Genesis of subglacial triangular‐shaped landforms (murtoos) formed by the Fennoscandian Ice Sheet

The purpose of this paper is to describe the internal structure and composition of recently discovered subglacial landforms called “murtoos” in order to interpret their formative processes and depositional environment. So far, murtoos have only been reported from Finland and Sweden, but they probably exist in all areas covered by past ice sheets. Murtoos mainly occur in fields along subglacial meltwater routes or corridors in close relation to eskers and ribbed moraine tracts. Murtoos were excavated perpendicular to the long axes of the triangular murtoo heads in six locations. Murtoos were found to be composed of silt/clay‐poor, sandy and gravelly diamictons interbedded with sorted sediments, and are suggested to be produced by pulsed, highly sediment‐concentrated flows during weak glaciotectonic deformation, indicating effective pressure close to zero. Murtoos can be divided into three main depositional units: 1) the core, 2) the murtoo body and 3) the murtoo mantle. The initial deposition of murtoos took place in a network of low canals and conduits or cavities with fluctuating stream flow likely over 50 km from the ice margin. Murtoos reveal an increasing influence of subglacial meltwater flow in rapidly widening broad and low conduits with increasing sediment transport over short distances. The results of this work suggest that murtoos were formed time‐transgressively over yearly meltwater cycles in a semi‐distributed drainage system not recognised before, in which was high‐pressure porewater conditions with rapid mobilisation of subglacial saturated sediments, a critical factor in the development of semi‐efficient drainage. Murtoos are suggested as missing element between distributed and channelised drainage systems not included in current glaciohydrological models or even in the theoretical basis of glacial hydrology.

the rapid melting and retreat of the FIS north of the Younger Dryas ice marginal zone in Finland (Ojala et al., 2019).
Murtoos are mostly less than 5 m in height, characterised by sandy diamictons interbedded with sorted sediments. The triangular heads of triangular-type murtoos, and also the heads of chevron-type and lobate-type murtoos (Ojala et al., 2019), point down-flow and are thus indicative of the local ice-flow direction. Previously, murtoos have either been mapped as hummocky moraines without mentioning their triangular heads, or just left unmapped as part of the ground moraine cover. A characteristic landscape feature of murtoo fields is the abundance of large boulders on their surfaces. This paper focuses on the internal architecture and sediment composition of murtoos formed during the rapid Holocene melting and deglaciation of the FIS in Finland, this in order to interpret their formative processes and environment. These results are compared with the sedimentary characteristics from the south-Swedish murtoos investigated by Peterson-Becher & Johnson (2021).

| Murtoo distribution
The distribution of well-defined triangular-shaped murtoos in Finland had been mapped by Ojala et al. (2019) and Ahokangas et al. (2021) ( Figure 1). Murtoos mainly occur in three larger clusters, two of these located along the track of the Finnish Lake District Ice-Lobe (FLDIL) (Figure 1), while a third cluster in SW Finland is associated with the eastern margin of the Baltic Sea Ice Lobe (BSIL). Murtoos typically occur as fields along subglacial meltwater routes or corridors F I G U R E 1 Distribution of murtoos (Ojala et al., 2019), ice-stream lobes and excavated study sites (stars)   . The study sites presented here ( Figure 1, Table 1) include the main clusters of murtoo fields in southern Finland, which represent areas with diverse glaciodynamic settings.
The Mikkeli site is located in an area that experienced a supra-aquatic deglaciation (Ojala et al., 2013), whereas the other sites are located in  areas where the ice-stream margin ended in a proglacial water body of the Baltic Sea basin at deglaciation. All excavation sites are located within the Svecofennian basement rocks domain, mainly composed of biotite paragneiss and tonalite (Lehtinen et al., 2005).

