Active temperate glacial landsystem evolution in association with outwash head/depositional overdeepenings

The response of temperate glaciers to rapid climate warming is reflected in the geomorphology (landsystem) resulting from snout recession. This paper develops a locally diverse process‐form model of active temperate glaciers through mapping and quantification of historical landscape change on the Fjallsjökull‐Hrútárjökull foreland, SE Iceland. Quantification of volumetric and areal changes on the foreland are based on time series of archival aerial images for the period 1945–1998, high‐resolution satellite imagery for 2014 and digital elevation models of differences derived from time series of UAV imagery for the years 2014, 2016, 2018, 2019 and 2022. Landscape change and glacier snout behaviour since 1945 highlights the importance of azonal and potentially intrazonal signatures in temperate glacial landsystems, particularly: (1) the development and collapse of partially supraglacial outwash fans to produce outwash heads fronting depositional overdeepenings; and (2) the emergence of ice‐cored eskers that record the evolution of englacial drainage networks operating over overdeepenings. Such landform assemblages are manifested as substantial ice‐cored/hummocky terrains, a characteristic of deglaciating forelands that is likely to be widely replicated wherever ice‐contact glacifluvial assemblages create outwash heads that act as depositional overdeepenings. Due to its significantly greater supraglacial debris cover, complete de‐icing of the Hrútárjökull snout in response to post‐1945 warming was delayed until around 2014. This constitutes a prime example of incremental stagnation, which in a rapidly warming climate has resulted in significant landscape change (land surface elevation collapse of 0.8 m a−1) over the last 8 years.

Spatial and temporal changes in the landsystem signatures of some glacier snouts in southern Iceland are trending towards a set of process-form regimes indicative of the accelerated climate warming of the late 20th to early 21st centuries. This involves an apparent switch from annual push moraine formation typical of active temperate glaciers to downwasting and calving in expanding proglacial lakes (Bennett & Evans, 2012;Bennett et al., 2010;Phillips et al., 2013).
However, these changes are not wholly climate-driven, as they have also been conditioned by snout recession into overdeepenings. In southeast Iceland, retreating piedmont lobes are rapidly uncovering depositional overdeepenings, or depressions formed inside the former ice-contact slopes of sandur fans (outwash heads; sensu Benn et al., 2003;Kirkbride, 2000), which may coincide with erosional overdeepenings (e.g. Bennett & Evans, 2012;Spedding & Evans, 2002). In some locations these depressions, especially where they are associated with erosional overdeepenings (e.g. Kvíárjökull, Heinabergsjökull and Hoffellsjokull; Bennett & Evans, 2012;Evans & Orton, 2015;Evans, Ewertowski, et al., 2019), are large enough to switch whole snouts into floating and calving mode. However, this may not constitute intrazonal landsystem change but rather azonal change, whereby local topographic conditions switch the process-form regime temporarily.
Such switching of process-form regimes is well demonstrated on the foreland of Fjallsjökull, where rapid calving into proglacial lake Fjallsarl on presently characterizes the northern part of the glacier snout (Dell et al., 2019). In contrast, on the south foreland, depositional overdeepenings have been evolving in the former coalescence zone of Fjallsjökull and Hrútárjökull since around 1945 Evans et al., 2009;Guðmundsson et al., 2019).
During the production of these overdeepenings and their associated landform-sediment assemblages, the glacier margins have continued in active temperate mode (Chandler et al., 2016a(Chandler et al., , 2016bEvans et al., 2009), with the snout of Hrútárjökull also developing a significant supraglacial debris cover on its right lateral margin.
The localized influence of depositional overdeepenings on the Fjallsjökull-Hrútárjökull foreland has provided an azonal landsystem signature on a historical and therefore measurable timescale. The aim of this study was to map these emerging glacial landystem signatures to quantify landscape change on the Fjallsjökull-Hrútárjökull foreland, enabling the development of locally diverse process-form models that reflect the response of individual temperate glaciers to rapid climate warming. Specifically, it provides a detailed analysis of the complex assemblage of ice-cored landforms that have emerged during glacier snout downwasting and recession inside outwash heads (depositional overdeepenings) in the former coalescence zone of Fjallsjökull and Hrútárjökull.
