Rediscovery of Cyperus flavescens (Cyperaceae) on the northeast periphery of its range in Europe

Introduction

Cyperus flavescens L. (= Pycreus flavescens (L.) P.Beauv. ex Rchb.) is a species with a broad cosmopolitan range (Meusel, Jäger & Weinert, 1965; Jiménez-Mejıas, Luceño & Peterson, 2011). It is listed as an endangered taxon in many places in Europe, and is extinct in some regions (Bilz et al., 2011; Lansdown, 2013). In Europe, habitats of C. flavescens are widespread, especially in the south of Europe, in places with a high concentration of nitrogen compounds and mineral salts that are usually classified to the Isoëto-Nanojuncetea class (Ter Braak & Smilauer, 2002). C. flavescens usually grows on the edges of lakes and periodically flooded places, where the water level is variable and probably suppresses competition. It also occurs on wet and marshy meadows, on the banks of rivers, waterholes and on old rice fields (e.g., Moor, 1936; Pietsch, 1973; Brullo & Minissale, 1998).

Central Europe represents the northeastern periphery of this taxon’s actual range in Europe (Hultén & Fries, 1986). It is mentioned on red lists of Central European countries: it is critically endangered in the Czech Republic (Grulich, 2012), strongly endangered in Germany (Ludwig & Schnittler, 1996), and vulnerable in Slovakia (Maglocký & Feráková, 1993). The northernmost occurrences of Cyperus flavescens are in Poland, where it is classified as a species vulnerable to extinction (Popiela, 2005; Popiela, Łysko & Konopska, 2014), similar to its status in Germany and the Czech Republic (Bettinger et al., 2013; Dřevojan & Šumberová, 2016). In Poland, the species is associated with natural habitat 3130, namely flooded muddy river banks, and on sandy, periodically flooded, often pastured and trampled banks of rivers, oxbow lakes and lakes. Phytocenoses with C. flavescens are rarely recorded in Central Europe and this habitat is treated as disappearing (Pietsch, 1973; Pietsch & Müller-Stoll, 1974; Popiela, 1997; Šumberová & Hrivnák, 2013).

In recent years, three large populations of Cyperus flavescens were found by the authors of this research (J.M. P.M., Ł.K.). They are the richest species occurrences in over 30 years. Thus, the aim of this research is to determine (1) the habitat factors that determine the mass occurrence of C. flavescens, and (2) the present situation of this endangered species and its habitat in Central Europe measured against its potential range.

Methods

Ecological research

In 2019 three populations of Cyperus flavescens were studied (Fig. 1):

(1) B¹ - located in the village of Bindużka (21.3350 long., 52.7921 lat..) on the extensive floodplain terrace of the Narew River (NE Poland), partly occupied by a large oxbow lake connected to the river. The area is used as an extensive pasture for cows. The population of C. flavescens occupies an area of 300 m². The estimated population size is more than 10,000 individuals. C. flavescens occurs in a zone, strongly disturbed by grazing and periodic breakdown due to overdriving by community of the Cyperetum flavescentis.

(2) GH¹ - Grobla Honczarowska (22.5550 long., 53.3055 lat.), in the Biebrza River marsh, Biebrzański National Park (NE Poland), many thousands of C. flavescens individuals occupy 7 km along a heavily used road;

(3) DG¹ - Dąbrowa Górnicza (S Poland, (19.2370 long., 50.3606 lat.): (a) Kuźnica Warężyńska sand pit, in the eastern part, within the boundaries of PLH240037 “Lipienniki w Dąbrowie Górniczej”, there are many thousands of individuals on sandy roads, roadsides and other periodically flooded sites, populations also occur on the mineral shores of an artificial lake, (b) a beach on the northeast shore of the lake Pogoria I - in a sandpit that was flooded around 1943. The site was proposed for protection as a 3130 habitat in the Natura 2000 site “Lipienniki w Dąbrowie Górniczej”; total area (a) and (b) around 3,400 m² with tens of thousands of Cyperus flavescens specimens in patches of Cyperetum flavescentis.