| Geomorphological setting of study sites
LiDAR The Laihia murtoo is located next to an esker within a murtoo field about 2.0 km 2 in size that is not part of a subglacial meltwater route but lies proximal to an area that has been interpreted as a cold-based area within the Suupohja district (Pitkäranta, 2009). The Mikkeli (Mi) site is situated along a subglacial meltwater route next to an esker and has small sporadic murtoo clusters between streamlined terrains of the FLDIL. The excavated murtoo in Sievi (Si) is situated within a 0.5-to 1.0-km-wide subglacial meltwater route that cuts through a field of ribbed moraine and subglacial hummocky moraine within the margin of a drumlinised ice-flow corridor along the trunk of the FLDIL (Ahokangas & Mäkinen, 2014;Vérité et al., 2021). The distances from the excavated murtoos to the nearest esker vary between 0.8 and 3.5 km for those sites (Kullaa, Laihia, Mikkeli, Sievi) where the murtoo route occurs alongside an esker, but between 8.2 and 27.2 km for the sites (Kämmäkkä and Kynäsjärvi) that join an esker down-ice. In addition to the main trenches, smaller test pits were excavated into the proximal side of murtoos in Sievi and Mikkeli, and into the front of the tip in Mikkeli. The trenches were first located with a hand-held GPS and levelled by using a levelling instrument in order to draw cross-sections in the field. The trenches were marked with a ground-fixed measuring tape to locate the observations and measure-  Figure 3d).
These sediments have a basal contact to ice-abraded bedrock.

| Depositional unit 2
The lower part of Unit 2 (subunit 2a) has similar gravels to those in Unit 1, but is sandier, including partly preserved interbeds (5-20 cm thick) of silty fine-to medium-grained sand with deformed lamination T A B L E 3 Roundness (Powers, 1953) of sieved murtoo-sediment clasts, 8-64 mm in size.

| Kynäsjärvi (Ky)
The Kynäsjärvi murtoo was excavated close to its eastern head with a 40-m long section, the northern part of it cutting the northern margin of the murtoo, whereas the southern part of the section follows its southern margin (Figure 6a). In addition was a 7-m long trench  The pebble-sized clasts in Units 1 and 2 are dominantly angular to subangular (similar to Kullaa clasts, Table 3), whereas larger clasts in the section are subangular to subrounded. According to measurement F I G U R E 5 Picture of a pit located between the murtoo and murtoo-related ridge at the NNE end of section Ku2 (see Figure 3b). The pit reveals a silty bouldery diamicton (till) at the bottom, overlain by partly preserved unit 1 and subunit 2a composed of trough-shaped diamictons of varying structure topped by poorly preserved bed of massive mud (Fm). The massive/matrix-supported diamicton of unit 3 with discontinuous bed of crudely stratified gravel in the lower part of the unit is altered by forest bed processes.

| Depositional unit 3
Unit 3 is composed of a loose and poorly sorted sandy massive/ matrix-supported diamicton with boulders, and is mostly disturbed by forest bed processes. The diamicton is partly underlain by a poorly preserved layer of stratified fine-to medium-grained sand. The lower

| Laihia (La)
At Laihia, a 40-m long and 2.0-to 2.5-m high section was excavated through a chevron type murtoo (cf. Ojala et al., 2021, Figure 9a). The SW part of the trench was excavated 4 m deeper, revealing a 2.0-m thick, clearly differently coloured, grey and silty massive/matrixsupported diamicton (till) (Figure 10a The diamictons in Unit 2 contain only 0.7-1.4% clay and 3.8-9.5% mud. The clasts in Units 1 and 2 are dominantly subangular to subrounded. Clast fabrics in Unit 2 show a strong flow-parallel alignment towards the murtoo tip in the middle part of the murtoo, whereas clast fabrics become more divergent closer to the murtoo margins ( Figure 9c). However, clast fabrics in the middle part (fabrics 7 and 9) may also show a less preferred orientation when measured close to boulders or trough margins (Figure 9c).

| Depositional unit 3
Unit 3 is composed of a loose, sandy and poorly sorted massive/ matrix-supported diamicton with boulders (Figure 9c,d). It is about 0.5 m thick and the lower contact with Unit 2 is sharp or amalgamated. The contact contains some boulders that are slightly pressed down into Unit 2 and overlain by sediment showing a crude fissility.
In places, a poorly preserved, stratified, medium-to coarse-grained sand layer covers subunit 2b. It displays intrusions and mixing into the overlying massive diamicton. The density of surface boulders is approximately 5 boulders/100 m 2 , but the chevron limbs are more bouldery.

