Recent High-Arctic glacial sediment redistribution: A process perspective using airborne lidar

Geomorphology 125 (2011) 27–39

Contents lists available at ScienceDirect

Geomorphology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / g e o m o r p h

Recent High-Arctic glacial sediment redistribution: A process perspective using airborne lidar T.D.L. Irvine-Fynn a,⁎, N.E. Barrand b,1, P.R. Porter c, A.J. Hodson a, T. Murray d a

Department of Geography, University of Sheffield, Winter Street, Sheffield, S10 2TN, UK Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, Alberta, T6G 2E3, Canada Division of Geography and Environmental Sciences, University of Hertfordshire, College Lane, Hatfield, AL10 9AB, UK d Department of Geography, University of Wales-Swansea, Singleton Park, Swansea, SA2 8PP, UK b c

a r t i c l e

i n f o

Article history: Received 19 April 2010 Received in revised form 20 August 2010 Accepted 23 August 2010 Available online 6 September 2010 Keywords: Arctic Lidar Sediment flux Moraine degradation Terrain relaxation processes

a b s t r a c t Progressive glacier thinning, retreat and mass loss in the High-Arctic is increasingly exposing forefield sediments to processes of mobilisation and redistribution. In this paper, we quantify forefield sediment redistribution at Midtre Lovénbreen, Svalbard, using repeat light detection and ranging (lidar) surveys conducted in 2003 and 2005 in combination with field-based observations. Average surface lowering of the forefield over the observation period identified from lidar surveys is −0.05 ma−1; and two primary areas of sediment reworking are identified: active fluvial incision of proglacial streams by ~ 2 m and lateral moraine downwasting of similar magnitude. Multivariate analysis of fluvial and climatological field data indicates that observed forefield sediment mobilisation is driven primarily by discharge forcing, but with contributions from thermoerosive processes and stochastic, autogenic sediment supply. During the period of observation, disparity between sediment loss in forefield fluvial systems as calculated from lidar data (3000– 4000 × 103 kg) and monitoring of fluvial sediment load (1600–3500 × 103 kg) suggests the likely presence of significant quantities of buried ice beneath a thick debris mantle, as evidenced by field observations. Relatively uniform lowering of the moraine crest identified from our repeat lidar surveys indicates thermoerosion of an ice core. However, simple debris layer thickness modelling indicates an increase in variation of debris layer thickness at lower elevations, providing support for the assertion that moraine disintegration is driven by complex combinations of both thermal and mechanical processes. This study demonstrates the viability of using lidar in conjunction with field monitoring to better understand sedimentary deglaciation dynamics and processes, and also highlights the significance of forefield areas in controlling the sediment yield from deglaciating catchments. © 2010 Elsevier B.V. All rights reserved.

1. Introduction The Arctic has been shown to be particularly sensitive to climatic forcing (ACIA, 2004), and contemporary observational evidence indicates that a warming trend at the high latitudes has dominated regional climate during the last century (e.g., Serreze et al., 2000). The High-Arctic archipelago of Svalbard represents one of the most significant northern hemisphere repositories of glacier ice outside Greenland, covering an area of ~ 36,000 km2 (Hagen et al., 1993). Research has generally accepted that glaciers in Svalbard reached their maximum Holocene extent at the end of the “Little Ice Age” (LIA), early in the twentieth century (Werner, 1993; Svendsen and Mangerud, 1997); and recent geodetic analyses have shown glaciers across much of Svalbard have exhibited negative mass balances, ⁎ Corresponding author. Tel.: + 44 114 222 7900. E-mail address: t.irvine-fynn@sheffield.ac.uk (T.D.L. Irvine-Fynn). 1 Present address: British Antarctic Survey, High Cross, Madingley Road, Cambridge, CB3 0ED, UK. 0169-555X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.geomorph.2010.08.012

equating to an annual average of −0.30 m water-equivalent (w.e.) since 1936 (Nuth et al., 2007). However, Kohler et al. (2007) also demonstrated that glacier thinning rates have accelerated throughout the last seven decades in western Svalbard, with values reaching up to −0.69 m w.e. per year between 2003 and 2005. These mass balance trends broadly reflect the region's climatic variations over the same time frame (e.g., Førland et al., 1997; Hanssen-Bauer and Førland, 1998; Nordli and Kohler, 2003). The accelerated thinning of many of Svalbard's glaciers is particularly noted over ablation areas and is readily coupled to coincident terminus retreat; this reflects the stagnation and downwasting of many of the region's smaller glaciers and affirms suggestions that areas in the archipelago evidence a latter stage of transformation from a glacierised to partially glaciated landscape (Etzelmüller, 2000; Lyså and Lønne, 2001; Sletten et al., 2001; Ziaja, 2001, 2005). The net result of recent progressive ablation and deglaciation at lower elevations in Svalbard is the increasing exposure of forefield, morainic sediments, and readily destabilised diamictons susceptible to processes of reworking and redistribution (Bennett et al., 2000;

