Download.aspx.pdf

Environmental Geotechnics Botanical and geotechnical characteristics of blanket peat at three Irish bogflows --Manuscript Draft-Manuscript Number:

ENVGEO-D-17-00089R1

Full Title:

Botanical and geotechnical characteristics of blanket peat at three Irish bogflows

Article Type:

Themed Issue: Geotechnical aspects of peatland restoration and management

Abstract:

Systematic investigations of instability of blanket peat began in the 1990s and quickly identified the potential importance of botanical controls on the properties and behaviour of the peat involved in the failures. During 2010-12 investigations of the blanket peat at three bogflows in northwest Ireland were done with the aim of establishing relationships between botanical characteristics and standard physical and geotechnical properties, assuming the latter to be meaningful but recognising that this may not be the case. In-situ measurements and investigations at all three sites were followed by laboratory characterisation of small core, block and monolith samples. The botanical composition of the peat could not be fully determined due to the high degree of decomposition. However, analysis of macrofossils allowed distinct depthrelated patterns of several botanical indicators to be determined. In particular the monocotyledon fragments, dominated by Eriophorum vaginatum, showed distinct and potentially useful distributions throughout the peat. Overall results showed that the basal peat at one site was discernibly different from the other two sites having fewer monocotyledons, fewer fibres, higher dry bulk density and higher saturated hydraulic conductivity. This approach therefore offers a potential basis for developing a means of assessing peat mass characteristics from small auger samples.

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation

Copyright form

Journal Publishing Agreement It is our policy to ask authors to assign the copyright of articles accepted for publication to the Publisher. Exceptions are possible for reasons of national rules or funding. Please tick the relevant options below. In assigning copyright to us, you retain all proprietary rights including patent rights, and the right to make personal (non-commercial) use of the article, subject to acknowledgement of the journal as the original source of publication. By signing this agreement, you are confirming that you have obtained permission from any co-authors and advised them of this copyright transfer. Kindly note that copyright transfer is not applicable to authors who are opting to publish their papers as Open Access. Open Access authors retain copyright of their published paper.

Please complete the form below and include it when you submit your article online (list of websites http://www.icevirtuallibrary.com/page/authors/submitting-your-article).

Environmental Geotechnics name:..........................................................................................................................................

Journal Article

Botanical and geotechnical characteristics of blanket peat at three Irish bogflows ................................................................................................................................................

title:

ENVGEO-S-17-00102 Manuscript reference number: ......................................................................................................................... Authors: Your

E S Foteu Madio and A P Dykes .................................................................................................................................................... Alan Dykes ...............................................................................................................................................

name:

19 March 2018 Signature and date:............................................... .............................................................................

Please tick either one option from part A or one option from part B. Please complete part C. A. Copyright

X

I hereby assign and transfer the copyright of this paper to Thomas Telford Ltd. British Crown Copyright: I hereby assign a non-exclusive licence to publish to Thomas Telford Ltd. I am a US Government employee: employed by (name of agency) ………………………………………………. I am subject to the national rules of (country) ……………………………… and confirm that I meet their requirements for copyright transfer or reproduction (please delete as appropriate)

B. Authors with open access funding requirements. Please specify the Creative Commons license version required. CC-BY (for full details click here Creative Commons Attribution (CC BY) 4.0 International License) C. Please confirm that you have obtained permission from the original copyright holder. For ICE Publishing’s copyright policy, please click here. ICE Publishing is a signatory to the STM Guidelines. X

I have obtained permission from the original copyright holder for the use of all subsidiary material included in this paper (E.g. for borrowed figures or tables).

I have the original file and am not using the published version (Fig. 1A and 1B)

Thomas Telford Ltd is wholly owned by the Institution of Civil Engineers ICE Publishing, 1 Great George Street, Westminster, London, SW1P 3AA, UK Telephone: +44 (0)20 7665 2242 | Fax: +44 (0)20 7665 2189 | Web: www.icevirtuallibrary.com

Response to Reviewer and Editor Comments

COMMENTS FROM ASSESSOR On page 8: “The Fm and Rm fibre fractions were estimated in the field according to the von Post system”: Elaborate how this was done. Text added: The Fm and Rm fibre fractions were estimated in the field based on the von Post system as presented by Hobbs (1986, p.79): Fine fibres (Fm) are ‘fibres and stems smaller than 1 mm in diameter or width’ and coarse fibres (Rm) are ‘fibres, stems, and rootlets greater than 1 mm in diameter or width’. To both of these definitions we added ‘or any plant particle’ and took the size boundary as ‘< or > 1 mm in all directions’. The von Post scheme uses a four point scheme from 0 ‘nil’ to 3 ‘high content’ but without a microscope it is difficult to be certain that there are no fragments of fibre present. Consequently we removed ‘0’ and assessed the quantity according to a five point scale: 1 = very low content (VL), 2 = low content (L), 3 = medium content (M), 4 = high content (H) and 5 = very high content (VH). “The bi-linear correction of the deviator stress due to membrane stiffness was applied to the results”: Give suitable reference. This statement was inserted in error has been corrected with added text: The bi-linear correction of the deviator stress due to membrane stiffness was not applied to the results because although the effect may be significant, (i) there is no consensus on appropriate corrections given complex peat-membrane interactions, and (ii) this was primarily a comparative study that was not necessarily expected to determine the exact value of the shear strength of peat. On page 12: “Field shear vane readings from depths between 1.25 and 2.00 m were between 6.6 and 14.0 kPa at all sites …..”: What vane correction factor, if any, was applied in determining these values. Text added: Uncorrected field shear vane readings from depths between 1.25 and 2.00 m were between 6.6 and 14.0 kPa at all sites and there were insufficient results from which to identify any patterns in the data. No corrections were applied because this was intended as a comparative study and the shear vane is known to be inappropriate for the determination of the undrained strength of peat due to the effects of fibres, although it can be used to identify patterns of peat strength variation with depth. On page 15: “constant head method”: the methodology was not presented in manuscript: please include in section 3: Methods The ‘constant head’ method for measuring permeability/saturated hydraulic conductivity is sufficiently routine that a detailed account of the method is surely not warranted? However, text has been added: For the constant head method we used a laboratory permeameter arrangement as described by, for example, Klute and Dirksen (1986) or Head (1994). Undisturbed core samples collected in thin-walled tubes 50 mm long (‘L’) × 50.5 mm diameter were trimmed to size, saturated in tap water and mounted vertically to form a permeameter maintaining a constant head of 0.15 m of water on the top of the sample. Water that passed through the sample was collected underneath and measured. The constant head saturated hydraulic conductivity was calculated according to Darcy’s Law from: k = VL / (tA.ΔH) where k = permeability = saturated hydraulic conductivity (m s–1) V = volume of water collected (m3) t = time (s) of water collection A = cross-sectional area of peat sample (2.00×10–3 m2) L = length of peat sample (0.050 m) ΔH = head difference across peat sample length (0.200 m)

On page 17: “We are not convinced that peat strength would reduce with increasing fibre content as shown in Fig. 11F”: check this statement against the figure, and confirm OK On reflection, although a hint of such a trend can be discerned from Fig. 11F, it is perhaps so vague that it is not sufficiently ‘real’ to be worth commenting on at all. Therefore the sentence has been deleted. On page 19: “Previous studies have found that higher fractions of coarse fibres had no effect on measured strength compared with lower coarse fibre contents (Zhang and O’Kelly 2014; Hendy et al. 2014);…..” Please check these referenced sources: At least for the first mentioned study, the testing undertaken was consolidated drained (CD) triaxial compression, and the purpose of this investigation was to demonstrate, among other things, that CD triaxial testing of (fibrous) peat produces meaningless strength results, since the test specimen in essentially undergoing 1D consolidation and not shearing in the conventional sense. In Section 5 of their paper (bottom of p.46 to p.48 inclusive), Zhang and O’Kelly (2014) present results from their tests that show, surprisingly, no real differences between the three peat preparations. They then comment on the unsuitability of the particular tests given the observed sample responses, but our comment is valid since we refer only to the measured strength and make no inference or assumptions regarding the validity of the results that those measurements represent. Figure 11: Include R2 values for correlations shown in the figures All values added to Figs. 12A-E inclusive.

REVIEWER #1 General comment: The Authors present an interesting study, with appreciable investigation effort and experimental honesty. The conclusions are not very strong but the problematic approach of the paper results in a discussion very useful for researchers. In my opinion, the paper is suitable for publication on the journal, but it should be improved by addressing the following comments. The absence of significant correlations between botanical features and mechanical parameters should be better discussed in order to address specific guidelines for the stability analyses of this kind of geotechnical context, for example by stating, due to the impossibility of using good empirical correlations to deduce the peat strength parameters, the importance of planning specific laboratory geotechnical tests for each site. Text added to the Discussion (last-but-one paragraph): However, given that the measured strengths at this site were no different from the others, we cannot say whether measurement of those characteristics would be useful for peats formed from significantly different plant assemblages. It is not possible to generalise any implications of our results for peatlands in general, and notwithstanding previous comments we cannot assume that any of our correlations between botanical and geotechnical characteristics will apply throughout Ireland. There is thus a clear necessity for carefully planned laboratory testing of peat from the site of any proposed development, probably requiring excavation of trial pits for the extraction of appropriate undisturbed samples. However, general recommendations for the most appropriate tests – and testing procedures suitable for peat – will probably take some time to emerge from ongoing research programmes. In this respect, the section describing the three sites should be renamed and integrated by adding a specific discussion on the most typical landsliding/failure mechanism for these peaty slopes. The needed of deducing undrained strength values should be related to the observed/expected kinetic of the in-situ failure processes: why are you interested in the undrained strength? Did you always detect fast undrained failure mechanism for these landslides? Explain and discuss.

Section 2 has been renamed ‘Blanket Bog Failures in Ireland’ with the existing ‘Study Sites’ text now comprising subsection 2.1. An additional page of text has been inserted at Section 2 to address the several points raised by the reviewer in this comment. In this respect, you must specify if the Mohr’s Circles you show are in terms of effective stresses or not. Since you mention the effect of pore pressures on their size, the reader could only suspect you have represented effective stresses, but you must specify it both in the text and in the caption of the figure. Otherwise, if the circles are in total stresses, make attention to the comment on the gas effect and explain better what you mean. Text added: Given that the assumed effect of pore pressures in undrained shear tests is to reduce the friction towards zero because the water is incompressible, then the presence of compressible gas within some pore spaces could allow some (additional) frictional resistance to arise during testing. Hanrahan (1954) found that the gas content of Irish Sphagnum peat may be considerably in excess of 5% of the volume and that significant volumes of gases such as sulphuretted and phosphoretted hydrogen (phosphine), as well as methane, could be emitted during construction involving the compression of peat. Therefore the possibility of gas affecting both permeability and pore pressures must be allowed for when interpreting results. Provide some additional hypotheses for the causes of the profile of tensile strength: does the reduction with depth depend on the coarse fibres profile only? Could you invoke a diagenetic process of “crusting” giving cohesion to the material in the upper soil? If not, please discuss in terms of chemical and physical environment. Text added: At individual sites it is possible that such a trend of decreasing tensile strength with depth may not always be found, although there are insufficient relevant data to be able to comment further. Helenelund (1967) suggested that the fibre contents, types and orientations – which depend on the morphology and the mode of growth of the original plant assemblage that formed the peat – may have major influences on the tensile strength. The macrofossil analyses of peats from our study sites revealed remains of sedges, the degree of humification of which increase with depth. In such monocotyledon peat, fibres are the remains of vascular bundles formed from the root systems that grow perpendicularly to the ground surface. The resulting tensile strength will therefore be related to the resisting force produced by the fibres, the frequency of which decreases with depth and is inversely proportional to the degree of humification. The tensile strength results obtained by Helenelund (1967) from Sphagnum bog peat, which has very few fibres, are comparable with the lowest of our results, showing that the monocotyledon peats at our sites generally have higher tensile strengths than Sphagnum bog peat. Due to the effect of compression during the accumulation of the peats, some fibres that were originally distributed vertically through the peat become squashed progressively into a horizontal alignment as pressure increases. The degree of inclination of these fibres toward the horizontal plane should therefore also increase with depth. The tensile strength values presented in this study were measured in a horizontal plane, intended to represent the effect of the peat mass pulling apart above a basal (shear?) failure zone. The effect of fibre orientation should be to increase the tensile strength with depth since horizontal breaking up of a failing peat mass is resisted by sometimes significant lengths of fibres adhering to amorphous colloidal matrix material. However, the role of living and minimally decomposed roots within the near-surface acrotelm layer combined with the very high degree of humification below the acrotelm appears to entirely override the fibre orientation effect. Rename the final section as “Conclusions and Future Work” and put in this section the lines 498-510. Done.

Further detailed observations: Avoid abbreviation of “sampling point” with SP: it creates confusion with the abbreviation of the sites. All instances changed to ‘Sampling Point’.

Table 1- Change the column name “Context” into “Geomorphological context”. Done. Rename 4.2: “Physical and mechanical properties of the peat”. Done. Line 292: Make attention to the meaning of “pre-consolidation” in Geotechnics: do you mean simply “consolidation”? Sorry, yes – we meant consolidation prior to strength determination. Amended accordingly. The caption of Figure 12 is not clear: do you mean “Different estimates of thickness…..” ? Effectively, yes, that is what we mean. Amended to your suggested wording. When you mention the stability analysis with Safety Factor equal to 1 for deducing mobilized strength values, use the term “back analysis”. Amended accordingly.

REVIEWER #2 Line 58: It would be useful to expand the introduction to present some key findings of previous studies on ''botanical controls on peat properties and instability''. The whole point is that there aren’t any such previous studies, and the few that do present some relevant data arrived at contradictory findings (O’Kelly 2017), hence our added text: The need for research into botanical controls on peat properties was further emphasised with respect to blanket peat instability by Kirk (2001), Dykes (2008a) and O’Kelly (2017) in response to findings from their investigations of physical and geotechnical properties thus far. Indeed, O’Kelly (2017) highlighted the scarcity of published works on the topic and the contradictory findings from the few such studies. The present research (Foteu Madio 2012) arose directly from this dearth of previous studies. Line 151: In the literature, there has been much debate on the accuracy of water content values determined using oven drying for organic soils such as peat and the appropriate temperature in this method. Could you comment on why you have chosen oven-drying for 24 h at 105°C for water content determination? Text added: There has been some debate in recent years regarding the appropriate drying temperature for water content determination, including evidence of the possibility of charring of the peat at temperatures higher than 8090°C (O’Kelly 2014). Further, O’Kelly (2014) found experimentally that the possible additional loss of mass due to charring is negligible compared with the mass of any retained water due to incomplete drying, particularly intracellular water within peat fibres that may constitute a significant proportions of the peat mass (Foteu Madio 2013). O’Kelly (2014) therefore recommended following the standard specification for mineral soils of 105°C, as used by many previous workers including Skempton and Petley (1970) and Hobbs (1986). We adopted the latter approach for the demonstrated reasons of standardisation and comparability of results.

Line 233: Water content is a key parameter to understand the geotechnical properties of earth structures. Seasonal changes of soil water content can cause significant seasonal changes in pore pressures, which affect soil strength. Have you considered (or would you recommend) pore pressure as a key controlling parameters for triggering landslides in peat covered slops? This is a difficult issue to address without substantial additional text. Therefore we have used two short paragraphs at different points in the manuscript to try to provide an adequate response. Text inserted into Section 2: It is likely that the peaty-debris slides are triggered by pore pressure effects, in part due to subsurface storm runoff being confined beneath a saturated and effectively impermeable peat cover. Peat slides (interface failures) probably occur for the same reason. Failure within the peat is a more complex issue because of the dual influences of effectively impermeable and normally saturated but weak catotelm peat material and the internal structure of the peat mass (sensu ‘rock mass’ considerations) that may experience high turbulent flows and even artesian conditions within networks of natural peat pipes and (relict) desiccation cracks (Dykes and Warburton 2007a, 2008a; Gilman and Newson 1980; Holden and Burt 2003). Text inserted at the end of the Discussion: Finally, the very low shear strength indicated above demands some consideration with respect to water conditions within the peat. Blanket bogs in the British Isles may experience water table variations of up to 0.5–1.0 m, but these are occasional reductions below the surface during warm periods of summer weather (Evans et al. 1999; Holden and Burt 2003). The usual condition for these deposits is to be fully saturated to the surface, i.e. with normal effective stress ≈ 0 and maximum pore water pressure most of the time. Periods of summer drying may increase the normal effective stress by a few kPa due to the reduced pore water pressure, i.e. temporarily increasing the effective shear strength. Failure within the peat cannot, therefore, be the result of raised pore water pressures throughout the peat matrix due to heavy rainfall (although it could due to external loading). The hydraulic effects of water-filled pipes, cracks and other voids (e.g. Dykes, this volume – in review) may play significant roles in the initiation of failure, i.e. peat mass effects, are thought to be more important than simply the peat matrix (shear) strength, but much more research is needed to test this hypothesis. Line 258: Some slides showing the samples used for laboratory testing could be useful for describing the testing procedures and the results. We assumed from the reference to ‘Line 258’ that the comment was concerned with illustrating the tensile strength methodology. Therefore a new Figure has been added that incorporates the only photos we have of this. Added text: Figure 3. Measuring the tensile strength of the peat: (A) the two sets of steel ‘fingers’ that are pushed through the centre of a cut block of undisturbed peat 100 mm high × 100 mm wide and 40-60 mm thick; (B) one half of a sample following tensile failure, still adhering to one set of ‘fingers’; (C) the testing apparatus, showing: centre – the ‘fingers’ assembly installed (without a sample); right – the force proving ring; lower far left – the hanger for applying the weights that apply the load just visible beside the end of the cupboards. Details of the design and development of this apparatus are provided in Dykes (2008c). Line 287: "The basic physical properties of the peat at the three landslides are summarised in Table 4. These are consistent with previous results obtained from Straduff Townland (the 1997 bogflow adjacent to ST) and SA by Yang and Dykes (2006)''. It would be good to add values of physical properties from the mentioned studies for comparison. Inserted into Table 4.

REVIEWER #3 The paper aims to present an original study of instability of blanket peat at three bogflows in northwest Ireland. In addition, were performed investigations between botanical characteristics and standard physical and geotechnical properties. The manuscript is balanced and clearly well-written in scientific and technical soundness. The topic is very interesting, emerging and challenging. I suggest only some minor corrections on the manuscript submitted: i) page 2, line 47 (PDF) enter a space - "… until the late 1990sthe occasional" to "until the late 1990s the occasional…" Done. ii) figure 1: missed geographical coordinates and a scale bar in Ireland map; Done. as well as a brief explanation about the map (grey colour is ? strips are?) … is a morphological map? Please improve. Explanatory details added to the caption. My congratulations to the authors for the outstanding manuscript presented. Thank you!

Original file showing changes

1

v. 03 November 2017

2 3

Botanical and geotechnical characteristics of blanket peat at three Irish

4

bogflows

5 6

Eliane S. Foteu Madio †

7

BSc(Hons), MSc, PhD, MCIWM, Director at Envigma Ltd., Clacton-on-Sea, UK

8

School of Geography and Geology, Kingston University, Kingston upon Thames, UK. †

9 10

Alan P. Dykes*

11

BSc(Hons), PhD, FRGS, FHEA, FGS, CGeol, Associate Professor at Kingston University, Kingston upon Thames, UK

12

Department of Civil Engineering, Kingston University, Kingston-upon-Thames, UK.

13

ORCID: 0000-0003-0327-0498

14 15

*Corresponding author:

16

Centre for Engineering, Environment and Society Research, SchoolDepartment of Civil Engineering, Kingston

17

University, Penrhyn Road, Kingston- upon- Thames, KT1 2EE, UK, Tel.: +44 (0)208 417 701, Email:

18

[email protected]

19 20

† Present

21

10 Crossfield Road, Clacton-on-Sea, CO15 3QT, UK, Tel.: +44 (0)1255 428787, Email: [email protected]

22

Address at the time of this research:

23

School of Geography and Geology, Kingston University, Kingston upon Thames, UK

address: Senior Engineer,

Formatted: Font: Not Italic

24 25

5419 words (excluding Figure captions and Tables), 132 Figures, 4 Tables

26 27

Abstract

28

1

Formatted: Font: Italic

29

Systematic investigations of instability and failure of peat covered hillslopes began in the late 1990s and quickly

30

identified the potential importance of botanical controls on the properties and behaviour of the blanket peat involved in

31

the failures. However, attempts to unravel some of these controls did not begin for several years. During 2010-12

32

investigations of the blanket peat at three relatively recent bogflows in northwest Ireland were done with the aim of

33

establishing some form of relationship between botanical or paleoecological characteristics and standard physical and

34

geotechnical properties, assuming the latter to be meaningful but recognising that this may not be the case. In-situ

35

measurements and investigations at all three sites were followed by extensive laboratory characterisation of small core,

36

block and monolith samples.

37 38

The botanical composition of the peat could not be fully determined due to the very high degree of decomposition.

39

However, analysis of macrofossils allowed distinct depth-related patterns of several key botanical indicators to be

40

determined. In particular the monocotyledon fragments, dominated by Eriophorum vaginatum, showed distinct and

41

potentially useful distributions throughout the peat profiles. Overall results showed that the basal peat at one of the sites

42

was discernibly different from the other two sites having fewer monocotyledons, fewer fibres, higher dry bulk density

43

and higher saturated hydraulic conductivity. This approach therefore offers a potential basis for developing a means of

44

assessing peat mass characteristics from small auger samples.

45 46

Key words

Fabric/structure of soils, Landslides, Strength & testing of materials

47 48

1. Introduction

49 50

Records of failures of peat bogs go back around 500 years to the collapse of Chat Moss near Manchester,

51

northwest England, in 1526 (Crofton 1902). However, until the late 1990s the occasional studies of isolated

52

examples of peatland failures were largely descriptive with estimates of geometric characteristics and

53

occasionally reports of the living plant assemblages present at the time of failure. Systematic investigations

54

began into the stability of blanket peat-covered slopes following significant peat landslides in Northern

55

Ireland (Dykes and Kirk 2001; Kirk 2001) and northern England (Mills 2002; Warburton et al. 2003, 2004).

56

The potential importance of the botanical composition as a controlling factor for the properties and

57

geotechnical behaviour of peat was highlighted earlier by Hobbs (1986), not least because of the widespread

58

adoption by engineers of the von Post scheme for classifying peat deposits (e.g. Landva and Pheeney 1980;

59

Carlsten 1993) which requires the estimation of relative frequencies of fibres and wood/shrub fragments as 2

Formatted: Font: 11 pt

60

well as the degree of decomposition of the plant matter (i.e. the humification). The need for research into

61

botanical controls on peat properties was further emphasised with respect to blanket peat instability by Kirk

62

(2001), Dykes (2008a) and O’Kelly (2017). in response to findings from their investigations of physical and

63

geotechnical properties thus far. Indeed, O’Kelly (2017) highlighted the scarcity of published works on the

64

topic and the contradictory findings from the few such studies. The present research (Foteu Madio et al.

65

2012; Foteu Madio 2013) arose directly from this dearth of previous studies.

66 67

The aim of this paper is to examine whether physical and geotechnical properties of Irish blanket peat can be

68

causally associated with measurable botanical characteristics. It does so by presenting and analysing data

69

representing the properties and characteristics of the peat at the sites of three significant bogflows in

70

northwest Ireland, obtained from a combination of field and laboratory investigations. The importance of this

71

study is to provide the basis for more efficient and reliable methods for assessing the stability of peat with

72

respect to planned interventions such as construction of access roads for windfarms or other purposes.

73 74

2. Blanket Bog Failures in IrelandStudy Sites

75 76

The topic of peat mass movements (as distinct from geotechnical engineering of peat) emerged from an

77

esoteric scientific by-way to become a mainstream theme in engineering geology and geomorphology

78

following several major events in late 2003. On 19 September 2003, two entirely independent extreme

79

rainfall events in Co. Mayo, Ireland, and South Shetland, Scotland, triggered multiple failures of peat-

80

covered hillslopes (Dykes and Warburton 2007a, 2008a,b). More significantly for civil engineering, four

81

weeks later the 450,000 m3 Derrybrien Windfarm landslide occurred (Lindsay and Bragg 2005). By that time

82

it had already become clear that the Irish blanket bogs were failing in several slightly different ways, giving

83

rise to morphologically distinctive types of failures (Dykes and Warburton 2007b). Most involve shearing of

84

mineral soil beneath the peat (‘peaty-debris slides’), shearing at the peat-mineral interface (‘peat slides’), or

85

shearing entirely within the basal peat (‘bog slides’). ‘Peat flows’, a term reserved for failures resulting

86

primarily from head-loading, appear to be effectively bearing capacity failures with small areas of shear

87

surface within the basal peat having been observed in the Derrybrien and Ballincollig Hill landslides (Long

88

20054; Dykes and Jennings 2011). 3

Formatted: Superscript

89 90

All of the available evidence relating to ‘bog bursts’ and ‘bogflows’ indicate that these failures involved

91

some sort of in-situ liquefaction of the lower or basal peat, with this (semi-)liquid peat slurry then breaking

92

out from beneath a stronger confining acrotelm layer (or from cut faces through the margins of raised bogs)

93

(Dykes and Warburton 2007b). The precise mechanisms of strength loss are unknown. One hypothesis, for

94

example, is that the basal peat fails like ‘quick clay’ with an initial small shear failure creating a disturbance

95

that propagates rapidly. As such a bogflow may simply be a bog slide involving weaker and wetter peat – but

96

there is a clear distinction because these two types of failure have different peat depth vs. gradient

97

characteristics (A P Dykes, unpublished data). A parallel hypothesis is that in some of these failures the

98

lower layer of the peat deposit was always a fluid body, for example if peat grew over a large pond so as to

99

eventually entirely bury it.

