OR/14/043 Discussions

Farr, G, Graham, J, and Stratford, C. 2014. Survey, characterisation and condition assessment of Palustriella dominated springs 'H7220 Petrifying springs with tufa formation (Cratoneurion)' in Wales. British Geological Survey Internal Report, OR/14/043.

Paulustriella dominated vegetation: grouping

The vegetation at 14 sites was sampled (comprising 29 sample sites in total). Of these, 2 sites (Cors Erdderiniog Sites 7a and 7f) were excluded as, although they were part of a larger site with Palustriella dominated springs, they individually were not dominated by Palustriella and therefore deemed not to qualify.

Figure 108 lists all species (in decreasing order of frequency) recorded from Palustriella dominated vegetation for these 27 samples along with their frequency (DAFOR). A total of 130 species are recorded comprising 77 flowering plants (Angiosperms), 46 bryophytes (Bryophyta), a club moss (Lycophyta), 2 Leptosporangiate (True) ferns, 3 horsetails (Calamophyta), 2 blue-green algae (Cynophyta) and a single charophyte (Chlorophyta).

Based on the characteristics of sampled Palustriella dominated vegetation, the 27 sites have been loosely divided into 2 main groups and 4 subgroups (highlighted in red and blue in Figure 112).

A first broad Group 1 (comprising 16 samples) is characterised by being dominated by Palustriella falcata and associated mainly with upland spring sites on open hills. The flushed vegetation resulting from these upland springs is particularly species rich and typically has a significant flowering plant component (especially Juncus and Carex species). Two samples dominated by Palustriella commutata and in the absence of P. falcata (Hen Allt Common and Tarren yr Esgob) have been included in this first group as they closely share wider species characteristics.

A second broad Group 2 (comprising 11 samples) is characterised by being dominated exclusively by Palustriella commutata and associated mainly with lowland sites. Apart from Cors Erddriniog (Anglesey), most of these sites are generally not open (being partially shaded) and often rocky (typically occurring within partly wooded sites, wooded gorges or associated with rock outcrops and cliffs). The Palustriella commutata dominated vegetation occurring at these sites is notably less species rich than the Palustiella falcata dominated vegetation of open upland sites, is dominated by bryophytes, has few associated flowering plants generally and often includes flowering plants characteristic of woodland habitat (a consequence of shading). The fewer species overall is likely to be partly a consequence of the rocky nature of many of these sites and an associated high frequency of bare rock or tufa.

Figure 112    Grouping of Palustriella dominated vegetation — 27 samples as two main groups 1 and 2 (highlighted red) and 4 subgroups A–D (highlighted blue).

Subgroup A - 13 samples in 9 sites (Foel Fawr, Moel Garnach, Brest Rhiw, Pont Clydach, Blaen Onnen, Llyn-yr-Fan Fach, Waun Ddu (Craig-yr-Cilau), Hen Allt Common and Tarren-yr Esgob)
Flushed vegetation associated with upland spring sites on open hills almost exclusively dominated by Palustriella falcata. Palustriella commutata may also be present alongside Palustriella falcata at these sites but typically occurs as smaller patches, often associated with spring heads, especially where these are rocky. These flushes are species-rich with a significant flowering plant component (including Juncus acutiflorus, Juncus effuses, Carex flacca, Carex demisa, Anagallis tenella, Cardamine pratensis, Ranuncus flammula, Ficaria verna, Galium palustre ssp. palustre) as well as a broad suite of associated bryophytes (including Fissidens adianthoides, Ctenidium molluscum, Scorpidium cossonnii, Leicolea bantriensis, Campylium stellatum, Jungermannia exsertifolia ssp. cordifolia, Bryum pseutotriquetrum).

Immediate spring heads often have a slightly modified flora typically with a greater proliferation of Palustriella commutata and lawns of dominant Cratoneuron filicinum, Philonotis fontana. Elsewhere smaller hollows or runnels are characterised by a greater domination of Campylium stellatum, dominant stands of Scorpidium cossonii and occasonal stands of Scorpidium scorpioides. Deeper central runnels (especially those with a perminant flow of water) regularly support emergent flowering plants (including Apium nodiflorum, Nasturtium officinale s.l, Potamogeton polygonifolius, Veronicca beccabunga, Mentha aquatica, Glyceria fluitans and occasionally Hypericum elodes.)