28

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

Etzelmüller et al., 2000; Laffy and Mercier, 2002; Lukas et al., 2005; Mercier and Laffly, 2005). Such changes result in modification of hydrological pathways and biogeochemical and ecological processes (e.g., Moreau et al., 2008; Rutter et al., submitted for publication); and they also have the potential to increase sediment availability, as reflected in model predictions of sediment delivery to Arctic oceans and water bodies (e.g., Syvitski, 2002). The close coupling between periglacial and fluvioglacial processes gives rise to both considerable complexity in the sedimentary record and a lack of well-preserved glacial landforms in response to reworking and the degradation of ice cores present within many of the geomorphic landforms associated with deglaciation (Hambrey, 1984; Etzelmüller, 2000; Etzelmüller and Hagen, 2005; Lukas et al., 2005). Furthermore, complex reworking of glaciogenic sediments is likely to affect the morphology and distribution of sedimentary facies (Bennett et al., 2000; Sletten et al., 2001; Lønne and Lyså, 2005; Lukas et al., 2005). These adjustments to the sedimentary record have, therefore, generated much debate regarding palaeoglacial sedimentary units and their use in environmental reconstruction (e.g., Dyke and Savelle, 2000; Woodward et al., 2002; Lukas, 2005, 2007; Graham et al., 2007). Within this context, it is critical to better constrain the rates of reworking and redistribution of morainic sediments for a better appreciation as to, firstly, the validity of utilising High-Arctic observations as contemporary analogues to Pleistocene/Quaternary processes and, secondly, understanding the expected timeframes of post-glacial terrain relaxation. While conceptual understanding of ice-marginal and proglacial terrain degradation has proliferated (Johnson, 1971; Szponar, 1975; Hambrey, 1984; Mattson and Gardner, 1991; Lyså and Lønne, 2001; Sletten et al., 2001; Lønne and Lyså, 2005), quantification of the rates of downwasting remains limited in the High-Arctic (Etzelmüller, 2000; Lukas et al., 2005; Schomacker and Kjær, 2008) when compared to other latitudes (e.g., McKenzie, 1969; Driscoll, 1980; Johnson, 1992; Krüger and Kjær, 2000; Everest and Bradwell, 2003; Schomacker and Kjær, 2007). Schomacker (2008) presented a thorough review of the degradation processes and rates associated with ice-cored glacial features over a range of latitudes. In this paper, we present results from the forefield of Midtre Lovénbreen, Svalbard, where repeat airborne lidar surveys were conducted in 2003 and 2005, with coincident field-based observations during 2004 and 2005, allowing detailed exploration of the terrain relaxation mechanisms and process rates at an unprecedented spatial scale and resolution. This enables an innovative demonstration and quantification of the complex progression of landscape change in a deglaciating Arctic catchment. 2. Midtre Lovénbreen Midtre Lovénbreen is a 5-km2 glacier situated in a north-facing catchment on Brøggerhalvøya, northwest Svalbard (78°50′ N., 12° E.; Fig. 1). The catchment's geology includes schists, quartzite, phyllite, and some sedimentary rocks (Hjelle, 1993) and is underlain by permafrost of ~ 150 m in depth at sea level (Liestøl, 1976). The glacier itself spans an elevation range of 50 to 650 m above sea level (masl) (Björnsson et al., 1996), but has been retreating since the close of the LIA (Lefauconnier et al., 1999; Hambrey et al., 2005) and has shown increasing rates of thinning (Kohler et al., 2007). Since 1968, the Norwegian Polar Institute has monitored the glacier's mass balance that has averaged −0.36 m w.e. (Kohler et al., 2003; Hodson et al., 2005). The associated retreat has exposed a poorly consolidated forefield currently totalling an area of ~2 km2. Moraine complexes, braid plains, angular debris, and multifarious diamictons have been related to englacial, subglacial, and supraglacial processes and their associated dynamical changes during the recent retreat and are described in full elsewhere (e.g. Hambrey et al., 1999; Glasser and Hambrey, 2001; Hansen, 2003; Hambrey et al., 2005). However,

further complexity is apparent because proglacial channel mobility and outlet migration have led to the reworking of glaciogenic sediment throughout the recent period of glacial retreat (Mercier and Laffly, 2005). Currently, ablation season runoff from Midtre Lovénbreen is routed to Kongsfjorden via two principal streams (MLE and MLW, hereafter) draining from the lateral margins of the glacier (Fig. 1). Because the glacier is polythermal (Björnsson et al., 1996; Rippin et al., 2003), subglacial drainage is sourced from the temperate ice zone near the glacier bed in the accumulation and upper ablation areas. These subglacial waters typically emerge in the early part of the ablation season as either singular or multiple upwellings close to the glacier snout (Rippin et al., 2003; Hodson et al., 2005; Irvine-Fynn et al., 2005); further, the glacier's internal drainage system is evidenced by the annual formation of proglacial icing(s) in the glacier forefield during winter. Around 30% of the summer runoff from Midtre Lovénbreen is routed through a subglacial drainage system (Hodson et al., 2005; Irvine-Fynn, 2008); however, subtle changes in position of both upwellings and icings between differing years are thought to suggest variations in the glacier's active hydrological system (Rippin et al., 2003; Hodson et al., 2005). Importantly, annual observations since 1994 (A. J. Hodson, University of Sheffield, unpublished data) show that in any given year turbid subglacial waters appear to drain either via MLE or MLW, but not simultaneously in both. During 2003 and 2004, the subglacial upwelling was located on the western side of the glacier snout, and subglacial (sediment-rich) waters passed solely down MLW. However, a unique combination of circumstances during summer 2004 prompted the research presented here: firstly, on 30th July (DOY210) MLE exhibited significant channel reorganisation from a well-established reach (perennially active since 1998) to courses that stabilised by the end of 2004 and were reoccupied in 2005 (see Fig. 1), and, secondly, during the 2004 ablation season significant degradation of the lateral moraine structure to the north-west of Midtre Lovénbreen's main tongue was observed. These observations, coincident with repeated airborne lidar surveys of the glacier and forefield during 2003 and 2005, enabled examination of glaciofluvial and thermal erosion rates and mechanisms. 3. Research methods 3.1. Airborne lidar survey and processing Previously, airborne lidar (light detection and ranging) data have been utilised to produce high resolution measurements for a variety of environmental applications including glacier volume change (e.g., Geist et al., 2005; Hopkinson and Demuth, 2006; Kohler et al., 2007), fluvial geomorphology (e.g., Charlton et al., 2003), geomorphological mapping (e.g., Smith et al., 2006) and landslide detection (e.g., McKean and Roering, 2004). Differentially processed aircraft trajectories from airborne and ground-based GPS are combined in postprocessing with aircraft inertial navigation system and laser ranging data to derive accurate and precise topographic data. These individual data points may be interpolated into continuous terrain surfaces or digital elevation models (DEMs). Lidar offers a number of advantages over traditional photogrammetric methods of airborne topographic data collection (e.g., Finsterwalder, 1954; Etzelmüller et al., 1993; Baltsavias et al., 2001; Schiefer and Gilbert, 2007; Hugenholtz et al., 2008). These include subdecimeter measurement accuracy (Krabill et al., 2002), an entirely digital processing stream enabling rapid collection of dense datasets, surveys that do not require ground control point data, and the ability to operate irrespective of ambient light sources (i.e., at night or in areas of shadow). In this study, lidar data were acquired using an Optech ALTM3033 laser scanner recording first and last return and laser intensity information on two survey flights over Midtre Lovénbreen and

