Plant replacement trend in soft-water lakes with isoetids

The study was conducted between 1998 and 2018 in 70 soft-water lakes of northwestern Poland (Fig. 1), where 12 702 plant samples and 38 106 water samples were collected (the latter were collected in places with plants). In each of these lakes, one transect perpendicular to the shoreline was delineated and divided into strips parallel to the shore at 1 m depth intervals. They are referred to as depth zones throughout this study (Fig. 2). In each of the depth zones, a diver randomly delineated 10 to 20 square fields of 0.1 m², registered all plants present in these fields, determined their abundance (expressed as % of a sample). Data from such fields are called plant samples in this study.

Figure 1

Figure 2

From each of the bottom zones or depth zones with 1 m intervals, the diver collected three water samples (a total of 1.5 dm³) 5 cm above the plants or sediment. The water samples were analyzed chemically and physically according to the methods proposed by Hermanowicz et al. (1999) and Eaton et al. (2005). The concentration of calcium, total nitrogen, total phosphorus and humic acids, as well as the hardness of water and its color, pH, redox and conductivity were determined. During the fieldwork, photosynthetically active radiation (PAR), water clarity, water temperature and oxygenation were measured.

The environmental factors were determined in the depth zones as per the following parameters:

The obtained data were entered into a spreadsheet allowing also for species abundance in plant samples and data describing the environmental factors in the depth zones. For plant samples, species diversity was calculated using the H’ index (Shannon & Weaver 1949; Huston 1996) according to the formula: H’ = –Σ[(mj/m) × log mj/m], where: mj – abundance of species j in a sample [%], m – abundance of all plants in a sample. The frequency of species in plant samples (F, in %; 100% = 12 702 samples), as well as the frequency of species groups (mosses, stoneworts, vascular plants, isoetids) and the contribution of specific species to the vegetation as a quotient of the abundance of a given species and the abundance of other species in samples (S, %) were also calculated.

Relationships between the relative abundance of a given species in the vegetation and pH of water near the sediment were determined by the Kruskal–Wallis non-parametric test (Sokal & Rohlf 2012). The frequency of species was calculated from the quotient of their occurrence in the samples. The relationship between the frequency and abundance of plants and the environmental factors was determined by multiple regression and canonical analysis (Gittins 1985; ter Braak 1987). The data included in the analyses were standardized in CANOCO 4.5 (ter Braak & Šmilauer 2002).

Detrended correspondence analysis (DCA) showed that all data on plant abundance and water properties were linear in their structure and represented only a part of the Gaussian distribution (eigenvalues < 2 SD), which required the use of redundancy analysis RDA (ter Braak & Šmilauer 2002). By using the redundancy analysis (RDA; ter Braak 1986; 1987; 1988), we correlate the frequency and abundance of plants with the physicochemical properties of water. Before the analysis, Spearman’s rank correlation coefficient (R) was calculated to determine strongly correlated variables. And thus, the following parameters were removed from the RDA analysis: water color, as it was strongly correlated with water clarity (R=0.78), depth and temperature, as they were strongly correlated with PAR (R = 0.74), calcium concentration and water hardness, as they were strongly correlated with water conductivity (0.85 and 0.89, respectively). The data were standardized and focused, yet not transformed, on species. The statistical significance of the environmental variables was tested using the Monte Carlo permutation test with 999 permutations.

A total of 71 species of macrophytes were identified in 12 702 plant samples collected from 70 soft-water lakes with isoetids: 41 vascular plants, including four isoetids, 19 mosses and 11 stoneworts (Table 1). Twenty-two plant species occurred with a frequency (F) higher than 1% in the total number of samples (Table 1). The most frequent were vascular plants (present in 6481 samples; F = 51.0%), followed by isoetids (4159; F=32.7%), bryophytes (3789; F=29.8%) and stoneworts (2432; F=19.1%).

Mosses are the main constituent of submerged vegetation in acidic lakes (pH 4.0–5.5), with usually one species prevailing in a sample (1.0 ± 1.0 species 0.1 m⁻²). The moss cover is very dense and may reach more than 90%, especially when pH is low. When pH of water is 4, the moss cover is 30.2±41.0%. In waters with neutral to alkaline pH, the number of moss species and their cover gradually decrease. Moreover, species of the genus Sphagnum are replaced by other mosses, such as Fontinalis antipyretica (Figs 3 & 4). Bryophytes do not occur when pH of water becomes higher than 8.

Figure 3

Figure 4

Vascular plants occur throughout the whole pH range (4.0–9.5). In acidic lakes, the number of species is small, on average 0.6±0.4 species 0.1 m⁻². However, as pH increases, the number of species gradually increases and reaches the highest level in alkaline lakes (pH>8.5). The number of species in such water bodies increases up to 1.9 ±1.1 species 0.1 m⁻². The cover of vascular plants in acidic waters is similar to other plant groups and increases only in alkaline waters, reaching the highest values (40.8±37.8%) at pH=9.

Like other vascular plants, isoetids occur at pH 4.0–9.5, but the largest number of species (0.7 ± 0.8 species 0.1 m⁻²) occurs in the pH range of 5.5–6.7. One of the isoetid species (Isoëtes echinospora) occurs only in acidic water (pH 4.5–6.0), while the others occur in a much wider pH range. One of the four identified isoetid species, Littorella uniflora, forms large populations also in highly alkaline conditions (pH 9.0; Figs 4 & 5).

