impact of climate and land use change on the hydrology of a large-scale agricultural catchment. Water Resources Research, 45, W00A15. https://doi. org/10.1029/2007WR006760 Sabbatini, M. R., & Murphy, K. J. (1996a). Response of Callitriche and Potamogeton to cutting, dredging and shade in English drainage channels. Journal of Aquatic Plant Management, 34, 8–12. Sabbatini, M. R., & Murphy, K. J. (1996b). Submerged plant survival strategies in relation to management and environmental pressures in drainage channel habitats. Hydrobiologia, 340, 191–195. https://doi. org/10.1007/BF00012754 Sand-Jensen, K. (1997). Macrophytes as biological engineers in the ecology of Danish streams. In K. Sand-Jensen, & O. Pedersen (Eds.), Freshwater biology – priorities and development in Danish research (pp. 74–101). Copenhagen, Denmark: G.E.C. Gad. Sand-Jensen, K. (2003). Drag and reconfiguration of freshwater macrophytes. Freshwater Biology, 48, 271–283. https://doi.org/10.1046/j. 1365-2427.2003.00998.x Sand-Jensen, K., Jeppesen, E., Nielsen, K., van der Bijl, L., Hjermind, A.-L., Nielsen, A.-L., & Iversen, T. M. (1989). Growth of macrophytes and ecosystem consequences in a Danish lowland stream. Freshwater Biology, 15, 15–32. https://doi.org/10.1111/j.1365-2427.1989.tb01080.x Sand-Jensen, K., & Mebus, J. (1998). Influence of submerged macrophytes on sediment composition and near-bed flow in lowland BAATTRUP-PEDERSEN ET AL. streams. Freshwater Biology, 39, 663–679. https://doi.org/10.1046/j. 1365-2427.1998.00316.x Schwarz, A. M., & Snelder, T. (1999). Integrated submerged aquatic vegetation management in an urban New Zealand river. Hydrobiologia, 415, 235–241. https://doi.org/10.1023/A:1003839402075 Steffen, K., Becker, T., Herr, W., & Leuschner, C. (2013). Diversity loss in the plant vegetation of northwest German streams and rivers between the 1950s and 2010. Hydrobiologia, 713, 1–17. https://doi. org/10.1007/s10750-013-1472-2 Svenning, J.-C. (2002). A review of natural vegetation openness in North-Eastern Europe. Biological Conservation, 104, 133–148. https://doi.org/10.1016/S0006-3207(01)00162-8 Teixeira-de Mello, F., Meerhoff, M., Gonzalez-Bergonzoni, I., Kristensen, E. A., Baattrup-Pedersen, A., & Jeppesen, J. (2015). Influence of riparian forests on fish assemblages in temperate lowland streams. Environmental Biology of Fishes, 99, 133–144. https://doi.org/10.1007/ s10641-015-0462-9 Thodsen, H., Baattrup-Pedersen, A., Andersen, H. E., Jensen, K. M. B., Andersen, P. M., Bolding, K., & Ovesen, N. B. (2014). Climate change effects on lowland stream flood regimes and riparian rich fen vegetation communities in Denmark. Hydrological Sciences Journal, 51, 1–15. https://doi.org/10.1080/02626667.2014.990965 Vereecken, H., Baetens, J., Viaene, P., Mostaert, E., & Meire, P. (2006). Ecological management of aquatic plants: Effects in lowland streams. Hydrobiologia, 570, 205–210. https://doi.org/10.1007/s10750-0060181-5 Vogelsang, L. (2016). Practical experience with environmentally friendly stream management in Europe to address the dilemma with increased flooding risk. Report of an internship at Deltares (Utrecht, the Netherlands). €ring, U., Filetti, M., Brux, H., & Herr, W. (2014). LongWiegleb, G., Bro term dynamics of aquatic plant dominance and growth-form types in two north-west German lowland streams. Freshwater Biology, 59, 1012–1025. https://doi.org/10.1111/fwb.12323 How to cite this article: Baattrup-Pedersen A, Ovesen NB, Larsen SE, et al. Evaluating effects of weed cutting on water level and ecological status in Danish lowland streams. Freshwater Biol. 2018;00:1–10. https://doi.org/10.1111/ fwb.13101