The cushion forming moss Hymenostylium recurvirostrum, the liverworts Jungermannia atrovirens, Presisia quadrata and alga Rivularia haematites are often associated with tufa occurring at these sites.

A number of interesting plants were recorded from these sample sites including Pinguicula vulgaris, Moerckia flotoviana, Wahlenbergia hederacea, Philonotis calcarea, Equisetum cf variegatum, Eriophorum cf latifolium and Carex lepidocarpa.

Subgroup B - 3 samples from a single site (Nant Peris)
Flushed vegetation associated with upland spring sites on open hills at altitude (681–689 maOD, Snowdon range) exclusively dominated by Palustriella falcata. These 3 sample sites are very similar to each other and share some species similarities with subgroup A but hava a number of additional northern (and more upland) vascular plant species includimg Montia fontana, Saxifraga sp. and the invasive alien Epilobium brunescens. The 3 sample sites are unique within this survey by comprising dome shaped floating mats of Palustriella falcata dominated vegetation above a spring.

A number of typical Palustriella associated bryophytes occur (i.e. those typically associated with base-rich water) including Crataneron filicinum, Campylium stellatum, Scorpidium revolvens, Fissidens adianthoides, Bryum pseudotriquetrum, Scorpidium, scorpioides, Philonotis fontana despite the water quality actually having a very low base status. In contrast, a number of less markadely calcicole species (Rhizomnium punctatum, Brachythecium rivulare, Plagiomnioum undulatum) were also recorded. These species are not unusual within upland flushes (and indeed were recorded in small quantity from a number of other sites during this survey). However, the presence of Straminergon stramineum (a marked calcifuge) in good quantity mixed with Palustriella falcata and Brachythecium rivulare at 2 of the sites was exceptional.

The floating vegetation mat present at these sites was estimated to be at least 25 cm thick and may of developed quickly as stems of both Straminergon stramineum and Bryum pseudotriquetrum of this length were pulled from the surface that still had some active leaves at their base.

Subgroup C - 7 samples from 4 sites (Cwm Clydach, Ddol, Mc Walters Dingle, Anglesey Beach)
Palustriella commutata dominated vegetation on rocky and shaded sites (including rock outcrops in woodland, wooded river gorges and coastal cliffs) that is typically not species-rich. Other common associated include Pellia endiviifolia, Cratoneuron filicinum, Conocephalum species and Chrysosplenium oppositifolium. The shaded and rocky nature of these sites favour woodland species that occur in small quantity including ferns (Asplenium scolopendrium, Asplenium trichomanes ssp. trichomanes) as well as Hedera helix, Rubus fruticosus, Geranium robertianum and at one site each Carex sylvatica, Hypericum androsaemum.

Many of these sites have tufa formations. However, they have more of a lowland plant assembledge with tufa supporting extensive fruiting cushions of Eucaldium verticillatum, Didymodon tophaceus, Fissidens taxifolius var. taxifoluis, Leicolea turbinata, occasional Bryoerythophyllum recurvirostrum (contrasting with the Hymenostylium, Gymnostomum aeruginosum, Rivularia haemaetes communities on upland tufa — subgroup A). In addition, the only Jungermannia species recorded from these sites was J .atroviens and typically (again in contrast to upland springs — subgroup A) usually occcuerd only in small quantity.

Subgroup D - 4 samples from a single site (Cors Erdderiniog)
The 4 sample sites (B–D) are associated with springs running into a calcareous fen from the east down a gentle slope and sometimes springs running out from woodland at a single site (Cors Erdderiniog).

The spring heads have dominant patches of Palustriella commutata with Conocephalum salebrosum (especially where rocky or wooded) and the Palustriella vegetation then continues along the margins of the spring runnels before quickly being replaced by Camplylium stellatum, Bryum psedotriquetrum and then a rapid transition into tall fen vegetation (dominated by Juncus species) or open water communities. The Palustriella commutata dominated vegetation is associated with Campylium stellatum, Fissidens adianthoides, Cratoneuron filicinum, Ctenidium molluscum, Scorpidium cossonii and locally Scorpidium scorpioides. Palustiella or Campylium stellatum dominated vegetation ofen grows at ground level between tussocks of Schoenus nigricans within the zone of flushing by calcareous spring water. In contrast, the upper tussocks of Molinia (and occasionally Schoenus) received a greater proportion of rain water than calcareous spring water, are without Palustriella and support a small number of calcifuge species (including Erica tetralix, Succisa, Potentilla erecta).