| Depositional unit 2
The lower part of Unit 2 is composed of sandy massive/matrixsupported to crudely-stratified diamicton with deformed patches of sand (subunit 2a, Figure 12). The eastern side of subunit 2a is topped by a massive bed of mud (27.3% clay, 71.1% silt, 1.6% sand), approximately 20-25 cm thick (Figures 11d and 12b). It is partly mixed into subunit 2b, but mostly eroded within the main section (K2). The mud layer was also found in Pits Tk1 and Tk2 in front of the murtoo (Figure 11d).
Subunit 2b is composed of gravelly massive/matrix-to clastsupported diamicton that occurs as several trough-shaped sets with

| Sievi (Si)
The excavated murtoo at Sievi forms the head of a composite murtoo, the NE side of which consists of a few small murtoos sharing the same long erosional margin (Figure 13a). A 33-m long trench (Si1) was excavated across the 4-m high murtoo, the tip of which is joined by a small ridge about 1-3 m high and 25 m wide (Figure 13a,b). One

| Depositional unit 3
Unit 3 is about 0.5 m thick and composed of a loose, sandy massive/ matrix-supported diamicton with boulders ( Figure 13c). The contact to subunit 2b is amalgamated but contains a discontinuous clast layer, one clast thick. The density of surface boulders is 6-10/100 m 2 .

| INTERPRETATION OF DEPOSITIONAL UNITS AND THEIR SEDIMENTS
All excavated murtoos show similar sedimentological characteristics that have been divided into three main architectural/depositional units: Unit 1 (the core), Unit 2 (the murtoo body) and Unit 3 (the murtoo mantle) ( Figure 14, Table 4). There are some variations in the composition of the units between the sites, but characteristically, all of them reveal diamicton in trough-shaped sets in the upper part of the murtoo body. All sites exhibit poor rounding of clasts, ranging between angular to subrounded (dominantly subangular), which indicates short transport distances within the subglacial system. The Kullaa and Kynäsjärvi sites record the highest quantity of angular pebbles but also the most sorted and coarse-grained cores, as well as the best-stratified lower part of the murtoo body. Furthermore, both sites are located within the best-developed meltwater routes of this study, implying that these murtoos show the highest influence of meltwater activity. Thus, the higher number of angular pebbles could be explained by higher erosion rates of the substratum and related short transport distances.
The clay content in diamictons of Units 1 and 2 is between 0.7 and 1.5% (except 2.5% in one diamicton in Sievi, Figure 15), indicating that the clay component of sediment in transport was passing through the subglacial drainage system. The maximum clast size (sieve diameter) in all investigated murtoo sediment is 1.0 m. This is interpreted to indicate the maximum height of the flow space within the subglacial conduit system, whereas the largest boulders remained transported by the ice.
Data from the Geological Survey of Finland (GTK) indicate that regional tills are more homogeneous, have a higher clay and silt content than murtoo bodies (Unit 2), and have boulder sizes exceeding 1.0 m in diameter (S2,S3). Murtoo diamictons are also more sandy and gravelly and contain silty, sandy and gravelly intraclasts ( Figure 15).
Based on the classification of poorly sorted sediments by Hambrey (1994), the diamictons sampled here could be classified as either poorly sorted gravels or clast-rich sandy diamictons ( Figure 15). The till data of GTK from the Kynäsjärvi murtoo field indicate a highly similar composition to the murtoos studied here, whereas tills (mainly hummocky moraine) surrounding the murtoo field have higher mud and clay contents (S3).

Murtoo mantle
Laminated medium to coarse sand, some gravelly beds (incl. troughs) overlain by loose, poorly sorted diamicton mixed with underlying sands (with intrusions). Some trough-shaped clast layers and weak fissility in places. Forest bed with podsol soil.
Sheet flow deposits followed by rapidly melting ice bottom and concomitant meltwater flow. Clogged conduits. Disturbed by forest bed activity. Eroded by shore terraces in places.