In each population, eight 1-m² research plots were chosen, as far apart as possible. The plots were selected in such a way that: two plots were 60%–100% covered with Cyperus flavescens, two plots where the species covered 20%–60%, two plots where the C. flavescens coverage was below 20%, and two plots without C. flavescens. In each plot a relevé was made. In addition, with a soil stick, five probes were collected in each plot, then combined in a bag as a collective sample.

The soil samples were air-dried, disaggregated and sieved through a 2-mm mesh to separate skeleton and earth mass. In the earth mass of the samples (without skeleton) the following properties were analyzed by methods commonly used in soil science: texture (particle size composition) was determined using the aerometric method of Casagrande in the Prószyński modification; the textural class names were determined using the classification of the USDA (United States Department of Agriculture); pH in KCl was determined by the potentiometric method; electric conductivity (EC) of a soil suspension with a 1:2.5 soil/water ratio was determined by conductometric method; the content of organic carbon (C org) was determined by the Tiurin method; the content of total nitrogen (C tot) was determined by the Kjeldahl method; the content of available forms of phosphorus and potassium (P₂O₅, K ₂O) was determined by the Egner–Riehm method; the content of available forms of magnesium (Mg) was determined by the Schachtschabel method.

To determine the correlation between the occurrence and abundance of species and the chemical parameters of the soil, Juice v.7.0.213 software and the Detrended Correspondence Analysis (DCA) and Canonical Correspondence Analysis (CCA) methods created by Canoco 4.52. were used. Initially DCA analyses were performed. Because of the small dataset, a selection method by segments without transformation method was used. Cyperus flavescens received a higher weight in the analysis (2.0), and other species received a 1.0.

The unimodal CCA method with inter-species distance and biplot scaling (Lá) was used to check the relationship between the occurrence of species and environmental variables. Because of the small dataset, species were not transformed, and C. flavescens had a high weight applied in the analysis (2.0 whereas other species had a weight of 1.0). To indicate the weight of individual environment variables, Monte Carlo permutation tests (n = 499) with restricted spatial structure were used.

Moreover, the Detrended Correspondence Analysis (DCA) of environmental Ellenberg values was made for all relevés published in Poland (45 relevés) that were performed in the years 1961–2018 (Śliwińska, 1961; Faliński, 1966; Fijałkowski & Kozak, 1970; Hereźniak, 1972; Fijałkowski, 1978; Ochyra, 1985; Cabała, Paul & Piątek, 2004; Nobis & Michalewska, 2004), and personal observations of the authors (J Marciniuk, P Marciniuk 2018–2019., Ł Krajewski, 2018–2019; see Table 4) using Juice 7.0 software. In addition, 213 and R statistical environment with the Vegan package were used. For interpretation of the main environmental gradients, average non-weighted Ellenberg indicator values for light, temperature, continentality, soil reaction, moisture and nutrients (Ellenberg et al., 1992) were used as supplementary variables.

Results

In jackknife MaxEnt analysis based on bioclimatic variables to the year 2017, the environmental variable with highest gain when used in isolation is BIO1 (Annual Mean Temperature), which therefore appears to have the most useful information by itself.

The average training AUC for the replicate runs is 0.8868, and the average standard deviation is 0.0037 (Fig. 2).

The result of the analysis for future bioclimatic variables using models from the years 2021 to 2040 is similar. The environmental variable with highest gain when used in isolation is BIO1, which therefore appears to have the most useful information by itself (AUC 0.8875, average standard deviation is 0.0036).

A map of the potential distribution of Cyperus flavescens in Central Europe, created according historical and future data, shows an extension of the range of potential habitats to the north and east (Fig. 2).