The altitude ranges covered by each vegetation group and subgroup are shown below in Table 111.

Table 111    Altitude ranges for vegetation groups/subgroups

	Minimum maOD	Maximum maOD	Average moAD
Group 1	275	690	426
Subgroup A	275	495	366
Subgroup B	682	690	685
Group 2	5	402	85
Subgroup C	5	402	97
Subgroup D	65	66	65

Conservation value of palustriella dominated springs

Palustriella dominated spring vegetation is an interesting habitat because it requires specific geological and hydrogeological settings and is therefore more restricted than many other vegetation types. In addition, this type of vegetation occurs in small, intricate and often difficult to map stands.

When compared to other vegetation types (notably lowland or calcareous types of grassland) Palustriella dominated springs cannot be considered species-rich. During this survey group 1 sites had a total number of species (22–52), group 2 has a total number of species (6–24) and no species recorded were listed in a category above Least Concern in The vascular plant red data list for Great Britain (Cheffings & Farrell 2005). However, a number of bryophytes associated with these flushes (including Philonotis fontana, Philonotis calcarea, Scorpidium cossonii, Scorpidium revolvens, Scorpidium scorpioides), while still of relatively wide occurrence within Wales, have been lost from many lowland sites (principally heathland sites) in England since 1950 as a result of urban development and land drainage.

Comparison of palustriella dominated samples with the national vegetation classification (NVC)

The European Commission has defined the Annex I habitat H7220 Petrifying springs with tufa formation (Cratoneurion) as ‘springs with active formation of tufa’ and many sources interpret this habitat to be analogous with M37 Cratoneuron commutatum-Festuca rubra and M38 Cratoneuron commutatum-Carex nigra spring communities in the UK (as defined by Rodwell, 1998).

The distinctions between M37 and M38 are subtle and difficult to separate for the following reasons:

• M37 and M38 are bryophyte dominated communities and field data for which the NVC is based did not comprehensively sample bryophytes
• Field data for which the NVC is based does not separate the two species Paulustriella commutata and Palustriella falcata
• Stands of M37 and M38 are typically small in extent and both communities are under sampled within the UK
• Characterisation of M37 and M38 has a strong basis on sward structure (including openness and grazing pressure) which is subjective and more difficult to measure compared with species composition of the sward

However, based on a combination of field observation, literature search, NVC descriptions and floristic tables for M37 and M38 (Rodwell) and a summary of M37 and M38 (JNCC, 2001), a number of generalisations have been made to separate the two communities. These are summarised in Table 112.

Table 112    Generalised differences between M37 and M38.

The generalisations for M37 and M38 (Table 112) along with TableFit (statistical computer software developed by Dr Mark Hill CEH) have been applied to the full data set for Palustriella dominated vegetation (Table 108) and there is strong correlation between Group 1 and NVC community M38 and between Groups 2 and NVC community M37 (as displayed in Table 113). However, it must be stressed that this is a generalised assessment as the collected species data is based on walk over site surveys (DAFOR) and did not include quadrat data.

Two other NVC communities were noted as occurring in small extent within Palustriella dominated vegetation but were not mapped separately. M10 Carex dioica  — Pinguicula vulgaris mire was noted in small quantity in at least 2 sites: with Pinguicula vulgaris at the margins of Palustriella dominated vegetation at Foel Fawr Site 2; with Pinguicula vulgaris and Carex lepidocarpa at the margins of Palustriella dominated vegetation at Hen Allt Common. Also, M29 Hypericum elodes  — Potamogeton polygonifolius soakway was noted associated with open water of central runnels within Palustriella dominated vegetation at several sites: with Potamogeton polygonifolius at Blaenonnen and Wun Ddu (Craig y-Cilau): with Potamogeton polygonifolius and Hypericum elodes at Pont Clydach Site 3.

Table 113    Generalised differences between M37 and M38.