Channels
Channel bottom with open-work boulder concentrations (empty spaces up to 0.5-1.0 m deep) or boulder lags.
Meltwater channel erosion after murtoo formation, transition to tunnels?

Surface boulders
Boulders up to 3 m in diameter. Boulders concentrated on murtoo surfaces.
Melt-out of large boulders, earlier concentration by ice flow towards the widening conduits.

| The core (unit 1)
The broadly arched, better sorted and stratified, and generally less Massive gravels indicate pulses of higher flow velocities and can be interpreted to record meltwater flow along the growing conduit or tunnel walls (Brennand, 1994;Shreve, 1972) and rapid deposition from heterogeneous suspension (Mäkinen, 2003). Crudely stratified gravels indicate that the largest clasts within the flow were transported as bed load. Amalgamated contacts (similar matrix) indicate closely spaced depositional pulses. According to Collinson & Thompson (1993), gravels with "massive and crude bedding may involve rapidly fluctuating episodes of sedimentation in which the sediment concentration is high, freezing of load takes place and individual events are hard to distinguish." A higher proportion of the coarsest core sediments occur in association with murtoos that reveal central boulder concentrations or bedrock knobs at their base (e.g., Kullaa and Sievi sites). This might indicate that conduit development was associated with reworking of existing till ridges or with cavities on the distal side of bedrock protrusions.
The overall weak or partial periodic deformation of Unit 1 indicates limited conduit closure by the overriding ice. However, the finegrained core in Kämmäkkä is interpreted to show upwards increasing deformation in the small bedrock trough just before the widening of the conduit. The scattered small pebbles and out-sized clasts were probably entrained from the melting ice roof, creating the common spotted texture, especially in sand-rich beds.
It is difficult to estimate the time frame for core deposition, but the development of the conduit and cavity system before the deposition of the murtoo body (Unit 2) indicates that murtoos formed in an environment where the drainage system gradually enlarged in order to accommodate increased meltwater and sediment flow, however, tunnels did not form (Hooke, 2019).