2 | FJALLSJÖKULL-HRÚTÁRJÖKULL: STUDY AREA AND PREVIOUS RESEARCH The piedmont lobes of Fjallsjökull and Hrútárjökull are two of many that are nourished by ice flowing out from the Öraefi caldera ( Figure 1). Formerly coalescent until $2010 (Hannesd ottir et al., 2015), they descend rapidly from their accumulation zone at $1000 m to near sea level over a distance of ca. 6 km. These outlet glaciers flow over a series of bedrock steps to form some of the most spectacular ice falls in Iceland (Figure 1). While the dimensions of the erosional overdeepenings beneath Fjallsjökull and Hrútárjökull have been reported (Magnusson et al., 2012) and their influence on ice margin recession evaluated (Hannesd ottir et al., 2014(Hannesd ottir et al., , 2015Figure 1b), no analysis has been undertaken on the emergence and development of the depositional overdeepenings in the area of downwasting and decoupling of the two snouts to the south (Figures 1 and 2).
The geomorphology of this area of the foreland has been represented by  in their map based upon imagery for the year 2012 ( Figure 3). This depicts an area mapped as overdeepening deposits (on the south Fjallsjökull margin) and icecored hummocky terrain (along the Hrútárjökull margin), together with small areas of kame and kettle topography, which are increasingly common landform-sediment associations in recently deglaciated forelands in southern Iceland (e.g. Bennett & Evans, 2012;Bennett et al., 2010;Evans et al., 2017aEvans et al., , 2017bEvans et al., , 2018Evans et al., , 2009). The sediment assemblages in such areas are related not only to supraglacial moraine construction but also to the widespread development of complex ice-contact glacifluvial processes, whereby englacial esker networks feeding into ice-marginal and proglacial outwash (sandur) fans progressively enlarge and bury downwasting snouts. This results in a hummocky kame and kettle topography that, since the benchmark process-form studies of Price (1969,1971,1973) and Howarth (1971), has been acknowledged as the product of widespread de-icing/melt-out of buried glacier ice (e.g. Evans & Orton, 2015;Evans & Twigg, 2002;Evans et al., 2009Evans et al., , 2018Storrar et al., 2015). An appreciation of the greater importance of glacifluvial rather than supraglacial topographic reversal origins of such hummocky terrain on deglaciating Icelandic forelands has alerted glacial geomorphologists to the limitations on, and localized nature of, supraglacial hummocky moraine generation more generally and the potential for its consequent over-representation in palaeoglaciological reconstructions (cf. Bennett & Evans, 2012;Eyles, 1979Eyles, , 1983Kjaer & Krüger, 2001;Slomka & Eyles, 2015;Spedding & Evans, 2002).
The present study builds on the modern landsystem analysis of Evans et al. (2009) and. More specifically, we focus on surveying, charting and quantifying: (1) the development and collapse of partially supraglacial outwash fans that culminate in the appearance of outwash heads fronting depositional overdeepenings (Bennett & Evans, 2012;Price, 1969); and (2) the genesis of the ice-cored eskers that record the evolution of englacial drainage networks operating over such depositional overdeepenings (Bennett & Evans, 2012;Price, 1969;Spedding & Evans, 2002). This is significant more widely in that such localized, azonal and potentially intrazonal landsystem signatures (sensu Evans, 2013) are creating substantial ice-cored/hummocky terrains and this is likely to be replicated in the near future on deglaciating forelands in Iceland as well as further afield where ice-contact glacifluvial assemblages create outwash heads that act as depositional overdeepenings.