Species altitiude ranges

Elevation data in meters above Ordnance Datum (maOD) was collected at each site and a simple comparison with the species data allows Box plots to be created. Two Box plots are shown, the first (Figure 113) shows that Palustriella falcata is absent from all lowland sites, this corresponds with the definition of Group 1. Palustriella commutata inhabits a wider range of altitudes in this study, occurring in both lowland and upland sites.

These different altitude preferences for Palustriella falcata and Palustriella commutata appear to be an important ecological distinction between vegetation groups 1 and 2 and corresponding NVC communities M38 and M37 although NVC floristic tables for NVC communities do not separate the two species. In addition, much existing bryophyte literature highlights the different ecology of these two species in terms of altitude as well as other parameters such as ‘openness’ of habitat and degree of base enrichment of habitat. The following literature is cited as examples of this:

P. falcata (Hypnum falcatum) occurs on ‘bogs, principally at higher elevations than H. commutatum’, Dixon 1896

P. falcata ‘would appear to be less exacting in requirements’ (than P. commutata) ‘occurring more generally in moorland flushes and mountain springs’, Watson 1968

P. falcata ‘is more common’ (than P. commutata) ‘in open habitats and shuns the gorges and woodlands where P. commutata sometimes grows’, Bosanquet et. al. 2005

P. falcata occurs in ‘weakly acidic habitats’ compared with P. commutata which occurs ‘in basic habitats’, Pedrotti 2006

P. falcata ‘may occur in slightly less calcareous habitats than P. commutata’, Atherton et. al. 2010

Figure 114 displays two Box plots, one for species that occur within Group 1 and one for Group 2. The upper and lower recorded elevations for all 130 bryophyte and flowing plant species recorded during this survey are plotted on the graphs, although the names of the 130 individual species have been omitted from the x axis for clarity. Where a species occurs in both Group 1 and Group 2 its total range is displayed, hence the plots show total ranges of the individual species recorded during this survey. The plots clearly show that the higher elevation Group 1 (Palustriella falcata) is more species rich and the lower elevation Group 2 (Palustriella commutate) is species poor.

Figure 113    Box plot showing elevation of Palustriella falcata and Palustriella commutata recorded during this study.

Figure 114    Box plot showing elevation range of all species for Group 1 and Group 2.

Comparison of water quality

In total 32 complete samples were collected during the project with 3 additional sites (BR1.1, CE1.1 and CE2.2) sampled for just for nutrients and not major ions. Tables for each site have been included within the main body of the text and the aim of this section is to bring together the entire dataset to look for similariities and differences within groups of data based on ecological observations and groupings. Before the data is plotted on any graphs, less than (<) values are removed and a figure that is 50% of the value of the limit of detection value is used. For instance if a site is reported to have <0.2 mg/l of ammoniacal nitrogen then the figure used in the statistics is 0.1 mg/l. We will use two main types of graph to display the data; Piper Plots are trilinear diagrams and allow the user to plot cations on the bottom left handside triangle and anions on the bottom right hand side triangle, information from both the triangles are then projected into the upper diamond and enables comparison of waters. Box Plots are also used to show the ranges (5th and 95th percentiles, median and outliers) of various data groups for instance the range of calcium values for all sites, or the range of nitrate values for all sites. All graphs have been prepared on SigmaPlot.

The first Piper plot (Figure 115) shows all the water quality samples collected from sites that represent H7220. The majority of samples cluster together and can be described as being of calcium-bicarbonate facies (CaHCO₃). Two samples circled in the plot, BR1.1 and NP1.1 fall slightly outside of this representing a greater contribution of chloride and less bicarbonate at Nant Peris possibly an indication of recent rainfall recharge and a greater concentration of sulphate and chloride at Brest Rhiw.

Figure 115    Piper diagram for all sites designated as H7220.

The same dataset as displayed on the piper plot in Figure 115 can also be displayed on a Box plot (Figure 116). The Box plot shows that the greatest variation is within calcium and bicarbonate and that there is less variation within the other major ions. Circled are the samples from Nant Peris.

Figure 116    Box Plot of major ions at all sites designated as H7220.