| The murtoo body (unit 2)
The murtoo body ( Subunit 2a records a higher variability in sediment composition; it is usually sandier and has a more stratified structure with frequent interbeds or patches of sand compared with subunit 2b. This is interpreted to indicate a widening of the subglacial conduit over the underlying core (Figure 14, stage 3). In general, dominantly laminated silty sands with occasional ripples were deposited by tranquil stream flows. However, repeated and more powerful pulses of sedimentconcentrated flows produced rapidly deposited beds of massive to crudely stratified gravels and massive/matrix-supported sandy diamictons. These beds include rafts, chaotic streaks and dislocated F I G U R E 1 5 The grain-size distribution curves of diamictons (upper) and sandy interlayers (lower) from the murtoo body (unit 2). Diamictons have <1.0-1.5% clay (except <2.5% in one diamicton in Sievi), 3-20% mud, 30-76% gravel, and d 50 values between medium sand and medium gravel. Sandy interbeds have mainly <5% clay, 18-73% mud, <2.5% gravel, and d 50 value ranging from coarse silt to fine sand (dominantly very fine sand). The fine-grained bed between units 2a and 2b in Sievi with fissile structure is clearly more poorly sorted compared to other fine-grained units. Thick lines indicate the average grain-size distribution.
patches of sand, suggesting erosion and subsequent rapid deposition from these sediment-concentrated flows. Patches of open-work gravel are indicative of deposition by meltwater flow. The bimodal parallel and transverse to flow clast fabric might be explained by flow conditions close to Lawson type III sediment flows (Lawson, 1979).
Moreover, the more stratified marginal trough in Kullaa (Ku2) with backflow ripples also provides evidence for turbulent, lateral stream flows during the formation of the lower murtoo body. Sandy beds and diamictons spotted by granules and pebbles, and associated oversized clasts are interpreted as related to the melting of the ice roof.
Furthermore, the sediments were periodically (after each depositional event) deformed by ductile deformation, interpreted to be due to shear stresses induced from the overriding ice and coupled to low shear strength due to high porewater pressures (Lesemann et al., 2010;Peterson-Becher & Johnson, 2021).
Boulder concentrations within the murtoo margins might be due to the development of deeper water with higher energy dissipation as well as higher marginal meltwater flow along the margins of a broad and low conduit (Hooke, 2019). Boulder concentrations were probably formed when the conduit was not able to widen further and became filled with sediments.
The dominantly sandy and gravelly, poorly sorted, matrixsupported diamictons, and their common occurrence in trough-shaped sets within the upper part of the murtoo body (subunit 2b), suggests repeated, often channelised flow events of slurries with high sediment-concentrations, in which the viscosity of the flows was controlled by the water content (Menzies, 1989). Importantly, the sudden changes in the character of the diamictons in the murtoo section at Kullaa (Ku1, Figure 3c) suggest rapid flow transformation of sediment flows (Lawson, 1981). The often poorly defined troughs and amalgamated contacts refer to close-in-time repeated flow events. Sharp, erosional trough bottoms with some signs of stratification indicate channelised stream flow preceding more sediment-concentrated flows. Crude clast layers one clast thick that often form trough bottoms and marginal low-to medium-angle clast layers probably represent lags from erosional currents (Collinson & Thompson, 1993;Lawson, 1981). Rafts, streaky stratification and dislocated patches of laminated silty sand within the lower parts of the diamictons suggest entrainment of the underlying sediments and shear-induced deformation and mixing by the overriding sediment flow (Phillips, 2006). The vertical distribution of clast sizes is determined by the grainsize distribution in individual slurries. The lack of inverse grading might refer to incremental deposition and aggradation by multiple debris flows "in which a number of momentum transport processes operate involving both solid and fluid forces" (Sohn et al., 1999, p. 112). Furthermore, as stated by Sohn et al. (1999, p.112), it is difficult to categorise sediment flows, "because individual flows may comprise more than one flow type at an instant in time and are subject to a series of flow transformation during transport" (Lawson, 1981). This chaotic process might lead to heterogeneity of the murtoo diamictons, where individual flow events are difficult to separate. The diamicton properties and common occurrence of trough-shaped sets described here suggest sediment flows close to Lawson type III sediment flows.
Such sediment flows are associated with flow channelisation, the shear zone encompassing the entire mass with laminar flow and localised temporary turbulence (Lawson, 1979(Lawson, , 1981. Moreover, thinly laminated silty sands in association with diamictons are common and imply meltwater flow over sediment flow surfaces (Lawson, 1981). The flow-parallel clast fabric in diamictons indicates high rates of laminar shear within individual flows (Sohn et al., 1999).
The chaotic deformation of fine-grained laminated sediments and their streaky appearance, larger clasts cutting through the sand beds, the dominant ductile soft-sediment deformation structures, and downward-protruding diamicton pockets (load structures) or injections of mobile debris with larger clasts can all be related to shear stresses induced by overriding ice and related debris flows (Menzies, 1989). The periodic deformation with a lack of apparent dewatering structures lends support to an effective pressure close to zero at the ice-bed contact (Lesemann et al., 2010). The downward protruding diamicton pockets with large clasts are typically located in the upper parts of subunit 2b, implying increased overburden pressure during the final deposition of the murtoo body.
The subunits of murtoo bodies (subunits 2a-2b) separated by muddy glaciofluvial deposits (subunit 2ab) suggest that they were deposited during similar flow variations after the formation of the core (Unit 1). This type of sediment successions can hardly be deposited over several years, but rather during systematic variations in water input. Therefore, we consider that the lower part of the murtoo body (subunit 2a) formed during the spring peak flow event at reestablishment of the subglacial drainage system, leading to rapid widening of the conduit above the core. Later, the muddy interlayer (subunit 2ab) was deposited at lower flow velocities as the drainage system became wider and better connected over the base of the ice (Vérité et al., 2021). The upper part of the murtoo body (subunit 2b) would then been deposited at an even later stage of the meltwater peak originating from increased ice surface melt. Such meltwater would be subglacially distributed within an already well-connected drainage system with rapidly upwards expanding conduits and cavities. Our scenario thus suggests that the murtoo bodies were built during one single melt season with two meltwater peaks upstream of an efficient drainage system (Chandler et al., 2013). However, some spatial and temporal differences in murtoo body development should be expected, depending on how the pressure conditions and meltwater flow velocities change during a melt season.
The dip direction of more well-preserved beds, current and backflow ripples, and trough-shaped diamicton beds with their clast fabrics all together suggest a lobe-shaped or divergent flow pattern within the murtoo heads during deposition of murtoo body sediments ( Figure 16). This provides evidence that these sediments did not exist prior to their erosion, but were cut to triangular shape by flowing water along the murtoo body margins. The well-defined single murtoos in Laihia and Mikkeli can be used for roughly estimate the volume of the murtoo body sediments (Unit 2), as the topography of the base of the murtoo body can be more reliably estimated at these two sites. Based on the excavations, we estimate that the average murtoo body thicknesses are about 2 m, and the area of each murtoo is about 4,000 m 2 . This gives a sediment volume of about 8,000 m 3 .