| METHODS
Systematic repeat mapping and landform characterization of rapidly deglaciating proglacial areas are essential for the development of our understanding and reconstruction of glacial process-form models and their relationships to glacier dynamics. However, the annual patterns of recession and the relatively small areas exposed every year mean that carrying out regular aerial or satellite surveying is expensive and impractical. Moreover, some of the landforms are very subtle, making it impossible to recognize them even with high-resolution satellite imagery . Only recently, advances in technology have enabled the development of low-cost alternatives for traditional aerial surveys Hugenholtz et al., 2013;Smith et al., 2016;Westoby et al., 2012;Whitehead et al., 2014). Small uncrewed aerial vehicles (UAVs or drones) can be used to acquire high-resolution (several centimetre ground sampling distance [GSD]) low-altitude photographs. These UAV-based photographs can subsequently be processed using structure-from-motion (SfM) photogrammetry to generate detailed orthophotos and digital elevation models (DEMs). This very detailed, but spatially restricted, data can be combined with a coarser, but still high-resolution, satellite imagery to extend our interpretation to the scale of the whole forelands.

| Datasets and data processing
Historical development of the Fjallsjökull-Hrútárjökull foreland was reconstructed based on time series of archival aerial images covering F I G U R E 1 Location maps and glaciology of Fjallsjökull-Hrútárjökull: (a) LiDAR image of the Fjallsjökull-Hrútárjökull foreland with inset map showing the location of the study area in southern Iceland; (b) the Öraefi caldera and its ice cover with ice-marginal recession patterns and ages of the outlet glaciers mapped; (c) cross profiles (located in b) through the underlying topography and surface changes of the two snouts over time (from Hannesd ottir et al., 2015) [Color figure can be viewed at wileyonlinelibrary.com] the period 1945-1998 and a high-resolution satellite image from WorldView-2 satellite captured on 23 September 2014. The satellite image was an ortho-ready standard product projected in ETRS89 UTM zone 28 N projection. The dataset was orthorectified and pansharpened to generate a four-band 0.5 m GSD multispectral product.  (Table 1). Flight operations were conducted at the end of August and the beginning of September each year. During each survey, ground control points (GCPs) were surveyed using differential GNSS Topcon Hiper II, with prominent stable features such as isolated boulders and small stone circles targeted as GCPs. The 2022 surveys were flown with a DJI Phantom 4 RTK, and the local GNSS base station was run simultaneously to enable future processing of RTK camera positions.
Images captured by UAVs were processed using an SfM approach in Metashape Photoscan 1.8.4. We adopted the general approach as proposed by Evans et al. (2016) and generated dense point clouds, DEMs and orthomosaics. In addition, individual processing parameters were set up as indicated in the 'optimal' workflow for processing UAV data in Metashape (Śledź & Ewertowski, 2022).

| Quantification of landscape change
Quantification of landscape change focused on four key areas of the foreland, demarcated by boxes in Figure 4, in order to develop evolutionary sequences of landform-sediment assemblages (landsystem facets) as they relate to englacial meltwater processes and supraglacial outwash fan construction and collapse. Time series of UAV-generated DEMs were resampled to 0.1 m to ensure consistency of datasets from different years. DEMs of difference (DoDs) were constructed using Geomorphic Change Detection software (https:// gcd.riverscapes.net/) as proposed by Wheaton, Brasington, Darby, Merz, et al. (2010) and . This tool has been applied in previous studies in glacial environments (e.g. Midgley et al., 2018;Tonkin et al., 2016). Errors were estimated based on stable surfaces (i.e. no changes were observed between sequential DEMs), with several areas selected for DEM pairs and used to investigate elevation differences. This provided an indicator of uncertainty, from which a 0.5 m elevation surface difference was established as a minimum level of change detection (minLoD).