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

29

Fig. 1. Map showing location of the Brøggerhalvøya site (inset) and the Midtre Lovénbreen forefield, illustrating the LIA moraine limit (dashed line), key drainage routes (passing monitoring stations MLE and MLW) observed in 2003 (solid lines) and 2005 (dotted lines). The western moraine ridge crest referred to later in the paper and in Fig. 5 is indicated with a dash–dot line, and is measured from its northern-most extent. Contours are shown at 10-masl intervals, and the position of the glacier AWS is indicated. Locations A–D, designated with a “b” indicate the approximate standpoint (apex) and viewshed of images presented as Figs. 4A, 4B, 7A and 7B, respectively.

surrounds on 9th August 2003 (DOY 221) and 5th July 2005 (DOY 186). The survey aircraft was flown at a ground speed of ~77 ms−1 and an average height of 1600 masl. The instrument operated at a scanning rate of 13 Hz, which provides 13 complete bidirectional scans per second at a maximum scanning angle of 18° off nadir. This corresponds to an average scan width of 783 m, point sample density of 1 per 1.83 m2, and along- and across-track point spacing of 1.38 and 1.33 m calculated for flat terrain at the mean equilibrium line altitude of the glacier (~ 400 masl). While data density varies according to flying height, topography, and the presence of swath overlaps (where data have approximately twice the spatial resolution of non overlapping points), resolution of the entire data set is calculated as 1.15 points per m2 and 1.10 per m2 over the forefield. Additional details of the lidar survey are found in Barrand (2007). Digital elevation models at 1-m horizontal resolution were processed for the entire data sets from raw lidar points using an adapted Delauney triangulation gridding algorithm. Elevation residuals calculated between lidar DEM cell values and 3242 independently surveyed differential GPS check data points give a root mean square (RMS) elevation difference of ±0.14 m for check points collected on the glacier surface and forefield either side of survey days (Barrand et al., 2009). The

spatial resolution and elevation accuracy of these lidar-derived DEMs is significantly greater than those derived from digitised contour maps or photogrammetry used in previous studies of proglacial morphological change (e.g., Etzelmüller, 2000; Schiefer and Gilbert, 2007; Hugenholtz et al., 2008; Schomacker and Kjær, 2008). 3.2. Field data and processing Hydrological data from the Midtre Lovénbreen catchment were collected over a 55-day period during the summer of 2004, corresponding to ~70% of the year's hydrologically active ablation season. Gauging stations were established on the two principal proglacial streams in stable, confined reaches where the streams pass through the LIA end moraine ~1 km from the glacier snout (Fig. 1). Continuous, hourly averaged records of water stage and stream turbidity were collated using Druck pressure transducers and Partech IR15C turbidity probes, respectively, following standard methods (e.g., Hodson and Ferguson, 1999). Stage-discharge relationships were ascertained from discrete discharge (Q) measurements using velocity–area and salt dilution techniques. Similarly, suspended sediment concentration (SSC) was estimated from the rating curve

30

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

describing water turbidity and sediment concentration (g/L) derived from discrete, manual ~ 500 mL gulp samples filtered through Whatman 0.45 μm cellulose nitrate membranes. Short-term instrumental failure required occasional missing data to be estimated: linear leastsquares regression between MLE and MLW was used to approximate Q data, while SSC values were estimated independently using nonlinear trigonometric regression assuming a diurnal cycle (λ = 24 h) such that: 2

SSCi = a + bðiÞ + Aði Þ + Bsin½2πði−f Þ = λ

ð1Þ

where SSCi is the value to be estimated at time i; and values of a, b, A, B, and f are coefficients determined in the regression procedure. Probabilistically determined errors in Q and SSC were ±12–19% and up to ±26%, respectively, which are dominantly those related to the forecasting procedure associated with the use of rating curves, as reported in similar studies (e.g., Hodson and Ferguson, 1999; Hodgkins et al., 2003). Air temperature, incident short-wave radiation, and net all-wave radiation data were collected from an automatic weather station (AWS) positioned ~150 masl on Midtre Lovénbreen's centre line (Fig. 1). As with the stream data, meteorological data on the glacier were recorded as hourly averages (with maximum uncertainties of ± 0.35 °C). Additional air temperature data were obtained from an AWS in NyÅlesund (~3 km north-west of the glacier) with a recording interval of 6 h. 4. Analytical results 4.1. Lidar image analysis To assess morphometric change over time, a raster difference layer was created by subtracting the 2003 and 2005 DEMs. However, in order to improve our estimates of net terrain (sediment) change within the Midtre Lovénbreen catchment, exclusion of areas of glacier, icing/naled ice, standing water, and snow patches from both elevation models was desirable. Further, additional processing was therefore applied to the DEMs. The entire area for each lidar survey was cropped to remove the ice surface and surrounding mountains, leaving only the glacier forefield. We delineated the forefield on the eastern side of the glacier using LIA glacial maximum moraine extents, on the western side by the extent of a large lateral moraine structure, and used the shore of Kongsfjorden as the northern limit (Fig. 1). Patches of standing water, icing, and snow from the 2003 DEM were identified using orthorectified vertical aerial photographs collected simultaneously with the lidar and controlled using three-dimensional coordinates of raw lidar points (e.g., James et al., 2006; Barrand et al., 2009). As no such imagery exists for the 2005 lidar data, we removed regions of standing water and snow using a novel histogram-based approach utilising laser pulse intensity return information. As liquid water and snow have, respectively, very low and very high laser signal reflectivity (from signal absorption and reflectance), we excluded lidar elevation data points with the lowest and highest 5% of intensity return values (based on a 256-bit brightness scale). Comparison with time-coincident field photography showed that this method was successful in removing regions of standing water and snow from the 2005 DEM. The same method applied to the 2003 data yielded no additional areas to those identified using the orthorectified images. A spatial plot of the difference between the processed 2003 and 2005 DEMs for the Midtre Lovénbreen forefield, excluding the areas of standing water and icing, is shown in Fig. 2. Several areas of the DEM difference model showed positive elevation changes up to ~ 2 m; these changes were most notable for the deserted, former course of MLE and a patch within the sandur fed by MLE (Fig. 2). The two areas indicate the importance of fluvial processes where deposition occurs when sediment transport efficacy decreases because of changing