100 101

In almost all cases of failures of (blanket) peat-covered slopes in Ireland, landslide morphologies and runout

102

characteristics display clear evidence of relatively rapid development of failure associated with very high

103

volumes of rainwater, with eyewitness accounts of some recent events (e.g. the Derrybrien peat flow in 2003,

104

the Croaghan peat slide in 2014) corroborating these interpretations. Warburton et al. (2004) discussed the

105

various hydrological processes giving rise to, or controlling, such failures. It is likely that the peaty-debris

106

slides are triggered by pore pressure effects, in part due to subsurface storm runoff being confined beneath a

107

saturated and effectively impermeable peat cover. Peat slides (interface failures) probably occur for the same

108

reason. Failure within the peat is a more complex issue because of the dual influences of effectively

109

impermeable and normally saturated but weak catotelm peat material and the internal structure of the peat

110

mass (sensu ‘rock mass’ considerations) that may experience high turbulent flows and even artesian

111

conditions within networks of natural peat pipes and (relict) desiccation cracks (Dykes and Warburton

112

2007a, 2008a; Gilman and Newson 1980; Holden and Burt 2003). The critical factor here is that rates of

113

deformation and then movement are likely to greatly exceed the saturated hydraulic conductivity of the intact

114

peat mass through which any shear surface may develop. Consequently, the focus of our research is on the

115

undrained strength characteristics of basal peat.

116 117

Formatted: Font: Bold, Italic

2.1 Study sites 4

118

We identified three locally significant bogflows (sensu Dykes and Warburton 2007b) for this study, located

119

within the same region of northwest Ireland: Straduff Townland (hereafter referred to as ‘ST’), Slieve

120

Anierin (‘SA’) and Slieve Rushen (‘SR’) (Fig. 1). Site and landslide characteristics are summarised in Table

121

1 and illustrated in Fig. 2. All were relatively recent, thus limiting the degree of post-failure degradation, and

122

two sites (though not the same landslide at one of these sites) had been investigated previously which

123

provided a cross-check for the peat characterisation results from this study. Furthermore, although the peat at

124

all three sites was generally very similar, one site (Slieve Anierin, below) was noted by Yang and Dykes

125

(2006) to be slightly but nevertheless distinctly different from others including a bogflow at Straduff

126

Townland adjacent to the one used for this study. We anticipated that the results of this new research would

127

also show this.

128 129

Table 1. Summary of site details and characteristics of the study bogflows. Bogflow County Latitude Longitude

Elevation Geology Geomorphological Length Slope Deptha Volume (m) (Carboniferous) Context (m) (°) (m) (m)

Formatted: Line spacing: single Formatted: Line spacing: single

ST

Sligo 54°7.2’N 8°12.9’W

405

Lackagh Sandstone

SR

Cavan 54°8.9’N 7°38.5’W

390

Glenade Sandstone

SA

Leitrim 54°6.3’N 7°58.7’W

440

Lackagh Sandstone

Escarpment failure

200

Basin slope failure 175 Escarpment failure

190

5.5 (top) 3 (mid) 6 (lower)

2.5

35,000

Formatted: Line spacing: single

5.5

2.0

20,000

Formatted: Line spacing: single

4

2.2

22,000

Formatted: Line spacing: single

130 131 132

Notes:

Formatted: Underline

a

Formatted: Line spacing: 1.5 lines

133

< FIGURE 1 >

134

Figure 1. Location of the study area in northwest Ireland, showing the distribution of peatlands (after Hammond 1979).

135

The outlined rectangle is enlarged to show the locations of the three bogflows: (left to right) ST = Straduff Townland,

136

SA = Slieve Anierin, SR = Slieve Rushen. Modified from Yang and Dykes (2006).

Indicative average depth of in-situ peat immediately adjacent to the landslide source area

137 138

< FIGURE 2 >

139

Figure 2. General views of the three study areas. (A) Straduff Townland bogflow, looking downslope from above the

140

head (July 2010). (B) Slieve Rushen bogflow, looking across at the failed slope from the other side of the peat basin

141

into which its displaced peat flowed (July 2010). (C) Slieve Anierin bogflow from the air (Nov. 1998, photo by APD). 5

142 143

ST, the Straduff Townland landslide, occurred overnight or early morning on 14 August 2008 during very

144

heavy rain. The dominant morphology is that of a bogflow. However, a basal shear surface around 20 mm

145

above the base of the peat was visible in two small parts of the source area (Dykes 2009; Dykes and Jennings

146

2011). Although the latter observation corresponds with a ‘bogslide’ (Dykes and Warburton 2007b), we will

147

refer to this failure as a bogflow. It involved an area of intact blanket peat between the source areas of

148

bogflows dating from 1945 and 1991, leaving narrow strips of minimally displaced peat separating the

149

failures. The physical characteristics of the peat at the 1991 bogflow, just a few metres from the margin of

150

the later failure, were determined by Yang and Dykes (2006). SA (the Slieve Anierin bogflow) is thought to

151

have occurred during 1998, based on its visible condition when first seen from a light aircraft in November

152

1998 and a conversation with a local resident in 2011. It was described, and the physical characteristics of

153

the peat reported, by Yang and Dykes (2006). The date of SR (the Slieve Rushen bogflow) is uncertain, but

154

the condition of the failure when first inspected in September 2004 was consistent with an age of only a few

155

years, i.e. it most likely occurred during the 1990s (Dykes 2008b).

156 157

3. Methods

158 159

The three bogflows were investigated using the same general methodology as previous studies of peat

160

landslides (e.g. Yang and Dykes 2006). All had previously been surveyed in detail by Dykes (2008b, 2009).

161

The focus for this study was to obtain samples for laboratory testing from a carefully prepared and fully

162

described vertical profile through the full depth of undisturbed in situ peat. Most peat failures leave irregular

163

sub-vertical peat profiles with varying amounts of peat debris covering the lower layers, around several parts

164

of the source area margins. A single study profile (hereafter referred to as the ‘study profile’ or ‘sSampling

165

pPoint’, SP) was selected at each landslide according to the feasibility of creating a clean vertical profile

166

through the full thickness of the peat, i.e. involving the minimum manual excavation of loose peat debris, but

167

ensuring the in-situ peat was undisturbed and not within a few metres of any tension cracks. Safety was

168

ensured by having wide open access to the prepared profile from within the evacuated source area of each

169

landslide, with one person maintaining active watch over the cut face while the other person worked there.

170 6

171

3.1 Field investigations

172

Around each bogflow source area, stratigraphic and topographic surveys were carried out in order to estimate

173

the morphology of the peat deposit and the variability of the peat within it prior to failure. The stratigraphy

174

and maximum depth of the in situ peat was determined on a coarse but regular grid using a 20 mm diameter

175

gouge auger. Maximum peat depths were measured at additional locations by probing with a metal rod. Peat

176

surface elevations were then surveyed at all the stratigraphy and peat depth measurement points by levelling,

177

with the mineral surface elevations being measured within the landslide source areas and calculated for the

178

peat-covered areas around the source areas from the measured peat depths.

179 180

Prior to sampling, a detailed description of the full thickness of the peat at each SPSampling Point was

181

recorded according to the von Post (von Post, 1922, as presented by Landva and Pheeney 1980, and Hobbs

182

1986) and Troels-Smith (Troels-Smith 1955) peat classification schemes. Several sets of samples were

183

obtained from each landslide. Most of the physical/geotechnical samples were obtained from the basal peat

184

at each Sampling Point. Samples were also obtained from the surface and middle peat at the Sampling Point

185

at bogflow ST to provide some indication of depth variations, assuming it to be representative of all three

186

sites. In addition, a Geonor H-60 field shear vane was used to measure the ‘field vane strength’ (FVS) of the

187

in-situ basal peat ~1 m behind each Sampling Point. For palaeoecological (or simply ‘botanical’) analyses,

188

monolith samples were extracted that included most of the thickness of the peat profile at each Sampling

189

Point (missing the uppermost part at SR and SA). In addition, 10 mm cubes of peat were carefully cut from

190

the auger samples from SR at approximately 200 mm depth intervals (Fig. 5 in Section 4 shows results from

191

this component of the work). Table 2 summarises the samples collected.

192 193

Table 2. Samples extracted from the study landslidesbogflows. Landslide / Physical properties – Tensile strength – position in small cores 50 mm blocks 120×120×70 peat profile dia. × 51 mm length mm

Triaxial – Shear strength (direct 38 mm dia. shear) – blocks cores 120×120×70 mm

Monoliths for botanical data 730×100×100 mm

Formatted: Line spacing: single

Formatted: Line spacing: single

ST see right ST basea SA 300-1030

6 at ~650 mm depth 3 at 10-80 mm depth 6 at ~1150 mm depth 3 at 890-960 mm depth

--

3 at 10-80 mm depth 1 at 400-1130 mm 3 at 890-960 mm depth depth

9

6

6

12

2

--

--

--

--

1

7

Formatted: Line spacing: single

Formatted: Line spacing: single

Formatted: Line spacing: single

mm depth SA base a SR 100-830 mm depth SR base b

9

6

6

12

2

--

--

--

--

1

9

6

6

12

2

Formatted: Line spacing: single

Formatted: Line spacing: single

Formatted: Line spacing: single

194 195 196 197

Notes:

Formatted: Underline

a

1600-1700 mm depth below the surface of the peat, 970-1700 mm depth for the lower monoliths

Formatted: Line spacing: 1.5 lines

b

1900-2000 mm depth, 1270-2000 mm depth for the lower monoliths

198

3.2 Laboratory testing

199

Some physical properties of peat can give a rough indication of the state or condition of the peat (Hobbs,

200

1986). Therefore, standard methods were used to determine the water content and bulk density (oven-drying

201

for 24 h at 105°C: O’Kelly 2017), loss on ignition (550°C for 3h: Skempton and Petley 1970; Andrejko et al.

202

1983; Jarrett 1983; Hobbs 1986) and saturated hydraulic conductivity (‘constant head’ method e.g. or Head

203

1994) to provide reference details for correlation with the results of the botanical and geotechnical analyses.

204

There has been some debate in recent years regarding the appropriate drying temperature for water content

205

determination, including evidence of the possibility of charring of the peat at temperatures higher than 80-

206

90°C (O’Kelly 2014). Further, O’Kelly (2014) found experimentally that the possible additional loss of mass

207

due to charring is negligible compared with the mass of any retained water due to incomplete drying,

208

particularly intracellular water within peat fibres that may constitute a significant proportions of the peat

209

mass (Foteu Madio 2013). O’Kelly (2014) therefore, and so recommended following the standard

210

specification for mineral soils of 105°C, as used by many previous workers including Skempton and Petley

211

(1970) and Hobbs (1986). We adopted the latter approach for the demonstrated reasons of standardisation

212

and comparability of results.

213 214

For the constant head method we used a laboratory permeameter arrangement as described by, for example,

215

Klute and Dirksen (1986) or Head (1994). Undisturbed core samples collected in thin-walled tubes 50 mm

216

long (‘L’) × 50.5 mm diameter were trimmed to size, saturated in tap water and mounted vertically to form a

217

permeameter maintaining a constant head of 0.15 m of water on the top of the sample. Water that passed

8

Formatted: Line spacing: Double

218

through the sample was collected underneath and measured. The constant head saturated hydraulic

219

conductivity was calculated according to Darcy’s Law from: k = VL / (tA.ΔH)

220 221

where k = permeability = saturated hydraulic conductivity (m s–1)

222

V = volume of water collected (m3)

223

t = time (s) of water collection

224

A = cross-sectional area of peat sample (2.00×10–3 m2)

225

L = length of peat sample (0.050 m)

226

ΔH = head difference across peat sample length (0.200 m)

227 228

The proportion of intracellular and interparticle water depends upon the structure and morphology of the

229

various plants present and on the degree of humification of peat (Hobbs 1986). Microfossils, including pollen

230

grains, may not represent the original in situ vegetation because they are small enough to be transported by

231

the wind, possibly over long distances. Therefore we investigated the fibres, macrofossil content and the

232

degree of humification of the peat. The latter influences the water holding capacity, pore sizes and fibre

233

quantities and properties, all of which could influence the peat strength. The fibres and macrofossils are

234

likely to directly affect the strength and other geotechnical properties.

235 236

3.2.1 Humification

237

Humification was quantitatively determined in the laboratory followed a modified version of the Bahnson

238

colorimetric method (Aaby and Tauber 1974; Blackford and Chambers 1993; Chambers et al. 1997).

239

Subsamples taken contiguously at every 10 mm from the monoliths were tested. The measurements were

240

obtained using a Hatch 2500 spectrometer set up at 540 nm. Results are expressed as ‘raw’ percentages of

241

light transmission through the diluted peat solution. The more light passes through the peat solution, the less

242

humified the sample.

243 244

3.2.2 Fibres

245

The fibre content (F) is an important characteristic that influences peat stability (Long and Jennings, 2006) as

246

it affects the peat structure and its strength properties. To explore this effect, we firstly re-defined the 9

247

different fractions as follows: (i) a ‘fine fibre’ (Fm) is a fragment or piece of plant tissue between 0.15 and

248

1.00 mm in any dimension including length; and (ii) a ‘coarse fibre’ (Rm) is a fragment or piece of plant

249

tissue > 1 mm in any dimension. In line with the ASTM’s (2008) standard for determining the fibre content

250

of peat, the ‘total fibre fraction’ (Ft) is been defined as all fibres ≥ 0.15 mm in any dimension. The humus

251

fraction (Fh) is defined as all particles < 0.15 mm in any dimension. We recognise that it is difficult to

252

determine a specific shape of some fibres and that, depending on the orientation of the fibre, any dimension

253

of a fibre or particle of a particular shape (e.g. elongated fibres) can prevent it passing through a hole in the

254

sieve, so further refinements to this methodology are likely to be needed in the future.

255 256

The Fm and Rm fibre fractions were estimated in the field according tobased on the von Post system as

257

presented by Hobbs (1986, p.79): Fine fibres (Fm) are ‘fibres and stems smaller than 1 mm in diameter or

258

width’ and coarse fibres (Rm) are ‘fibres, stems, and rootlets greater than 1 mm in diameter or width’. To

259

both of these definitions we added ‘or any plant particle’ and took the size boundary as ‘< or > 1 mm in all

260

directions’. The von Post scheme uses a four point scheme from 0 ‘nil’ to 3 ‘high content’ but without a

261

microscope it is difficult to be certain that there are no fragments of fibre present. Consequently we removed

262

‘0’ and assessed the quantity according to a five point scale: 1 = very low content (VL), 2 = low content (L),

263

3 = medium content (M), 4 = high content (H) and 5 = very high content (VH). All of the other fractions

264

defined above were determined in the laboratory and recorded using a similar 5-point scheme:

265

1 = fibre content ≤ 40%

266

2 = fibre content > 40 and ≤ 60%

267

3 = fibre content > 60 and ≤ 80%

268

4 = fibre content > 80 and ≤ 95%

269

5 = fibre content > 95%

270

Differentiating peat in this way enabled field estimates to be corrected with measurements obtained from

271

laboratory tests, and this simple 5-point scale allowed cluster analyses of the results to be carried out in a

272

consistent way.

273 274

For the laboratory determinations, duplicate subsamples of known masses were taken every 70 mm from

275

along the lowest monolith sample from each site. These were analysed differently in order to separate the 10

276

peat into different fractions, the initial part of the procedure following ASTM (2008) but with a much

277

smaller initial sample mass. Thus the fibre contents of the lowest 0.7 m of the peat profile at each landslide’s

278

Sampling Point were fully quantified. The first subsample was soaked in a dispersing agent (5% sodium

279

hexametaphosphate) for approximately 15 hours and then the peat was gently washed through a 0.15 mm

280

mesh size sieve using tap water. The fibrous material retained on the sieve was washed through a further 1

281

mm sieve and the fine fraction that passed through was collected. The fibres retained on the 1 mm sieve

282

comprised the coarse fraction. Both fractions were oven-dried at 105°C until constant masses were achieved.

283

The masses of fine and coarse fibres were combined to obtain the total mass of fibres. The mass of humus

284

was obtained from the difference between the mass of total fibres and the initial dry mass of peat determined

285

from the second subsample. The second subsample was dried at 105ºC for 24 h and the mass ratio of dry to

286

wet peat determined. The duplicate peat samples had slightly different masses and assuming that their

287

respective mass ratios of dry to ‘field wet’ peat were equal, the corresponding initial mass of the sample used

288

for fibre content testing was established. The fibre (Ff)/humus (Fh) fractions (without any mineral matter)

289

were then expressed as percentages of the initial dry mass (Ms) as follows:

Formatted: Font: 11 pt Formatted: Font: 11 pt

290

F f / h = (M f / h / M s) × 100

291

where Ff/h is the fibre/humus fraction (%), Mf/h is the mass (g) of the fibre/humus fraction after drying at

292

105°C to constant mass then subtracting the mass of ash, and Ms is the mass (g) of the initial peat sample

293

after drying at 105°C to constant mass less the mass of ash.

294 295

3.2.3 Macrofossils

296

The heterogeneity of peat is due to the variability of factors and environmental gradients that influence its

297

initiation and development (Moore, 1984; Charman, 2002). The original plant composition of peat

298

influences its structure and is assumed to affect its geotechnical properties. We used macrofossil analysis to

299

assess these botanical factors. 10 mm cubes of peat were obtained from the along the length of each monolith

300

sample, with 40–80 mm separation except within the basal peat where the cubic subsamples were

301

contiguous. Analysis was undertaken using the ‘Quadrat and Leaf Count Macrofossil Analysis technique’

302

(QLCMA) developed at the Southampton Palaeoecology Laboratory (Barber et al. 1994). The method

303

estimates the percentage coverage of all macrofossil types with the aid of a 10 × 10 grid graticule in the

304

eyepiece of a stereomicroscope. Monocotyledon epidermis tissues and Sphagnum branch leaves were 11

305

examined further at a magnification of ×400 under transmitted light. Daniels and Eddy (1990) (for

306

Sphagnum), Smith (2004) (for other bryophytes), Grosse-Brauckmann (1972) and Katz et al. (1977) (for

307

vascular plants) were used to identify the remains.

308 309

The additional small cubic samples obtained from SR were further investigated using the method developed

310

by Walker and Walker (1961), in which on a scale of 0 to 5, 0 indicates absence and 5 indicates that the

311

sample consisted largely of a particular macrofossil. This was done to check that peat at the Sampling Point

312

was representative of the entire blanket bog.

Formatted: Font: 11 pt Formatted: Font: 11 pt

313 314

3.2.4 Shear strength

315

The mechanism of failure of in-situ peat in natural landslides is uncertain and indeed there may be different

316

mechanisms operating in different contexts. Examples of these are outlined in Section 2. Our tests focused

317

on undrained shear strength because of the documented rapid development of failures compared with the

318

measured very low permeability of Irish catotelm peat (e.g. 10–6 to 10–9 m s–1: see Section 4.2.1). We used a

319

direct shear apparatus with a 100 mm × 100 mm shearbox to try to obtain reproducible values of shear

320

strength using normal stresses representing in-situ conditions, i.e. typically less than 5 kPa (after Dykes

321

2008a). Samples were sheared at normal stresses of 0.7, 1.2, 1.7, 2.2, 3.2, 4.2 and 5.6 kPa using the method

322

outlined by Dykes (2008a), i.e. no pre-consolidation, but with a slightly higher shear rate of 1 mm min–1

323

(2×10–5 m s–1) to represent moderate failure (IUGS, 1995) with the associated likelihood of some undrained

324

shearing effects..

Formatted: Superscript Formatted: Superscript

325 326

The triaxial tests were intended to give an indication of the undrained shear strength and associated stress-

327

strain behaviour of the peat by means of rapid unconsolidated-undrained tests on standard 38 mm diameter ×

328

76 mm high samples (carried out according to Head (1994)) at a range of cell pressures at the lowest end of

329

what was possible with the available equipment, i.e. 50, 100 and 200 kPa. To minimise membrane effects we

330

used thinner membranes, samples were allowed to saturate before the axial load was applied at 1 mm min–1

331

as for the direct shear tests. After each test, the sample was visually inspected to assess the failure

332

mechanism or any other deformation. The bi-linear correction of the deviator stress due to membrane

333

stiffness was not applied to the results because although the effect may be significant, (i) there is no 12

334

consensus on appropriate corrections given complex peat-membrane interactions, and (ii) this was primarily

335

a comparative study that was not necessarily expected to determine the exact value of the shear strength of

336

peat.

337 338

3.2.5 Tensile strength

339

The tensile strength of block samples of undisturbed peat was measured using the equipment (Fig. 3C) and

340

procedure described by Dykes (2008c). This involved applying a tensile load, in 100 g increments, to half of

341

the cross-sectional area of each 100 × 100 mm test sample by means of five 10 mm wide steel fingers (Fig.

342

3A), the tensile resistance being provided by the four 12.5 mm wide strips of peat between the fingers (Fig.

343

3B). Tensile stress and strain were recorded 30 s after the application of each load increment until the sample

344

failed. Although results obtained using this method have been found to be reproducible and consistent, two

345

key limitations are recognised: (i) the apparatus does not allow a vertical load to be placed on the sample to

346

replicate the condition of the basal peat in-situ; and (ii) significant sample disturbance may occur during

347

installation due to large fibres or woody fragments (Dykes 2008c).

348 349

< FIGURE 3 >

350

Figure 3. Measuring the tensile strength of the peat: (A) the two sets of steel ‘fingers’ that are pushed through the

351

centre of a cut block of undisturbed peat 100 mm high × 100 mm wide and 40-60 mm thick; (B) one half of a sample

352

following tensile failure, still adhering to one set of ‘fingers’; (C) the testing apparatus, showing: centre – the ‘fingers’

353

assembly installed (without a sample); right – the force proving ring; lower far left – the hanger for applying the

354

weights that apply the load just visible beside the end of the cupboards. Details of the design and development of this

355

apparatus are provided in Dykes (2008c).

356 357 358

4. Results

359 360

4.1 Field descriptions of the peat

361

The peat at the three landslides showed remarkably little variability in terms of structure and macrofossil

362

content. Four major stratigraphic units were identified at each site. The first unit (starting at the top),

13

Formatted: Font: Not Bold

363

including the living roots near the surface, comprised slightly humified peat, with each unit below being

364

progressively more humified and the fourth unit at the base having highly humified and/or greasy peat,

365

sometimes with bitumen or sludge like patches (Table 3; e.g. Fig. 34). The identifiable plant material in the

366

peat was predominantly monocotyledon (‘monocot’) remains (‘Turfa herbacea’ in the Troels-Smith scheme)

367

including undifferentiated roots, stems and leaves (Fig. 45). This general lack of variation between (Table 3)

368

and within (Fig. 45) sites allowed us to consider one Sampling Point as being broadly representative of each

369

landslide site.

370 371

< FIGURE 34 >

372

Figure 34. Peat stratigraphy across the slope above the head of the Slieve Rushen bogflow. This linear transect was

373

located 7.5 m upslope of the source area head at the closest point. Modified from Foteu Madio (2013).

374 375

< FIGURE 45 >

376

Figure 45. Results of macrofossil analyses of samples obtained from across the Slieve Rushen bogflow. Labels A1, A5,

377

etc. refer to sampling positions: A1 to A9 are shown in Fig. 45; E1/E3 and E4/E7 are located either side of the

378

downslope extent of the source area. The materials found at each position are from, and in the same order as, this list:

379

Charcoal (0.5-1 mm); Charcoal (less than 0.5 mm); Ericales; Eriophorum vaginatum; Monocot fragments (Monocot

380

leaves at E4), Roots; Sphagnum; Unidentified organic matter. Source: Foteu Madio (2013).

381 382

4.2 Physical and mechanical properties of the peat

383 384

4.2.1 Geotechnical characteristics

385

The basic physical properties of the peat at the three landslides are summarised in Table 4. These are broadly

386

consistent with previous results obtained from Straduff Townland (the 1997 bogflow adjacent to ST) and SA

387

by Yang and Dykes (2006). FUncorrected field shear vane readings from depths between 1.25 and 2.00 m

388

were between 6.6 and 14.0 kPa at all sites and there were insufficient results from which to identify any

389

patterns in the data. No corrections were applied because this was intended as a comparative study and the

390

shear vane is known to be inappropriate for the determination of the undrained strength of peat due to the

391

effects of fibres, although it can be used to identify patterns of peat strength variation with depth.