| Geomorphological mapping
Geomorphological mapping of the former Fjallsjökull-Hrútárjökull coalescence zone (Figure 4) was undertaken following protocols outlined by Chandler et al. (2018) and conforming to cartographic styles of previous Icelandic foreland maps (see Evans, 2009 for an overview and references for case studies). Genetic classifications of landforms and sediments were based on interpretations of repeat aerial and satellite imagery as well as extensive field checking. The historical aerial photography was also employed to reconstruct the sequence of ice marginal change, lake development, meltwater and lake drainage and overall foreland evolution ( Figure 2). The surficial geology units and landform symbology are compatible with previous maps by Evans et al. (2009)  , but with greater emphasis on the landform details of the depositional overdeepenings. The map (Figure 4) was produced at a scale of 1:1800 based on a 2014 orthophoto generated from UAV images.
F I G U R E 2 Oblique views over the Fjallsjökull-Hrútárjökull foreland for various dates using the available aerial photography and showing the changes in ice margins and proglacial landforms (aerial photographs by Landmaelingar Islands). HO = Hrútárjökull overdeepening; FO = Fjallsjökull overdeepening; TM = thrust moraine; FR = Fitjaoldur remnant F I G U R E 3 Extract from the surficial geology and geomorphology map of , showing the location of the study area in Figure  In order to place the case studies presented here in the context of the wider foreland of Fjallsjökull-Hrútárjökull, we now provide a brief overview of historical glacier recession from the LIA maximum and the landforms charting that recession, as previously reported by  and Guðmundsson and Evans et al. (2022)  and mountains to the southwest in the river Múlakvísl (Guðmundsson & Evans, 2022). Overall glacier recession by around 1945 was beginning to uncover the large, arcuate and broad ridge of Fitjaöldur, which is up to 80 m high and superimposed by recessional moraines (Figure 2). The sedimentary core of Fitjaöldur is exposed in a spillway gorge, as documented by  and, together with its asymmetrical cross profile, reveals that this large arcuate feature is an overridden outwash head dating to pre-LIA times. The ridge likely formerly continued in a second arc across the Hrútárjökull foreland to join the hillslopes to the southwest. However, it has been dissected by glacial erosion during the LIA by the two gla- After 1994, significant changes took place at the margins of both glaciers as, like most other snouts in southern Iceland, they readvanced and constructed prominent moraines (Evans & Hiemstra, 2005;Evans et al., 2009Evans et al., , 2016Evans et al., , 2017a. This appears to have been driven by a combination of a ca. 5 year-long, positive North Atlantic Oscillation index and a post-1965-1990s long-term cooling trend (Evans, Guðmundsson, et al., 2019;Sigurđsson et al., 2007). readvance moraine complex was gradually partially consumed by buried ice melt-out and lake development (see case study 2 below) but the proglacially thrust outwash continues to be recognizable by its surface relict channels. The development of surface tension cracks and collapse pits is also increasingly evident, indicative of an ice core and hence a supraglacial origin for the former sandur corridor.
In 1998, the remnants of the south Fjallsjökull proglacial lake appear to be partially interlinked and are predominantly supraglacial ponds, lying mostly within a small push moraine that is linked to the

| Evolution of the glacial geomorphology of the Fjallsjökull-Hrútárjökull foreland and overdeepenings
Quantification of landform evolution in the Fjallsjökull-Hrútárjökull coalescence zone since 2014 is now exemplified through four case study areas (Figures 4 and 8  Fitjaöldur gorge drained the lake in the south Fjallsjökull overdeepening ( Figure 6); a maximum of 5 years.
Two further complex networks of minor esker ridges and interlinked crevasse infills (short linear, gravel-filled ridges), predominantly aligned north-south, occur on the adverse slope of the outwash head ( Figure 9ii) and on the inner floor of the overdeepening (Figure 9i).
The former appear to record glacier sub-marginal drainage from the ice-contact fan surface back under the thinning snout and are therefore effectively subglacially or englacially engorged eskers (sensu Evans et al., 2018;Mannerfelt, 1945Mannerfelt, , 1949 , 14 and 15). This is mostly replacing large areas of debriscovered ice that had developed at the right lateral and frontal margin of Hrútárjökull and that had thickened and readvanced in the early 1990s (Figures 2 and 7b). The outwash deposits to the southwest were previously formed in an ice-marginal braided stream lying at the base of a steep debris-covered ice cliff and now constitute a kame terrace after the removal of the debris-covered ice; this terrace links directly to the sandur surface to the south.