| The murtoo mantle (Unit 3)
Much of the original sedimentological characteristics of Unit 3 have been altered or erased during later soil forming processes (podsol) and frost activity in these quite thin, uppermost sediments. In spite of this, the Unit 3 diamicton often show a weak stratification and have associated clast horizons (sometimes trough shaped) and interbedded openwork pebble gravel beds, suggesting influence of subglacial water flow also after the formation of the murtoo body ( Figure 14, stages 6-7). The poorly preserved arched beds of sorted and stratified sediments (mostly medium and coarse-grained sands) between Units 2 and 3, indicate that murtoo body formation (Unit 2) was followed by diminished stream flow velocities over the whole murtoo. Moreover, the Laihia murtoo clearly shows how these sorted sediments intrude and mix into the overlying diamicton (Unit 3). This might partly explain the common sandy character of Unit 3.
At places the contact between Units 2 and 3 includes ploughedin boulders, at which the diamicton conformably overlying that contact often reveal a weak fissility, which we relate to shearing of the ice flow (Piotrowski et al., 2001). Importantly, the murtoo mantle (Unit 3) is not only confined to murtoos, but also drapes adjacent murtoo-related ridges (e.g., in Kullaa, Kämmäkkä and Mikkeli), indicating deposition over a wider area of the ice bed.
In   (2021) proposed that the subglacial environment of murtoo formation "is within the distributed system where the bed receives meltwater from repeated influxes of supraglacially derived meltwater." The fact that each murtoo has a stratified core with sorted sediments, sometimes interbedded with diamictons, interpreted as deposited from subglacial slurry flows, suggests initial deposition in a network of low canals or conduits and cavities with fluctuating and variable stream flows ( Figure 16). This type of environment suggests formation of murtoo fields within a zone of wet/warm-based conditions along the ice/bed interface and under high overburden pressure.
As shown by the sedimentological evidence from our sections, murtoo sediment deposition is characterised by rapid mobilisation of watersaturated sediment and transformation into sediment-concentrated flows within channels or canals incised both down into basal sediment and up into the ice (Eyles, 2006;Nienow et al., 2017).
It is evident that a distributed/linked cavity system versus tunnels forms the end members of subglacial drainage patterns (Hooke, 2019). Therefore, the depositional environment of murtoos, preceded by deposition of murtoo cores within a widening conduit, probably relates to a semi-efficient transitional system in a highpressure environment not recognised before (Figure 16, cf. Fudge et al., 2008). The proposed semi-efficient drainage system responsible for murtoo formation is suggested to have occurred 40-60 km from the ice margin (cf. Bartholomew et al., 2011;Chandler et al., 2013;Greenwood et al., 2016;Hooke & Fastook, 2007;Ojala et al., 2019;Peterson-Becher & Johnson, 2021;Vérité et al., 2021), but its extension further up-ice remains unknown. Dow et al. (2014) suggested that subglacial conduits at about 70 km from the Greenland Ice Sheet margin and further upglacier are improbable to maintain because of the high overburden pressure (Figure 16).
Murtoos often occur up-flow from eskers formed within the R-channels of the channelised drainage system closer to the ice margin (Ojala et al., 2019) and connect to them via erosional escarpments (Ojala et al., 2022).
Murtoos are primarily depositional landforms with erosional triangular heads (Peterson-Becher & Johnson, 2021). The depositional units of murtoos suggest an increasing influence of subglacial meltwater flow in rapidly widening broad and low conduits with increasing sediment input and transport. We thus suggest that murtoos are indicative of saturated subglacial bed conditions with hyperconcentrated flows or slurries that dominated the last depositional stage of their formation. Eyles (2006) described the transformation of overpressured till into hyper-concentrated slurries as a highly effective erosional tool due to the recharge of meltwater into the ice sheet bed. Brodzikowski & Loon (1991) stated that subglacial conditions, particularly in an environment with an irregular topography, and that water-saturated sediments and instable density gradients, favour the development of mass-transport processes. Closely spaced, narrow and channelised hyper-concentrated flows were probably initiated as subglacial flood pulses under rapidly changing water pressure conditions (Vérité et al., 2021).
Based on the sedimentological evidence presented here, it is proposed that the murtoo body (Unit 2) was deposited during a single melt season because of seasonal variation in meltwater input and to that related changes in the subglacial drainage system. Extra input of meltwater to drive this semi-efficient system must have come from the ice surface or via the drainage of subglacial lakes (Shackleton et al., 2018). The source area of murtoo sediments was undoubtedly upstream of the murtoo head, where the creep of saturated till into a conduit tended to constrict it ( Figure 16). Such a zone would include high water pressure and low shear stresses induced from the ice. If till creep was faster than ice flow velocity, subglacial channels cut into the sediment (canals) could exist (Walder & Fowler, 1994). It is usually considered that ice surface slope of the FIS was low (Shackleton et al., 2018). In such a situation, the development of a distributed subglacial canal system would be a stable hydrological configuration with an effective pressure close to zero at the ice/bed interface (Greenwood et al., 2016). The observed weak to moderate, periodic deformation of murtoo sediments induced by overriding ice indicates that the subglacial water pressure (Pw) and the overburden pressure (Pi) were constantly close to each other (Lesemann et al., 2010). This would relate to moderately efficient drainage that could occur through a system close to the overburden pressure (Greenwood et al., 2016).
Murtoo body formation (Unit 2) was most probably initiated by the rapid clogging of the preceding conduits, throttling the flow ( Figure 16). In a steady state, the flow of saturated till into the conduit must be balanced by meltwater erosion and transport of the same sediment (Hooke, 2019). Increased water pressures and down-ice-mobilised sediment led to up-ice-migrating lateral erosion of the till and to lateral meltwater flow and erosion around the conduit fill, as reflected by the typical morphology of murtoos   (Figure 16). During the murtoo formation, the vertical closure rate was probably highest in the middle and decreased towards the margins (Ng, 1999). Accordingly, melt rates were higher at conduit margins, with the deepest water and greater energy dissipation (Shreve, 1985). In conditions where Pw increased as meltwater discharge (Q) increases, the drainage system would tend to remain braided or distributed (Hooke, 2019). The continuum from murtoo mantle deposition to lateral channel erosion and the final draping of murtoos by boulder blanket suggests development towards route-wide discharge and increased basal melting rates. This could also explain the good preservation of murtoo morphology and suggests a time-transgressive origin for the murtoo fields.

