Habitat requirements of Elodea canadensis Michx. in Polish rivers

The study was based on a countrywide survey conducted in Poland, with a dataset for 1135 river sites (Fig. 1). We analyzed 206 sites with Elodea canadensis, located in 173 water courses (the full list is presented in Appendix 1). The database was completed between 2003 and 2014. The basic parameters describing the studied rivers containing E. canadensis are presented in Table 1 (altitude, width and depth of a riverbed, bottom sediments, flow types, river valley land use, physical and chemical water parameters, and hydromorphological parameters) and in Table 2 (concentrations of macroelements and trace metals in water).

Figure 1

Table 1

Descriptive statistics of major habitat characteristics of the surveyed rivers (N = 206)

Table 2

Descriptive statistics of concentrations of elements in water of selected Elodea canadensis rivers (based on data from the State Environmental Monitoring)

Macrophyte surveys were carried out during the intensive growth of most aquatic plants (from mid-June to mid-September). Field surveys were conducted using the Macrophyte Method for River Assessment (Szoszkiewicz et al. 2010b). This method is currently an official monitoring approach to rivers in Poland (Dziennik Ustaw 2016). The macrophyte assessment was based on the presence of algae, mosses, horsetails, liverworts, monocotyledonous and dicotyledonous plant species that are biological indicators of water quality. All submerged, free floating, semi-terrestrial and emergent plants were considered. The assessment also included macrophytes attached to or rooted in parts of the river bank substrate where they were likely submerged for most of the year. In wadeable survey sites, an aquascope was used to support the observations. The macrophyte survey was conducted over reaches of 100 m length. The survey includes a list of species and estimated ground cover of plants. The presence of each species was recorded with their percentage cover using the following nine-point scale: < 0.1%, 0.1-1%, 1-2.5%, 2.5-5%, 5-10%, 10-25%, 25-50%, 50-75% and > 75%. Based on the collected field data, the numerical index MIR (Macrophyte Index for Rivers) was computed (Szoszkiewicz et al. 2010b). It reflects river degradation, especially eutrophication and ranges from 10 (most degraded rivers) to 100 (highest quality). To assess the ecological status of the river, the calculated values of the MIR index were referenced to the current standards (Dziennik Ustaw 2016).

During plant and hydromorphological surveys, three subsamples of water were randomly collected from each river site in the river midstream at a depth of 0.5-1 m. Water samples were not collected during rainy weather or periods with heavy runoff; if necessary, an additional visit was organized. Prior to analysis, all water samples were filtered using Sartorius cellulose filters with a nominal pore size of 0.45 μm, except for those used for the determination of total phosphorus. Water samples were cooled below 10°C and all parameters were analyzed in a laboratory within a 12h period. Electrical conductivity and pH were measured by digital potentiometers. Alkalinity was measured with sulfuric acid to the end point pH of 4.5 in the presence of methyl orange. Concentrations of phosphate (molybdenum blue method), total phosphorus (molybdenum blue method after microwave mineralization in MARS 5X), nitrate nitrogen (cadmium reduction method), and ammonium nitrogen (Nessler’s method) were determined using a spectrophotometer HACH DR/2800.

Information about concentrations of 21 elements in water was obtained from the State Environmental Monitoring. The research was carried out in laboratories accredited by the Polish Centre for Accreditation. Table 2 shows the average annual values as a mean of twelve measurements (Ca²⁺, Mg²⁺, Cl⁻, F⁻, and SO₄²⁻), or as a mean of four measurements (Al, As, B, Ba, Cd, Cr, Cu, Fe, Hg, K, Mn, Na, Ni, Pb, Se, and Zn).

Calcium and magnesium were determined by titration with EDTA (PN-ISO 6058:1999, PN-ISO 6059:1999) or by atomic absorption spectrometry (PN-EN ISO 7980:2002); sodium and potassium were measured by atomic absorption spectrometry (PN-ISO 9964-2:1994); chlorides were analyzed using the Mohr method, i.e. titration with silver nitrate in the presence of chromate as indicator (PN-ISO 9297:1994); sulfur was determined gravimetrically with barium chloride (PN-ISO 9280:2002); fluorides were measured using ion chromatography (PN-EN ISO 10304-1:2009). Trace elements were determined by atomic emission spectrometry with inductively coupled plasma (PN-EN ISO 11885:2009) and by atomic absorption spectrometry with a graphite tube (PN EN ISO 15586:2005) or with flame atomization (PN ISO 8288:2002). Mercury was determined by atomic fluorescence spectrometry with amalgamation enrichment (PN-EN ISO 17852:2009).