gradient, braiding or reductions in discharge. Conversely, visible regions of similar magnitude surface lowering, indicative of debris mobilisation were found along the western moraine and in the area of the new eastern proglacial stream course at the glacier margin, running north toward the LIA moraine limit. With observations suggesting sediment redistribution occurring within the MLE stream channels and along the western moraines, considering these areas independently from the DEMs was necessary in order to, firstly, gain an improved estimate for forefield surface lowering rates and, secondly, to determine rates of fluvial erosion and ice-cored moraine degradation. The two proglacial stream reaches were delineated in order to examine the magnitude of volume changes along their courses. The shift in course of the east river draining the glacier allowed the delineation, from the 1-m resolution DEMs, of both the relic channel (active in 2003, eastward, see Fig. 1) and the current channel course (active in 2005, westward, see Fig. 1). Both courses were defined by a 5 m buffer either side of the channel centre from the glacier snout up to their confluence point proximate to the LIA moraine limit. Additionally, we identified an area of sandur beyond the LIA moraines to examine the magnitude of volume change from aggradation. Table 1 presents the morphometric change results for the individual areas of interest, as well as the forefield as a whole (which excludes the areas of most marked change highlighted above). Volume change errors were calculated as the root sum of squares (RSS) of the ±0.14 m errors inherent to each DEM surface (see Barrand et al., 2009). The given rates of change assume the changes that occurred were distributed over the full 23 months that elapsed between lidar surveys; we note, because of the brevity of the melt season in Arctic latitudes and the timing of the surveys, these rates may be underestimates. The total spatially averaged degradation rate of the forefield as a whole is −0.05 (±0.2) ma−1. The most notable region of downwasting was the western moraine (−0.7 (±0.2) ma−1); such change is likely indicative of thermoerosion of an ice core. The data presented in Table 1 show that within the fluvial system, despite deposition in the sandur, there is net export of sediment from the forefield that may be available for coastal progradation (Mercier and Laffly, 2005). 5. Process modelling and identification To examine the details of the processes involved in the observed reworking of Midtre Lovénbreen's forefield, we focus on two key areas of change clear in the lidar imagery: firstly, the proglacial fluvial reach of MLE incised in 2004, and, secondly, the western lateral moraine feature. 5.1. Adjustment of fluvial topography With the continuous time-series of both Q and SSC (Fig. 3), it was possible to estimate the suspended load resulting from the channel change that occurred in 2004. This was facilitated by the fact the subglacial upwelling occurred in MLW, ensuring that the sediment transported in MLE is principally of proglacial origin. This was further corroborated by field observations revealing negligible sediment transport by supraglacial streams feeding MLE. Table 2 summarises the hydrological data for 2004 and provides a comparison to identical data collected in 1998 (A. J. Hodson, University of Sheffield, unpublished data), when stable drainage through MLE was also unaffected by subglacial discharge contributions. During 1998, higher Q did not equate to elevated SSC levels; while in contrast, during 2004, SSC was heightened even with a lower mean Q in pre- and postchannel change time-series. To identify the likely processes of sediment mobilisation along MLE during 2004, we adopt the multivariate modelling approach of Irvine-Fynn et al. (2005). To objectively identify hydrometeorological subseasons, principal components analysis (PCA) was applied to

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

437000

437500

438000

438500

8761500

8761500

8762000

436500

8762000

436000

31

8760500

8760500

8761000

8761000

(ii)

(iv)

(iii)

Difference (m) >2m

8760000

8760000

(i)

2-1m 0.5 - 0 m 0 - –0.5m

8759500

8759500

1 - 0.5 m

–0.5 - –1 m

8759000

–2 - –4 m < –4 m 436000

436500

437000

437500

438000

8759000

–1 - –2 m

438500

Fig. 2. Map of DEM difference for the forefield extent. Note the key regions of interest: (i) fluvial deposition in the deserted MLE stream course, (ii) the zone of deposition in the proglacial sandur, (iii) the degrading western lateral moraine, and (iv) fluvial incision apparent in the reach of MLE that opened up during 2004. Note that areas of glacier ice, residual naled, and standing water have been removed from the difference raster.

decompose time-series of Q, SSC, forefield air temperature (Tf), and incident and net radiation (IR and NR, respectively) into the first “seasonal” component (PC1) that explained 49% of the variance. The scores from PC1 were then subject to a change point analysis (CPA) using Change Point Anlayzer 2.3. To avoid the numerical problems resulting from autocorrelation readily apparent in diurnally cyclical

Table 1 Table of lidar-determined volumes of morphometric change; elevation change rates have been calculated assuming the 23 months between lidar surveys. Highlighted area (Fig. 2)

i ii iii iv

Location

Area (km2)