| A comparison between the sedimentary characteristics of Swedish and Finnish murtoos
Our study presents the typical sedimentary characteristics of triangular-shaped murtoos in the Finnish area of the FIS, which can be compared with those recently excavated and described from Sweden (Peterson-Becher & Johnson, 2021).
Despite some spatial and temporal differences, there are many similarities in sedimentological composition and internal architecture.
The most important similarities for both areas include the following: 1) investigated murtoos are characteristically composed of sandy and gravelly diamictons, interbedded with sorted sediments beds (silt, sand and gravel), interpreted as deposited in subglacial position; 2) the murtoo diamictons are relatively loose and semi-sorted in character and with low amounts of finer grain sizes (silt/clay). This deviates from typical subglacial tills, such as traction and basal melt-out tills; 3) in most cases, the sediment in murtoos is found on top of more compact and finer-grained subglacial till; 4) the dominant ductile deformation is restricted to separate beds indicating the periodic nature of deformation events; 5) the contacts between diamictons and sorted beds are mainly gradational and sediments are mostly mixed; 6) there is a frequent, patchy occurrence of sand intra-clasts, which often carry outsized gravel clasts and small boulders; 7) there is a higher abundance of sorted sediments in the lower part of each murtoo, and coarser sediments (including diamicton beds) in the upper part; and 8) in all cases, murtoo surfaces carry a high frequency of large boulders.
A difference, however, is that the Finnish murtoos lack the folding, large-scale convolution and clastic dykes, as described in some of the Swedish murtoos. Swedish murtoos also appear to have been more influenced by liquefaction and exhibit water-escape structures.
In general, Finnish murtoos appear to be less deformed, and reveal a tri-partite architecture often with trough-shaped (channelised) diamictons in their upper parts. Finnish murtoos also show current ripples, so far not recorded from Sweden.
It appears that Finnish murtoos were deposited in subglacial environment with an effective pressure close to zero, in which the meltwater input occurred as repeated, short-lived pulses on top of generally high meltwater input. Thus, because of high meltwater pressure, deformation from ice remained low. In spite of this, the sedimentary records from the Swedish examples points to an overall similar sediment composition and internal architecture with the exception of a larger degree of deformation structures, suggesting that the depositional processes of murtoos were highly similar.