An examination of the correlation between the occurrence and abundance of species and the chemical parameters of the soil showed that the gradient length represented by the first consulting axis is 3.506 SD. This means that along this gradient, the species does not decompose in the full spectrum of the Gauss curve. This result showed that the dataset can be analyzed using linear and unimodal techniques (Jongman, Ter Braak & Van Tongeren, 1987; Brullo & Minissale, 1998). The first axis is the most important; it takes into account the differences in sets and explains 20.3% of the variability; the second axis explains 10.8% (Table 1, Fig. 3).

The analyses made shown that the substrate for Cyperus flavescens growth and development on the three examined locations was loose sand (sand acc. USDA). The contents of the granulometric fractions were: 91%–98% for the sand fraction, 1%–8% for the silt fraction, and 0%–2% for the clay fraction. Sands are very dry substrates; they are periodically wet if they are located in terrain with depressions, valleys, or floodplain areas. Soil pH values most often indicated an alkaline character, and less often, a neutral or slightly acidic character (Table 2). The EC values were quite variable within the locations; however, they indicated a low level of soil salinity. The soils were very poor in available potassium, and often poor in available phosphorus. However, the available magnesium content was often very high and high.

Canonical correspondence analysis (CCA) as well as the Monte Carlo permutation show that the Ec and C org variables had the most correlation. Therefore, Ec was deleted from further analysis. Only pH and P₂O₅ were statistically significant, which explains the variability of the species in individual phytosociological images (p ≤ 0.05; pH = 13.46% variation; P₂O₅ = 8.33%) (Table 2).

The result of the CCA analysis, which evaluates the relationship between species and soil chemical parameters, indicates that the most important variables for the analyzed dataset are pH and P₂O₅, which indicate strong alkalization of the habitat. However, the P₂O₅ and EC associated with the first consulting axis are responsible for highest species variability in the habitats in total; this axis explains 41% of species diversity in the study areas. The second consulting axis is most strongly associated with pH and with K₂O and Ntot. This axis accounts for 22% of variation (Table 3). Cyperus flavescens is placed very close to the center of the axis, which indicates that all analyzed traits are important for this species and none has an advantage over another. However, two groups of species were distinguished: one was strongly associated with variable pH and the other was associated with P₂O₅ content (Fig. 4).

The result of DCA analysis on Ellenberg values indicates that the longest gradients are temperature, moisture and nutrients (SD of the axis: DCA1: 4.5; DCA2: 4.2). Values of the species are clustered and concentrated mostly with the first ordination axis (moisture), and with the second axes (temperature and nutrients) (Fig. 5).

The analysis of the vegetation data available from Central Europe and involving Cyperus flavescens (Table 4) indicate that this species occurs mainly in the company of Juncus bufonius and Plantago intermedia. Other species of the Isoëto-Nanojuncetea class rarely appear. The list of species is complemented by taxa associated with the classes Molinio-Arrhenatheretea, Bidentetea, Phragmito-Magnocaricetea and Scheuchzerio palustris-Caricetea fuscae, where the highest constancy is achieved by Juncus articulatus and Agrostis stolonifera.