Elevation change rate (ma−1)

Volume change (×105 m3)

Net forefield (excluding ponds, snow, icing etc) Eastern proglacial stream (relict) Sandur Western moraine ridge Eastern proglacial stream (active)

4.66

−0.05

± 0.2

−4.59

± 9.32

0.01

+ 0.01

± 0.2

+ 0.01

± 0.001

0.09 0.17 0.04

+ 0.06 −0.65 −0.13

± 0.2 ± 0.2 ± 0.2

+ 0.05 −2.19 −0.02

± 0.19 ± 0.34 ± 0.09

data (Fig. 3), it was necessary to group the hourly PC1 scores into two 12-h periods per day (0600–1700 and 1800–0500) and use 100,000 bootstraps within the analysis for heightened confidence and repeatability in the detected change points. Having defined 10 subseasons within the 55-day monitoring period, Q and SSC data were log-transformed to achieve normality; and bivariate regression between the two variables showed a range in the bivariate coefficient of determination (r2) values between 0.01 and 0.86 (see Table 3). Sediment supply processes not directly linked to Q forcing can be identified by incorporating additional proxy variables into a model of SSC (Hodson and Ferguson, 1999). Table 4 gives full details of the multivariate predictors used in the analyses here. Because the subseasonal lags between Q and SSC identified by cross correlation were b2 h, time lags were ignored. Three groups of models for SSC were examined: (i) purely hydrological; (ii) hydrological and energy related; and (iii) hydrology, energy and residual SSC. A stepwise regression was used to develop the models of SSC, and Table 3 gives details of the most significant predictor variables for each. In all cases, the use of additional predictors improved the coefficients of determination (R2) such that, for the best multivariate models, R2 ranged between 0.73 and 0.98 (Table 3).

32

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

Fig. 3. Time-series of air temperature at the glacier AWS (~150 masl) through the 2004 ablation season and proglacial discharge (Q) and suspended sediment concentration (SSC) recorded in MLE where the stream passes through the LIA moraine ridge. Note the distinct change in temporal patterns in SSC following the inception of channel change on DOY209. The hydrometeorological subseasons identified for modelling purposes using the PCA-CPA method are indicated using dashed lines.

For MLE, the results of the multivariate modelling show that for 5 of the 10 subseasons, sediment transport was by direct mobilisation through near-instantaneous forcing. Prior to the initiation of MLE's channel change on DOY 210 from the course occupied in previous years (Fig. 1), direct discharge forcing (logQ) dominated subseasons 1 and 3, autogenic processes of sediment supply (R) such as bank collapse were weak, while thermoerosion (Q2) and increasing sediment supply area (sQ+) were apparent in subseason 2. Following the channel change that was partially prompted by a combination of meteorological conditions and increased discharge, the importance of autogenic sediment release appeared heightened, with R2 increasing by up to 0.5 by the inclusion of R. Nonetheless, predictor variables logQ, Q2, Tf, and sQ+ still remained significant throughout the latter subseasons. Such results strongly indicate that sediment mobilisation during the period following channel initiation was driven by a combination of discharge forcing and thermoerosion, which would enlarge the sediment contributing area and/or indirectly enable stochastic processes such as bank collapse or channel migration. However, despite the clear fluvial and periglacial processes invoked in sediment mobilisation and redistribution in the forefield, a discrepancy arises: that of sediment volumes. The lidar data showed a volume of ~ 2000 m3 was eroded from the new reach of MLE; and assuming a typical moraine density of between 1.5 and 2.0 g/cm3, this equates to 3000–4000 (±18,000) × 103 kg. Contrastingly, the total SSC load was only 1100 (±370) × 103 kg over the monitoring period. This value was adjusted to 1600 (± 570) × 103 kg by assuming observations occurred over only 70% of the active hydrological season, thereby crudely accounting for close-of-season transport and commencement of non-subglacial runoff in 2005. Further, the suspended load only represents a fraction of the total load: the bedload component of the total sediment yield for glacial rivers has been quoted from b30% to N75% throughout a wide range of environments (Pearce et al., 2003). Here, bedload was estimated by assuming a value equivalent to 60% of the total stream load as reported by Gurnell et al. (1988) for comparably sized glaciers with broadly similar discharge

Table 2 Summary statistics of hydrological variables with comparison to similar data collected for 1998 when runoff through MLE did not include turbid subglacial waters. Data

3

Q (m /s) SSC (g/L) Yield (kg/day)

1998

2004

Mean (σ)

Mean (σ)

0.62 (0.24) 0.13 (0.10) 7660 (4520)

0.40 (0.24) 0.41 (0.58) 21,100 (3380)