Measurement accuracy was determined by comparing the results of the determination of three separated portions of each sample, which were analyzed using the identical methods. Blank samples were digested in the same manner.

Hydromorphological evaluation was conducted at each site according to the River Habitat Survey (RHS) method (Environment Agency 2003; Szoszkiewicz et al. 2012). The RHS data were collected from 500 m stretches of rivers. The RHS surveys were performed in ten profiles (spot checks) distributed at 50 m intervals. The macrophyte survey section was located inside each RHS site, always between the 6th and 8th spot check. Four numerical metrics based on the RHS protocol were produced: Habitat Quality Assessment– HQA, Habitat Modification Score – HMS (Raven et al. 1998; Szoszkiewicz et al. 2012), River Habitat Quality – RHQ, and River Habitat Modification – RHM (Tavzes & Urbanic 2009). The range of variability of the analyzed hydromorphological indices is given in Table 1. High values of HQA and RHQ indicate an extensive presence of a number of natural river features and high landscape diversity along the river. High HMS and RHM values indicate extensive anthropogenic alteration such as bank and channel re-sectioning and reinforcement or other river engineering constructions. The grain size composition and flow types were derived from the RHS database. Six flow types and six types of riverbed material were distinguished (Table 1). We also calculated the granulometry index and the flow type index (Jusik et al. 2015).

The granulometry index (GM_index) reflects the average grain size composition of the riverbed associated with the kinetic energy of the flow. It is based on the parameter “dominant channel substrate in spot checks” assessed using the RHS method (section E).

The flow type index (FT_index) reflects the average riverbed hydraulic characteristics associated with parameters such as slope, flow velocity and depth. It is based on the parameter “dominant flow type in spot checks” by the RHS method (section E).

Factor analysis (principal components analysis –PCA) with varimax normalized rotation was used to uncover the structure of environmental matrices and reveal environmental gradients. The ecological amplitude of E. canadensis was identified based on descriptive statistics of environmental variables (minimum, median, maximum and coefficient of variation). All calculations were performed with the Statistica 12.5 software (StatSoft Inc. 2014). Taxonomic diversity of macrophytes accompanying E. canadensis was analyzed using detrended correspondence analysis (DCA) from CANOCO for Windows version 4.55 (TerBraak & Smilauer 2002). Rare taxa found at up to three sampling sites were excluded from the analysis. DCA analysis of the biological data revealed that the first gradient length was 3.201 SD (standard deviation), indicating that the biological data exhibited unimodal responses to underlying environmental variables.

Principal components analysis resulted in a simplified habitat description of the analyzed matrix. The first three factors were responsible for 43.5% of the sample variance. Table 3 presents three principal components and their corresponding eigenvalues after varimax normalized rotation; each of the three eigenvalues was responsible for more than 10% of the variance. The first principal component was strongly correlated with human impact, especially the degree of hydromorphological modification. It was negatively correlated with a percentage of urban areas, the RHM index, and the HMS index. The second principal component was strongly correlated with physicochemical water parameters reflecting the eutrophication. It was strongly positively correlated with total phosphorous, reactive phosphorous, and ammonia nitrogen. Finally, the third principal component was strongly correlated with high hydromorphological naturalness, forests as a land-use type and shading of a riverbed. The PCA results show that the analyzed database was characterized by a strong human impact gradient associated with hydromorphological modification of a riverbed and changes in land use and eutrophication. In the studied rivers, pH was the most stable environmental variable (CV = 0.04) and ammonium nitrogen – the most diverse one (CV = 2.78) (Table 2).

Table 3

Factor loadings of the first three principal components

E. canadensis occurs in lowland rivers throughout Poland (Fig. 1). It rarely occurs in upland rivers, and only up to an altitude of about 500 m a.s.l. (Table 1). The species was not found in mountain rivers. It prefers small and medium sandy lowland rivers (78% of the analyzed sites). Moreover, it occurs in gravelly lowland streams (14%), gravelly and sandy upland (7%) and stony siliceous upland watercourses (only three sites located in the Zadrna and Sopot rivers). The ecological status of the studied river sites, based on the MIR index, was as follows: 45% classified as good, 40% moderate, 8% very good and 7% poor. No Canadian waterweed sites were classified into bad ecological status.