Discussion

The results of the potential coverage modeling analysis of Cyperus flavescens show that Central Europe is at its northern end of potential occurrence (average AUC, > 0.75). It is assumed that species located on the edge of their area have a narrower ecological amplitude than the central populations (Brown, 1984). This could be a result of a “relative constant habitat rule” (Walter & Walter, 1953), i.e., a change of habitat as a compensation of climatic changes and maintenance of possibly constant environmental conditions. That change, or because of a lower genetic variability of the border populations, leads to a narrowing of the ecological amplitude and of accessible ecological niches (Van Valen, 1965). However, the results of some research enable the formulation of a different hypothesis: taxa with a lager distribution have broader ecological amplitudes on the northern borders of their geographical areas, which seems to be associated with reduced competition (Diekmann & Lawesson, 1999). In addition, higher extinction risks of species at the edge of their range are often mentioned (e.g., Bahn et al., 2006). Over the last 100 years, for reasons that are unclear, C. flavescens has lost about 90% of its locations in Central Europe (Bettinger et al., 2013; Popiela, Łysko & Konopska, 2014; Dřevojan & Šumberová, 2016). From descriptions in the literature and information on herbarium specimens’ labels (Popiela, 1996), it appears that in Poland, mainly sites on the lake shores disappeared, because of what could be linked with the progressive eutrophication of waters during the 20th century (Fig. 1). Moreover, nowadays, cattle do not regularly graze near the lakes, which results in a lack of trampling, initial places, deprived of rushes reaching the water. Such local extinction could mean the beginning of the disappearance of the species throughout Europe past of its range. Another problem is the difficulties in the field research: the taxon is very small and annual, hard to find, and its growth areas fluctuate (it appears in different locations); hence, it can be easily overlooked by researchers. For these reasons, C. flavescens is included in the IUCN red list of endangered species in Europe with the category of Lc (the lowest risk) because of its wide range and lack of identified global threats (Lansdown, 2013). However, extinction of this species has been observed in some European countries, especially where it has always been rare (Hodálová, Feráková & Procházka, 1999; Lansdown, 2013; Nikolić & Topić, 2005; Popiela, Łysko & Konopska, 2014; Kaźmierczakowa et al., 2016).

The number of Cyperus flavescens ‘ individuals within populations existing in Poland was always low (from a few to 300), and the species occupied small areas (Popiela, Łysko & Konopska, 2014; Krawczyk et al., 2016). In this context, the currently studied populations (B, HG, DG) might be the largest in the northeast periphery of the potential range of the taxon; hence, the current research is important for a better understanding of the ecological requirements and preservation of this species.

In the light of this research, none of the examined soil factors is decisive for the mass occurrence of Cyperus flavescens. The species is placed very close to the center of the axis, which indicates that all analyzed traits are important for this species and none has an advantage over another (Fig. 4). However, temperature, moisture and nutrients are important factors for the environment in which it occurs (Fig. 5). This may mean that the environmental factor affecting the occurrence of C. flavescens is different from the tested factors. It may turn out that anthropogenic factors, as well as soil properties (see the Results section) have a positive effect on the occurrence of the species. Population B occurs on roads periodically driven by vehicles and on the shores of a large oxbow of the Narew River, which are regularly edged and grazed by cattle. The habitat is flooded during the spring swells of the river and then remains dry during the summer and autumn. C. flavescens occurs in large numbers, achieving maximum density in quite dry and disturbed places. In GH, C. flavescens occurs abundantly, but the population develops before the mowing season of the neighboring sedge meadows (late summer); at the time of the mowing, the route is intensively used. Similarly, in the DG, the species is enhanced by anthropogenic disturbances: it occurs mainly in the sands on dirt roads that are periodically inundated after heavy rains and disturbed by vehicles, thus developing only a thin mud layer, at most. This environment allows for patches of pioneer vegetation. C. flavescens is also present on the sandy shore of an artificial water reservoir in an old sand quarry that is regularly used as a recreational beach (city swimming place); the taxon is most abundant here before and after the summer holidays (see also Krajewski, 2016).

The occurrence of the Cyperus flavescens seems to be strongly dependent on zoopressure, which maintains the initial nature of its habitat. It might be that under completely natural conditions, the occurrence of C. flavescens is limited to the rather heavily pitched waterways used by large herbivorous animals; the positive effect of herbivores on the population of C. flavescens was observed also by Taylor et al. (1994). In B, the replacement for wild herbivores is cattle. In the whole Narew River bend, grazing animals have wide access to the oxbow lake and the river, which means that the pressure is probably too weak. Therefore, the population of C. flavescens is limited to the small portion of the habitat used as a periodic (only at very low water levels) path through the oxbow lake. In B, this additional pressure seems to be a necessary factor for the stability of the population. It seems, that zoochory/anthropochory is also an important factor of spreading. For example, in GH, C. flavescens occurs along several kilometers of road, but it was not observed in neighboring plant communities. Similarly, in DG, the taxon appears most often at sandy roadsides (Krajewski, 2016). The species benefits from accidental transport of seeds to new habitats along roads, and in optimal conditions, the seeds subsequently germinate.