14

392 393

Results from the experimental low-stress direct shear tests (without pre-consolidation prior to shearing) are

394

shown in Fig. 56. These are consistent with results obtained from basal peat at another landslide in

395

northwestern Ireland (identified as ‘E6’ by Kirk (2001): Dykes 2008a). The surface peat at ST clearly

396

demonstrates a higher strength due to the greater density, and probably strength, of less humified fibres.

397

Samples inevitably consolidate under even these small loads as shearing takes place. Straight line

398

approximations in the normal stress range 2-5 kPa would all give cohesion intercepts of 1-4 kPa (Fig. 56). If

399

the peat was overconsolidated by up to 10-15 kPa as seems to be the general case (O’Kelly 2017), such low

400

shear stress values within this range of applied normal loads should not be expected.

Formatted: Font: 11 pt

401 402

Results from the unconsolidated-undrained triaxial tests similarly demonstrate the inherently low shear

403

strength of the basal peat with all three sites in the range 1.5-2.5 kPa (Fig. 67). Slight variations in the

404

diameters of the Mohr’s circles arise from the heterogeneity of the peat mass, as also observed in raised bog

405

peat by Hanrahan (1954), but may also result from gas in the peat causing variations in pore water pressures

406

within the samples. Given that the assumed effect of pore pressures in undrained shear tests is to reduce the

407

friction towards zero because the water is incompressible, then the presence of compressible gas within some

408

pore spaces could allow some (additional) frictional resistance to arise during testing. Hanrahan (1954) found

409

that the gas content of Irish Sphagnum peat may be considerably in excess of 5% of the volume and that

410

significant volumes of gases such as sulphuretted and phosphorated hydrogen (phosphine), as well as

411

methane, could be emitted during construction involving the compression of peat. Therefore the possibility

412

of gas affecting both permeability and pore pressures must be allowed for when interpreting results.

413 414

Fig. 78 shows the tensile strengths obtained from this and previous studies using the same methodology

415

(Dykes 2008c). The tensile strengths of the basal peat at the three landslides in this study are all less than 3

416

kPa except where locally reinforced by matted woody fragments.

417

With the exception of two outliers, which arose from the respective samples containing significant fragments

418

of decomposing roots or woody stems, there is an apparent trend of reducing tensile strength with depth.

419

Although this trend arises from combined results from several locations in Ireland, the similarities of all

420

other measured peat properties between all of these sites (Dykes 2008c; Dykes and Warburton 2008a; Dykes 15

Formatted: Font: Italic

421

and Jennings 2011) means that this general trend is probably real. The tensile strengths of the basal peat at

422

the three landslides in this study are all less than 3 kPa except where locally reinforced by matted woody

423

fragments.At individual sites it is possible that such a trend of decreasing tensile strength with depth may not

424

always be found, although there are insufficient relevant data to be able to comment further.

425 426

Helenelund (1967) suggested that the fibre contents, types and orientations – which depend on the

427

morphology and the mode of growth of the original plant assemblage that formed the peat – may have major

428

influences on the tensile strength. The macrofossil analyses of peats from our study sites revealed remains of

429

sedges, the degree of humification of which increase with depth. In such monocotyledon peat, fibres are the

430

remains of vascular bundles formed from the root systems that grow perpendicularly to the ground surface.

431

The resulting tensile strength will therefore be related to the resisting force produced by the fibres, the

432

frequency of which decreases with depth and is inversely proportional to the degree of humification. The

433

tensile strength results obtained by Helenelund (1967) from Sphagnum bog peat, which has very few fibres,

434

are comparable with the lowest of our results, showing that the monocotyledon peats at our sites generally

435

have higher tensile strengths than Sphagnum bog peat. Due to the effect of compression during the

436

accumulation of the peats, some fibres that were originally distributed vertically through the peat become

437

squashed progressively into a horizontal alignment as pressure increases. The degree of inclination of these

438

fibres toward the horizontal plane should therefore also increase with depth. The tensile strength values

439

presented in this study were measured in a horizontal plane, intended to represent the effect of the peat mass

440

pulling apart above a basal (shear?) failure zone. The effect of fibre orientation should be to increase the

441

tensile strength with depth since horizontal breaking up of a failing peat mass is resisted by sometimes

442

significant lengths of fibres adhering to amorphous colloidal matrix material. However, the role of living and

443

minimally decomposed roots within the near-surface acrotelm layer combined with the very high degree of

444

humification below the acrotelm appears to entirely override the fibre orientation effect.

445 446

< FIGURE 56 >

447

Figure 56. Results from experimental low-stress direct shear tests of basal peat from all three landslides and from

448

around 10-60 mm depth at Straduff Townland. Previous results from bog slide ‘E6’ at Cuilcagh Mountain, Co. Cavan,

449

obtained using the same methodology, are also shown. Modified from Foteu Madio et al. (2012), after Dykes (2008a).

16

Formatted: English (United States)

450 451

< FIGURE 67 >

452

Figure 67. Mohr’s Circles (total stresses) obtained from unconsolidated-undrained triaxial tests on peat samples from

453

the three landslides: (A) ST– Straduff Townland; (B) SR – Slieve Rushen; (C) SA – Slieve Anierin. Source: Foteu

454

Madio (2013).

455 456

< FIGURE 78 >

457

Figure 78. Tensile strength results obtained from the three landslides in this study and from previous studies using the

458

same methodology. MHA-00s refers to the Maghera bogflow, Co. Galway; SDF-08 is bogflow ST in this study; BHW-

459

08 is the Ballincollig Hill peat flow, Co. Kerry; DCM-03 is the collective reference for the 40 landslides that occurred

460

on Dooncarton Mountain, Co. Mayo on 19 September 2003, the results here being obtained from peat slide ‘SE5’.

461

Modified from Foteu Madio (2013).

462 463

4.2.2 Humification and fibres

464

The results of the quantitative determination of humification, recorded as the ‘raw’ percentage of light

465

transmission through the peat, showed no significant differences between the mean values for the three sites.

466

However, only the results from ST showed a clear reduction in light transmission (i.e. increase in degree of

467

humification) with depth (Fig. 89).

468 469

Table 3. Summary description of the peat at each landslide Sampling Point (Foteu Madio 2013). The four major

470

stratigraphic units are separated by the solid lines of the Table.

Depth (m) 0.00-0.40 0.40-0.78 0.78-1.22 1.22-1.60 1.60-1.80 >1.80

Peat profile description at ST Light brown fibrous peat, slightly humified, mainly monocotyledon fine fibres and low amorphous material, moderate horizontal tensile strength. Black and moderately humified, mainly monocotyledon fine fibre peat and moderate amorphous material, moderate horizontal tensile strength. Light brown with dark patches, very weak and moderately humified peat. Monocotyledon fine fibre limited. Low horizontal tensile strength. Brown, moderately to strongly humified peat. Monocotyledon fine fibre present. Low horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Rare and very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1on the Munsell soil colour chart.

Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single

Depth (m)

Peat profile description at SR1

Formatted: Line spacing: single

0.00-0.15

Brown fibrous peat with moderately humified, mainly monocotyledon fine fibres and low amorphous material, moderate horizontal tensile strength.

Formatted: Line spacing: single

17

0.15-0.36

Brown, less fibrous peat with moderately humified, mainly monocotyledon fine fibre peat and moderate amorphous material, moderate horizontal tensile strength.

Formatted: Line spacing: single

0.36-0.58

Dark brown humified peat with monocotyledon fragments. Low horizontal tensile strength.

Formatted: Line spacing: single

0.58-0.88

Dark brown decomposing peat with monocotyledon fragments. Low horizontal tensile strength.

Formatted: Line spacing: single

>1.64

Dark grey, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1 on the Munsell soil colour chart.

Depth (m)

Peat profile description at SA

0.88-1.58 1.58-1.64

0.00-0.76 0.76-1.56 1.56-1.76 1.76-1.78 >1.78

471 472

Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single

Dark fibrous peat, slightly humified, mainly monocotyledon fine fibres, low amorphous material and moderate horizontal tensile strength. Light brown less fibrous peat with moderately humified, mainly monocotyledon fine fibre peat, moderate amorphous material and moderate horizontal tensile strength. Black humified peat with monocotyledon fragments. Low horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1on the Munsell soil colour chart.

Note 1

Formatted: Line spacing: single

Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Line spacing: 1.5 lines

Recorded in July 2010 prior to the moorland fire.

473 474

Table 4. Summary of physical properties of peat at the three landslides., including previous data from Yang and Dykes

475

(2006)*. Water content a,b (% mass fraction)

Loss on Ignition c (%)

Saturated bulk density e (Mg m–3)

Dry bulk density e Saturated (Mg m–3) hydraulic conductivity h,i (m s–1)

700-900 *620-860 600-700 600-700 *600-740

94.5-95.6 *97.8-98.8 94.6-97.2 d 95.0 *97.7-98.5

1.00 f *1.06 1.00 1.00 *1.05

0.10-0.20 *0.13 0.10-0.20 0.20 g *0.15

Formatted: Line spacing: single

Formatted: Line spacing: single

ST SR SA

476 477 478 479 480 481 482 483 484 485 486

10–9 to 10–8 * < 10–11 10–9 to 10–6 10–8 to 10–6 * < 10–11

Notes

Formatted: Line spacing: single Formatted: Superscript Formatted: Line spacing: single Formatted: Line spacing: single Formatted: Superscript

a

There was negligible difference between field-wet and saturated water contents at all sites

b

Indicative ranges of mean values from 256-319 samples per site

c

Indicative ranges of mean values from 123-196 samples per site

d

The basal peat at Slieve Rushen was noticeably higher in organic matter than any other sampled peat

e

Mean values from 20-26 samples per site

f

The basal peat at Straduff Townland was noticeably higher (~1.10 Mg m–3) than any other sampled peat

g

The peat at Slieve Anierin had higher dry bulk densities throughout its depth

h

Indicative ranges of mean values from 19-26 samples per site, obtained using a ‘constant head’ method

i

Results obtained using a ‘falling head’ method were consistently 102-103 m s–1 higher than the respective ‘constant

head’ values

487

18

Formatted: Line spacing: 1.5 lines

488

The mean ‘total fibre fraction’ (Ft) of the lowest 0.7 m of the peat profile at each landslide Sampling Point,

489

based on 70 depth-consecutive measurements per site, was 68% at ST, 71% at SR and 56% at SA. As Fig.

490

910 shows, the latter appears to indicate a small but consistent difference from the other two, having slightly

491

fewer coarse fibres throughout the sampled depth range. At all three sites there is a general trend of reducing

492

coarse fibre content with depth but the fine fibre content seems to increase slightly towards the base of SR.

493 494

< FIGURE 89 >

495

Figure 89. ‘Raw’ percentage of light transmission at the Straduff Townland bogflow (ST).

496 497

< FIGURE 910 >

498

Figure 910. Depth variations of fibre contents throughout the lower half of the peat profile at each landslide: (A)

499

Straduff Townland, (B) Slieve Rushen, (C) Slieve Anierin. Source: Foteu Madio (2013).

500 501 502

4.3 Peat stratigraphy according to macrofossil results

503

The blanket bog at the Sampling Point at each landslide mostly comprised the remains of monocotyledon

504

plants, particularly E. vaginatum (Fig. 1011). Monocotyledon contents were lowest within the basal peat

505

zones (as defined by cluster analysis) and, at ST, immediately above the basal zone.

506 507

< FIGURE 1011 >

508

Figure 1011. Macrofossil content of the peat monolith from the three landslides: (a) ST, (b) SR, (c) SA. Parameter

509

values are raw counts for charcoal and E. vaginatum spindles, otherwise percentages. The figure shows the dendrogram

510

produced from unconstrained incremental sum square cluster analysis of strata analysed. Dashed lines separate clusters

511

corresponding to zones in the diagram. Source: Foteu Madio (2013).

512 513

4.4 Comparing botanical and geotechnical characteristics

514

The results were examined in order to identify any statistical associations (using Pearson’s ‘r’ correlation

515

coefficient) between physical/geotechnical parameters, and then between geotechnical characteristics and

516

botanical results, that may have physical explanations potentially exploitable for predictive purposes. In this

19

517

study, only significant (p < 0.05) correlations with |r| > 0.7 at all three landslides were interpreted as possibly

518

indicating a causal relationship because the study was based on a single monolith per study site.

519

Furthermore, the full depth of the peat at each site was not analysed for most of the parameters investigated.

520

The only significant associations with |r| > 0.7 that were found between physical/geotechnical parameters at

521

all three sites were between: (i) the humus fraction, Fh, and the total fibre content, Ft (Fig. 1112A); (ii) the

522

total fibre content, Ft, and the coarse fibre fraction, Rm (Fig. 1112B); (iii) the humus fraction, Fh, and the

523

coarse fibre fraction, Rm; and (iv) the coarse fibre fraction, Rm – and therefore also the total fibre content

524

and the humus fraction – and the field water content (Fig. 1112C).

525 526

Figs. 1112D and 1112E show the only consistently high correlations (p < 0.05) between macrofossil data and

527

physical/geotechnical properties of peat, i.e. between: (i) the total fibre content and the proportion of

528

monocot fragments; and (ii) the von Post degree of humification and the percentage of unidentified organic

529

matter. This may arise from the QLCMA method used for macrofossil analyses probably being more

530

appropriate for Sphagnum peat with small leaves that can be easily counted, compared with monocotyledon

531

peat with larger original plant fragments. The general lack of strong or consistent associations correlations

532

between the physical/geotechnical and botanical parameters at the three landslides suggests that these

533

physical properties cannot be used as indicators of peat mass structure and, thus, of potential peat instability.

534

However, the method used to quantify the fibre contents (Section 3.2.2, above) may be useful for

535

investigating relationships between the structural properties of failed Irish blanket peats in order to classify

536

peat for stability assessments.

537 538

The macrofossil analyses at the three landslides showed that the original plant assemblage was

539

predominantly monocotyledons, especially Eriophorum vaginatum. Therefore, the undrained strengths

540

obtained at the three landslides were plotted against the other properties (e.g. coarse fibre content in Fig.

541

12F) in order to investigate any possible relationship that may exist. The statistical analyses revealed no

542

significant correlation coefficients (p < 0.05). We are not convinced that peat strength would reduce with

543

increasing fibre content as shown in Fig. 11F and we suspect that this may be an artefact of the limited data.

544

20

545

Fig. 1213 shows the thickness of a weak basal layer at each site identified by cluster analyses of the results

546

(e.g. Fig. 1011). If all three landslides failed in a similar manner (i.e. by initial basal shearing), then it

547

appears that field observations of shear surfaces within a few tens of mm above the peat-mineral interface

548

can be explained in terms of formation of a failure zone (a) within the weakest layer of the peat profile, and

549

(b) at the lowest elevation within that weakest layer giving a continuous plane above the level of any large

550

stones or woody remnants that would resist shearing within the basal peat. The mean thickness of this layer

551

based on cluster analyses of the data (e.g. Fig. 1011) is around 170 mm, but this is clearly overestimated

552

because of the lack of a clear depth-related trend in the quantitative humification results (‘raw’ % light

553

transmission) from Straduff Townland and is probably less than 140 mm in reality.

554 555

< FIGURE 1112 >

556

Figure 1112. Correlations between physical/geotechnical parameters and between botanical characteristics of the peat.

557

(A) Total fibre content vs. humus fraction. (B) Total fibre content vs. coarse fibre fraction. (C) Coarse fibre fraction vs.

558

field water content. (D) Monocot fragments vs. total fibre content. (E) Unidentified organic matter vs. von Post

559

humification. (F) Undrained shear strength (including ‘field vane strength’) vs. total fibre content. In (A) to (E), solid

560

line = ST, long dashed line = SR and the thin broken line = SA. After Foteu Madio (2013).

561 562

< FIGURE 1213 >

563

Figure 1213. Variation of mean thickness of basal peat depths according to specific physical properties at all three

564

landslides.

565 566

5. Discussion

567 568

The three sites investigated for this study were remarkably similar in terms of the characteristics of their

569

blanket peat. Slieve Anierin had a lower fraction of identifiable monocot fragments and a correspondingly

570

higher fraction of unidentified organic matter, but this may simply reflect greater decomposition of the same

571

plants rather than being evidence of different constituents. The smaller proportion of coarse fibres throughout

572

the peat at this site, and particularly towards the base, supports the interpretation of more advanced

573

decomposition. However, the higher dry bulk density and slightly higher saturated hydraulic conductivities 21

574

(Table 4) perhaps indicate a very slightly different composition. One tensile strength measurement at this site

575

was significantly out of line with the others (Fig. 78) due to a high density of woody remains within one test

576

sample, but the other measures of shear strength were entirely consistent with the other two sites. Therefore

577

we suggest that this site has essentially the same palaeoenvironmental history of peat accumulation as the

578

others. Furthermore, the similarity between these results and some obtained from other landslide sites

579

throughout northwestern and western Ireland and Northern Ireland (e.g. Kirk 2001; Yang and Dykes 2006;

580

Dykes 2008c; Dykes and Warburton 2008a; Dykes and Jennings 2011; Dykes, this issue) and indeed eastern

581

Ireland (e.g. Boylan and Long 2010) strongly suggests that the general geotechnical characteristics of upland

582

blanket peat throughout the island of Ireland are very similar everywhere.

583 584

Much of the present vegetation of Ireland’s blanket bogs is dominated by sedges (e.g. E. vaginatum),

585

heathers (Ericacae, including Calluna vulgaris) and some Sphagnum and other mosses. These are all

586

represented in the analyses, with the sedges dominating the identifiable macrofossils (Fig. 910). In many

587

places there are the remains of trees at the base of the peat, which act like fragments of weathered bedrock to

588

resist movement of the peat over the in situ ground. However, at these three sites, separated by up to 20 km,

589

there is a weak basal layer around 150 mm thick that can be clearly distinguished from the peat above on the

590

basis of the properties measured for this study. Intriguingly, a higher proportion of the macrofossils can be

591

identified as monocot fragments in this layer, which somewhat contradicts the idea of greater decomposition.

592

On the other hand, fibre contents reduce sharply towards this basal layer (Fig. 910). O’Kelly (2017)

593

suggested that the properties of fibrous peat depend on the fibre content, but we suggest that these Irish

594

blanket peats cannot be considered to be ‘fibrous’ in the same sense, since even the acrotelm layer may

595

contain relatively few identifiable fibres. The issue is in any case unclear. Previous studies have found that

596

higher fractions of coarse fibres had no effect on measured strength compared with lower coarse fibre

597

contents (Zhang and O’Kelly 2014; Hendy et al. 2014); Price et al. (2005) found that fibre content was not

598

related to compressibility, and Lee et al. (2015) concluded that the effect of fibre orientation on frictional

599

shearing resistance was not clear. However, Boylan and Long (2010) undertook a quantitative analysis of

600

fibre contents adjacent to peat slides in Co. Wicklow and found lower fibre contents with depth. We

601

therefore conclude that the occurrence of failure in upland Irish blanket bogs must be at least in part due to

602

the lower fibre content, as well as higher overall degree of decomposition, towards the base of the peat. 22

603 604

We found some relationships between measured properties of the peat we analysed. The very strong

605

association ( |r| > 0.95) between the humus fraction, coarse fibre fraction and total fibre content at the three

606

landslides mean that only one of these parameters may be needed to investigate other properties of peat. This

607

association can be explained by the fact that with increasing plant decomposition, the size and amount of

608

organic particles decrease, resulting in low fibre contents (Fig. 1112A). When the fibre content decreases, the

609

water content also decreases (Fig. 1112C) because the voids within the fibres, which contain the largest

610

amount of water (MacFarlane and Radforth 1968), also decrease. The coarse fibres influence peat structure

611

and possibly strength (see above) and may be used for stability assessments given that at all three sites they

612

were similarly abundant and showed high ( |r| > 0.9) correlations with other properties. However, the

613

apparent uniformity of the peat across these sites precludes any suggestion that this may form the basis of a

614

generalised approach, in the absence of further studies from different peatlands (e.g. Northern England or

615

Scotland). Figs. 1112D and 1112E merely highlights the effect of humification in that if there are more fibres

616

remaining then there should also be more macrofossils that have not yet decomposed too far to be identified.

617

Fig. 1112F shows that whichever method of strength determination is used (excluding the field vane), the

618

(shear) strength of the basal peat appears to be around 2 kPa. This is consistent with stability analyses of

619

landslides involving failure within the peat (i.e. bog slides, bogflows and some peat flows sensu Dykes and

620

Warburton 2007b) as reported by Dykes (2008c), Dykes and Jennings (2011) and Farrell (2012) and with test

621

results obtained from other similar studies in Ireland (e.g. Dykes 2008c; Dykes et al. 2008).

622 623

The very low shear strength indicated above demands some consideration with respect to water conditions

624

within the peat. Blanket bogs in the British Isles may experience water table variations of 0.5–1.0 m, but

625

these are occasional reductions below the surface during warm periods of summer weather. The usual

626

condition for these deposits is to be fully saturated to the surface, i.e. with normal effective stress ≈ 0 and

627

maximum pore water pressure most of the time. Periods of summer drying may increase the normal effective

628

stress by a few kPa due to the reduced pore water pressure, i.e. temporarily increasing the effective shear

629

strength. Failure within the peat cannot, therefore, be the result of raised pore water pressures throughout the

630

peat matrix due to heavy rainfall (although it could due to external loading). The hydraulic effects of water-

631

filled pipes, cracks and other voids may play significant roles in the initiation of failure, i.e. peat mass 23

632

effects, are thought to be more important than simply the peat matrix (shear) strength, but much more

633

research is needed to test this hypothesis.

634 635

Two of the characteristics identified as being slightly different at Slieve Anierin, i.e. the monocot content and

636

the coarse fibre content, can be readily determined from small auger samples because they are quantified

637

with respect to the dry mass. A hand auger capable of cutting ‘intact’ core samples, notwithstanding issues of

638

sample deformation due to compression or fibres not being cut cleanly (Long and Boylan 2013; Hendy et al.

639

2014), could in principle provide samples for simple determination of dry bulk density and possibly saturated

640

hydraulic conductivity, i.e. the other two slightly distinctive characteristics. However, given that the

641

measured strengths at this site were no different from the others, we cannot say whether measurement of

642

those characteristics would be useful for peats formed from significantly different plant assemblages. It is not

643

possible to generalise any implications of our results for peatlands in general, and notwithstanding previous

644

comments we cannot assume that any of our correlations between botanical and geotechnical characteristics

645

will apply throughout Ireland. There is thus a clear necessity for comprehensive laboratory testing of peat

646

from the site of any proposed development, probably requiring excavation of trial pits for the extraction of

647

appropriate undisturbed samples. However, general recommendations for the most appropriate tests – and

648

testing procedures suitable for peat – will probably take some time to emerge from ongoing research

649

programmes.

650 651

Finally, the very low shear strength indicated above demands some consideration with respect to water

652

conditions within the peat. Blanket bogs in the British Isles may experience water table variations of up to

653

0.5–1.0 m, but these are occasional reductions below the surface during warm periods of summer weather

654

(Evans et al. 1999; Holden and Burt 2003). The usual condition for these deposits is to be fully saturated to

655

the surface, i.e. with normal effective stress ≈ 0 and maximum pore water pressure most of the time. Periods

656

of summer drying may increase the normal effective stress by a few kPa due to the reduced pore water

657

pressure, i.e. temporarily increasing the effective shear strength. Failure within the peat cannot, therefore, be

658

the result of raised pore water pressures throughout the peat matrix due to heavy rainfall (although it could

659

due to external loading). The hydraulic effects of water-filled pipes, cracks and other voids (e.g. Dykes, this

660

volume – in review) may play significant roles in the initiation of failure, i.e. peat mass effects, are thought 24

661

to be more important than simply the peat matrix (shear) strength, but much more research is needed to test

662

this hypothesis.

663

Dykes (2008a, p.344) wrote: ‘The research priorities are therefore to investigate the botanical controls on the

664

geotechnical properties of peat, to establish a reliable method for determining the shear strength of peat, and

665

to identify or develop a reliable method for analysing the stability of blanket bog covered slopes.’ More

666

recently, O’Kelly (2017, p.21) stated that: ‘More extensive testing of peats with different botanical

667

compositions is recommended to confirm relationships between tensile strength, other strength parameters

668

and humification level’. All of these issues are now starting to be addressed more systematically by a few

669

researchers in several countries. However, more extensive integrative research is needed, perhaps involving

670

palaeoecologists alongside geotechnical engineers, to explore the causes and geotechnical effects of different

671

peat accumulation scenarios. Detailed measurements of all possible characteristics, such as presented in this

672

study, are required for several known sites of peat landslides in each of several different biogeographical

673

zones such as Dartmoor (SW England), North Pennines (N England), Isle of Skye (W Scotland), Shetland

674

Islands (N Scotland), ideally including full depth variations at each study location in order to generate

675

sufficient data for reliable statistical analyses.