The evolution of this large lake and its surrounding landforms is continued into the ice-cored hummocky terrain as well as the early 1990s push-moraine ridge and pitted thrust moraine on the east side of the lake (Figure 17). This indicates that the sandur of pre-1990s readvance age was ice-cored and deposited over shallow snout ice during a period of relative snout stability before the readvance occurred. As highlighted above, this stable period started around the mid-1960s and culminated in the early 1990s readvance (Figures 2   and 7b; Evans, Guðmundsson, et al., 2019;Sigurđsson et al., 2007) and followed a phase of apparently rapid recession from the late 1930s to 1945 (Guðmundsson & Evans, 2022).
The evolution of this area of outwash collapse is quantified using  Table 2). In 2014, the 1990s thrust moraine covered an area T A B L E 2 Calculations of total volume change using DoDs as a surrogate for buried ice lost to melt-out over time since 2014 (panel a) and average elevation loss (normalized by area) = surface lowering volume divided by case study area (m 3 /m 2 ) (panel b)   1945,1964,1982,1994,1998

| RECONSTRUCTION OF THE SPATIAL AND TEMPORAL CHANGES AT THE FJALLSJÖKULL-HRÚTÁRJÖ KUL L OVERDEEPENINGS
Based upon the observations, quantification and interpretations outlined above, we now present reconstructions of the events in the two depositional overdeepenings (Figures 19 and 20). These reconstructions constitute glacial process-form models that are invaluable because they are based upon events captured in historical imagery and ongoing annual field surveys, which both validate and expand upon similar research programmes initiated in southern Iceland in the 1960s (Howarth, 1971;Price, 1969Price, , 1971Price, , 1980Welch & Howarth, 1968).
The outwash head that fronts the south Fjallsjökull local overdeepening (case study area 1) was created by the development of icemarginal drainage around the front of, and over, a shallowing ice lobe in the early 1990s, as evidenced by the pattern of relict drainage channels on the fan surface (Figures 7a and 20a). The origin of the fan's steep and pitted proximal slope, topped by a minor push F I G U R E 2 0 Conceptual models of outwash head/depositional overdeepening development and their associated landform-sediment associations for Fjallsjökull (a) and Hrútárjökull (b). Locations of transects are demarcated in Figure 19. See text for details [Color figure can be viewed at wileyonlinelibrary.com] moraine, is coeval with the later (1994)(1995)(1996)(1997)(1998) construction of the thrust moraine in the ice-cored outwash overlying the overdeepening.
The switch of marginal meltwater drainage from the kame terrace to a tunnel through the outer thrust block to the fan is recorded by the highest-elevation esker on the west side of the overdeepening ( Figure 9). As the thrust blocks and debris-covered ice in the overdeepening melted out and downwasted and the resulting supraglacial lake level fell in response to spillway incision in 2003-2004 (Figure 6), the complex array of eskers was produced. The arcuate, inset series of altitudinally decreasing eskers relate to tunnels draining through the ice and increasingly directed across the overdeepening towards the spillway. The englacial/subglacial engorged eskers were fed by meltwater flowing between the downwasting thrust block margins and the proximal fan slope and then re-entering the ice, partially directed by fractures but also driven by the hydraulic gradient towards the spillway. The highly fractured stagnant ice that remained on the north part of the overdeepening floor as late as 2016 was tunnelled and its fractures infilled by the runoff in the Hrútá (Figures 9i and 12). further verifies process-form models compiled by Price (1965Price ( , 1966Price ( , 1969 and Howarth (1971) and developed by Gustavson and Boothroyd (1987), Evans and Twigg (2002) and Storrar et al. (2015. The recent emergence of the Hrútárjökull depositional overdeepening (case study areas 2, 3 and 4) has occurred through the production of the large lake as a result of rapid melting of the extensive area of debris-covered snout on the south margin of Hrútárjökull (Figures 7b, 19 and 20b). This part of the snout was driven forward during the 1990s, forming a minor push moraine fronted by thrust  7b). This prograding outwash buries not only the remaining glacier ice but also the older outwash head, which presumably lies beneath the post-1968 outwash head (Figures 19 and 20b). This superimposition of outwash heads and buried ice helps to explain the abnormally large area and depth of sandur collapse due to ice melt-out on the Hrútárjökull foreland.