| CONCLUSIONS
We conclude here that murtoos are depositional landforms with erosional triangular heads. Typically, a murtoo consist of a core with sorted and stratified sediment, followed by deposition of coarser sorted sediment and diamictons, related to mobilisation of saturated sediment during increased meltwater input. Murtoos are thus characteristically composed of clay/silt-poor and semi-sorted heterogeneous diamictons interbedded with coarse gravels, all produced by highly sediment-concentrated flows with intervening low-velocity stream flows and with periodic deformation by the overriding ice. The weak to moderate periodic deformation by the ice indicates a basal effective pressure close to zero due to a high porewater pressure.
The murtoo bodies are proposed to indicate the deposition of the described sediment successions during just one melt season within rapidly widening broad and low conduits or cavities. Murtoo body sediments and the underlying subglacial murtoo core infills together represent a semi-efficient drainage system not recognised before. We suggest that this happened within a zone of high porewater pressure, tentatively about 40-60 km from the ice margin. Our results also suggest that, during the final deglaciation and after murtoo body formation, the subglacial drainage reached route-wide discharge. During this stage was murtoo mantle deposition as well as the final shaping of the murtoos with lateral channel erosion. The boulder-rich murtoo surfaces represent increased basal melting closer to the ice margin.
This succession of events, suggesting a time-transgressive development of murtoo tracts, might also explain their good preservation.
Importantly, we propose that murtoo tracts within meltwater corridors are a missing element between distributed and channelised drainage systems, not recognised in current glaciohydrological models or even in the theoretical basis of glacial hydrology. The rapid mobilisation of saturated sediment is probably a critical factor in the development of the suggested semi-efficient drainage system. Therefore, understanding of murtoo processes and genesis will have a crucial impact on modelling approaches, as well as on resolving uncertainties associated with the presence of wet subglacial sediment at the ice-terrain boundary and the related development of the drain- remarkably helped to improve the manuscript. The authors have no conflict of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.