2004 means for pre/post-channel change (DOY210) 0.46/0.38 0.18/0.47 8020/24,700

ranges, although in a temperate alpine setting: the total load passing MLE during the period of observations was estimated to be 2560 (±912) × 103 kg. This represents only 64–85% (±22%) of the volume indicated by the lidar. Such a disparity suggests that the value for moraine density may be overestimated, implying a volume of buried ice exists within the forefield perhaps below a sediment mantle of N1 m. Field observations support this (Fig. 4), and such buried ice has been found in the forefields of many proximate glaciers on Brøggerhalvøya (Hambrey, 1984; Hoelzle, 1993) and elsewhere in Svalbard (Kozarski, 1982; Gibas et al., 2005). The ongoing thermoerosion of this buried ice would ensure sediment availability in MLE was indeed governed by Q and thermal processes, as was found with the presented models of SSC. Interestingly, the volume of deposition within the sandur (Table 1) suggested sediment mobilised within the LIA moraine limit was subsequently redeposited. This indicates the rates of coastal progradation reported by Mercier and Laffly (2005) may be more strongly influenced by sediment exported from Midtre Lovénbreen's subglacial system than solely from subaerial fluvial redistribution. 5.2. Adjustment of moraine topography A total of 10,120 points confined to the lateral moraine were used to extract elevation values from the two DEMs, including along the ridge line, cross profiles, and randomly generated points throughout the area of interest. As the plots of the ridge centre line from north to south, several cross profiles, and a histogram of surface elevation change (Figs. 5 and 6A) show, a relatively uniform lowering was found over the entire moraine feature. Based on our sample data, the mean elevation change of the moraine was −1.17 m between 2003 and 2005: a rate of −0.56 ma−1. This latter value compares well to the annual, area-averaged glacier ice ablation estimates for Midtre Lovénbreen of b0.45 m w.e. for 1997–2002 (Hodson et al., 2005) and 0.51 m w.e. for 2003–2005 (Barrand et al., 2009). However, using the 2003–2005 lidar difference model, Barrand et al. (2009) showed that ice ablation at the glacier snout was between 2.5 and 5 m, equivalent to elevation changes of the order of −1.3 to −2.6 ma−1. Comparison of the time-coincident elevation changes indicate that downwasting of the moraine feature is less than that of the glacier ice, highlighting the role of debris as an insulator. Above 90 masl, the elevation change (dZ) does show a weak altitude dependence, indicating downwasting may be temperature dependent. The clear, marked departures from the general uniformity of surface lowering shown in Fig. 5 could be readily linked to active processes of erosion evidenced by field observations, which confirmed the moraine was ice-cored. At a location ~300 m along the

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

33

Table 3 Summary of numerical modelling of SSC in MLE using linear, hydrology (H), energy-hydrology (EH), and energy-hydrology-residual-SSC (EHR) regression models (RMs).a Subseason

1 2 3 4 5 6 7 8 9 10

EH-MRMs

ΔR2

logSSC = f(logQ)

H-MRMs

r2

R2

P1

P2

P3

R2

P1

P2

P3

R2

P1

P2

P3

0.86 0.53 0.41 0.43 0.01 0.56 0.14 0.23 0.62 0.57

0.91 0.83 0.61 0.49 0.46 0.93 0.61 0.49 0.79 0.60

logQ Q2 logQ logQ sQ logQ sQ logQ logQ logQ

dQ sQ− sQ− sQ Q2 sQ− Q2 sQ sQ sQ−

hQ dQ hQ

0.95 0.87 0.61 0.49 0.47 0.90 0.61 0.58 0.79 0.60

logQ Q2 logQ logQ Tf logQ sQ logQ logQ logQ

IR sQ− sQ− sQ

hQ IR hQ

IR Q2 sQ sQ sQ−

Tf−

0.96 0.91 0.81 0.98 0.74 0.93 0.80 0.79 0.86 0.94

logQ Q2 logQ R Tf logQ sQ R logQ logQ

IR sQ R logQ R Q2 R logQ sQ R

R R sQ− sQ dQ hQ Q2 sQ R sQ−

logQ− hQ dQ

EHRn − 1-MRMs

IR

0.01 0.03 0.20 0.49 0.27 0.03 0.20 0.22 0.06 0.34

a Key multivariate predictor variables in order of descending importance are shown (P1, P2, P3), while coefficients of determination (r2 and R2 for bivarate and multivariate models, respectively) are given. For interpretation of proxy variables (e.g., logQ) see Table 4. Where SSC appeared inversely (negatively) related to a variable, a “−” appears next to the predictor (e.g., sQ−). The influence of inclusion of residual SSC (R) are shown by change in coefficient of determination value (ΔR2).

moraine crest, a very significant area showed elevation changes in excess of −3 m; these were taken to reflect backwasting from the western moraine margin potentially associated with the standing water present (Figs. 2 and 7A). At the position 650–670 m along the moraine crest (~90 masl), the small lateral stream fed by a remnant hanging glacier to the west of the moraine was found to breach the moraine ridge during 2004, cutting down by up to ~10 m to the local level of the proglacial river (Fig. 7B). At 790–890 m (~ 120 masl) along the moraine crest, mass failures were observed, likely resulting in more significant changes in elevation as underlying ice was then exposed to further ice melt (Fig. 7B). These large-scale changes are reflected in the standard deviation of elevation change, which otherwise is remarkably consistent over the lateral moraine's full elevation range (Fig. 6A). The predominantly uniform change in surface elevation implies decay of the ice core within the moraine feature. To examine the more detailed processes of such thermal erosion, we use a simple approach to model the debris thickness following Driscoll (1980), Nakawo and Young (1982), and Nicholson and Benn (2006). The subglacial melt rate (M in m/s) may be expressed as M = Q = ρi L f

ð2Þ

for which ρi and Lf are, respectively, the density of glacier ice (900 kg/m3 at Midtre Lovénbreen (J. Kohler, Norsk Polarinstitutt, personal commu-

Table 4 Summary of proxy multivariate predictors and interpretations used in modelling SSC (for full details see Hodson and Ferguson, 1999; Irvine-Fynn et al., 2005). Predictor Justification logQ dQ sQ hQ

Q2

Tf

IR R

Log-transformed Q, allows for direct relationship between sediment mobilisation and river water flux. Rate of change in Q hour-to-hour, allows for hysteretic effects such as the dynamics in Q forces sediment entrainment. Cumulative sum of Q, reflecting long-term increase (if positive coefficient) or exhaustion (if negative coefficient) of sediment sources. Time since current Q was last equalled or exceeded, suggesting periodic mobilisation as heightened Q accesses previously unavailable sediment sources. Square of Q interpreted to allow for small changes in Q resulting in large changes in SSC, potentially occurring when new sediment sources are made accessible or when thermoerosion of frozen sediments occurs. Air temperature in the forefield, corrected from the glacier AWS using a seasonal mean lapse rate. Relates to thaw related sediment mobilisation when coefficient is positive. Incident radiation, reflects direct energy receipt resulting in thaw or meltout releasing sediment. Previous hours suspended sediment concentration (SSCn − 1) effectively accounting for autocorrelation within the SSC series, implying autogenetic sediment release (i.e., stochastic events elevating SSC over short- to medium-term timeframes).