676 677

6. Conclusions and Future Work

678 679

The upland blanket bogs of northwestern Ireland appear to be formed from essentially the same assemblages

680

of plant species, dominated by sedges (mostly represented by Eriophorum vaginatum), and therefore having

681

similar physical and botanical characteristics. The data describing those characteristics show a statistically

682

distinct basal layer around 150 mm thick characterised by, in particular, a sharp reduction in the coarse – and

683

total – fibre content. Tensile strength, experimental low stress direct shear and unconsolidated undrained

684

triaxial compression measurements of peat strength converge on a value of around 2 kPa which is consistent

685

with stability back-analyses requiring undrained shear strengths of around 2 kPa for FS = 1.0. Contrary to

686

some published accounts, it appears that the lack of coarse fibres may be a contributory factor in the

687

incidence of peat slope failures. Some relationships between measured properties suggest that there may be

688

usable indicators of peat strength and stability conditions, possibly obtainable by means of samples from

25

689

hand augers, but the apparent uniformity of the peat at these three locations precludes any definitive proposal

690

of useful new methodologies at present.

691 692

Dykes (2008a, p.344) wrote: ‘The research priorities are therefore to investigate the botanical controls on the

693

geotechnical properties of peat, to establish a reliable method for determining the shear strength of peat, and

694

to identify or develop a reliable method for analysing the stability of blanket bog covered slopes.’ It has been

695

recognised for some time that the development of methods for reliably estimating the shear strength of peat

696

is likely to require some detailed investigations of botanical controls on relevant geotechnical properties (e.g.

697

Dykes 2008a). More recently, O’Kelly (2017, p.21) stated that: ‘More extensive testing of peats with

698

different botanical compositions is recommended to confirm relationships between tensile strength, other

699

strength parameters and humification level’. All of these issues are now starting to be addressed more

700

systematically by a few researchers in several countries. However, more extensive integrative research is

701

needed, perhaps involving palaeoecologists alongside geotechnical engineers, to explore the causes and

702

geotechnical effects of different peat accumulation scenarios. Detailed measurements of all possible

703

characteristics, such as presented in this study, are required for several known sites of peat landslides in each

704

of several different biogeographical zones such as Dartmoor (SW England), North Pennines (N England),

705

Isle of Skye (W Scotland), Shetland Islands (N Scotland), ideally including full depth variations at each

706

study location in order to generate sufficient data for reliable statistical analyses.

707 708

Acknowledgements

709 710

This work was funded by Kingston University’s Centre for Earth and Environmental Science Research (CEESR)

711

studentship support fund. EF thanks Prof M Waller (Kingston University), Dr P Hughes (University of Southampton)

712

and Dr M Grant (Kingston University/Wessex Archaeology) for advice and assistance with palaeoecological research

713

techniques. We are grateful to Prof E Bromhead for redrawing Figure 34, and to Mr C Somerfield for assistance with

714

the triaxial testing.

715 716

References

717

Aaby B and Tauber H (1974) Rates of peat formation in relation to degree of humification and local environment as

718

shown by studies of a raised bog in Denmark. Boreas (Oslo) 4(1), 1-18. 26

719

Andrejko MJ, Fiene F and Cohen AD (1983) Comparison of ashing techniques for determination of the inorganic

720

content of peats. In Testing of Peats and Organic Soils (Jarrett PM (ed)). ASTM Special Technical Publication, 820.

721

American Society for Testing and Materials, Philadelphia, pp.5-20.

722 723 724 725 726 727 728 729 730 731 732 733

ASTM (2008) D1997-91: Standard test method for laboratory determination of the fiber content of peat samples by dry mass. American Society for Testing and Materials, Philadelphia. Barber KE, Chambers FM, Maddy D, Stoneman R and Brew JS (1994) A sensitive high resolution record of late Holocene climatic change from a raised bog in northern England. The Holocene 4, 198-205. Blackford JJ and Chambers FM (1993) Determining the degree of peat decomposition for peat based palaeoclimatic studies. International Peat Journal 5, 7-24. Boylan N and Long M (2010) An investigation of two peat slope failures in the Wicklow mountains. Biology and Environment: Proceedings of the Royal Irish Academy 110B (3), 173-184. Carlsten P (1993) Peat - Geotechnical Properties and Up-to-Date Methods of Design and Construction. State-of-the-ArtReport. Linköping: Swedish Geotechnical Institute. Chambers FM, Barber KE, Maddy D and Brew JS (1997) A 5500-year proxy-climate and vegetation record from blanket mire at Talla Moss, borders, Scotland. The Holocene 7, 391-399.

734

Charman DJ (2002) Peatlands and Environmental Change. Wiley, Chichester.

735

Crofton HT (1902) How Chat Moss broke out in 1526. Transactions of the Lancashire and Cheshire Antiquarian

736 737 738

Society XX, 139-144. Daniels RE and Eddy A (1990) Handbook of European Sphagna. Institute of Terrestrial Ecology, Natural Environment Research Council. HMSO, London.

739

Dykes AP (2008a) Properties of peat relating to instability of blanket bogs. In Landslides and Engineered Slopes,

740

Volume 1 (Chen ZY, Zhang J, Li Z, Wu A and Ho K (eds)). Taylor and Francis, London, UK, pp.339-345.

741

Dykes AP (2008b) Geomorphological maps of Irish peat landslides created using hand-held GPS. Journal of Maps

742 743 744 745 746 747 748 749 750

v2008, 258-276. Dykes AP (2008c) Tensile strength of peat: laboratory measurement and role in Irish blanket bog failures. Landslides 5(4), 417-429. Dykes AP (2009) Geomorphological maps of Irish peat landslides created using hand-held GPS - Second Edition. Journal of Maps v2009, 179-185. Dykes AP and Jennings P (2011) Peat slope failures and other mass movements in western Ireland, August 2008. Quarterly Journal of Engineering Geology and Hydrogeology 44(1), 5-16. Dykes AP and Kirk KJ (2001) Initiation of a multiple peat slide on Cuilcagh Mountain, Northern Ireland. Earth Surface Processes and Landforms 26, 395-408. 27

751

Dykes AP (this issue) New insights from a recent peat slide at Croaghan, Co. Antrim, Northern Ireland.

752

Dykes AP and Warburton J (2007a) Significance of geomorphological and subsurface drainage controls on failures of

753

peat-covered hillslopes triggered by extreme rainfall. Earth Surface Processes and Landforms 32, 1841-1862.

754 755 756 757 758 759 760 761 762 763 764

Dykes AP and Warburton J (2007b) Mass movements in peat: A formal classification scheme. Geomorphology 86(1-2), 73-93. Dykes AP and Warburton J (2008a) Failure of peat-covered hillslopes at Dooncarton Mountain, Co. mayo, Ireland: Analysis of topographic and geotechnical factors. Catena 72, 129-145. Dykes AP and Warburton J (2008b) Characteristics of the Shetland Islands (UK) peat slides of 19 September 2003. Landslides 5, 213-226. Dykes AP, Gunn J and Convery (Née Kirk) KJ (2008) Landslides in blanket peat on Cuilcagh Mountain, northwest Ireland. Geomorphology 102, 325-340. Evans MG, Burt TP, Holden J, Adamson JK (1999) Runoff generation and water table fluctuations in blanket peat: evidence from UK data spanning the dry summer of 1995. Journal of Hydrology 221, 141-160. Farrell ER (2012) Organics/peat soils. In ICE Manual of Geotechnical Engineering: Volume 1, Geotechnical

765

Engineering Principles, Problematic Soils and Site Investigation (Burland J, Chapman T, Skinner H and Brown M

766

(eds)). ICE Publishing, London, UK, pp.463-479.

767 768

Foteu Madio ES (2013) Botanical and geotechnical influences on peat instability. Unpublished PhD thesis, Kingston University, UK.

769

Foteu Madio ES, Dykes AP, Waller MP, Hughes P and Grant MJ (2012) Botanical and geotechnical influences on peat

770

instability. In Landslides and Engineered Slopes: Protecting Society through Improved Understanding (Vol. 2)

771

(Eberhardt E, Froese C, Turner AK and Leroueil S (eds)). Proceedings of the 11th International and 2nd North

772

American Symposium on Landslides and Engineered Slopes. CRC Press, London, pp.421-427.

773 774 775 776

Gilman K, Newson MD (1980) Soil Pipes and Pipeflow: a Hydrological Study in Upland Wales, British Geomorphological Research Group Monograph 1. Geobooks: Norwich. Grosse-Brauckmann G (1972) Über pflanzliche Makrofossilien mitteleuropäischer Torfe - I. Gewebereste krautiger Pflanzen und ihre Merkmale. Telma2, 19-55.

777

Hammond RF (1979) The Peatlands of Ireland. Survey Bulletin No. 35. An Foras Talúntais, Dublin

778

Hanrahan ET (1954) An investigation of some physical properties of peat. Géotechnique 4, 108-123.

779

Head KH (1994) Manual of Soil Laboratory Testing, Permeability, Shear Strength and Compressibility Testing.

780 781 782

Volume 2. Wiley, New York. Helenelund KV (1967) Vane tests and tension tests on fibrous peat. Proceedings of the Geotechnical Conference, Oslo, Vol. 1, 199–203. 28

783

Hendy MT, Barbour SL and Martin CD (2014) Evaluating the effect of fiber reinforcement on the anisotropic undrained

784

stiffness and strength of peat. Journal of Geotechnical and Geoenvironmental Engineering, published on -line at

785

doi:10.1061/(ASCE)GT.1943-5606.0001154.

786 787 788 789 790 791 792 793

Hobbs NB (1986) Mire morphology and the properties and behaviour of some British and foreign peats. Quarterly Journal of Engineering Geology 19, 7-80. Holden J, Burt TP (2003) Hydrological studies on blanket peat: the significance of the acrotelm–catotelm model. Journal of Ecology 91: 103–113. IUGS – International Union of Geological Sciences Working Group on Landslides (1995) A suggested method for describing the rate of movement of landslides. Bulletin International Association of Engineering Geology 52, 75-78. Jarrett PM (1983) Summary. In Testing of Peats and Organic Soils (Jarrett PM (ed)). ASTM Special Technical Publication, 820. American Society for Testing and Materials, Philadelphia, pp.233-237.

794

Katz NJ, Katz SV and Skobeyeva EI (1977) Atlas of plant remains in peat soil. Nedra [In Russian].

795

Kirk KJ (2001) Instability of blanket bog slopes on Cuilcagh Mountain, N.W. Ireland. Unpublished PhD thesis.

796 797

University of Huddersfield, UK. Klute A, Dirksen C (1986) Hydraulic conductivity and diffusivity: laboratory methods. Methods of Soil Analysis

798

Part 1. Physical and Mineralogical Methods. Agronomy Monograph (2nd edition), Vol. 9. Soil Science Society

799

of America, Madison, pp. 687– 734. Landva AO and Pheeney PE (1980) Peat fabric and structure. Canadian Geotechnical Journal 17(3), 416-435.

801

Lee J-S, Seo S-Y and Lee C (2015) Geotechnical and geophysical characteristics of muskeg samples from Alberta,

803 804 805 806 807 808

Canada. Engineering Geology 195, 135-141. Lindsay R, Bragg O (2005) Wind farms and blanket peat: a report on the Derrybrien bog slide (2nd edition). Derrybrien Development Cooperative Ltd., Gort. Long M (2005) Review of peat strength, peat characterisation and constitutive modelling of peat with reference to landslides. Studia Geotechnica et Mechanica XXVII, 67-90. Long M and Boylan N (2013) Predictions of settlement in peat soils. Quarterly Journal of Engineering Geology and Hydrogeology 46, 303-322.

809

Long M and Jennings P (2006) Analysis of the peat slide at Pollatomish, County Mayo, Ireland. Landslides 3(1), 51-61.

810

MacFarlane IC and Radforth NW (1968) Structure as a basis of peat classification. National Research Council of

811

Canada (Ottawa). Reprinted from Proceedings, Third International Peat Congress held in Quebec, Canada, 18-23

812

August 1968, pp.91-97.

813

Formatted: Font: 10 pt Formatted: Font: 10 pt Formatted: Font: 10 pt Formatted: Font: 10 pt

800

802

Formatted: Indent: Left: 0", Hanging: 0.2", Line spacing: Double, Adjust space between Latin and Asian text, Adjust space between Asian text and numbers

Mills AJ (2002) Peat slides: morphology, mechanisms and recovery. Unpubl. PhD thesis. University of Durham, UK.

29

Formatted: Font: 10 pt Formatted: Font: 10 pt

814 815 816 817 818 819 820 821 822 823

Moore PD (1984) The classification of mires: An introduction. In European Mires (Moore PD (ed)). Academic Press, London, pp.1-10. O’Kelly BC (2014) Drying temperature and water content–strength correlations. Environmental Geotechnics 1 (EG2), 81-95. O’Kelly BC (2017) Measurement, interpretation and recommended use of laboratory strength properties of fibrous peat. Geotechnical Research, published on-line at http://dx.doi.org/10.1680/jgere.17.00006. Price JS, Cagampan J and Kellner E (2005) Assessment of peat compressibility: is there an easy way? Hydrological Processes 19, 3469-3475. Skempton AW and Petley DN (1970) Ignition loss and other properties of peats and clays from Avonmouth, Kings Lynn and Cranberry Moss. Geotechnique 20(4), 343-356.

824

Smith AJE (2004) The moss flora of Britain and Ireland (2nd edition). Cambridge University Press, Cambridge.

825

Troels-Smith J (1955) Karakterisering af lose jordater (Characterisation of unconsolidated sediments). Denmarks

826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845

Geologiske Undersogelse 4(3), 1-73. von Post L (1922) Sveriges geologiska undersoknings torvinventering och nagre av dess hittills vunna resultat, sr. mosskulturfor. Tidskr 1, 1-27. Walker D and Walker PM (1961) Stratigraphic evidence of regeneration in some Irish bogs. Journal of Ecology 49, 169-185. Warburton J, Higgit D and Mills A (2003) Anatomy of a Pennine peat slide, northern England. Earth Surface Processes and Landforms 28(5), 457-473. Warburton J, Holden J and Mills AJ (2004) Hydrological controls of superficial mass movements in peat. Earth-Science Reviews 67(1-2), 139-156. Yang J and Dykes AP (2006) The liquid limit of peat and its application to the understanding of Irish blanket bog failures. Landslides 3(3), 205-216. Zhang L and O’Kelly BC 2014 The principle of effective stress and triaxial compression testing of peat. Geotechnical Engineering 167, 40-50.

List of Figures

Figure 1. Location of the study area in northwest Ireland, showing the distribution of peatlands (after Hammond 1979). The outlined rectangle is enlarged to show the locations of the three bogflows: (left to right) ST = Straduff Townland, SA = Slieve Anierin, SR = Slieve Rushen. Modified from Yang and Dykes (2006). Figure 2. General views of the three study areas. (A) Straduff Townland bogflow, looking downslope from above the head (July 2010). (B) Slieve Rushen bogflow, looking across at the failed slope from the other side of the peat

30

846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886

basin into which its displaced peat flowed (July 2010). (C) Slieve Anierin bogflow from the air (Nov. 1998, photo by APD). Figure 3. Measuring the tensile strength of the peat: (A) the two sets of steel ‘fingers’ that are pushed through the centre of a cut block of undisturbed peat 100 mm high × 100 mm wide and 40-60 mm thick; (B) one half of a sample following tensile failure, still adhering to one set of ‘fingers’; (C) the testing apparatus, showing: centre – the ‘fingers’ assembly installed (without a sample); right – the force proving ring; lower far left – the hanger for applying the weights that apply the load just visible beside the end of the cupboards. Details of the design and development of this apparatus are provided in Dykes (2008c). Figure 34. Peat stratigraphy across the slope above the head of the Slieve Rushen bogflow. This linear transect was located 7.5 m upslope of the source area head at the closest point. Modified from Foteu Madio (2013). Figure 45. Results of macrofossil analyses of samples obtained from across the Slieve Rushen bogflow. Labels A1, A5, etc. refer to sampling positions: A1 to A9 are shown in Fig. 45; E1/E3 and E4/E7 are located either side of the downslope extent of the source area. The materials found at each position are from, and in the same order as, this list: Charcoal (0.5-1 mm); Charcoal (less than 0.5 mm); Ericales; Eriophorum vaginatum; Monocot fragments (Monocot leaves at E4), Roots; Sphagnum; Unidentified organic matter. Source: Foteu Madio (2013). Figure 56. Results from experimental low-stress direct shear tests of basal peat from all three landslides and from around 10-60 mm depth at Straduff Townland. Previous results from bog slide ‘E6’ at Cuilcagh Mountain, Co. Cavan, obtained using the same methodology, are also shown. Modified from Foteu Madio et al. (2012), after Dykes (2008a). Figure 67. Mohr’s Circles obtained from unconsolidated-undrained triaxial tests on peat samples from the three landslides: (A) ST– Straduff Townland; (B) SR – Slieve Rushen; (C) SA – Slieve Anierin. Source: Foteu Madio (2013). Figure 78. Tensile strength results obtained from the three landslides in this study and from previous studies using the same methodology. MHA-00s refers to the Maghera bogflow, Co. Galway; SDF-08 is bogflow ST in this study; BHW-08 is the Ballincollig Hill peat flow, Co. Kerry; DCM-03 is the collective reference for the 40 landslides that occurred on Dooncarton Mountain, Co. Mayo on 19 September 2003, the results here being obtained from peat slide ‘SE5’. Modified from Foteu Madio (2013). Figure 89. ‘Raw’ percentage of light transmission at the Straduff Townland bogflow (ST). Figure 910. Depth variations of fibre contents throughout the lower half of the peat profile at each landslide: (A) Straduff Townland, (B) Slieve Rushen, (C) Slieve Anierin. Source: Foteu Madio (2013). Figure 1011. Macrofossil content of the peat monolith from the three landslides: (a) ST, (b) SR, (c) SA. Parameter values are raw counts for charcoal and E. vaginatum spindles, otherwise percentages. The figure shows the dendrogram produced from unconstrained incremental sum square cluster analysis of strata analysed. Dashed lines separate clusters corresponding to zones in the diagram. Source: Foteu Madio (2013). Figure 1112. Correlations between physical/geotechnical parameters and between botanical characteristics of the peat. (A) Total fibre content vs. humus fraction. (B) Total fibre content vs. coarse fibre fraction. (C) Coarse fibre fraction vs. field water content. (D) Monocot fragments vs. total fibre content. (E) Unidentified organic matter vs. von Post humification. (F) Undrained shear strength (including ‘field vane strength’) vs. total fibre content. In (A) to (E), solid line = ST, long dashed line = SR and the thin broken line = SA. After Foteu Madio (2013). Figure 1213.Variation of mean thickness of basal peat depths according to specific physical properties at all three landslides.

31

887

888 889 890 891 892 893 894 895 896 897 898 899

Figure 1. Location of the study area in northwest Ireland, showing the distribution of peatlands (grey shading, after Hammond 1979). The outlined rectangle is enlarged, right, to show the relative locations of the three bogflows and their upland contexts: (left to right) ST = Straduff Townland, SA = Slieve Anierin, SR = Slieve Rushen. In this study area map, grey shading is land above 300 m elevation, horizontal stripes indicate water bodies and the solid black line is the international border. Modified from Yang and Dykes (2006).

32

900

901

902 903 904 905 906

Figure 2. General views of the three study areas. (A) Straduff Townland bogflow, looking downslope from above the head (July 2010). (B) Slieve Rushen bogflow, looking across at the failed slope from the other side of the peat basin into which its displaced peat flowed (July 2010). (C) Slieve Anierin bogflow from the air (Nov. 1998, photo by APD).

33

Formatted: Centered

907 908 909 910 911 912 913

914 915 916

Figure 3. Measuring the tensile strength of the peat: (A) the two sets of steel ‘fingers’ that are pushed through the centre of a cut block of undisturbed peat 100 mm high × 100 mm wide and 40-60 mm thick; (B) one half of a sample following tensile failure, still adhering to one set of ‘fingers’; (C) the testing apparatus, showing: centre – the ‘fingers’ assembly installed (without a sample); right – the force proving ring; lower far left – the hanger for applying the weights that apply the load just visible beside the end of the cupboards. Details of the design and development of this apparatus are provided in Dykes (2008c).

Figure 34. Peat stratigraphy across the slope above the head of the Slieve Rushen bogflow. This linear transect was located 7.5 m upslope of the source area head at the closest point. Modified from Foteu Madio (2013).

34

917 918 919 920 921 922

Figure 45. Results of macrofossil analyses of samples obtained from across the Slieve Rushen bogflow. Labels A1, A5, etc. refer to sampling positions: A1 to A9 are shown in Fig. 45; E1/E3 and E4/E7 are located either side of the downslope extent of the source area. The materials found at each position are from, and in the same order as, this list: Charcoal (0.5-1 mm); Charcoal (less than 0.5 mm); Ericales; Eriophorum vaginatum; Monocot fragments (Monocot leaves at E4), Roots; Sphagnum; Unidentified organic matter. Source: Foteu Madio (2013).

8.0 7.0

Shear stress (kPa)

6.0 5.0 4.0 3.0

ST ST near surface SR SA E6 replicate (Dykes 2008a) E6 (Dykes 2008a)

2.0 1.0 0.0 0.0

923 924 925 926 927 928

1.0

2.0

3.0

4.0 5.0 6.0 7.0 8.0 Normal stress (kPa)

9.0

10.0 11.0 12.0

Figure 56. Results from experimental low-stress direct shear tests of basal peat from all three landslides and from around 10-60 mm depth at Straduff Townland. Previous results from bog slide ‘E6’ at Cuilcagh Mountain, Co. Cavan, obtained using the same methodology, are also shown. Modified from Foteu Madio et al. (2012), after Dykes (2008a).

35

929

930

931 932 933 934

Figure 67. Mohr’s Circles obtained from unconsolidated-undrained triaxial tests on peat samples from the three landslides: (A) ST– Straduff Townland; (B) SR – Slieve Rushen; (C) SA – Slieve Anierin. Source: Foteu Madio (2013).

0

1

2

3

4

Tensile strength (kPa) 5 6 7 8 9 10 11 12 13 14 15 16 17

0 0.2 0.4

Depth (m)

0.6 0.8 1 1.2 1.4 1.6 1.8

935 936 937 938 939 940

2

ST SR SA MHA-00s (Dykes 2008c) SDF-08 (Dykes and Jennings 2011) BHW-08 (Dykes and Jennings 2011) DCM-03: SE5 (Dykes and Warburton 2008)

Figure 78. Tensile strength results obtained from the three landslides in this study and from previous studies using the same methodology. MHA-00s refers to the Maghera bogflow, Co. Galway; SDF-08 is bogflow ST in this study; BHW08 is the Ballincollig Hill peat flow, Co. Kerry; DCM-03 is the collective reference for the 40 landslides that occurred on Dooncarton Mountain, Co. Mayo on 19 September 2003, the results here being obtained from peat slide ‘SE5’. Modified from Foteu Madio (2013).

941 36

942 943 944 945 946 70

y = 53.03 - 19.05x R² = 0.5405

Transmission (%)

60 50 40 30 20 10 0

947 948

0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 Depth below ground surface (m) Figure 89. ‘Raw’ percentage of light transmission at the Straduff Townland bogflow (ST).

949 950 951 952 Percentage (%) of the initial dry mass of peat

A 0

10

20

30

40

50

60

70

80

90

100

0.9 Humus fraction (<0.15 mm) Fine fibres (0.15-1mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

1.0

Depth (m)

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8

953

1.9

37

B 0

10

20

Percentage (%) of the initial dry mass of peat 30 40 50 60 70 80

90

100

0.9 1.0 1.1

Depth (m)

1.2 1.3 1.4 1.5 1.6 Humus fraction (<0.15 mm) Fine fibres (0.15-1 mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

1.7 1.8 1.9

954 C

0

10

20

Percentage (%) of the initial dry mass of peat 30 40 50 60 70 80

90

100

0.9 1.0 1.1

Depth (m)

1.2 1.3 1.4 1.5 1.6

Humus fraction (<0.15 mm) Fine fibres (0.15-1 mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

1.7 1.8

955 956 957

1.9

Figure 910. Depth variations of fibre contents throughout the lower half of the peat profile at each landslide: (A) Straduff Townland, (B) Slieve Rushen, (C) Slieve Anierin. Source: Foteu Madio (2013).

958 959 960 961

38

962

963

964 965 966 967 968 969

Figure 1011. Macrofossil content of the peat monolith from the three landslides: (a) ST, (b) SR, (c) SA. Parameter values are raw counts for charcoal and E. vaginatum spindles, otherwise percentages. The figure shows the dendrogram produced from unconstrained incremental sum square cluster analysis of strata analysed. Dashed lines separate clusters corresponding to zones in the diagram. Source: Foteu Madio (2013).

39

(A)

(B)

100 90 80 70 60 50 40 30 20 10 0

Total fibre content (%)

R2 = 0.98 R2 = 0.93 R2 = 0.86

971 972

100 90 80 70 60 50 40 30 20 10 0

ST SR SA

Total fibre content (%)

970

(C)

(D) Monocotyledon fragments (%)

100 90 80 70 60 50 40 30 20 10 0

Field water content (%)

1000 900 800 700 600 R2 = 0.66

500

R2 = 0.67

400

R2 = 0.41

300

ST SR SA

0 10 20 30 40 50 60 70 80 90 100 Coarse fibre fraction (%)

R2 = 0.09

SR

R2 = 0.33

SA

R2 = 0.32

(F) 120

ST

R2 = 0.14

14

100

SR

R2

12

Triaxial test

SA

R2 = 0.50

10

Tensile str.