Canadian waterweed has high light requirements and occurs mainly in unshaded sections of shallow rivers (often flowing through meadows) with an average depth of about 0.6 m (Table 1). It occurs in places with a dominant laminar flow (smooth) or slight turbulence (rippled). The median of the FT_index is 1.51. It avoids highly turbulent flow of high kinetic energy and marginal dead-water zones (Fig. 2). Waterweed occurs mostly in sandy bottom sections (median GM_index = 2.09), with some admixture of coarse mineral (gravel/pebble) and fine organic fraction (silt) (Fig. 3). It prefers moderately natural (HQA = 41 ± 14, RHQ = 307 ± 76) and moderately transformed (HMS = 32 ± 30, RHM = 40 ± 40) river sections, usually with a straightened planform, a uniform bank profile, re-sectioned and reinforced by fascines. It was not found in strongly transformed river sections (reinforced by concrete, sheet piling, cladding, cobblestones, and gabion) and those with reinforced banks and bottoms. E. canadensis was most often found in grasslands or forest river sections (Fig. 4), but moderately shady, on average 10-20% (Table 1).

Figure 2

Figure 3

Figure 4

In the present study, the pH of water from E. canadensis stands varied from 6.90 to 9.30 and was higher than the optimum range (pH 6-7) for the photosynthesis of E. nuttallii (Jones et al. 2000), which has very similar ecological and physiological requirements (Barrat-Segretain 2001). This is consistent with Santos et al. (2011), who stated that biomass of Canadian waterweed is positively correlated with pH. According to Santos et al. (2011), E. canadensis shows a higher growth rate and higher cover in areas with high conductivity, thus higher salinity. Also in the present study, Canadian waterweed prefers moderately mineralized water (545 ± 329 μS cm⁻¹), rich in calcium and magnesium carbonates (174 ± 63 mg CaCO₃ l⁻¹, 84.1 ± 31.4 mg Ca²+ l⁻¹ and 11.1 ± 6.4 mg Mg²⁺ l⁻¹), with moderate concentrations of chlorides and sulfates (38.9 ± 59.1 mg Cl^- l⁻¹ and 62.3 ± 50.9 mg SO₄²⁻ l⁻¹). The average concentrations of Cl^- and SO₄²⁻ in the surveyed rivers were similar to that in the water of the Wielkopolska district (Poland) where E. canadensis stands were examined by Samecka-Cymerman & Kempers (2003). It also tolerates considerable salinity of the water (up to 247 mg Na l⁻¹, 440 mg Cl⁻ l⁻¹ and 343 mg SO₄²⁻ l⁻¹).

In terms of nutrients, E. canadensis prefers water from mesotrophic to eutrophic (0.70 ± 1.25 mg PO₄^3- l⁻¹, 1.17 ± 1.73 mg N_NO3− l⁻¹, and 0.44 ± 1.21 mg N_NH4+ l⁻¹). This is in accordance with Thiébaut (2005), Kuhar et al. (2010) and Zehnsdorf et al. (2015) who reported that E. canadensis grows successfully under a wide range of environmental conditions, ranging from mesotrophic to eutrophic ones. Also Kolada & Kutyła (2016) showed that E. canadensis is most frequently accompanied by phytocoenoses of submerged macrophytes typical of mesotrophic and meso-eutrophic conditions. On the other hand, Meilinger et al. (2005); Dodkins et al. (2012) as well as Samecka-Cymerman & Kempers (2003) reported that Canadian waterweed belongs to nutrient tolerant taxa and is associated with high nutrient enrichment. Thriving in eutrophic water is probably enabled by its tolerance to low light intensity, even though it also tolerates high light without photoinhibition (Madsen et al. 1991).

In the studied rivers, concentrations of trace metals were generally at low levels (Table 2) and were in ranges characteristic of clean water bodies (Kabata-Pendias 2011), although the content of metals in water from some study sites was very high, e.g.: 3.28 μg Hg l⁻¹ (the Meszna river, site Kąty, 2006), 7.70 μg Cd l⁻¹, 32.5 μg As l⁻¹, 33.0 μg Pb l⁻¹, and 101 μg Cu l⁻¹ (the Kania river, site Gostyn, 2006), 245 μg Zn l⁻¹, 409 μg B l⁻¹, and 492 μg Mn l⁻¹ (the Przemsza river, site Będzin, 2006). These metal concentrations were significantly higher than those determined in water of waterweed stands in Poland (Samecka-Cymerman & Kempers 2003) and in north-eastern France (Thiébaut et al. 2010).

The Detrended Correspondence Analysis (DCA) ordination of the macrophyte species is presented in Fig. 5 The first axis (λ₁ = 0.303) can be identified with the kinetic energy of water (current velocity), which directly affects the degree of riverbed material fragmentation and the observed types of flow. Aquatic bryophytes and some filamentous algae (e.g. Hildenbrandia rivularis) as well as vascular macrophytes associated with the community of Ranunculion fluitantis (genera Batrachium and Callitriche) prefer more turbulent and fast flows, while floating-leaved macrophytes and pleustophytes – no flow (Fig. 5). The second axis (λ₂ = 0.248) was correlated with the concentration gradient of nutrients in the water of the surveyed rivers.