The previous graphs display the data for all sites designated as H7220 within this report. The next set of graphs display the data split into two groups using the vegetation groupings (see Chapter 8.1) proposed for Palustriella dominated vegetation, Group 1 and Group 2. Figure 117 shows water quality samples that have been grouped on two separate Piper Plots. Both groups show a preference to waters dominated by calcium bicarbonate however Group 1 has a larger range, with the samples from Nant Peris (NP1.1) and Brest Rhiw (BR1.1) both being projected further into the upper diamond a factor of more dominance of chloride and sulphate in the respective samples. The same data is presented in the adjacent Box Plots (Figure 118) that illustrates the dominance of calcium bicarbonate in Group 1 and to a lesser degree Group 2. The lack of data for Group 2 is the reason that the Box Graph could not display percentile values. In summary there are no major differences between the water chemistry of Group 1 or Group 2 based on the collected within this study although two samples from Group 2 (BR1.1 and NP1.1) display slightly more chloride and sulphate.

Figure 117    Piper diagrams. Left is ‘Group 1’ vegetation and right is ‘Group 2’ vegetation.

Figure 118    Box Plot of major ions. Left is 'Group 1' and right is 'Group 2'.

Nutrient samples were taken at 35 locations and allow an initial range of values to be presented for both Group 1 and Group 2 and all of the data is presented together in Figure 119, in addition orthophosphate and phosphate are displayed on an additional graph (Figure 120). Overall the suite of nutrients were very low, often less than their limits of detection (LOD), and when this was the case a value of 50% was used to represent these samples statistically. Ammoniacal Nitrogen, Total Oxidized Nitrogen, Nitrate as N and Nitrite were all for the majority of samples very low (i.e <1 mg/l) or less than their limit of detection. Only samples from Cors Erdderiniog, Fedw Fawr Beach, McWalters Dingle were >1 mg/l with the highest readings from Cors Erddreiniog 1.7–7.4 mg/l Nitrate as N. Orthophosphate was generally close to or at its LOD. Phospahte as P varied between <0.02 the lowest limit of detection and 0.07 ug/l (LyF2.4 — Llyn y Fan Fach). Further division of the nutrients into Group 1 (dominantly upland) and Group 2 (dominantly lowland) shows a clear division (see Figure 121) with the higher nutrient levels reported for lowland sites and this is a clear reflection of the more agricultural landuse in lower elevation land compared to higher elevation land in Wales.

Figure 119    Box Plot. Nutrients (mg/l) for Group 1 and Group 2.

Figure 120    Box Plot. Phosphate (μg/l) for Group 1 and Group 2.

Figure 121    Box Plot of nutrients, to the left is Group 1 and right is Group 2.

Geological and hydrogeological controls

Palustriella dominatedsprings occurred across a range of geologies including both bedrock, superficial and made ground, however active or historic tufa deposition is not associated with all the locations that were included within this study.

Where bedrock aquifers are considered the primary source of water they include the Carboniferous Limestone and associated sandstones and site include; Clydach Gourge, Foel Garnach, Waun Ddu-Craig y Cilau, Cors Erddreiniog and Fedw Fawr 1, 2 and 3. H7220 is associated with Devonian Old Red Sandstone — where significant calcretes horizons occur (e.g Hen Allt and Tarren y Esgob) and to a lesser degree where they are not present (e.g Llyn y Fan Fach). Interestingly H7220 also occurs in association with the Ordovician Snowdon Volcanic Group (Nant Peris).

The interaction of water from superficial aquifers is hard to quantify however it may have an influence at sites including, but not limited to Pont Clydach 1, 2 and 4, Llyn y Fan Fach 1 and 2. The remainder of the sites are closely associated with made ground in the form of lime spoil tips from Carboniferous Limestone quarries and include, Foel Fawr 1 and 2, Brest Rhiw and Blaen Onnen.

Although the association of carbonate dominated geological units (e.g. Carboniferous Limestone) are well illustrated, importantly the occurrence of H7220 is not restricted only to these environments, sites such as Nant Peris are an exception to this rule.

Visual estimates of groundwater discharge and flow from individual springs, seepages faces and within runnels were made at each site. It should be noted that these estimates are very crude and do not account for total groundwater discharge to a site rather for specific runnels or areas adjacent to Palustriella vegetation. Visual estimation flow rates varied from <0.1–0.5 l/s at Group 1 and <0.1 to 1 l/s in Group 2. Both Group 1 and Group 2 sites shared similar characteristics in that the occurrence of very shallow, but also flowing water was associated with each site and this may well be a hydrological requirement for supporting of H7220 habitats. Although observations have not been made throughout the year (especially during dry periods) many of the sites may experience discharge of water for the majority of the year.