80

= 0.30

Peat strength (kPa)

Unidentified organic matter (%)

ST

0 10 20 30 40 50 60 70 80 90 100 Total fibre content (%)

(E)

60 40 20

Field vane

Direct shear

8 6 4 2

0

0

4

975 976 977 978 979 980

ST SR SA

0 10 20 30 40 50 60 70 80 90 100 Coarse fibre fraction (%)

0 10 20 30 40 50 60 70 80 90 100 Humus fraction (%)

1100

973 974

R2 = 0.96 R2 = 0.96 R2 = 0.98

5 6 7 8 9 10 Degree of humification (von Post)

0

10 20 30 40 50 60 70 80 90 100

Coarse fibre fraction (%)

Figure 1112. Correlations between physical/geotechnical parameters and between botanical characteristics of the peat. (A) Total fibre content vs. humus fraction. (B) Total fibre content vs. coarse fibre fraction. (C) Coarse fibre fraction vs. field water content. (D) Monocot fragments vs. total fibre content. (E) Unidentified organic matter vs. von Post humification. (F) Undrained shear strength (including ‘field vane strength’) vs. total fibre content. In (A) to (E), solid line = ST, long dashed line = SR and the thin broken line = SA. After Foteu Madio (2013).

981 982 40

Thickness of basal layer (mm)

600

ST SR

500

SA

400 300 200 100 0

983 984 985

Field description

'Raw' % light transmission

Fibre content

Macrofossil content

Quantitative fibre content

Site average

Figure 1213. VariationDifferent estimates of mean thickness of basal peat depths according to specific physical properties at all three landslides.

41

Main Text

1 12 2 3 43 5 6 4 7 8 95 10 11 12 6 13 14 7 15 16 8 17 18 19 9 20 2110 22 2311 24 2512 26 2713 28 2914 30 3115 32 3316 34 3517 36 37 18 38 39 19 40 41 20 42 43 21 44 45 4622 47 4823 49 5024 51 5225 53 5426 55 5627 57 5828 59 6029 61 62 63 64 65

Click here to download Main Text Foteu Madio and Dykes R1 FINAL.docx

v. 17 March 2018

Botanical and geotechnical characteristics of blanket peat at three Irish bogflows

Eliane S. Foteu Madio † BSc(Hons), MSc, PhD, MCIWM, Director at Envigma Ltd., Clacton-on-Sea, UK

Alan P. Dykes* BSc(Hons), PhD, FRGS, FHEA, FGS, CGeol, Associate Professor at Kingston University, Kingston upon Thames, UK ORCID: 0000-0003-0327-0498

*Corresponding author: Centre for Engineering, Environment and Society Research, Department of Civil Engineering, Kingston University, Penrhyn Road, Kingston upon Thames, KT1 2EE, UK, Tel.: +44 (0)208 417 701, Email: [email protected]

†

Present address:

10 Crossfield Road, Clacton-on-Sea, CO15 3QT, UK, Tel.: +44 (0)1255 428787, Email: [email protected] Address at the time of this research: School of Geography and Geology, Kingston University, Kingston upon Thames, UK

7422 words (excluding Figure captions and Tables), 13 Figures, 4 Tables

Abstract

Systematic investigations of instability and failure of peat covered hillslopes began in the late 1990s and quickly identified the potential importance of botanical controls on the properties and behaviour of the blanket peat involved in the failures. However, attempts to unravel some of these controls did not begin for several years. During 2010-12 investigations of the blanket peat at three relatively recent bogflows in northwest Ireland were done with the aim of

1

30

establishing some form of relationship between botanical or paleoecological characteristics and standard physical and

31

geotechnical properties, assuming the latter to be meaningful but recognising that this may not be the case. In-situ

1 2 32 3 4 33 5 6 734 8 935 10 1136 12 1337 14 1538 16 1739 18 1940 20 2141 22 2342 24 2543 26 2744 28 2945 30 3146 32 33 47 34 35 3648 37 3849 39 4050 41 4251 43 44 4552 46 4753 48 4954 50 5155 52 53 5456 55 5657 57 5858 59 6059 61 62 63 64 65

measurements and investigations at all three sites were followed by extensive laboratory characterisation of small core, block and monolith samples.

The botanical composition of the peat could not be fully determined due to the very high degree of decomposition. However, analysis of macrofossils allowed distinct depth-related patterns of several key botanical indicators to be determined. In particular the monocotyledon fragments, dominated by Eriophorum vaginatum, showed distinct and potentially useful distributions throughout the peat profiles. Overall results showed that the basal peat at one of the sites was discernibly different from the other two sites having fewer monocotyledons, fewer fibres, higher dry bulk density and higher saturated hydraulic conductivity. This approach therefore offers a potential basis for developing a means of assessing peat mass characteristics from small auger samples.

Key words

Fabric/structure of soils, Landslides, Strength & testing of materials

1. Introduction

Records of failures of peat bogs go back around 500 years to the collapse of Chat Moss near Manchester, northwest England, in 1526 (Crofton 1902). However, until the late 1990s the occasional studies of isolated examples of peatland failures were largely descriptive with estimates of geometric characteristics and occasionally reports of the living plant assemblages present at the time of failure. Systematic investigations began into the stability of blanket peat-covered slopes following significant peat landslides in Northern Ireland (Dykes and Kirk 2001; Kirk 2001) and northern England (Mills 2002; Warburton et al. 2003, 2004). The potential importance of the botanical composition as a controlling factor for the properties and geotechnical behaviour of peat was highlighted earlier by Hobbs (1986), not least because of the widespread adoption by engineers of the von Post scheme for classifying peat deposits (e.g. Landva and Pheeney 1980; Carlsten 1993) which requires the estimation of relative frequencies of fibres and wood/shrub fragments as well as the degree of decomposition of the plant matter (i.e. the humification). The need for research into botanical controls on peat properties was further emphasised with respect to blanket peat instability by Kirk (2001), Dykes (2008a) and O’Kelly (2017) in response to findings from their investigations of physical and

2

60 161 2 362 4 563 6 7 64 8 9 1065 11 1266 13 1467 15 1668 17 18 1969 20 2170 22 2371 24 2572 26 27 2873 29 3074 31 3275 33 3476 35 36 77 37 38 3978 40 4179 42 4380 44 45 81 46 47 4882 49 5083 51 5284 53 5485 55 56 5786 58 5987 60 6188 62 63 64 65

geotechnical properties thus far. Indeed, O’Kelly (2017) highlighted the scarcity of published works on the topic and the contradictory findings from the few such studies. The present research (Foteu Madio et al. 2012; Foteu Madio 2013) arose directly from this dearth of previous studies.

The aim of this paper is to examine whether physical and geotechnical properties of Irish blanket peat can be causally associated with measurable botanical characteristics. It does so by presenting and analysing data representing the properties and characteristics of the peat at the sites of three significant bogflows in northwest Ireland, obtained from a combination of field and laboratory investigations. The importance of this study is to provide the basis for more efficient and reliable methods for assessing the stability of peat with respect to planned interventions such as construction of access roads for windfarms or other purposes.

2. Blanket Bog Failures in Ireland

The topic of peat mass movements (as distinct from geotechnical engineering of peat) emerged from an esoteric scientific by-way to become a mainstream theme in engineering geology and geomorphology following several major events in late 2003. On 19 September 2003, two entirely independent extreme rainfall events in Co. Mayo, Ireland, and South Shetland, Scotland, triggered multiple failures of peatcovered hillslopes (Dykes and Warburton 2007a, 2008a,b). More significantly for civil engineering, four weeks later the 450,000 m3 Derrybrien Windfarm landslide occurred (Lindsay and Bragg 2005). By that time it had already become clear that the Irish blanket bogs were failing in several slightly different ways, giving rise to morphologically distinctive types of failures (Dykes and Warburton 2007b). Most involve shearing of mineral soil beneath the peat (‘peaty-debris slides’), shearing at the peat-mineral interface (‘peat slides’), or shearing entirely within the basal peat (‘bog slides’). ‘Peat flows’, a term reserved for failures resulting primarily from head-loading, appear to be effectively bearing capacity failures with small areas of shear surface within the basal peat having been observed in the Derrybrien and Ballincollig Hill landslides (Long 2005; Dykes and Jennings 2011).

All of the available evidence relating to ‘bog bursts’ and ‘bogflows’ indicate that these failures involved some sort of in-situ liquefaction of the lower or basal peat, with this (semi-)liquid peat slurry then breaking 3

89 190 2 391 4 592 6 7 93 8 9 1094 11 1295 13 1496 15 1697 17 18 1998 20 2199 22 23 100 24 25 101 26 27 102 28 29 103 30 31 32 104 33 34 105 35 36 106 37 38 107 39 40 108 41 42 43 109 44 45 110 46 47 111 48 49 112 50 51 52 113 53 54 114 55 56 115 57 58 116 59 60 61 117 62 63 64 65

out from beneath a stronger confining acrotelm layer (or from cut faces through the margins of raised bogs) (Dykes and Warburton 2007b). The precise mechanisms of strength loss are unknown. One hypothesis, for example, is that the basal peat fails like ‘quick clay’ with an initial small shear failure creating a disturbance that propagates rapidly. As such a bogflow may simply be a bog slide involving weaker and wetter peat – but there is a clear distinction because these two types of failure have different peat depth vs. gradient characteristics (A P Dykes, unpublished data). A parallel hypothesis is that in some of these failures the lower layer of the peat deposit was always a fluid body, for example if peat grew over a large pond so as to eventually entirely bury it.

In almost all cases of failures of (blanket) peat-covered slopes in Ireland, landslide morphologies and runout characteristics display clear evidence of relatively rapid development of failure associated with very high volumes of rainwater, with eyewitness accounts of some recent events (e.g. the Derrybrien peat flow in 2003, the Croaghan peat slide in 2014) corroborating these interpretations. Warburton et al. (2004) discussed the various hydrological processes giving rise to, or controlling, such failures. It is likely that the peaty-debris slides are triggered by pore pressure effects, in part due to subsurface storm runoff being confined beneath a saturated and effectively impermeable peat cover. Peat slides (interface failures) probably occur for the same reason. Failure within the peat is a more complex issue because of the dual influences of effectively impermeable and normally saturated but weak catotelm peat material and the internal structure of the peat mass (sensu ‘rock mass’ considerations) that may experience high turbulent flows and even artesian conditions within networks of natural peat pipes and (relict) desiccation cracks (Dykes and Warburton 2007a, 2008a; Gilman and Newson 1980; Holden and Burt 2003). The critical factor here is that rates of deformation and then movement are likely to greatly exceed the saturated hydraulic conductivity of the intact peat mass through which any shear surface may develop. Consequently, the focus of our research is on the undrained strength characteristics of basal peat.

2.1 Study sites We identified three locally significant bogflows (sensu Dykes and Warburton 2007b) for this study, located within the same region of northwest Ireland: Straduff Townland (hereafter referred to as ‘ST’), Slieve Anierin (‘SA’) and Slieve Rushen (‘SR’) (Fig. 1). Site and landslide characteristics are summarised in Table 4

118

1 and illustrated in Fig. 2. All were relatively recent, thus limiting the degree of post-failure degradation, and

119 1

two sites (though not the same landslide at one of these sites) had been investigated previously which

2 3 120 4 5 121 6 7 122 8 9 123 10 11 124 12 13 14 125 15 16 126 17 18 127 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 128 36 129 37 130 38 39 40 131 41 42 132 43 44 133 45 46 134 47 48 135 49 50 136 51 52 137 53 54 55 138 56 57 139 58 59 140 60 61 141 62 63 64 65

provided a cross-check for the peat characterisation results from this study. Furthermore, although the peat at all three sites was generally very similar, one site (Slieve Anierin, below) was noted by Yang and Dykes (2006) to be slightly but nevertheless distinctly different from others including a bogflow at Straduff Townland adjacent to the one used for this study. We anticipated that the results of this new research would also show this.

Table 1. Summary of site details and characteristics of the study bogflows. Bogflow County Latitude Longitude

Elevation Geology Geomorphological Length Slope Deptha Volume (m) (Carboniferous) Context (m) (°) (m) (m)

ST

Sligo 54°7.2’N 8°12.9’W

405

Lackagh Sandstone

SR

Cavan 54°8.9’N 7°38.5’W

390

Glenade Sandstone

SA

Leitrim 54°6.3’N 7°58.7’W

440

Lackagh Sandstone

Escarpment failure

200

Basin slope failure 175 Escarpment failure

190

5.5 (top) 3 (mid) 6 (lower)

2.5

35,000

5.5

2.0

20,000

4

2.2

22,000

Note a

Indicative average depth of in-situ peat immediately adjacent to the landslide source area

< FIGURE 1 > Figure 1. Location of the study area in northwest Ireland, showing the distribution of peatlands (after Hammond 1979). The outlined rectangle is enlarged to show the locations of the three bogflows: (left to right) ST = Straduff Townland, SA = Slieve Anierin, SR = Slieve Rushen. Modified from Yang and Dykes (2006).

< FIGURE 2 > Figure 2. General views of the three study areas. (A) Straduff Townland bogflow, looking downslope from above the head (July 2010). (B) Slieve Rushen bogflow, looking across at the failed slope from the other side of the peat basin into which its displaced peat flowed (July 2010). (C) Slieve Anierin bogflow from the air (Nov. 1998, photo by APD). 5

142 143 1 2 3 144 4 5 145 6 7 146 8 9 147 10 11 148 12 13 14 149 15 16 150 17 18 151 19 20 152 21 22 23 153 24 25 154 26 27 155 28 29 156 30 31 32 157 33 34 158 35 36 159 37 38 160 39 40 161 41 42 43 162 44 45 163 46 47 164 48 49 165 50 51 52 166 53 54 167 55 56 168 57 58 169 59 60 61 170 62 63 64 65

ST, the Straduff Townland landslide, occurred overnight or early morning on 14 August 2008 during very heavy rain. The dominant morphology is that of a bogflow. However, a basal shear surface around 20 mm above the base of the peat was visible in two small parts of the source area (Dykes 2009; Dykes and Jennings 2011). Although the latter observation corresponds with a ‘bogslide’ (Dykes and Warburton 2007b), we will refer to this failure as a bogflow. It involved an area of intact blanket peat between the source areas of bogflows dating from 1945 and 1991, leaving narrow strips of minimally displaced peat separating the failures. The physical characteristics of the peat at the 1991 bogflow, just a few metres from the margin of the later failure, were determined by Yang and Dykes (2006). SA (the Slieve Anierin bogflow) is thought to have occurred during 1998, based on its visible condition when first seen from a light aircraft in November 1998 and a conversation with a local resident in 2011. It was described, and the physical characteristics of the peat reported, by Yang and Dykes (2006). The date of SR (the Slieve Rushen bogflow) is uncertain, but the condition of the failure when first inspected in September 2004 was consistent with an age of only a few years, i.e. it most likely occurred during the 1990s (Dykes 2008b).

3. Methods

The three bogflows were investigated using the same general methodology as previous studies of peat landslides (e.g. Yang and Dykes 2006). All had previously been surveyed in detail by Dykes (2008b, 2009). The focus for this study was to obtain samples for laboratory testing from a carefully prepared and fully described vertical profile through the full depth of undisturbed in situ peat. Most peat failures leave irregular sub-vertical peat profiles with varying amounts of peat debris covering the lower layers, around several parts of the source area margins. A single study profile (hereafter referred to as the ‘study profile’ or ‘Sampling Point’) was selected at each landslide according to the feasibility of creating a clean vertical profile through the full thickness of the peat, i.e. involving the minimum manual excavation of loose peat debris, but ensuring the in-situ peat was undisturbed and not within a few metres of any tension cracks. Safety was ensured by having wide open access to the prepared profile from within the evacuated source area of each landslide, with one person maintaining active watch over the cut face while the other person worked there. 6

171 172 1 2 3 173 4 5 174 6 7 175 8 9 176 10 11 177 12 13 14 178 15 16 179 17 18 180 19 20 181 21 22 23 182 24 25 183 26 27 184 28 29 185 30 31 32 186 33 34 187 35 36 188 37 38 189 39 40 190 41 42 43 191 44 45 192 46 47 193 48 49 194 50 51 52 195 53 54 196 55 56 197 57 58 198 59 60 61 199 62 63 64 65

3.1 Field investigations Around each bogflow source area, stratigraphic and topographic surveys were carried out in order to estimate the morphology of the peat deposit and the variability of the peat within it prior to failure. The stratigraphy and maximum depth of the in situ peat was determined on a coarse but regular grid using a 20 mm diameter gouge auger. Maximum peat depths were measured at additional locations by probing with a metal rod. Peat surface elevations were then surveyed at all the stratigraphy and peat depth measurement points by levelling, with the mineral surface elevations being measured within the landslide source areas and calculated for the peat-covered areas around the source areas from the measured peat depths.

Prior to sampling, a detailed description of the full thickness of the peat at each Sampling Point was recorded according to the von Post (von Post, 1922, as presented by Landva and Pheeney 1980, and Hobbs 1986) and Troels-Smith (Troels-Smith 1955) peat classification schemes. Several sets of samples were obtained from each landslide. Most of the physical/geotechnical samples were obtained from the basal peat at each Sampling Point. Samples were also obtained from the surface and middle peat at the Sampling Point at bogflow ST to provide some indication of depth variations, assuming it to be representative of all three sites. In addition, a Geonor H-60 field shear vane was used to measure the ‘field vane strength’ (FVS) of the insitu basal peat ~1 m behind each Sampling Point. For palaeoecological (or simply ‘botanical’) analyses, monolith samples were extracted that included most of the thickness of the peat profile at each Sampling Point (missing the uppermost part at SR and SA). In addition, 10 mm cubes of peat were carefully cut from the auger samples from SR at approximately 200 mm depth intervals (Fig. 5 in Section 4 shows results from this component of the work). Table 2 summarises the samples collected.

7

200 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 201 24 202 25 203 26 27 204 28 29 205 30 31 206 32 33 34 207 35 36 208 37 38 209 39 40 210 41 42 43 211 44 45 212 46 47 213 48 49 214 50 51 215 52 53 54 216 55 56 217 57 58 218 59 60 219 61 62 63 64 65

Table 2. Samples extracted from the study bogflows. Landslide / Physical properties – Tensile strength – position in small cores 50 mm blocks 120×120×70 peat profile dia. × 51 mm length mm ST see right ST basea SA 300-1030 mm depth SA base a SR 100-830 mm depth SR base b

Triaxial – Shear strength (direct 38 mm dia. shear) – blocks cores 120×120×70 mm

6 at ~650 mm depth 3 at 10-80 mm depth 6 at ~1150 mm depth 3 at 890-960 mm depth

--

Monoliths for botanical data 730×100×100 mm

3 at 10-80 mm depth 1 at 400-1130 mm 3 at 890-960 mm depth depth

9

6

6

12

2

--

--

--

--

1

9

6

6

12

2

--

--

--

--

1

9

6

6

12

2

Notes a

1600-1700 mm depth below the surface of the peat, 970-1700 mm depth for the lower monoliths

b

1900-2000 mm depth, 1270-2000 mm depth for the lower monoliths

3.2 Laboratory testing Some physical properties of peat can give a rough indication of the state or condition of the peat (Hobbs, 1986). Therefore, standard methods were used to determine the water content and bulk density (oven-drying for 24 h at 105°C: O’Kelly 2017), loss on ignition (550°C for 3h: Skempton and Petley 1970; Andrejko et al. 1983; Jarrett 1983; Hobbs 1986) and saturated hydraulic conductivity (‘constant head’ method) to provide reference details for correlation with the results of the botanical and geotechnical analyses. There has been some debate in recent years regarding the appropriate drying temperature for water content determination, including evidence of the possibility of charring of the peat at temperatures higher than 80-90°C (O’Kelly 2014). Further, O’Kelly (2014) found experimentally that the possible additional loss of mass due to charring is negligible compared with the mass of any retained water due to incomplete drying, particularly intracellular water within peat fibres that may constitute a significant proportions of the peat mass (Foteu Madio 2013), and so recommended following the standard specification for mineral soils of 105°C as used by many previous workers including Skempton and Petley (1970) and Hobbs (1986). We adopted the latter approach for the demonstrated reasons of standardisation and comparability of results. 8

220 221 1

For the constant head method we used a laboratory permeameter arrangement as described by, for example,

2 3 222 4 5 223 6 7 224 8 9 225 10 11 226 12 13 14 227 15 16 228 17 18 229 19 20 230 21 22 23 231 24 25 232 26 27 233 28 29 234 30 31 32 235 33 34 236 35 36 237 37 38 238 39 40 239 41 42 43 240 44 45 241 46 47 242 48 49 243 50 51 52 244 53 54 245 55 56 246 57 58 247 59 60 61 62 63 64 65

Klute and Dirksen (1986) or Head (1994). Undisturbed core samples collected in thin-walled tubes 50 mm long (‘L’) × 50.5 mm diameter were trimmed to size, saturated in tap water and mounted vertically to form a permeameter maintaining a constant head of 0.15 m of water on the top of the sample. Water that passed through the sample was collected underneath and measured. The constant head saturated hydraulic conductivity was calculated according to Darcy’s Law from: k = VL / (tA.ΔH) where k = permeability = saturated hydraulic conductivity (m s–1) V = volume of water collected (m3) t = time (s) of water collection A = cross-sectional area of peat sample (2.00×10–3 m2) L = length of peat sample (0.050 m) ΔH = head difference across peat sample length (0.200 m)

The proportion of intracellular and interparticle water depends upon the structure and morphology of the various plants present and on the degree of humification of peat (Hobbs 1986). Microfossils, including pollen grains, may not represent the original in situ vegetation because they are small enough to be transported by the wind, possibly over long distances. Therefore we investigated the fibres, macrofossil content and the degree of humification of the peat. The latter influences the water holding capacity, pore sizes and fibre quantities and properties, all of which could influence the peat strength. The fibres and macrofossils are likely to directly affect the strength and other geotechnical properties.

3.2.1 Humification Humification was quantitatively determined in the laboratory followed a modified version of the Bahnson colorimetric method (Aaby and Tauber 1974; Blackford and Chambers 1993; Chambers et al. 1997). Subsamples taken contiguously at every 10 mm from the monoliths were tested. The measurements were obtained using a Hatch 2500 spectrometer set up at 540 nm. Results are expressed as ‘raw’ percentages of

9

248

light transmission through the diluted peat solution. The more light passes through the peat solution, the less

249 1

humified the sample.

2 3 250 4 5 251

6 7 252 8 9 253 10 11 12 254 13 14 255 15 16 256 17 18 257 19 20 21 258 22 23 259 24 25 260 26 27 261 28 29 30 262 31 32 263 33 34 264 35 36 265 37 38 266 39 40 41 267 42 43 268 44 45 269 46 47 270 48 49 50 271 51 52 272 53 54 273 55 56 274 57 58 59 275 60 61 276 62 63 64 65

3.2.2 Fibres The fibre content (F) is an important characteristic that influences peat stability (Long and Jennings, 2006) as it affects the peat structure and its strength properties. To explore this effect, we firstly re-defined the different fractions as follows: (i) a ‘fine fibre’ (Fm) is a fragment or piece of plant tissue between 0.15 and 1.00 mm in any dimension including length; and (ii) a ‘coarse fibre’ (Rm) is a fragment or piece of plant tissue > 1 mm in any dimension. In line with the ASTM’s (2008) standard for determining the fibre content of peat, the ‘total fibre fraction’ (Ft) is been defined as all fibres ≥ 0.15 mm in any dimension. The humus fraction (Fh) is defined as all particles < 0.15 mm in any dimension. We recognise that it is difficult to determine a specific shape of some fibres and that, depending on the orientation of the fibre, any dimension of a fibre or particle of a particular shape (e.g. elongated fibres) can prevent it passing through a hole in the sieve, so further refinements to this methodology are likely to be needed in the future.

The Fm and Rm fibre fractions were estimated in the field based on the von Post system as presented by Hobbs (1986, p.79): Fine fibres (Fm) are ‘fibres and stems smaller than 1 mm in diameter or width’ and coarse fibres (Rm) are ‘fibres, stems, and rootlets greater than 1 mm in diameter or width’. To both of these definitions we added ‘or any plant particle’ and took the size boundary as ‘< or > 1 mm in all directions’. The von Post scheme uses a four point scheme from 0 ‘nil’ to 3 ‘high content’ but without a microscope it is difficult to be certain that there are no fragments of fibre present. Consequently we removed ‘0’ and assessed the quantity according to a five point scale: 1 = very low content (VL), 2 = low content (L), 3 = medium content (M), 4 = high content (H) and 5 = very high content (VH). All of the other fractions defined above were determined in the laboratory and recorded using a similar 5-point scheme: 1 = fibre content ≤ 40% 2 = fibre content > 40 and ≤ 60% 3 = fibre content > 60 and ≤ 80% 4 = fibre content > 80 and ≤ 95% 5 = fibre content > 95% 10

277

Differentiating peat in this way enabled field estimates to be corrected with measurements obtained from

278 1

laboratory tests, and this simple 5-point scale allowed cluster analyses of the results to be carried out in a

2 3 279 4 5 280 6 7 281 8 9 282 10 11 283 12 13 14 284 15 16 285 17 18 286 19 20 287 21 22 23 288 24 25 289 26 27 290 28 29 291 30 31 32 292 33 34 293 35 36 294 37 38 295 39 40 296 41 42 43 297 44 45 298 46 47 299 48 49 300 50 51 52 301 53 54 302 55 56 303 57 58 304 59 60 61 305 62 63 64 65

consistent way.