Habitats with occurrences of Cyperus flavescens are widespread in the south of Europe. The species occurs in places with a high concentration of nitrogen compounds and mineral salts, as well as on relatively short flooded and fairly compacted habitats, including anthropogenic ones, e.g., roads. The taxon occurs in many associations with the Isoëto-Nanojuncetea class (e.g., Brullo & Minissale, 1998; Ninot et al., 2000; Niculescu, 2016). In the Central Europe, unlike in its central range in Europe (see Brullo & Minissale, 1998), C. flavescens is not present in various associations but creates its own community, strongly referring to the Cyperetum flavescentis (Koch, 1926) association. Centaurium pulchellum is often missing in communities developing in Central Europe. In addition, the participation of the Radiolion linoidis alliance and the Eleocharition soloniensis alliance character species is small (Table 4) compared with the patches in the class’s central range (see Brullo & Minissale, 1998); like the other syntaxa from Isoëto-Nanojuncetea, Cyperetum flavescentis appears here in an impoverished form (Popiela, 2005). Libbert (1932) described patches of Carex viridula and Centaurium pulchellum from Western Pomerania (NW Poland), which he considered a very poor form of Cyperetum flavescentis.

Until now Cyperetum flavescentis has been rarely recorded in Central Europe, and this habitat has been treated as disappearing; patches of growth usually developed on periodically flooded lakes with crushed and grazed vegetation, peat pits, and the edges of ponds and streams (Pietsch, 1973; Pietsch & Müller-Stoll, 1974; Popiela, 1997). Among the 94 relevés listed in the synoptic (Table 4), 30 relevés (31%) were taken in Poland in the years 2004–2018 at new locations, meaning that habitat 3,130 is renewing and Poland can be the center of its preservation in this part of the continent. It should be noted that, so far, Cyperetum flavescentis were documented in Poland with only 10 relevés taken between 1932 and 1985 (Popiela, 1997; Popiela, 2005; Popiela, Łysko & Konopska, 2014). These communities probably no longer exist. Analysis of the ecological characteristics of this association occurring on the periphery of the potential range of Cyperus flavescens is problematic because the number of relevés is still small; however, analysis of Ellenberg indicator values has shown clear clustering of species mostly concentrating in areas with high moisture, temperature and nutrient axes (Fig. 5).

Conclusion

Our research did not allow us to definitively resolve the reason for the rebirth of Cyperus flavescens in Poland. None of the examined soil factors namely pH, P₂O₅, EC, Mg, K₂O, N tot, C org, C/N) is decisive for the mass occurrence of Cyperus flavescens. It may be an effect of anthropogenic changes (narrowing of the ecological amplitude) on the periphery of its potential range of the habitat, as has already been seen in other species of the Isoëto-Nanojuncetea class, namely Radiola linoides, Centunculus minimus, and Illecebrum verticillatum (Popiela, 2005). It is possible the species is enhanced by very warm summers with only short periods of heavy rain, which are now observed in Poland more often than ever before. Especially regarding the results of the MaxEnt analysis, the most important variable is BIO1 (Annual Mean Temperature). Over the years, the species has lost many habitats throughout Poland, and it is now limited to a small area of the country. The taxon’s range could be re-established in favorable climatic conditions and with certain forms of anthropopressure in the agricultural landscape. The problem of C4 species promotion caused by global warming has been raised many times (e.g., Lattanzi, 2010). Likewise, results of the comparative analysis of MaxEnt analysis using historical data and a predicted future model CM6-SSP3-7.0-bc indicate that the range of Cyperus flavescens in Europe will increase by 2040 as a result of global warming, especially in the north and east.

Supplemental Information