The combination of repeat aerial imagery, glacier snout oscillation records and field observations at Fjallsjökull-Hrútárjökull facilitates the reconstruction of two interlinked glacial process-form regimes, relating to outwash heads and depositional overdeepenings, which constitute azonal signatures in the active temperate glacial landsystem (Figures 19 and 20).
Of particular significance are the evolutionary stages of outwash head development, which occur through the progradation and aggradation of glacifluvial outwash deposits (sandur) around masses of debris-rich/debris-covered glacier snout. The occurrence of such slowly ablating parts of glacier margins appears crucial to the development of these sandar and the burying of unusually large ice masses, similar to those recorded for example at the Breiðamerkurjökull supraglacial fan sites (Price, 1969;Storrar et al., 2015Storrar et al., , 2020; in both settings the emergence of englacial eskers is diagnostic of buried snout ice. Phases of extensive burial of glacier ice by outwash, similar to that observed at Hrútárjökull, are more likely to occur where debriscovered ice becomes isolated from the receding snout (incremental stagnation; Bennett & Evans, 2012;Eyles, 1979). In the case of repeat events this can result in superimposition of outwash heads (Figures 19 and 20) or features such as dead-ice sinks and moats as proposed by Fleisher (1986). The uncovering of overdeepenings in Iceland has been an important development in the recognition of azonal landsystem signatures, especially those fronted by outwash heads and hence at least partially depositional in origin. This is a prime example of local topographic conditions temporarily switching the predominant process-form regime during overall ice recession and is being observed at an increasing number of Iceland glacier snouts (e.g. Bennett & Evans, 2012;Evans & Orton, 2015;Phillips et al., 2013). The creation of depositional overdeepenings involves the build-up of proglacial, ice-contact outwash fans or outwash heads as a glacier snout progressively overrides its own sandur (e.g. Evans, Ewertowski, et al., 2019;Thompson & Jones, 1986 Although pressurized subglacial drainage pathways entering the overdeepening from up-glacier may be forced to rise the adverse slope and thereby supercool (e.g. Larson et al., 2010;Roberts et al., 2002), englacial tunnels producing eskers in downwasting glaciers have been seen to operate at a level within the snout that is controlled by the relatively higher base level created by fan aggradation (e.g. englacial eskers emerging at Kviarjokull; Bennett & Evans, 2012;Phillips et al., 2017;Spedding & Evans, 2002). As emergence fountains are regularly observed around and on south Icelandic glacier snouts (e.g. Tweed et al., 2018), it is clear that subglacial meltwater is also rising through the snout ice to the higher levels of ice-marginal outwash fans and indeed may prograde fans over shallow snouts if the emergence occurs some distance up-snout (Gustavson & Boothroyd, 1987;Price, 1969). Additionally, the process of inwash from lateral meltwater steams and/or englacial tunnels, observed in both of the Fjallsjökull-Hrútárjökull overdeepenings, can result in more extensive burying of the snout, giving rise to collapsing fan surfaces and emerging eskers during later melt-out behind the outwash head (Evans & Twigg, 2002;Price, 1969;Storrar et al., 2015;Figures 19 and 20).
The process of inwash burying a shallowing snout was captured by Price (1971) when an ice-dammed lake along the west margin of  (Evans & Twigg, 2002;Price, 1969;Storrar et al., 2015), reveals not only the existence of englacial to subglacial drainage networks beneath outwash fans but also documents the supraglacial nature rather than jökulhlaup origins of extensively pitted or hummocky outwash often classified as kame and kettle topography.