For the laboratory determinations, duplicate subsamples of known masses were taken every 70 mm from along the lowest monolith sample from each site. These were analysed differently in order to separate the peat into different fractions, the initial part of the procedure following ASTM (2008) but with a much smaller initial sample mass. Thus the fibre contents of the lowest 0.7 m of the peat profile at each landslide’s Sampling Point were fully quantified. The first subsample was soaked in a dispersing agent (5% sodium hexametaphosphate) for approximately 15 hours and then the peat was gently washed through a 0.15 mm mesh size sieve using tap water. The fibrous material retained on the sieve was washed through a further 1 mm sieve and the fine fraction that passed through was collected. The fibres retained on the 1 mm sieve comprised the coarse fraction. Both fractions were oven-dried at 105°C until constant masses were achieved. The masses of fine and coarse fibres were combined to obtain the total mass of fibres. The mass of humus was obtained from the difference between the mass of total fibres and the initial dry mass of peat determined from the second subsample. The second subsample was dried at 105ºC for 24 h and the mass ratio of dry to wet peat determined. The duplicate peat samples had slightly different masses and assuming that their respective mass ratios of dry to ‘field wet’ peat were equal, the corresponding initial mass of the sample used for fibre content testing was established. The fibre (Ff)/humus (Fh) fractions (without any mineral matter) were then expressed as percentages of the initial dry mass (Ms) as follows: F f / h = (M f / h / M s) × 100 where Ff/h is the fibre/humus fraction (%), Mf/h is the mass (g) of the fibre/humus fraction after drying at 105°C to constant mass then subtracting the mass of ash, and Ms is the mass (g) of the initial peat sample after drying at 105°C to constant mass less the mass of ash.

3.2.3 Macrofossils The heterogeneity of peat is due to the variability of factors and environmental gradients that influence its initiation and development (Moore, 1984; Charman, 2002). The original plant composition of peat influences its structure and is assumed to affect its geotechnical properties. We used macrofossil analysis to 11

306

assess these botanical factors. 10 mm cubes of peat were obtained from the along the length of each monolith

307 1

sample, with 40–80 mm separation except within the basal peat where the cubic subsamples were

2 3 308 4 5 309 6 7 310 8 9 311 10 11 312 12 13 14 313 15 16 314 17 18 315 19 20 316 21 22 23 317 24 25 318 26 27 319 28 29 320 30 31 32 321 33 34 322 35 36 323 37 38 39 324 40 41 325 42 43 326 44 45 327 46 47 48 328 49 50 329 51 52 330 53 54 331 55 56 57 332 58 59 333 60 61 334 62 63 64 65

contiguous. Analysis was undertaken using the ‘Quadrat and Leaf Count Macrofossil Analysis technique’ (QLCMA) developed at the Southampton Palaeoecology Laboratory (Barber et al. 1994). The method estimates the percentage coverage of all macrofossil types with the aid of a 10 × 10 grid graticule in the eyepiece of a stereomicroscope. Monocotyledon epidermis tissues and Sphagnum branch leaves were examined further at a magnification of ×400 under transmitted light. Daniels and Eddy (1990) (for Sphagnum), Smith (2004) (for other bryophytes), Grosse-Brauckmann (1972) and Katz et al. (1977) (for vascular plants) were used to identify the remains.

The additional small cubic samples obtained from SR were further investigated using the method developed by Walker and Walker (1961), in which on a scale of 0 to 5, 0 indicates absence and 5 indicates that the sample consisted largely of a particular macrofossil. This was done to check that peat at the Sampling Point was representative of the entire blanket bog.

3.2.4 Shear strength The mechanism of failure of in-situ peat in natural landslides is uncertain and indeed there may be different mechanisms operating in different contexts. Examples of these are outlined in Section 2. Our tests focused on undrained shear strength because of the documented rapid development of failures compared with the measured very low permeability of Irish catotelm peat (e.g. 10–6 to 10–9 m s–1: see Section 4.2.1). We used a direct shear apparatus with a 100 mm × 100 mm shearbox to try to obtain reproducible values of shear strength using normal stresses representing in-situ conditions, i.e. typically less than 5 kPa (after Dykes 2008a). Samples were sheared at normal stresses of 0.7, 1.2, 1.7, 2.2, 3.2, 4.2 and 5.6 kPa using the method outlined by Dykes (2008a), i.e. no pre-consolidation, but with a slightly higher shear rate of 1 mm min–1 (2×10–5 m s–1) to represent moderate failure (IUGS, 1995) with the associated likelihood of undrained shearing effects.

The triaxial tests were intended to give an indication of the undrained shear strength and associated stressstrain behaviour of the peat by means of rapid unconsolidated-undrained tests on standard 38 mm diameter × 12

335

76 mm high samples (carried out according to Head (1994)) at a range of cell pressures at the lowest end of

336 1

what was possible with the available equipment, i.e. 50, 100 and 200 kPa. To minimise membrane effects we

2 3 337

used thinner membranes, samples were allowed to saturate before the axial load was applied at 1 mm min–1

4 5 338

6 7 339 8 9 340 10 11 341 12 13 14 342 15 16 343 17 18 344 19 20 345 21 22 23 346 24 25 347 26 27 348 28 29 349 30 31 32 350 33 34 351 35 36 352 37 38 353 39 40 354 41 42 43 355 44 45 356 46 47 48 357 49 50 358 51 52 359 53 54 360 55 56 361 57 58 362 59 60 363 61 62 63 64 65

as for the direct shear tests. After each test, the sample was visually inspected to assess the failure mechanism or any other deformation. The bi-linear correction of the deviator stress due to membrane stiffness was not applied to the results because although the effect may be significant, (i) there is no consensus on appropriate corrections given complex peat-membrane interactions, and (ii) this was primarily a comparative study that was not necessarily expected to determine the exact value of the shear strength of peat.

3.2.5 Tensile strength The tensile strength of block samples of undisturbed peat was measured using the equipment (Fig. 3C) and procedure described by Dykes (2008c). This involved applying a tensile load, in 100 g increments, to half of the cross-sectional area of each 100 × 100 mm test sample by means of five 10 mm wide steel fingers (Fig. 3A), the tensile resistance being provided by the four 12.5 mm wide strips of peat between the fingers (Fig. 3B). Tensile stress and strain were recorded 30 s after the application of each load increment until the sample failed. Although results obtained using this method have been found to be reproducible and consistent, two key limitations are recognised: (i) the apparatus does not allow a vertical load to be placed on the sample to replicate the condition of the basal peat in-situ; and (ii) significant sample disturbance may occur during installation due to large fibres or woody fragments (Dykes 2008c).

< FIGURE 3 > Figure 3. Measuring the tensile strength of the peat: (A) the two sets of steel ‘fingers’ that are pushed through the centre of a cut block of undisturbed peat 100 mm high × 100 mm wide and 40-60 mm thick; (B) one half of a sample following tensile failure, still adhering to one set of ‘fingers’; (C) the testing apparatus, showing: centre – the ‘fingers’ assembly installed (without a sample); right – the force proving ring; lower far left – the hanger for applying the weights that apply the load just visible beside the end of the cupboards. Details of the design and development of this apparatus are provided in Dykes (2008c).

13

364 365 1 2 3 366 4 5 367 6 7 368 8 9 369 10 11 370 12 13 14 371 15 16 372 17 18 373 19 20 374 21 22 23 375 24 25 376 26 27 377 28 29 378 30 31 32 379 33 34 380 35 36 381 37 38 382 39 40 383 41 42 384 43 44 45 385 46 47 386 48 49 387 50 51 388 52 53 389 54 55 390 56 57 391 58 59 392 60 61 62 63 64 65

4. Results

4.1 Field descriptions of the peat The peat at the three landslides showed remarkably little variability in terms of structure and macrofossil content. Four major stratigraphic units were identified at each site. The first unit (starting at the top), including the living roots near the surface, comprised slightly humified peat, with each unit below being progressively more humified and the fourth unit at the base having highly humified and/or greasy peat, sometimes with bitumen or sludge like patches (Table 3; e.g. Fig. 4). The identifiable plant material in the peat was predominantly monocotyledon (‘monocot’) remains (‘Turfa herbacea’ in the Troels-Smith scheme) including undifferentiated roots, stems and leaves (Fig. 5). This general lack of variation between (Table 3) and within (Fig. 5) sites allowed us to consider one Sampling Point as being broadly representative of each landslide site.

< FIGURE 4 > Figure 4. Peat stratigraphy across the slope above the head of the Slieve Rushen bogflow. This linear transect was located 7.5 m upslope of the source area head at the closest point. Modified from Foteu Madio (2013).

< FIGURE 5 > Figure 5. Results of macrofossil analyses of samples obtained from across the Slieve Rushen bogflow. Labels A1, A5, etc. refer to sampling positions: A1 to A9 are shown in Fig. 5; E1/E3 and E4/E7 are located either side of the downslope extent of the source area. The materials found at each position are from, and in the same order as, this list: Charcoal (0.5-1 mm); Charcoal (less than 0.5 mm); Ericales; Eriophorum vaginatum; Monocot fragments (Monocot leaves at E4), Roots; Sphagnum; Unidentified organic matter. Source: Foteu Madio (2013).

14

393

4.2 Physical and mechanical properties of the peat

394 1 2 3 395 4 5 396 6 7 397 8 9 398 10 11 399 12 13 14 400 15 16 401 17 18 402 19 20 403 21 22 23 404 24 25 405 26 27 406 28 29 407 30 31 32 408 33 34 409 35 36 410 37 38 411 39 40 412 41 42 43 413 44 45 414 46 47 415 48 49 416 50 51 52 417 53 54 418 55 56 419 57 58 420 59 60 61 421 62 63 64 65

4.2.1 Geotechnical characteristics The basic physical properties of the peat at the three landslides are summarised in Table 4. These are broadly consistent with previous results obtained from Straduff Townland (the 1997 bogflow adjacent to ST) and SA by Yang and Dykes (2006). Uncorrected field shear vane readings from depths between 1.25 and 2.00 m were between 6.6 and 14.0 kPa at all sites and there were insufficient results from which to identify any patterns in the data. No corrections were applied because this was intended as a comparative study and the shear vane is known to be inappropriate for the determination of the undrained strength of peat due to the effects of fibres, although it can be used to identify patterns of peat strength variation with depth.

Results from the experimental low-stress direct shear tests (without consolidation prior to shearing) are shown in Fig. 6. These are consistent with results obtained from basal peat at another landslide in northwestern Ireland (identified as ‘E6’ by Kirk (2001): Dykes 2008a). The surface peat at ST clearly demonstrates a higher strength due to the greater density, and probably strength, of less humified fibres. Samples inevitably consolidate under even these small loads as shearing takes place. Straight line approximations in the normal stress range 2-5 kPa would all give cohesion intercepts of 1-4 kPa (Fig. 6). If the peat was overconsolidated by up to 10-15 kPa as seems to be the general case (O’Kelly 2017), such low shear stress values within this range of applied normal loads should not be expected.

Results from the unconsolidated-undrained triaxial tests similarly demonstrate the inherently low shear strength of the basal peat with all three sites in the range 1.5-2.5 kPa (Fig. 7). Slight variations in the diameters of the Mohr’s circles arise from the heterogeneity of the peat mass, as also observed in raised bog peat by Hanrahan (1954), but may also result from gas in the peat causing variations in pore water pressures within the samples. Given that the assumed effect of pore pressures in undrained shear tests is to reduce the friction towards zero because the water is incompressible, then the presence of compressible gas within some pore spaces could allow some (additional) frictional resistance to arise during testing. Hanrahan (1954) found that the gas content of Irish Sphagnum peat may be considerably in excess of 5% of the volume and that significant volumes of gases such as sulphuretted and phosphorated hydrogen (phosphine), as well as 15

422

methane, could be emitted during construction involving the compression of peat. Therefore the possibility

423 1

of gas affecting both permeability and pore pressures must be allowed for when interpreting results.

2 3 424 4 5 425

6 7 426 8 9 427 10 11 428 12 13 14 429 15 16 430 17 18 431 19 20 432 21 22 23 433 24 25 434 26 27 435 28 29 436 30 31 32 437 33 34 438 35 36 439 37 38 440 39 40 441 41 42 43 442 44 45 443 46 47 444 48 49 445 50 51 52 446 53 54 447 55 56 448 57 58 449 59 60 61 450 62 63 64 65

Fig. 8 shows the tensile strengths obtained from this and previous studies using the same methodology (Dykes 2008c). The tensile strengths of the basal peat at the three landslides in this study are all less than 3 kPa except where locally reinforced by matted woody fragments. With the exception of two outliers, which arose from the respective samples containing significant fragments of decomposing roots or woody stems, there is an apparent trend of reducing tensile strength with depth. Although this trend arises from combined results from several locations in Ireland, the similarities of all other measured peat properties between all of these sites (Dykes 2008c; Dykes and Warburton 2008a; Dykes and Jennings 2011) means that this general trend is probably real. At individual sites it is possible that such a trend of decreasing tensile strength with depth may not always be found, although there are insufficient relevant data to be able to comment further.

Helenelund (1967) suggested that the fibre contents, types and orientations – which depend on the morphology and the mode of growth of the original plant assemblage that formed the peat – may have major influences on the tensile strength. The macrofossil analyses of peats from our study sites revealed remains of sedges, the degree of humification of which increase with depth. In such monocotyledon peat, fibres are the remains of vascular bundles formed from the root systems that grow perpendicularly to the ground surface. The resulting tensile strength will therefore be related to the resisting force produced by the fibres, the frequency of which decreases with depth and is inversely proportional to the degree of humification. The tensile strength results obtained by Helenelund (1967) from Sphagnum bog peat, which has very few fibres, are comparable with the lowest of our results, showing that the monocotyledon peats at our sites generally have higher tensile strengths than Sphagnum bog peat. Due to the effect of compression during the accumulation of the peats, some fibres that were originally distributed vertically through the peat become squashed progressively into a horizontal alignment as pressure increases. The degree of inclination of these fibres toward the horizontal plane should therefore also increase with depth. The tensile strength values presented in this study were measured in a horizontal plane, intended to represent the effect of the peat mass pulling apart above a basal (shear?) failure zone. The effect of fibre orientation should be to increase the tensile strength with depth since horizontal breaking up of a failing peat mass is resisted by sometimes 16

451

significant lengths of fibres adhering to amorphous colloidal matrix material. However, the role of living and

452 1

minimally decomposed roots within the near-surface acrotelm layer combined with the very high degree of

2 3 453 4 5 454 6 7 455 8 9 456 10 11 457 12 13 458 14 15 459 16 17 18 460 19 20 461 21 22 462 23 24 463 25 26 464 27 28 465 29 30 466 31 32 33 467 34 35 468 36 37 469 38 39 470 40 41 471 42 43 472 44 45 473 46 47 474 48 49 475 50 51 52 476 53 54 477 55 56 478 57 58 479 59 60 61 480 62 63 64 65

humification below the acrotelm appears to entirely override the fibre orientation effect.

< FIGURE 6 > Figure 6. Results from experimental low-stress direct shear tests of basal peat from all three landslides and from around 10-60 mm depth at Straduff Townland. Previous results from bog slide ‘E6’ at Cuilcagh Mountain, Co. Cavan, obtained using the same methodology, are also shown. Modified from Foteu Madio et al. (2012), after Dykes (2008a).

< FIGURE 7 > Figure 7. Mohr’s Circles (total stresses) obtained from unconsolidated-undrained triaxial tests on peat samples from the three landslides: (A) ST– Straduff Townland; (B) SR – Slieve Rushen; (C) SA – Slieve Anierin. Source: Foteu Madio (2013).

< FIGURE 8 > Figure 8. Tensile strength results obtained from the three landslides in this study and from previous studies using the same methodology. MHA-00s refers to the Maghera bogflow, Co. Galway; SDF-08 is bogflow ST in this study; BHW08 is the Ballincollig Hill peat flow, Co. Kerry; DCM-03 is the collective reference for the 40 landslides that occurred on Dooncarton Mountain, Co. Mayo on 19 September 2003, the results here being obtained from peat slide ‘SE5’. Modified from Foteu Madio (2013).

4.2.2 Humification and fibres The results of the quantitative determination of humification, recorded as the ‘raw’ percentage of light transmission through the peat, showed no significant differences between the mean values for the three sites. However, only the results from ST showed a clear reduction in light transmission (i.e. increase in degree of humification) with depth (Fig. 9). 17

481 482 1 2

483 3 4 484 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 485 48 486 49 50 487 51 52 488 53 54 489 55 56 57 490 58 59 491 60 61 62 63 64 65

Table 3. Summary description of the peat at each landslide Sampling Point (Foteu Madio 2013). The four major stratigraphic units are separated by the solid lines of the Table.

Depth (m) 0.00-0.40 0.40-0.78 0.78-1.22 1.22-1.60 1.60-1.80 >1.80

Peat profile description at ST Light brown fibrous peat, slightly humified, mainly monocotyledon fine fibres and low amorphous material, moderate horizontal tensile strength. Black and moderately humified, mainly monocotyledon fine fibre peat and moderate amorphous material, moderate horizontal tensile strength. Light brown with dark patches, very weak and moderately humified peat. Monocotyledon fine fibre limited. Low horizontal tensile strength. Brown, moderately to strongly humified peat. Monocotyledon fine fibre present. Low horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Rare and very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1on the Munsell soil colour chart. Peat profile description at SR1

Depth (m) 0.00-0.15 0.15-0.36

Brown fibrous peat with moderately humified, mainly monocotyledon fine fibres and low amorphous material, moderate horizontal tensile strength. Brown, less fibrous peat with moderately humified, mainly monocotyledon fine fibre peat and moderate amorphous material, moderate horizontal tensile strength.

0.36-0.58

Dark brown humified peat with monocotyledon fragments. Low horizontal tensile strength.

0.58-0.88

Dark brown decomposing peat with monocotyledon fragments. Low horizontal tensile strength.

>1.64

Dark grey, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1 on the Munsell soil colour chart.

Depth (m)

Peat profile description at SA

0.88-1.58 1.58-1.64

0.00-0.76 0.76-1.56 1.56-1.76 1.76-1.78 >1.78

Dark fibrous peat, slightly humified, mainly monocotyledon fine fibres, low amorphous material and moderate horizontal tensile strength. Light brown less fibrous peat with moderately humified, mainly monocotyledon fine fibre peat, moderate amorphous material and moderate horizontal tensile strength. Black humified peat with monocotyledon fragments. Low horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1on the Munsell soil colour chart.

Note 1

Recorded in July 2010 prior to the moorland fire.

18

492

Table 4. Summary of physical properties of peat at the three landslides, including previous data from Yang and Dykes

493 1

(2006)*.

2 3 4 5 6 7 8 9 10 11 12 494 13 14 495 15 496 16 17 497 18 19 498 20 499 21 22 500 23 501 24 25 502 26 503 27 28 504 29 505 30 31 32 506 33 34 507 35 36 508 37 38 509 39 40 510 41 42 43 511 44 45 512 46 47 513 48 49 514 50 51 52 515 53 54 516 55 56 517 57 58 518 59 60 519 61 62 63 64 65

ST SR SA

Water content a,b (% mass fraction)

Loss on Ignition c (%)

Saturated bulk density e (Mg m–3)

Dry bulk density e Saturated (Mg m–3) hydraulic conductivity h,i (m s–1)

700-900 *620-860 600-700 600-700 *600-740

94.5-95.6 *97.8-98.8 94.6-97.2 d 95.0 *97.7-98.5

1.00 f *1.06 1.00 1.00 *1.05

0.10-0.20 *0.13 0.10-0.20 0.20 g *0.15

10–9 to 10–8 * < 10–11 10–9 to 10–6 10–8 to 10–6 * < 10–11

Notes a

There was negligible difference between field-wet and saturated water contents at all sites

b

Indicative ranges of mean values from 256-319 samples per site

c

Indicative ranges of mean values from 123-196 samples per site

d

The basal peat at Slieve Rushen was noticeably higher in organic matter than any other sampled peat

e

Mean values from 20-26 samples per site

f

The basal peat at Straduff Townland was noticeably higher (~1.10 Mg m–3) than any other sampled peat

g

The peat at Slieve Anierin had higher dry bulk densities throughout its depth

h

Indicative ranges of mean values from 19-26 samples per site, obtained using a ‘constant head’ method

i

Results obtained using a ‘falling head’ method were consistently 10 2-103 m s–1 higher than the respective ‘constant

head’ values

The mean ‘total fibre fraction’ (Ft) of the lowest 0.7 m of the peat profile at each landslide Sampling Point, based on 70 depth-consecutive measurements per site, was 68% at ST, 71% at SR and 56% at SA. As Fig. 10 shows, the latter appears to indicate a small but consistent difference from the other two, having slightly fewer coarse fibres throughout the sampled depth range. At all three sites there is a general trend of reducing coarse fibre content with depth but the fine fibre content seems to increase slightly towards the base of SR.

< FIGURE 9 > Figure 9. ‘Raw’ percentage of light transmission at the Straduff Townland bogflow (ST).

< FIGURE 10 > Figure 10. Depth variations of fibre contents throughout the lower half of the peat profile at each landslide: (A) Straduff Townland, (B) Slieve Rushen, (C) Slieve Anierin. Source: Foteu Madio (2013).

19

520 521 1 2 3 522 4 5 523 6 7 524 8 9 525 10 11 526 12 13 14 527 15 16 528 17 18 529 19 20 530 21 22 531 23 24 532 25 26 533 27 28 29 534 30 31 535 32 33 536 34 35 537 36 37 538 38 39 40 539 41 42 540 43 44 541 45 46 542 47 48 49 543 50 51 544 52 53 545 54 55 546 56 57 58 547 59 60 61 62 63 64 65

4.3 Peat stratigraphy according to macrofossil results The blanket bog at the Sampling Point at each landslide mostly comprised the remains of monocotyledon plants, particularly E. vaginatum (Fig. 11). Monocotyledon contents were lowest within the basal peat zones (as defined by cluster analysis) and, at ST, immediately above the basal zone.

< FIGURE 11 > Figure 11. Macrofossil content of the peat monolith from the three landslides: (a) ST, (b) SR, (c) SA. Parameter values are raw counts for charcoal and E. vaginatum spindles, otherwise percentages. The figure shows the dendrogram produced from unconstrained incremental sum square cluster analysis of strata analysed. Dashed lines separate clusters corresponding to zones in the diagram. Source: Foteu Madio (2013).

4.4 Comparing botanical and geotechnical characteristics The results were examined in order to identify any statistical associations (using Pearson’s ‘r’ correlation coefficient) between physical/geotechnical parameters, and then between geotechnical characteristics and botanical results, that may have physical explanations potentially exploitable for predictive purposes. In this study, only significant (p < 0.05) correlations with |r| > 0.7 at all three landslides were interpreted as possibly indicating a causal relationship because the study was based on a single monolith per study site. Furthermore, the full depth of the peat at each site was not analysed for most of the parameters investigated. The only significant associations with |r| > 0.7 that were found between physical/geotechnical parameters at all three sites were between: (i) the humus fraction, Fh, and the total fibre content, Ft (Fig. 12A); (ii) the total fibre content, Ft, and the coarse fibre fraction, Rm (Fig. 12B); (iii) the humus fraction, Fh, and the coarse fibre fraction, Rm; and (iv) the coarse fibre fraction, Rm – and therefore also the total fibre content and the humus fraction – and the field water content (Fig. 12C).

20

548

Figs. 12D and 12E show the only consistently high correlations (p < 0.05) between macrofossil data and

549 1

physical/geotechnical properties of peat, i.e. between: (i) the total fibre content and the proportion of

2 3 550 4 5 551 6 7 552 8 9 553 10 11 554 12 13 14 555 15 16 556 17 18 557 19 20 558 21 22 23 559 24 25 560 26 27 561 28 29 562 30 31 32 563 33 34 564 35 36 565 37 38 566 39 40 567 41 42 43 568 44 45 569 46 47 570 48 49 571 50 51 52 572 53 54 573 55 56 574 57 58 575 59 60 61 576 62 63 64 65

monocot fragments; and (ii) the von Post degree of humification and the percentage of unidentified organic matter. This may arise from the QLCMA method used for macrofossil analyses probably being more appropriate for Sphagnum peat with small leaves that can be easily counted, compared with monocotyledon peat with larger original plant fragments. The general lack of strong or consistent associations correlations between the physical/geotechnical and botanical parameters at the three landslides suggests that these physical properties cannot be used as indicators of peat mass structure and, thus, of potential peat instability. However, the method used to quantify the fibre contents (Section 3.2.2, above) may be useful for investigating relationships between the structural properties of failed Irish blanket peats in order to classify peat for stability assessments.