nication, 2006)) and the latent heat of fusion (334,000 J/kg); Q represents the downward flux of energy, which using a crude simplification of a linear temperature gradient (k) between the upper and lower surfaces of the debris layer, can be estimated by Q = kðTm −Ti Þ = hd

ð3Þ

where Tm and Ti, respectively, represent temperatures (in K) at the air– moraine and debris–ice interfaces; and hd represents the debris thickness. Because Ti was assumed to be constant at 0 °C, Eqs. (2) and (3) were combined and simplified to estimate hd using temperature, in °C, at the moraine surface (Tms): hd = kTms = Mρi Lf

ð4Þ

An estimated value of k = 1.23 W/m/K was based on observations of debris thickness, temperatures, and melt rates at five locations on Larsbreen, Svalbard (L. Nicholson, Geographie Innsbruck, personal communication, 2007). The Larsbreen k-value was considered appropriate for Midtre Lovénbreen because of the broad similarity in characteristics between the two sites. The values for M at each of 10,120 random sample points were calculated from the vertical elevation difference between the 2003 and 2005 lidar surveys. Mean daily air temperatures (Ta) were estimated for each point by using the 6-hourly temperature records from Ny-Ålesund and the mean diurnal temperature lapse rate of −0.0068 °C/m as derived from an 89-day temperature time-series collated for the summer of 2004 for both AWSs (correlation between the two data series was significant, r = 0.89). The mean ablation season temperature at each point was calculated for the time span between lidar surveys: here, the ablation seasons were bounded by the period of consistently positive (N0 °C) air temperatures in NyÅlesund, but limited by the exclusion of the first 15 days, which represents the typical time taken for the elimination of the snowpack in the local area (Bruland et al., 2001). Finally, a correction factor of 80% was then used to adjust the mean ablation season air temperature (at 2 m) to the debris surface temperature (Tms) in following with Driscoll (1980). Estimates of hd were then calculated using Eq. (4). Importantly, this simple model of debris thickness is only valid for 0.0015 b hd b 1.5 m. (Østrem, 1959; Mattson et al., 1993). The lower limit is defined as the debris thickness resulting in maximum melt (Østrem, 1959; Mattson et al., 1993), below which ablation is reduced from increased surface albedo. However, the upper limit is adjusted to account for permafrost, as underlying ice will be protected from melt where debris cover is greater than the active layer depth, estimated to be between 0.7 and 1.5 m in the region (Hallet and Prestrud, 1986; J. Kohler, Norsk Polarinstitutt, personal communication, 2007). Whilst more complex models for ice ablation

34

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

A

1m

B

1m

Fig. 4. Field observations of buried ice within the forefield, near the current glacier terminus (A) as well as at more distal locations (B). Refer to Fig. 1 for image locations. Note the bedding within sediments overlying the ice mass, indicative of numerous reworking events leading to the burial and subsequent preservation of the ice feature. Vertical scales are approximate.

beneath debris accounting for nonlinear temperature gradients and heat storage within the debris layer have been presented (e.g., Han et al., 2006; Nicholson and Benn, 2006), as a first approximation for determination of hd (given the limited data available for the moraine) the above model was considered adequate. Of the sample points, those showing positive (increased) elevation change were omitted from modelling, as were points where hd was either b0.015 m or N1.5 m: a total of b6% of the sample points were thus discounted. The results of the simple model are presented in Fig. 6B, which shows the histogram of hd over the moraine and includes the standard deviation of each variable for each elevation bin. Broadly, the data suggest an increase in sediment mantle thickness as ice-cored moraine degradation nears completion, as found elsewhere (e.g., Small, 1987; Krüger and Kjær, 2000). The increased variance in hd at the lower elevations, given the reduced variability at higher

elevations, is suggestive that a variety of processes not directly related to the thermal degradation of the ice core govern the reworking of morainic debris here. The high variability in hd at the uppermost elevations along the moraine ridge likely reflects initial moraine stabilisation following meltout from the glacier. 6. Discussion and implications Schomacker (2008) asserted that lidar holds considerable potential for the quantification of ice-cored terrain relaxation rates at high resolution and over large areas. Here, we have presented the first analysis of proglacial forefield degradation using sequential DEMs derived from airborne lidar surveys in the High-Arctic. The vertical (±0.14 m) and horizontal (1 m) resolution afforded by the method is far greater than that provided using more traditional cartographic or

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

Elevation

A

105 103 101 99 97 95

0

10

20

155

235

153

233

151

231

149

229

147

227

145

30

225

1000 1010 1020 1030

35

1750 1760 1770 1780

Distance along moraine crest (m)

Elevation

B

255 235 215 195 175 155 135 115 95

i ii Fig 7B

Fig 7A

2005 iv 0

500

Elevation

1000

2000

1500

Distance along moraine crest (m)

North

C

2003

iii

255 235 215 195 175 155 135 115 95

South

i ii iii iv

-60

-40

-20

0

20

40

60

Distance from moraine crest (m) Fig. 5. Profiles illustrating the broadly uniform surface elevation change across the western moraine ridge at Midtre Lovénbreen as revealed from the two lidar DEMs for (A) locations at ~ 15, 1015, and 1760 m along the ridge crest from north to south (refer to Fig. 1); (B) the entire ridge crest, highlighting morphological changes at positions ~ 350, 680, and 725 m along the crest line as discussed in the text and with reference to Fig. 7; and (C) cross profiles at positions i–iv as indicated on profile B.

photogrammetric techniques (e.g., Etzelmüller et al., 1993; Etzelmüller, 2000; Schiefer and Gilbert, 2007; Schomacker and Kjær, 2008). Moreover, as shown, the quality of the data sets readily enables the coupling of remotely sensed data to field-based observations to provide

dZ (m)