Topography and slope

There is clear distinction between Groups 1 and 2 with group 1 sites occurring more generally in open upland locations (with gentler overall slope) and group 2 occurring more generally in less open and less upland locations (with a greater overall slope that can involve vertical rock faces).

At a micro level (individual vegetation patch or runnel), slope can vary greatly and this regularly results in a great range in the degree of base flushing of vegetation (and hence the type of vegetation) across a very small distance. Based on individual transect measurements, it was common to record vegetation with Sphagnum (pH 5–6, occasionally 4–5), Palustriella dominated vegetation (pH 6–7) and open runnels with Scorpidium cossonii (pH 7–8, occasionally 9) within less than 0.5 m distance. At a number of sites, particularly dramatic vegetation transitions over a very short distance were observed. These include Wan Ddu (Craig-y-Cilau) where stands of Sphagnum cuspidatum (at the margins of a bog) directly adjoin Palustriella dominated vegetation at the break of a small slope (pH transition 4.93 to 7.42 and EC 51.2 to 272 μs/cm) and Moel Gornoch where acid grassland and bog vegetation closely adjoins Palustriella dominated vegetation at the break of slope (pH transition 4.74 to 6.30).

In summary, the observed dramatic changes in vegetation type, water quality and supply, emphasise microtopography as being a important factor influencing the character and structure of Palustriella dominated springs.

Does the occurance of tufa matter?

Tufa is deposited when water supersaturated with calcium carbonate degasses on contact with the atmosphere and this process is associated with many, but importantly not all of the sites within this study. The argument is that a water can have elevated levels of calcium carbonate, enough to support Palustriella and the other species associated with the tufa forming springs habitat, but without being supersaturated and thus without depositing tufa. Although the geomorphology of tufa formation adds to the attractiveness of any given site it does not seem to be a limiting factor for the presence of the key bryophyte sepcies, i.e based on the sites in this survey you do not have to have active or historic tufa deposition at a site in order to have the characteristic tufa bryophytes.

All group 2 sites (generally considered to be NVC M37) had active tufa present (i.e. tufa associated with plant growth particularly bryophytes) and a proportion of group 1 sites (generally considered to be NVC M38) had similarly active tufa present. However, active tufa was observed at many sites that was not associated with plant growth including deposits on stones, branches and notably an old ten pence piece (Clydach Gorge). At Ddol, active tufa deposits were seen on fallen branches and twigs but not associated with plant growth. In addition, based on the field work, there was a lack of consistency in characteristic tufa bryophytes species (i.e. those frequently cited in literature such as Palustriella, Eucladium, Hymenostylium) as being associated with active tufa formation. Table 112 is in included as a record of species observed as being associated with tufa.

Table 114    Species associated with tufa deposits

Species associated with a ctive tufa formation (i.e. leaves and stems becoming petrified — like stone to the touch)	Species recorded from tufa but not active in tufa formation (i.e. secondary colonists on the surface of eroded tufa or within cracks and holes within tufa)
Bryum dichotomum	Asplenium scolopendrium
Bryum pseudotriquetrum	Asplenium trichomanes ssp. trichomanes
Campylium stellatum	Bryoerythrophyllum recurvirostrum
Chara vulgaris	Centaurea nigra
Didymodon tophaceus	Chrysosplenium oppositifolium
Eucladium verticillatum	Conocephalum conicum
filamentous algae (Chlorophyta)	Conocephalum salebrosum
Gymnostomum aeruginosum	Ctenidium molluscum
Hymenostylium recurvirostrum	Dicranella varia
Jungermannia atrovirens	Didymodon fallax
Palustriella commutata	Fissidens taxifolius var. taxifolius
Palustriella falcata	Geranium robertianum
Philonotis calcarea	Hedera helix
Rivularia haematites	Hypericum androsaemum
Scorpidium cossonii	Leicolea badensis
	Leicolea turbinata
	Pellia endiviifolia
	Potentilla steralis
	Pressia quadrata
	Rubus fruticosus agg.