The macrofossil analyses at the three landslides showed that the original plant assemblage was predominantly monocotyledons, especially Eriophorum vaginatum. Therefore, the undrained strengths obtained at the three landslides were plotted against the other properties (e.g. coarse fibre content in Fig. 12F) in order to investigate any possible relationship that may exist. The statistical analyses revealed no significant correlation coefficients (p < 0.05).

Fig. 13 shows the thickness of a weak basal layer at each site identified by cluster analyses of the results (e.g. Fig. 11). If all three landslides failed in a similar manner (i.e. by initial basal shearing), then it appears that field observations of shear surfaces within a few tens of mm above the peat-mineral interface can be explained in terms of formation of a failure zone (a) within the weakest layer of the peat profile, and (b) at the lowest elevation within that weakest layer giving a continuous plane above the level of any large stones or woody remnants that would resist shearing within the basal peat. The mean thickness of this layer based on cluster analyses of the data (e.g. Fig. 11) is around 170 mm, but this is clearly overestimated because of the lack of a clear depth-related trend in the quantitative humification results (‘raw’ % light transmission) from Straduff Townland and is probably less than 140 mm in reality.

21

577

< FIGURE 12 >

578 1

Figure 12. Correlations between physical/geotechnical parameters and between botanical characteristics of the peat. (A)

2

579 3

Total fibre content vs. humus fraction. (B) Total fibre content vs. coarse fibre fraction. (C) Coarse fibre fraction vs. field

4

580 5 6 581 7 8 582 9 10 583 11 12 13 584 14 15 585 16 17 586 18 19 587 20 21 22 588 23 24 589 25 26 27 590 28 29 591 30 31 592 32 33 593 34 35 594 36 37 38 595 39 40 596 41 42 597 43 44 598 45 46 47 599 48 49 600 50 51 601 52 53 602 54 55 56 603 57 58 604 59 60 61 62 63 64 65

water content. (D) Monocot fragments vs. total fibre content. (E) Unidentified organic matter vs. von Post humification. (F) Undrained shear strength (including ‘field vane strength’) vs. total fibre content. In (A) to (E), solid line = ST, long dashed line = SR and the thin broken line = SA. After Foteu Madio (2013).

< FIGURE 13 > Figure 13. Variation of mean thickness of basal peat depths according to specific physical properties at all three landslides.

5. Discussion

The three sites investigated for this study were remarkably similar in terms of the characteristics of their blanket peat. Slieve Anierin had a lower fraction of identifiable monocot fragments and a correspondingly higher fraction of unidentified organic matter, but this may simply reflect greater decomposition of the same plants rather than being evidence of different constituents. The smaller proportion of coarse fibres throughout the peat at this site, and particularly towards the base, supports the interpretation of more advanced decomposition. However, the higher dry bulk density and slightly higher saturated hydraulic conductivities (Table 4) perhaps indicate a very slightly different composition. One tensile strength measurement at this site was significantly out of line with the others (Fig. 8) due to a high density of woody remains within one test sample, but the other measures of shear strength were entirely consistent with the other two sites. Therefore we suggest that this site has essentially the same palaeoenvironmental history of peat accumulation as the others. Furthermore, the similarity between these results and some obtained from other landslide sites throughout northwestern and western Ireland and Northern Ireland (e.g. Kirk 2001; Yang and Dykes 2006; Dykes 2008c; Dykes and Warburton 2008a; Dykes and Jennings 2011) and indeed eastern Ireland (e.g.

22

605

Boylan and Long 2010) strongly suggests that the general geotechnical characteristics of upland blanket peat

606 1

throughout the island of Ireland are very similar everywhere.

2 3 607 4 5 608

6 7 609 8 9 610 10 11 611 12 13 14 612 15 16 613 17 18 614 19 20 615 21 22 23 616 24 25 617 26 27 618 28 29 619 30 31 32 620 33 34 621 35 36 622 37 38 623 39 40 624 41 42 43 625 44 45 626 46 47 627 48 49 628 50 51 52 629 53 54 630 55 56 631 57 58 632 59 60 61 633 62 63 64 65

Much of the present vegetation of Ireland’s blanket bogs is dominated by sedges (e.g. E. vaginatum), heathers (Ericacae, including Calluna vulgaris) and some Sphagnum and other mosses. These are all represented in the analyses, with the sedges dominating the identifiable macrofossils (Fig. 10). In many places there are the remains of trees at the base of the peat, which act like fragments of weathered bedrock to resist movement of the peat over the in situ ground. However, at these three sites, separated by up to 20 km, there is a weak basal layer around 150 mm thick that can be clearly distinguished from the peat above on the basis of the properties measured for this study. Intriguingly, a higher proportion of the macrofossils can be identified as monocot fragments in this layer, which somewhat contradicts the idea of greater decomposition. On the other hand, fibre contents reduce sharply towards this basal layer (Fig. 10). O’Kelly (2017) suggested that the properties of fibrous peat depend on the fibre content, but we suggest that these Irish blanket peats cannot be considered to be ‘fibrous’ in the same sense, since even the acrotelm layer may contain relatively few identifiable fibres. The issue is in any case unclear. Previous studies have found that higher fractions of coarse fibres had no effect on measured strength compared with lower coarse fibre contents (Zhang and O’Kelly 2014; Hendy et al. 2014); Price et al. (2005) found that fibre content was not related to compressibility, and Lee et al. (2015) concluded that the effect of fibre orientation on frictional shearing resistance was not clear. However, Boylan and Long (2010) undertook a quantitative analysis of fibre contents adjacent to peat slides in Co. Wicklow and found lower fibre contents with depth. We therefore conclude that the occurrence of failure in upland Irish blanket bogs must be at least in part due to the lower fibre content, as well as higher overall degree of decomposition, towards the base of the peat.

We found some relationships between measured properties of the peat we analysed. The very strong association ( |r| > 0.95) between the humus fraction, coarse fibre fraction and total fibre content at the three landslides mean that only one of these parameters may be needed to investigate other properties of peat. This association can be explained by the fact that with increasing plant decomposition, the size and amount of organic particles decrease, resulting in low fibre contents (Fig. 12A). When the fibre content decreases, the water content also decreases (Fig. 12C) because the voids within the fibres, which contain the largest amount 23

634

of water (MacFarlane and Radforth 1968), also decrease. The coarse fibres influence peat structure and

635 1

possibly strength (see above) and may be used for stability assessments given that at all three sites they were

2 3 636 4 5 637 6 7 638 8 9 639 10 11 640 12 13 14 641 15 16 642 17 18 643 19 20 644 21 22 23 645 24 25 646 26 27 647 28 29 648 30 31 32 649 33 34 650 35 36 651 37 38 652 39 40 653 41 42 43 654 44 45 655 46 47 656 48 49 657 50 51 52 658 53 54 659 55 56 660 57 58 661 59 60 61 662 62 63 64 65

similarly abundant and showed high ( |r| > 0.9) correlations with other properties. However, the apparent uniformity of the peat across these sites precludes any suggestion that this may form the basis of a generalised approach, in the absence of further studies from different peatlands (e.g. Northern England or Scotland). Figs. 12D and 12E merely highlights the effect of humification in that if there are more fibres remaining then there should also be more macrofossils that have not yet decomposed too far to be identified. Fig. 12F shows that whichever method of strength determination is used (excluding the field vane), the (shear) strength of the basal peat appears to be around 2 kPa. This is consistent with stability analyses of landslides involving failure within the peat (i.e. bog slides, bogflows and some peat flows sensu Dykes and Warburton 2007b) as reported by Dykes (2008c), Dykes and Jennings (2011) and Farrell (2012) and with test results obtained from other similar studies in Ireland (e.g. Dykes 2008c; Dykes et al. 2008).

Two of the characteristics identified as being slightly different at Slieve Anierin, i.e. the monocot content and the coarse fibre content, can be readily determined from small auger samples because they are quantified with respect to the dry mass. A hand auger capable of cutting ‘intact’ core samples, notwithstanding issues of sample deformation due to compression or fibres not being cut cleanly (Long and Boylan 2013; Hendy et al. 2014), could in principle provide samples for simple determination of dry bulk density and possibly saturated hydraulic conductivity, i.e. the other two slightly distinctive characteristics. However, given that the measured strengths at this site were no different from the others, we cannot say whether measurement of those characteristics would be useful for peats formed from significantly different plant assemblages. It is not possible to generalise any implications of our results for peatlands in general, and notwithstanding previous comments we cannot assume that any of our correlations between botanical and geotechnical characteristics will apply throughout Ireland. There is thus a clear necessity for comprehensive laboratory testing of peat from the site of any proposed development, probably requiring excavation of trial pits for the extraction of appropriate undisturbed samples. However, general recommendations for the most appropriate tests – and testing procedures suitable for peat – will probably take some time to emerge from ongoing research programmes.

24

663

Finally, the very low shear strength indicated above demands some consideration with respect to water

664 1

conditions within the peat. Blanket bogs in the British Isles may experience water table variations of up to

2 3 665 4 5 666 6 7 667 8 9 668 10 11 669 12 13 14 670 15 16 671 17 18 672 19 20 673 21 22 23 674 24 25 675 26 27 676 28 29 677 30 31 32 678 33 34 679 35 36 680 37 38 681 39 40 682 41 42 43 683 44 45 684 46 47 685 48 49 686 50 51 52 687 53 54 688 55 56 689 57 58 690 59 60 61 62 63 64 65

0.5–1.0 m, but these are occasional reductions below the surface during warm periods of summer weather (Evans et al. 1999; Holden and Burt 2003). The usual condition for these deposits is to be fully saturated to the surface, i.e. with normal effective stress ≈ 0 and maximum pore water pressure most of the time. Periods of summer drying may increase the normal effective stress by a few kPa due to the reduced pore water pressure, i.e. temporarily increasing the effective shear strength. Failure within the peat cannot, therefore, be the result of raised pore water pressures throughout the peat matrix due to heavy rainfall (although it could due to external loading). The hydraulic effects of water-filled pipes, cracks and other voids (e.g. Dykes, this volume – in review) may play significant roles in the initiation of failure, i.e. peat mass effects, are thought to be more important than simply the peat matrix (shear) strength, but much more research is needed to test this hypothesis.

6. Conclusions and Future Work

The upland blanket bogs of northwestern Ireland appear to be formed from essentially the same assemblages of plant species, dominated by sedges (mostly represented by Eriophorum vaginatum), and therefore having similar physical and botanical characteristics. The data describing those characteristics show a statistically distinct basal layer around 150 mm thick characterised by, in particular, a sharp reduction in the coarse – and total – fibre content. Tensile strength, experimental low stress direct shear and unconsolidated undrained triaxial compression measurements of peat strength converge on a value of around 2 kPa which is consistent with stability back-analyses requiring undrained shear strengths of around 2 kPa for FS = 1.0. Contrary to some published accounts, it appears that the lack of coarse fibres may be a contributory factor in the incidence of peat slope failures. Some relationships between measured properties suggest that there may be usable indicators of peat strength and stability conditions, possibly obtainable by means of samples from hand augers, but the apparent uniformity of the peat at these three locations precludes any definitive proposal of useful new methodologies at present.

25

691

It has been recognised for some time that the development of methods for reliably estimating the shear

692 1

strength of peat is likely to require some detailed investigations of botanical controls on relevant

2 3 693 4 5 694 6 7 695 8 9 696 10 11 697 12 13 14 698 15 16 699 17 18 700 19 20 701 21 22 23 702 24 25 703 26 27 704 28 29 705 30 31 706 32 33 707 34 35 708 36 37 709 38 39 710 40 41 711 42 43 712 44 45 713 46 47 714 48 49 715 50 51 716 52 53 717 54 55 718 56 57 719 58 59 60 61 62 63 64 65

geotechnical properties (e.g. Dykes 2008a). More recently, O’Kelly (2017, p.21) stated that: ‘More extensive testing of peats with different botanical compositions is recommended to confirm relationships between tensile strength, other strength parameters and humification level’. All of these issues are now starting to be addressed more systematically by a few researchers in several countries. However, more extensive integrative research is needed, perhaps involving palaeoecologists alongside geotechnical engineers, to explore the causes and geotechnical effects of different peat accumulation scenarios. Detailed measurements of all possible characteristics, such as presented in this study, are required for several known sites of peat landslides in each of several different biogeographical zones such as Dartmoor (SW England), North Pennines (N England), Isle of Skye (W Scotland), Shetland Islands (N Scotland), ideally including full depth variations at each study location in order to generate sufficient data for reliable statistical analyses.

Acknowledgements

This work was funded by Kingston University’s Centre for Earth and Environmental Science Research (CEESR) studentship support fund. EF thanks Prof M Waller (Kingston University), Dr P Hughes (University of Southampton) and Dr M Grant (Kingston University/Wessex Archaeology) for advice and assistance with palaeoecological research techniques. We are grateful to Prof E Bromhead for redrawing Figure 4, and to Mr C Somerfield for assistance with the triaxial testing.

References Aaby B and Tauber H (1974) Rates of peat formation in relation to degree of humification and local environment as shown by studies of a raised bog in Denmark. Boreas (Oslo) 4(1), 1-18. Andrejko MJ, Fiene F and Cohen AD (1983) Comparison of ashing techniques for determination of the inorganic content of peats. In Testing of Peats and Organic Soils (Jarrett PM (ed)). ASTM Special Technical Publication, 820. American Society for Testing and Materials, Philadelphia, pp.5-20. ASTM (2008) D1997-91: Standard test method for laboratory determination of the fiber content of peat samples by dry mass. American Society for Testing and Materials, Philadelphia.

26

720 721 1 2 722 3 4

723 5 6

724 7 8

725 9 10 726 11 12 727 13 14 728 15 16 729 17 18 730 19 20 731 21 22 732 23 24 733 25 26 734 27 28 735 29 30 736 31 32 737 33 34 35 738 36 37 739 38 39 740 40 41 741 42 43 742 44 45 743 46 47 744 48 49 745 50 51 746 52 53 747 54 55 748 56 57 749 58 59 750 60 61 62 63 64 65

Barber KE, Chambers FM, Maddy D, Stoneman R and Brew JS (1994) A sensitive high resolution record of late Holocene climatic change from a raised bog in northern England. The Holocene 4, 198-205. Blackford JJ and Chambers FM (1993) Determining the degree of peat decomposition for peat based palaeoclimatic studies. International Peat Journal 5, 7-24. Boylan N and Long M (2010) An investigation of two peat slope failures in the Wicklow mountains. Biology and Environment: Proceedings of the Royal Irish Academy 110B (3), 173-184. Carlsten P (1993) Peat - Geotechnical Properties and Up-to-Date Methods of Design and Construction. State-of-the-ArtReport. Linköping: Swedish Geotechnical Institute. Chambers FM, Barber KE, Maddy D and Brew JS (1997) A 5500-year proxy-climate and vegetation record from blanket mire at Talla Moss, borders, Scotland. The Holocene 7, 391-399. Charman DJ (2002) Peatlands and Environmental Change. Wiley, Chichester. Crofton HT (1902) How Chat Moss broke out in 1526. Transactions of the Lancashire and Cheshire Antiquarian Society XX, 139-144. Daniels RE and Eddy A (1990) Handbook of European Sphagna. Institute of Terrestrial Ecology, Natural Environment Research Council. HMSO, London. Dykes AP (2008a) Properties of peat relating to instability of blanket bogs. In Landslides and Engineered Slopes, Volume 1 (Chen ZY, Zhang J, Li Z, Wu A and Ho K (eds)). Taylor and Francis, London, UK, pp.339-345. Dykes AP (2008b) Geomorphological maps of Irish peat landslides created using hand-held GPS. Journal of Maps v2008, 258-276. Dykes AP (2008c) Tensile strength of peat: laboratory measurement and role in Irish blanket bog failures. Landslides 5(4), 417-429. Dykes AP (2009) Geomorphological maps of Irish peat landslides created using hand-held GPS - Second Edition. Journal of Maps v2009, 179-185. Dykes AP (this issue? – still in review) New insights from a recent peat slide at Croaghan, Co. Antrim, Northern Ireland. Dykes AP and Jennings P (2011) Peat slope failures and other mass movements in western Ireland, August 2008. Quarterly Journal of Engineering Geology and Hydrogeology 44(1), 5-16. Dykes AP and Kirk KJ (2001) Initiation of a multiple peat slide on Cuilcagh Mountain, Northern Ireland. Earth Surface Processes and Landforms 26, 395-408. Dykes AP and Warburton J (2007a) Significance of geomorphological and subsurface drainage controls on failures of peat-covered hillslopes triggered by extreme rainfall. Earth Surface Processes and Landforms 32, 1841-1862.

27

751 752 1 2 753 3 4

754 5 6

755 7 8

756 9 10 757 11 12 758 13 14 759 15 16 760 17 18 761 19 20 762 21 22 763 23 24 764 25 26 765 27 28 766 29 30 767 31 32 768 33 34 35 769 36 37 770 38 39 771 40 41 772 42 43 773 44 45 774 46 47 775 48 49 776 50 51 777 52 53 778 54 55 779 56 57 780 58 59 781 60 61 782 62 63 64 65

Dykes AP and Warburton J (2007b) Mass movements in peat: A formal classification scheme. Geomorphology 86(1-2), 73-93. Dykes AP and Warburton J (2008a) Failure of peat-covered hillslopes at Dooncarton Mountain, Co. mayo, Ireland: Analysis of topographic and geotechnical factors. Catena 72, 129-145. Dykes AP and Warburton J (2008b) Characteristics of the Shetland Islands (UK) peat slides of 19 September 2003. Landslides 5, 213-226. Dykes AP, Gunn J and Convery (Née Kirk) KJ (2008) Landslides in blanket peat on Cuilcagh Mountain, northwest Ireland. Geomorphology 102, 325-340. Evans MG, Burt TP, Holden J, Adamson JK (1999) Runoff generation and water table fluctuations in blanket peat: evidence from UK data spanning the dry summer of 1995. Journal of Hydrology 221, 141-160. Farrell ER (2012) Organics/peat soils. In ICE Manual of Geotechnical Engineering: Volume 1, Geotechnical Engineering Principles, Problematic Soils and Site Investigation (Burland J, Chapman T, Skinner H and Brown M (eds)). ICE Publishing, London, UK, pp.463-479. Foteu Madio ES (2013) Botanical and geotechnical influences on peat instability. Unpublished PhD thesis, Kingston University, UK. Foteu Madio ES, Dykes AP, Waller MP, Hughes P and Grant MJ (2012) Botanical and geotechnical influences on peat instability. In Landslides and Engineered Slopes: Protecting Society through Improved Understanding (Vol. 2) (Eberhardt E, Froese C, Turner AK and Leroueil S (eds)). Proceedings of the 11th International and 2nd North American Symposium on Landslides and Engineered Slopes. CRC Press, London, pp.421-427. Gilman K, Newson MD (1980) Soil Pipes and Pipeflow: a Hydrological Study in Upland Wales, British Geomorphological Research Group Monograph 1. Geobooks: Norwich. Grosse-Brauckmann G (1972) Über pflanzliche Makrofossilien mitteleuropäischer Torfe - I. Gewebereste krautiger Pflanzen und ihre Merkmale. Telma2, 19-55. Hammond RF (1979) The Peatlands of Ireland. Survey Bulletin No. 35. An Foras Talúntais, Dublin Hanrahan ET (1954) An investigation of some physical properties of peat. Géotechnique 4, 108-123. Head KH (1994) Manual of Soil Laboratory Testing, Permeability, Shear Strength and Compressibility Testing. Volume 2. Wiley, New York. Helenelund KV (1967) Vane tests and tension tests on fibrous peat. Proceedings of the Geotechnical Conference, Oslo, Vol. 1, 199–203. Hendy MT, Barbour SL and Martin CD (2014) Evaluating the effect of fiber reinforcement on the anisotropic undrained stiffness and strength of peat. Journal of Geotechnical and Geoenvironmental Engineering, published on-line at doi:10.1061/(ASCE)GT.1943-5606.0001154. 28

783 784 1 2 785 3 4

786 5 6

787 7 8

788 9 10 789 11 12 790 13 14 791 15 16 792 17 18 793 19 20 794 21 22 795 23 24 796 25 26 27 797 28 29 798 30 31 799 32 33 800 34 35 801 36 37 802 38 39 803 40 41 804 42 43 805 44 45 806 46 47 807 48 49 808 50 51 809 52 53 810 54 55 811 56 57 812 58 59 813 60 61 814 62 63 64 65

Hobbs NB (1986) Mire morphology and the properties and behaviour of some British and foreign peats. Quarterly Journal of Engineering Geology 19, 7-80. Holden J, Burt TP (2003) Hydrological studies on blanket peat: the significance of the acrotelm–catotelm model. Journal of Ecology 91: 103–113. IUGS – International Union of Geological Sciences Working Group on Landslides (1995) A suggested method for describing the rate of movement of landslides. Bulletin International Association of Engineering Geology 52, 75-78. Jarrett PM (1983) Summary. In Testing of Peats and Organic Soils (Jarrett PM (ed)). ASTM Special Technical Publication, 820. American Society for Testing and Materials, Philadelphia, pp.233-237. Katz NJ, Katz SV and Skobeyeva EI (1977) Atlas of plant remains in peat soil. Nedra [In Russian]. Kirk KJ (2001) Instability of blanket bog slopes on Cuilcagh Mountain, N.W. Ireland. Unpublished PhD thesis. University of Huddersfield, UK. Klute A, Dirksen C (1986) Hydraulic conductivity and diffusivity: laboratory methods. Methods of Soil Analysis Part 1. Physical and Mineralogical Methods. Agronomy Monograph (2nd edition), Vol. 9. Soil Science Society of America, Madison, pp. 687– 734. Landva AO and Pheeney PE (1980) Peat fabric and structure. Canadian Geotechnical Journal 17(3), 416-435. Lee J-S, Seo S-Y and Lee C (2015) Geotechnical and geophysical characteristics of muskeg samples from Alberta, Canada. Engineering Geology 195, 135-141. Lindsay R, Bragg O (2005) Wind farms and blanket peat: a report on the Derrybrien bog slide (2nd edition). Derrybrien Development Cooperative Ltd., Gort. Long M (2005) Review of peat strength, peat characterisation and constitutive modelling of peat with reference to landslides. Studia Geotechnica et Mechanica XXVII, 67-90. Long M and Boylan N (2013) Predictions of settlement in peat soils. Quarterly Journal of Engineering Geology and Hydrogeology 46, 303-322. Long M and Jennings P (2006) Analysis of the peat slide at Pollatomish, County Mayo, Ireland. Landslides 3(1), 51-61. MacFarlane IC and Radforth NW (1968) Structure as a basis of peat classification. National Research Council of Canada (Ottawa). Reprinted from Proceedings, Third International Peat Congress held in Quebec, Canada, 18-23 August 1968, pp.91-97. Mills AJ (2002) Peat slides: morphology, mechanisms and recovery. Unpubl. PhD thesis. University of Durham, UK. Moore PD (1984) The classification of mires: An introduction. In European Mires (Moore PD (ed)). Academic Press, London, pp.1-10. O’Kelly BC (2014) Drying temperature and water content–strength correlations. Environmental Geotechnics 1 (EG2), 81-95. 29

815 816 1 2 817 3 4

818 5 6

819 7 8

820 9 10 821 11 12 822 13 14 823 15 16 824 17 18 825 19 20 826 21 22 827 23 24 828 25 26 829 27 28 830 29 30 831 31 32 832 33 34 35 833 36 37 834 38 39 835 40 41 836 42 43 837 44 45 838 46 47 839 48 49 50 840 51 841 52 53 842 54 843 55 56 844 57 845 58 59 846 60 847 61 62 63 64 65

O’Kelly BC (2017) Measurement, interpretation and recommended use of laboratory strength properties of fibrous peat. Geotechnical Research, published on-line at http://dx.doi.org/10.1680/jgere.17.00006. Price JS, Cagampan J and Kellner E (2005) Assessment of peat compressibility: is there an easy way? Hydrological Processes 19, 3469-3475. Skempton AW and Petley DN (1970) Ignition loss and other properties of peats and clays from Avonmouth, Kings Lynn and Cranberry Moss. Geotechnique 20(4), 343-356. Smith AJE (2004) The moss flora of Britain and Ireland (2nd edition). Cambridge University Press, Cambridge. Troels-Smith J (1955) Karakterisering af lose jordater (Characterisation of unconsolidated sediments). Denmarks Geologiske Undersogelse 4(3), 1-73. von Post L (1922) Sveriges geologiska undersoknings torvinventering och nagre av dess hittills vunna resultat, sr. mosskulturfor. Tidskr 1, 1-27. Walker D and Walker PM (1961) Stratigraphic evidence of regeneration in some Irish bogs. Journal of Ecology 49, 169-185. Warburton J, Higgit D and Mills A (2003) Anatomy of a Pennine peat slide, northern England. Earth Surface Processes and Landforms 28(5), 457-473. Warburton J, Holden J and Mills AJ (2004) Hydrological controls of superficial mass movements in peat. Earth-Science Reviews 67(1-2), 139-156. Yang J and Dykes AP (2006) The liquid limit of peat and its application to the understanding of Irish blanket bog failures. Landslides 3(3), 205-216. Zhang L and O’Kelly BC 2014 The principle of effective stress and triaxial compression testing of peat. Geotechnical Engineering 167, 40-50.