Figure 5

E. canadensis occurs most often with vascular macrophytes associated with slow-flowing rivers with sandy bottom material, indicating mesotrophic and eutrophic water (Fig. 5). In the studied material, the Canadian waterweed occurred accompanied by 105 taxa of macrophytes, including 13 macroscopic algae, 11 bryophytes and 81 vascular plants. The most dominant emergent species were (Fig. 5): Alisma plantago-aquatica, Berula erecta, Butomus umbellatus, Iris pseudacorus, Lycopus europaeus, Lysimachia vulgaris, Mentha aquatica, Myosotis palustris, Phalaris arundinaceae, Ranunculus sceleratus, Rorippa amphibia, Scrophularia umbrosa, Sparganium emersum, S. erectum, and Veronica anagallis-aquatica. Among submerged plant species, Batrachium circinatum and Potamogeton obtusifolius were most common, while among pleustophytes, Lemna minor dominated. The filamentous algae Cladophora sp. also occurred, a tolerant taxon which is an indicator of eutrophication (Haury et al. 2006; Szoszkiewicz et al. 2010b).

The average contribution of E. canadensis in the lake phytolittoral varied between 2.3 and 5.5% (Kolada & Kutyła 2016) and was similar to its average contribution in the studied rivers (5.9%), although the species was more common in lakes (40% of the studied lakes) studied by Kolada & Kutyła (2016) than in rivers (26% of the studied river sites). The reason may be greater heterogeneity of habitat conditions in lotic ecosystems.

In terms of abiotic factors, E. canadensis prefers deeper and larger lakes, with a long water retention time, lower trophic status of waters and better ecological status (Kolada & Kutyła 2016). Other studies indicate that Canadian waterweed is common in water depths between 4 and 8 m (Nichols & Shaw 1986). However, Kłosowski et al. (2011) reported that E. canadensis phytocoenoses are associated with the shallowest parts of lakes. In the present study, it occurs more often in small and shallow streams (average depth of 0.6 m and width of 5.5 m). In terms of trophic and ecological status, the results obtained for the studied rivers and Portuguese rivers (Dodkins et al. 2012) were similar to those obtained for lakes – mesotrophic to eutrophic conditions and at least good ecological status (53% of all sites) (Kolada & Kutyła 2016). It should be added, however, that the conductivity and concentration of nutrients in water were higher in rivers than in lakes – the conductivity was on average 449 μS cm⁻¹ in the studied rivers and 303 μS cm⁻¹ in lakes (Kolada & Kutyła 2016), whereas the content of phosphorus was 0.19 mg P l⁻¹ in rivers and 0.043 mg P l⁻¹ (Kolada & Kutyła 2016) in lakes, respectively. Also alkalinity was slightly higher in rivers, on average 165 mg CaCO₃ l⁻¹, while in lakes it was 130 mg CaCO₃ l⁻¹ (Kolada & Kutyła 2016). On the other hand, average pH was higher in lakes – 8.3 (Kolada & Kutyła 2016) and 7.76 in rivers.

In the analyzed Polish lakes, 78 hydrophyte communities were identified. Elodea canadensis forms a compact plant community in the vicinity of other submerged phytocoenoses (elodeids and charids). The most common are plant communities such as: Charetum fragilis, Charetum rudis, Myriophyllum alterniflorum, Potametum alpinii, P.compressi, P.lucentis, P. obtusifolii, P. pusilli, Ranunculetum circinati (Kolada & Kutyła 2016). Macrophytes in rivers were assessed based on the Macrophyte Method for River Assessment, i.e. based on species rather than plant communities, so direct comparison with lakes is not possible. In addition, charids are a very important group in lakes (represented by e.g. Charetum fragilis and Charetum rudis), especially those with at least good ecological status, preferred by Elodea canadensis, while they are very rare in rivers. Only Chara globularis (syn. Chara fragilis) from this group was found in the studied rivers, at 4 sites. In total, 105 taxa of macrophytes were identified in rivers, including 81 vascular plants. Most of the studied rivers were dominated by relatively shallow sections (0.13-1.65 m, mean 0.59 m), therefore, the Canadian waterweed was accompanied by numerous emergent species – rushes, in addition to the submerged macrophytes (e.g. Batrachium circinatum, Potamogeton obtusifolius). The most important of them, occurring in at least 50% of the Elodea canadensis sites, were: Berula erecta, Mentha aquatica, Myosotis palustris, Sparganium emersum and Veronica anagallis-aquatica.