Figure 5

Few stonewort species occur in soft-water lakes (0.6 ± 0.7 species 0.1 m⁻²). The largest populations of these plants occur at pH 6.5–8.5 (Fig. 3). Stoneworts do not occur in low pH waters, while their relative abundance in submerged vegetation is negligible when pH is higher than 8.5. The abundance of stoneworts is small and they reach the highest cover (11.4 ± 25.6) in waters with neutral or alkaline pH (7–8.5).

Plant species diversity expressed by the H’ index is small (Fig. 6). The H’ values along the increasing pH gradient of water are contained within a narrow range of 0.2–0.4, but the differences are statistically significant (p<0.001). The most statistically significant differences occur in low pH values (pH from 4 to 6), and the number of statistically significant differences decreases with increasing pH (Table 2). However, there is no trend in the variation of H’ in the pH range of 4.0–9.5. The low value of the H’ index results mainly from the small and fairly similar number of species in the samples and the similar structure of submerged plant communities throughout the gradient, usually dominated by a single bryophyte or vascular species.

Figure 6

The pH of water is a characteristic of the aquatic environment of soft-water lakes with isoetids, which significantly affects the occurrence and abundance of submerged plant species (Fig. 7, Table 3). Other characteristics determining the conditions for the occurrence of plants in lakes, such as PAR, conductivity and water clarity, should also be noted.

Figure 7

The Monte Carlo Test showed that PAR and pH account for most of the variance (lambda 1 in Marginal Effects is 0.06 and 0.02, respectively), but once the environmental variables are successively included in the model (Conditional Effects), PAR accounts for 0.06 of the variation in the vegetation traits, temperature accounts for 0.05, while pH accounts for 0.02 of the variation. The remaining environmental characteristics, although statistically significant, are less important.

The occurrence of isoetids as a species group is largely affected by factors correlated with the first axis (mainly PAR), but it is the factors correlated with the second axis, mainly pH, that have a decisive impact on the occurrence of individual species. Littorella uniflora is positively correlated with pH, while Isoëtes lacustris correlates negatively.

As far as other species groups are concerned, mosses are associated with high water acidity, while stoneworts are associated with low acidity. The effect of pH on the number of mosses and stoneworts is greater than on the abundance (cover) of their populations. The above analysis indicates that the main course of species turnover along the water pH gradient involves the replacement of mosses by stoneworts, leading to a minor transformation within the group of vascular plants, including isoetids.

Mosses occur with the highest frequency in acidic (pH 4.0–6.5) water (Figs 3, 4). Sphagnum denticulatum dominates (frequency 69.3%) at pH 4.0–5.0 and is often accompanied by Warnstorfia exannulata, less frequently by Sphagnum fallax (frequency below 12%). Isoetids are common in such conditions, especially Isoëtes lacustris (F = 28.5%), Lobelia dortmanna (16.1%) and Luronium natans (2.9%).

The first signs of species turnover were observed at low water acidity – pH 6.6–7.5. In plant samples from such lakes, an increase in the frequency of mosses other than those mentioned above, mainly of Drepanocladus sordidus, Fontinalis antipyretica and F. delecarlica, was observed. Their frequency is high in the pH range of 6.0–7.5, while the contribution of mosses to the structure of submerged vegetation is relatively small, in terms of phytolittoral coverage, and does not exceed several percent.

The second stage of species replacement was identified in waters with relatively high alkalinity (pH 7.5–9.5). Almost all mosses, as well as isoetids and stoneworts were replaced (Figs 3, 4). Their habitats were occupied by vascular plants. At pH close to neutral, Elodea canadensis (F=17.2%) and Myriophyllum alterniflorum accounted for the largest relative abundance; their coverage can be as high as 54.9%, especially when pH of water is very high. At water pH ≥ 8, other species of vascular plants may occur, such as M. spicatum or two pondweed species (Table 1). Interestingly enough, Myriophyllum alterniflorum, as a species typical of lobelia lakes continues to dominates. Stoneworts occur when pH is 6.5–8.5, but their highest relative abundance is observed in waters with pH above 7. The frequency of Chara delicatula, Chara globularis and Nitella flexilis was very similar (below 12%), while the contribution of Nitellopsis obtusa was much smaller and did not exceed 5.9%. Nitella gracilis associated with neutral waters (pH 6–7.5) occurred with a low frequency (F ≤ 1.1%). Stoneworts are a group of plants occurring at pH ranging from 6.5 to 8.5, which implies their narrow ecological range. Only one of all isoetids is capable of growing in water with high reaction, i.e. Littorella uniflora, which occurs more often in waters with neutral or alkaline reaction and its frequency may be high (17.7%) even at water pH equal to 9. The occurrence of other isoetid species is limited to waters with low reaction. It is also noteworthy that the analyzed H’ index does not show the total species replacement along the pH gradient. Individual species of bryophytes and isoetids, such as Isoëtes lacustris and Littorella uniflora, are eliminated and replaced by eutrophilous species of vascular plants.