List of Figures

Figure 1. Location of the study area in northwest Ireland, showing the distribution of peatlands (after Hammond 1979). The outlined rectangle is enlarged to show the locations of the three bogflows: (left to right) ST = Straduff Townland, SA = Slieve Anierin, SR = Slieve Rushen. Modified from Yang and Dykes (2006). Figure 2. General views of the three study areas. (A) Straduff Townland bogflow, looking downslope from above the head (July 2010). (B) Slieve Rushen bogflow, looking across at the failed slope from the other side of the peat basin into which its displaced peat flowed (July 2010). (C) Slieve Anierin bogflow from the air (Nov. 1998, photo by APD).

30

848 849 1 850 2 3 851 4 852 5 6 853 7 854 8 9 855 10 856 11 12 857 13 858 14 15 859 16 860 17 18 861 19 862 20 21 863 22 864 23 24 865 25 866 26 27 867 28 868 29 30 869 31 870 32 33 871 34 35 872 36 873 37 38 874 39 875 40 41 876 42 877 43 44 878 45 879 46 47 880 48 881 49 50 882 51 883 52 53 884 54 885 55 56 886 57 58 59 60 61 62 63 64 65

Figure 3. Measuring the tensile strength of the peat: (A) the two sets of steel ‘fingers’ that are pushed through the centre of a cut block of undisturbed peat 100 mm high × 100 mm wide and 40-60 mm thick; (B) one half of a sample following tensile failure, still adhering to one set of ‘fingers’; (C) the testing apparatus, showing: centre – the ‘fingers’ assembly installed (without a sample); right – the force proving ring; lower far left – the hanger for applying the weights that apply the load just visible beside the end of the cupboards. Details of the design and development of this apparatus are provided in Dykes (2008c). Figure 4. Peat stratigraphy across the slope above the head of the Slieve Rushen bogflow. This linear transect was located 7.5 m upslope of the source area head at the closest point. Modified from Foteu Madio (2013). Figure 5. Results of macrofossil analyses of samples obtained from across the Slieve Rushen bogflow. Labels A1, A5, etc. refer to sampling positions: A1 to A9 are shown in Fig. 5; E1/E3 and E4/E7 are located either side of the downslope extent of the source area. The materials found at each position are from, and in the same order as, this list: Charcoal (0.5-1 mm); Charcoal (less than 0.5 mm); Ericales; Eriophorum vaginatum; Monocot fragments (Monocot leaves at E4), Roots; Sphagnum; Unidentified organic matter. Source: Foteu Madio (2013). Figure 6. Results from experimental low-stress direct shear tests of basal peat from all three landslides and from around 10-60 mm depth at Straduff Townland. Previous results from bog slide ‘E6’ at Cuilcagh Mountain, Co. Cavan, obtained using the same methodology, are also shown. Modified from Foteu Madio et al. (2012), after Dykes (2008a). Figure 7. Mohr’s Circles obtained from unconsolidated-undrained triaxial tests on peat samples from the three landslides: (A) ST– Straduff Townland; (B) SR – Slieve Rushen; (C) SA – Slieve Anierin. Source: Foteu Madio (2013). Figure 8. Tensile strength results obtained from the three landslides in this study and from previous studies using the same methodology. MHA-00s refers to the Maghera bogflow, Co. Galway; SDF-08 is bogflow ST in this study; BHW-08 is the Ballincollig Hill peat flow, Co. Kerry; DCM-03 is the collective reference for the 40 landslides that occurred on Dooncarton Mountain, Co. Mayo on 19 September 2003, the results here being obtained from peat slide ‘SE5’. Modified from Foteu Madio (2013). Figure 9. ‘Raw’ percentage of light transmission at the Straduff Townland bogflow (ST). Figure 10. Depth variations of fibre contents throughout the lower half of the peat profile at each landslide: (A) Straduff Townland, (B) Slieve Rushen, (C) Slieve Anierin. Source: Foteu Madio (2013). Figure 11. Macrofossil content of the peat monolith from the three landslides: (a) ST, (b) SR, (c) SA. Parameter values are raw counts for charcoal and E. vaginatum spindles, otherwise percentages. The figure shows the dendrogram produced from unconstrained incremental sum square cluster analysis of strata analysed. Dashed lines separate clusters corresponding to zones in the diagram. Source: Foteu Madio (2013). Figure 12. Correlations between physical/geotechnical parameters and between botanical characteristics of the peat. (A) Total fibre content vs. humus fraction. (B) Total fibre content vs. coarse fibre fraction. (C) Coarse fibre fraction vs. field water content. (D) Monocot fragments vs. total fibre content. (E) Unidentified organic matter vs. von Post humification. (F) Undrained shear strength (including ‘field vane strength’) vs. total fibre content. In (A) to (E), solid line = ST, long dashed line = SR and the thin broken line = SA. After Foteu Madio (2013). Figure 13.Variation of mean thickness of basal peat depths according to specific physical properties at all three landslides.

31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 887 25 888 26 889 27 28 890 29 30 891 31 892 32 33 893 34 894 35 36 37 38 39 40 41 42 43 44 45 46 47 48 895 49 50 51 52 53 54 55 56 57 58 59 60 61 896 62 63 64 65

Figure 1. Location of the study area in northwest Ireland, showing the distribution of peatlands (grey shading, after Hammond 1979). The outlined rectangle is enlarged, right, to show the relative locations of the three bogflows and their upland contexts: (left to right) ST = Straduff Townland, SA = Slieve Anierin, SR = Slieve Rushen. In this study area map, grey shading is land above 300 m elevation, horizontal stripes indicate water bodies and the solid black line is the international border. Modified from Yang and Dykes (2006).

32

1 2 3 4 5 6 7 8 9 10 897 11 12 898 13 899 14 15 900 16 901 17 18 902 19 903 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 904 47 48 905 49 906 50 51 907 52 908 53 54 909 55 910 56 57 911 58 59 60 61 62 63 64 65

Figure 2. General views of the three study areas. (A) Straduff Townland bogflow, looking downslope from above the head (July 2010). (B) Slieve Rushen bogflow, looking across at the failed slope from the other side of the peat basin into which its displaced peat flowed (July 2010). (C) Slieve Anierin bogflow from the air (Nov. 1998, photo by APD).

Figure 3. Measuring the tensile strength of the peat: (A) the two sets of steel ‘fingers’ that are pushed through the centre of a cut block of undisturbed peat 100 mm high × 100 mm wide and 40-60 mm thick; (B) one half of a sample following tensile failure, still adhering to one set of ‘fingers’; (C) the testing apparatus, showing: centre – the ‘fingers’ assembly installed (without a sample); right – the force proving ring; lower far left – the hanger for applying the weights that apply the load just visible beside the end of the cupboards. Details of the design and development of this apparatus are provided in Dykes (2008c).

33

1 2 3 4 5 6 7 8 9 10 912 11 913 12 13 914 14 915 15 16 916 17 917 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 918 41 42 919 43 920 44 921 45 46 922 47 923 48 49 924 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Figure 4. Peat stratigraphy across the slope above the head of the Slieve Rushen bogflow. This linear transect was located 7.5 m upslope of the source area head at the closest point. Modified from Foteu Madio (2013).

Figure 5. Results of macrofossil analyses of samples obtained from across the Slieve Rushen bogflow. Labels A1, A5, etc. refer to sampling positions: A1 to A9 are shown in Fig. 5; E1/E3 and E4/E7 are located either side of the downslope extent of the source area. The materials found at each position are from, and in the same order as, this list: Charcoal (0.5-1 mm); Charcoal (less than 0.5 mm); Ericales; Eriophorum vaginatum; Monocot fragments (Monocot leaves at E4), Roots; Sphagnum; Unidentified organic matter. Source: Foteu Madio (2013).

34

8.0 7.0 6.0

Shear stress (kPa)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 925 20 926 21 22 927 23 928 24 25 929 26 930 27 28 29 30 31 32 33 34 35 931 36 37 38 39 40 41 42 43 44 45 932 46 47 48 49 50 51 52 53 54 55 933 56 934 57 58 935 59 936 60 61 62 63 64 65

5.0 4.0 3.0

ST ST near surface SR SA E6 replicate (Dykes 2008a) E6 (Dykes 2008a)

2.0 1.0 0.0 0.0

1.0

2.0

3.0

4.0 5.0 6.0 7.0 8.0 Normal stress (kPa)

9.0

10.0 11.0 12.0

Figure 6. Results from experimental low-stress direct shear tests of basal peat from all three landslides and from around 10-60 mm depth at Straduff Townland. Previous results from bog slide ‘E6’ at Cuilcagh Mountain, Co. Cavan, obtained using the same methodology, are also shown. Modified from Foteu Madio et al. (2012), after Dykes (2008a).

Figure 7. Mohr’s Circles obtained from unconsolidated-undrained triaxial tests on peat samples from the three landslides: (A) ST– Straduff Townland; (B) SR – Slieve Rushen; (C) SA – Slieve Anierin. Source: Foteu Madio (2013).

35

0

2

3

4

Tensile strength (kPa) 5 6 7 8 9 10 11 12 13 14 15 16 17

0 0.2 0.4 0.6

Depth (m)

0.8 1 ST SR SA MHA-00s (Dykes 2008c) SDF-08 (Dykes and Jennings 2011) BHW-08 (Dykes and Jennings 2011) DCM-03: SE5 (Dykes and Warburton 2008)

1.2 1.4 1.6 1.8 2

Figure 8. Tensile strength results obtained from the three landslides in this study and from previous studies using the same methodology. MHA-00s refers to the Maghera bogflow, Co. Galway; SDF-08 is bogflow ST in this study; BHW08 is the Ballincollig Hill peat flow, Co. Kerry; DCM-03 is the collective reference for the 40 landslides that occurred on Dooncarton Mountain, Co. Mayo on 19 September 2003, the results here being obtained from peat slide ‘SE5’. Modified from Foteu Madio (2013).

70

y = 53.03 - 19.05x R² = 0.5405

60

Transmission (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 937 20 938 21 22 939 23 940 24 25 941 26 942 27 28 943 29 30 944 31 945 32 33 946 34 35 947 36 948 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 949 54 950 55 56 951 57 952 58 59 953 60 61 954 62 63 64 65

1

50 40 30 20 10 0 0

0.25 0.5 0.75 1 1.25 1.5 1.75 2 Depth below ground surface (m)

Figure 9. ‘Raw’ percentage of light transmission at the Straduff Townland bogflow (ST).

36

Percentage (%) of the initial dry mass of peat

A 0

10

20

30

40

50

60

70

80

90

100

0.9 Humus fraction (<0.15 mm) Fine fibres (0.15-1mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

1.0

Depth (m)

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

B 0

10

20

Percentage (%) of the initial dry mass of peat 30 40 50 60 70 80

90

100

0.9 1.0 1.1

Depth (m)

1.2 1.3 1.4 1.5 1.6 Humus fraction (<0.15 mm) Fine fibres (0.15-1 mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

1.7 1.8 1.9

C 0

10

20

Percentage (%) of the initial dry mass of peat 30 40 50 60 70 80

90

100

0.9 1.0 1.1 1.2

Depth (m)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 955 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 956 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 957 52 958 53 54 959 55 960 56 57 961 58 59 962 60 963 61 62 63 64 65

1.3 1.4 1.5 1.6

Humus fraction (<0.15 mm) Fine fibres (0.15-1 mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

1.7 1.8 1.9

Figure 10. Depth variations of fibre contents throughout the lower half of the peat profile at each landslide: (A) Straduff Townland, (B) Slieve Rushen, (C) Slieve Anierin. Source: Foteu Madio (2013).

37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 964 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 965 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 966 54 967 55 968 56 57 969 58 970 59 60 971 61 62 63 64 65

Figure 11. Macrofossil content of the peat monolith from the three landslides: (a) ST, (b) SR, (c) SA. Parameter values are raw counts for charcoal and E. vaginatum spindles, otherwise percentages. The figure shows the dendrogram produced from unconstrained incremental sum square cluster analysis of strata analysed. Dashed lines separate clusters corresponding to zones in the diagram. Source: Foteu Madio (2013).

38

(B)

Total fibre content (%)

R2 = 0.98 R2 = 0.93 R2 = 0.86

100 90 80 70 60 50 40 30 20 10 0

ST SR SA

Total fibre content (%)

100 90 80 70 60 50 40 30 20 10 0

R2 = 0.96 R2 R2

= 0.96 = 0.98

ST SR SA

0 10 20 30 40 50 60 70 80 90 100 Coarse fibre fraction (%)

0 10 20 30 40 50 60 70 80 90 100 Humus fraction (%)

(C)

(D)

1100

Monocotyledon fragments (%)

100 90 80 70 60 50 40 30 20 10 0

1000

Field water content (%)

900 800 700 600 500 400 300

R2 = 0.66

ST

R2

= 0.67

SR

R2 = 0.41

SA

0 10 20 30 40 50 60 70 80 90 100 Coarse fibre fraction (%)

ST

R2 = 0.09

SR

R2 = 0.33

SA

R2 = 0.32

0 10 20 30 40 50 60 70 80 90 100 Total fibre content (%)

(E)

(F) 120

ST

R2 = 0.14

14

100

SR

R2 = 0.30

12

Triaxial test

SA

R2

10

Tensile str.

80

Peat strength (kPa)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 973 15 16 974 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 975 33 34 976 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 977 51 52 978 53 979 54 55 980 56 981 57 58 982 59 983 60 61 984 62 63 64 65

(A)

Unidentified organic matter (%)

972

= 0.50

60 40 20

Field vane

Direct shear

8 6 4 2

0

0

4

5 6 7 8 9 10 Degree of humification (von Post)

0

10 20 30 40 50 60 70 80 90 100

Coarse fibre fraction (%)

Figure 12. Correlations between physical/geotechnical parameters and between botanical characteristics of the peat. (A) Total fibre content vs. humus fraction. (B) Total fibre content vs. coarse fibre fraction. (C) Coarse fibre fraction vs. field water content. (D) Monocot fragments vs. total fibre content. (E) Unidentified organic matter vs. von Post humification. (F) Undrained shear strength (including ‘field vane strength’) vs. total fibre content. In (A) to (E), solid line = ST, long dashed line = SR and the thin broken line = SA. After Foteu Madio (2013).

39

1 2 3 4 5 6 7 8 9 10 11 12 13 985 14 15 986 16 17 987 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

Thickness of basal layer (mm)

600

ST SR

500

SA

400 300 200 100 0 Field description

'Raw' % light transmission

Fibre content

Macrofossil content

Quantitative fibre content

Site average

Figure 13. Different estimates of mean thickness of basal peat depths according to specific physical properties at all three landslides.

40

Figure 1A

Click here to download Figure Foteu Madio and Dykes Fig 1a.tif

Figure 1B

Click here to download Figure Foteu Madio and Dykes Fig 1b.tif

Figure 2A

Click here to download Figure Foteu Madio and Dykes Fig 2a.tif

Figure 2B

Click here to download Figure Foteu Madio and Dykes Fig 2b.tif

Figure 2C

Click here to download Figure Foteu Madio and Dykes Fig 2c.tif

Figure 3

Click here to download Figure Foteu Madio and Dykes Fig 3.tif

Figure 4

Click here to download Figure Foteu Madio and Dykes Fig 4.tif

Figure 5

Click here to download Figure Foteu Madio and Dykes Fig 5.tif

Figure 6

8 7

Shear stress (kPa)

6 5 4 3 ST ST near surface SR SA E6 replicate (Dykes 2008a) E6 (Dykes 2008a)

2 1 0 0

1

2

3

4 5 6 7 Normal stress (kPa)

8

9

10

11

12

Figure 7A

Click here to download Figure Foteu Madio and Dykes Fig 7a.tif

Figure 7B

Click here to download Figure Foteu Madio and Dykes Fig 7b.tif

Figure 7C

Click here to download Figure Foteu Madio and Dykes Fig 7c.tif

Figure 8

Tensile strength (kPa) 0

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17

0 0.2 0.4

Depth (m)

0.6 0.8 1 1.2

ST SR

1.4

SA

1.6

MHA-00s (Dykes 2008c)

1.8

BHW-08 (Dykes and Jennings 2011)

2

SDF-08 (Dykes and Jennings 2011) DCM-03: SE5 (Dykes and Warburton 2008)

Figure 9

70

y = -18.819x + 52.83 R² = 0.5322

Transmission (%)

60 50 40 30 20 10 0 0

0.25

0.5

0.75

1

1.25

1.5

Depth below ground surface (m)

1.75

2

Figure 10

A 0

10

20

Percentage (%) of the initial dry mass of peat 30 40 50 60 70

0

10

20

Percentage (%) of the initial dry mass of peat 30 40 50 60 70

80

80

0.9 1.0

Depth (m)

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

B 0.9 1 1.1

Depth (m)

1.2 1.3 1.4 1.5 1.6

Humus fraction (<0.15 mm) Fine fibres (0.15-1 mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

1.7 1.8 1.9

C 0

10

20

Percentage (%) of the initial dry mass of peat 30 40 50 60 70

80

0.9 1 1.1

Depth (m)

1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

Humus fraction (<0.15 mm) Fine fibres (0.15-1 mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

80

90

100

Humus fraction (<0.15 mm) Fine fibres (0.15-1mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

80

90

100

Humus fraction (<0.15 mm) Fine fibres (0.15-1 mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

80

90

Humus fraction (<0.15 mm) Fine fibres (0.15-1 mm) Coarse fibres (>1 mm) Total fibre content (>0.15 mm)

100

Figure 11A

Click here to download Figure Foteu Madio and Dykes Fig 11a.tif

Figure 11B

Click here to download Figure Foteu Madio and Dykes Fig 11b.tif

Figure 11C

Click here to download Figure Foteu Madio and Dykes Fig 11c.tif

Figure 12

R2 = 0.98 R2 = 0.93

0

1000

Monocotyledon fragments (%)

Field water content (%)

0

D

900 800 700 600 R2 = 0.66

500 400 300 0

Unidentified organic matter (%)

100

R2

= 0.67

R2

= 0.41

ST SR SA

40 20 0

-20

5

6

0

F

R2 = 0.14 R2 = 0.30 R2 = 0.50

60

4

100 90 80 70 60 50 40 30 20 10 0

10 20 30 40 50 60 70 80 90 100 Coarse fibre fraction (%)

ST SR SA

80

100 90 80 70 60 50 40 30 20 10 0

10 20 30 40 50 60 70 80 90 100 Humus fraction (%)

1100

C

E

B

ST SR SA

Total fibre content (%)

100 90 80 70 60 50 40 30 20 10 0

7

8

9

10

10

14

Peat strength (kPa)

Total fibre content (%)

A

12 10 8 6 4 2 0

-40

0

Degree of humification (von Post)

10

R2 = 0.96 R2

= 0.96 R2 = 0.98 0

0

ST SR SA

10 20 30 40 50 60 70 80 90 100 Coarse fibre fraction (%)

ST

R2 = 0.09

SR

R2 = 0.33

SA

R2 = 0.32

10 20 30 40 50 60 70 80 90 100 Total fibre content (%)

Field vane Triaxial test Tensile str. Direct shear

10

20

30

40

50

60

70

Coarse fibre fraction (%)

80

90 100

Figure 13

Thickness of basal layer (mm)

600

ST

500

SR

400

SA

300 200 100 0 Field description 'Raw' % light transmission

Fibre content

Macrofossil content

Quantitative fibre content

Site average

Table 1

Click here to download Table Foteu Madio and Dykes - Table 1.docx

Table 1. Summary of site details and characteristics of the study bogflows. Bogflow County Latitude Longitude

Elevation Geology Geomorphological Length Slope Deptha Volume (m) (Carboniferous) Context (m) (°) (m) (m)

ST

Sligo 54°7.2’N 8°12.9’W

405

Lackagh Sandstone

SR

Cavan 54°8.9’N 7°38.5’W

390

Glenade Sandstone

SA

Leitrim 54°6.3’N 7°58.7’W

440

Lackagh Sandstone

Escarpment failure

Basin slope failure 175 Escarpment failure

Note a

200

Indicative average depth of in-situ peat immediately adjacent to the landslide source area

190

5.5 (top) 3 (mid) 6 (lower)

2.5

35,000

5.5

2.0

20,000

4

2.2

22,000

Table 2

Click here to download Table Foteu Madio and Dykes - Table 2.docx

Table 2. Samples extracted from the study bogflows. Landslide / Physical properties – Tensile strength – position in small cores 50 mm blocks 120×120×70 peat profile dia. × 51 mm length mm ST see right ST basea SA 300-1030 mm depth SA base a SR 100-830 mm depth SR base b

Triaxial – Shear strength (direct 38 mm dia. shear) – blocks cores 120×120×70 mm

6 at ~650 mm depth 3 at 10-80 mm depth 6 at ~1150 mm depth 3 at 890-960 mm depth

--

Monoliths for botanical data 730×100×100 mm

3 at 10-80 mm depth 1 at 400-1130 mm 3 at 890-960 mm depth depth

9

6

6

12

2

--

--

--

--

1

9

6

6

12

2

--

--

--

--

1

9

6

6

12

2

Notes a

1600-1700 mm depth below the surface of the peat, 970-1700 mm depth for the lower monoliths

b

1900-2000 mm depth, 1270-2000 mm depth for the lower monoliths

Table 3

Click here to download Table Foteu Madio and Dykes - Table 3.docx

Table 3. Summary description of the peat at each landslide Sampling Point (Foteu Madio 2013). The four major stratigraphic units are separated by the solid lines of the Table.

Depth (m) 0.00-0.40 0.40-0.78 0.78-1.22 1.22-1.60 1.60-1.80 >1.80

Peat profile description at ST Light brown fibrous peat, slightly humified, mainly monocotyledon fine fibres and low amorphous material, moderate horizontal tensile strength. Black and moderately humified, mainly monocotyledon fine fibre peat and moderate amorphous material, moderate horizontal tensile strength. Light brown with dark patches, very weak and moderately humified peat. Monocotyledon fine fibre limited. Low horizontal tensile strength. Brown, moderately to strongly humified peat. Monocotyledon fine fibre present. Low horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Rare and very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1on the Munsell soil colour chart. Peat profile description at SR1

Depth (m) 0.00-0.15 0.15-0.36

Brown fibrous peat with moderately humified, mainly monocotyledon fine fibres and low amorphous material, moderate horizontal tensile strength. Brown, less fibrous peat with moderately humified, mainly monocotyledon fine fibre peat and moderate amorphous material, moderate horizontal tensile strength.

0.36-0.58

Dark brown humified peat with monocotyledon fragments. Low horizontal tensile strength.

0.58-0.88

Dark brown decomposing peat with monocotyledon fragments. Low horizontal tensile strength.

>1.64

Dark grey, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1 on the Munsell soil colour chart.

Depth (m)

Peat profile description at SA

0.88-1.58 1.58-1.64

0.00-0.76 0.76-1.56 1.56-1.76 1.76-1.78 >1.78

Dark fibrous peat, slightly humified, mainly monocotyledon fine fibres, low amorphous material and moderate horizontal tensile strength. Light brown less fibrous peat with moderately humified, mainly monocotyledon fine fibre peat, moderate amorphous material and moderate horizontal tensile strength. Black humified peat with monocotyledon fragments. Low horizontal tensile strength. Dark grey, greasy, highly humified and amorphous peat. Very fine monocotyledon fragments. Low to zero horizontal tensile strength. Sandstone in clay matrix, 5YR 3/1on the Munsell soil colour chart.

Note 1

Recorded in July 2010 prior to the moorland fire.

Table 4

Click here to download Table Foteu Madio and Dykes - Table 4.docx

Table 4. Summary of physical properties of peat at the three landslides, including previous data from Yang and Dykes (2006)*.

ST SR SA

Water content a,b (% mass fraction)

Loss on Ignition c (%)

Saturated bulk density e (Mg m–3)

Dry bulk density e Saturated (Mg m–3) hydraulic conductivity h,i (m s–1)

700-900 *620-860 600-700 600-700 *600-740

94.5-95.6 *97.8-98.8 94.6-97.2 d 95.0 *97.7-98.5

1.00 f *1.06 1.00 1.00 *1.05

0.10-0.20 *0.13 0.10-0.20 0.20 g *0.15

10–9 to 10–8 * < 10–11 10–9 to 10–6 10–8 to 10–6 * < 10–11

Notes a

There was negligible difference between field-wet and saturated water contents at all sites

b

Indicative ranges of mean values from 256-319 samples per site

c

Indicative ranges of mean values from 123-196 samples per site

d

The basal peat at Slieve Rushen was noticeably higher in organic matter than any other sampled peat

e

Mean values from 20-26 samples per site

f

The basal peat at Straduff Townland was noticeably higher (~1.10 Mg m–3) than any other sampled peat

g

The peat at Slieve Anierin had higher dry bulk densities throughout its depth

h

Indicative ranges of mean values from 19-26 samples per site, obtained using a ‘constant head’ method

i

Results obtained using a ‘falling head’ method were consistently 10 2-103 m s–1 higher than the respective ‘constant

head’ values

Author photographs - Foteu Madio

Author photographs - Dykes