All of these circumstances result in eskers grading to fan apexes (e.g. Bennett & Evans, 2012;Price, 1969;Storrar et al., 2015) but the operation of tunnels feeding proglacial fans becomes more difficult once snout downwasting has resulted in the ice surface falling below the fan apex, thereby lowering the base level of the glaciohydraulic system below the immediate surrounding topography. What has been observed in such situations is the formation of ice-margin parallel esker formation along the outwash head cliff in tunnels draining towards the nearest area of lower topography (Boulton, 1986;Storrar et al., 2015Storrar et al., , 2020. In situations where meltwater cannot drain effectively from behind the outwash head to lower topography, proglacial lakes will form (Bennett & Evans, 2012;Bennett et al., 2010;Evans, Ewertowski, et al., 2019;Figures 19 and 20), giving rise to the draping of glacilacustrine sediments over, and inter-connecting with, icecontact glacifluvial assemblages and buried glacier ice.
The recent amount and rate of buried ice loss by melt-out on the south Fjallsjökull-Hrútárjökull foreland, as calculated using DoDs out at a much slower rate (cumulative volume loss increasing from $100 000 m 3 in 2015/2016 to $300 000 m 3 in 2022), but has nevertheless gradually exposed the former outwash head, with an average rate of buried ice loss of $21 000 m 3 per year over the last 3 years. This quantity and rate of buried ice melting is reflected also in the hummocky terrain/medial moraine area on the Fitjaöldur remnant. Also related to buried ice melt-out is the volume change of the 1990s thrust moraine, within which volume loss has increased at a similar rate, but with smaller totals, from $50 000 m 3 in 2015/2016 to $120 000 m 3 in 2022.
In contrast to the large volumes of buried ice loss over the Hrútárjökull overdeepening, landscape change has been more modest and volume change has stabilized in the esker network and adverse slope of the Fjallsjökull overdeepening since around 2016/2017. This no doubt reflects the significantly greater supraglacial debris cover and glacifluvial snout burial that has characterized the Hrútárjökull snout since 1945, which has retarded ablation and thereby delayed glacier ice melt. The higher rates of volume change on the inner part of the Fjallsjökull overdeepening since 2014 relate to the later date of ice recession from this area, amounting to buried ice removal at similar rates to those in the Hrútárjökull collapsed sandur and hummocky terrain/medial moraine, the latter being partly related to supraglacial melting of Fjallsjökull at its suture point with Hrútárjökull.
Each of the individual study area rates and totals for volume change reported here reflect the variability in paraglacial landscape response to deglaciation (cf. Ballantyne, 2002;Ballantyne & Benn, 1994;Bennett & Evans, 2012;Schomacker & Kjaer, 2007, wherein the debris-charged snout of Hrútárjökull has resulted in the retarded ablation that drives incremental stagnation (sensu Bennett & Evans, 2012;Eyles, 1983). This was exacerbated by the extensive burial of the downwasting Hrútárjökull snout by the sandur fan that emanated from the suture zone between the glacier snouts when they were coalescent. The large volume of outwash in this sandur was at least partially due to jökulhlaup drainage from the AErfjallsl on icedammed lake between the late 1930s until the mid-20th century and/or englacial tunnels, as well as proglacial outwash in the burial of large areas of debris-covered, downwasting glacier snouts in the production of extensive sandur collapse, is also recorded in these models; the Fjallsjökull-Hrútárjökull case study demonstrates that such depositional contexts are inextricably linked to incremental stagnation.
Moreover, there is clear evidence inherent within these models that drainage tunnel and esker development in particular is taking place englacially rather than subglacially, so that large volumes of meltwater drainage appear to be bypassing the deepest parts of depositional overdeepenings during advanced stages of glacier downwasting. This accumulation of significant volumes of glacifluvial sediment within and over glacier snouts, actively downwasting into depositional overdeepenings, is producing an outwash head landsystem signature that is diagnostic of active temperate glacier response to rapid climate change.