A

hd (m)

B

2 1.5 1 0.5 0 -0.5 -1 -1.5 -2 -2.5 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

50

70

90

110

130

150

170

190

210

230

170

190

210

230

Z (masl)

50

70

90

110

130

150

Z (masl) Fig. 6. Plots of average (A) surface elevation change (dZ) across the entire Midtre Lovénbreen moraine ridge, and (B) modelled debris thickness (hd) estimated from melt beneath a debris layer. Data are summarised using 20-m elevation bins, with the x-axis showing median elevation. The standard deviations (σ) for each binned data set are shown as diamonds on both plots.

detailed examination of the processes underlying morphometric change. Comparisons drawn between the lidar DEMs difference model and field observations revealed two primary areas of change in the forefield area of Midtre Lovénbreen: highly active fluvial incision of ~ 2 m and downwasting of a lateral moraine with a similar magnitude of change. In consideration of the first, detailed modelling using multivariate analyses of proglacial hydrological data indicated that the processes of sediment mobilisation were driven primarily by discharge forcing, with evidence for thermoerosional contributions and stochastic, autogenetic sediment supply. Similar results regarding sediment mobilisation have been found from analysis of other glaciofluvial data sets from Arctic glaciers lacking a subglacial meltwater component (e.g., Gurnell et al., 1994; Hodson and Ferguson, 1999; Irvine-Fynn et al., 2005) and thus highlights the coupling between periglacial and fluvial processes involved in proglacial sediment redistribution (e.g., Etzelmüller, 2000; Etzelmüller et al., 2000). The reworking of the glacier forefield is strongly linked to runoff and thermoerosional processes, as has been documented in previous studies (see IrvineFynn et al., 2005; Mercier and Laffly, 2005). However, the volumetric results indicate ice masses are present within the forefield, evidencing significant thicknesses of overlying debris. Here, the origin, volume, and form of the ice in front of Midtre Lovénbreen remains unclear: relict glacier ice may be of Holocene age, potentially draped with more modern sediments following the LIA advance; or the ice may be of naled or aufeis origin, simply buried beneath fluvially reworked sediments within the last century since the termination of the LIA. In terms of sediment budget, the sandur linked to MLE showed a gain of material, implying that currently there is no supply limitation as relative

36

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

A

B

Fig. 7. Images of the western moraine ridge at Midtre Lovénbreen, illustrating (A) backwasting along the western moraine margin apparently relating to fluvial activity and standing water and (B) fluvial incision and mass failure of the sedimentary mantle, both exposing the moraine's ice core. These images relate to the large differences identified in Fig. 5 and image positions are indicated on Fig. 1, at locations C and D, respectively.

fluxes from slopes and moraine complexes remain high (cf. Beylich et al., 2009) and despite the presence of buried ice in front of Midtre Lovénbreen. In consideration of the moraine feature to the west, there was evidence for a broad similarity between High-Arctic and lower latitude sites in terms of degradation rates for ice-cored moraine features (Table 5; Schomacker, 2008). The relatively uniform lowering across the morphologic feature strongly implicates thermally driven downwasting of underlying buried ice. This process could be better understood using thermal imagery (e.g., Lougeay, 1974) coupled with lidar surveys (e.g., Hopkinson et al., in press) which reveal spatial variations in thermoerosion that may be more closely linked to field observations and models of sediment thickness and redistribution. However, as indicated by Schomacker (2008), moraine ridge centre-

line plots reveal that degradation of ice-cored landforms are not simply a function of thermal erosion. The presence of some standing water to the west of the moraine (Fig. 6A) indicates that the downwasting may also be accelerated through basal melt and backwasting (e.g., Pickard, 1983). Moreover, modelling of the debris thickness showed an increase in the variation in debris cover at the lowest elevations, indicative of the increased prevalence of nonthermal mass movements. Such findings reaffirm the supposition that moraine degradation is driven by a complex combination of melt and mechanical processes (Driscoll, 1980). The field observations made at Midtre Lovénbreen of site-specific, large-scale moraine degradation support early suggestions highlighting the importance of mass failure of the moraine mantle and subsequent melting of exposed ice (Johnson, 1971) and of fluvial action (Watson,

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39 Table 5 Summary of some published ranges of vertical ice-cored moraine degradation rates. Vertical moraine degradation ratea (ma−1)

Source reference

b 4.8 0.048–0.08 0.1 0.003–0.3 b5 1.0 0.4–4.8 0.05 0.1–0.4 0.3–2.5 0.4–3.5 1.1 0.9 0.7

Østrem (1965) McKenzie (1969) Ross (1976) Driscoll (1980) Mattson and Gardner (1991) Johnson (1992) Mattson et al. (1993) King and Volk (1994)* Etzelmüller (2000)* Krüger and Kjær (2000) Lukas et al. (2005)* Hugenholtz et al. (2008)+ Schomacker and Kjær (2008)* this study *

a Where only daily rates are given, figures have been adjusted assuming an 80-day, snow-free melt season is applicable to all sites. Sites in Svalbard are indicated with * while + indicates a rate for moraine creep, which may not necessarily involve an ice core.

1980) in determining the downwasting of ice-cored topography. Thus, although topography may play a role in controlling the rates of ice-cored landscape degradation (Schomacker, 2008), the processes of sediment redistribution and deglaciated landscape relaxation are strongly conditioned and controlled by fluvial processes (Lukas et al., 2005) and, more significantly, perhaps by fluvial magnitude–frequency events (Etienne et al., 2008). Such events may provide some control of sediment mobilisation and yield from the fluvial–periglacial system of deglaciating basins and the significant moraine stores (Beylich and Gintz, 2004; Etienne et al., 2008), which can represent up to 60% of sediments stored within alpine-style, deglaciating catchments (e.g., Otto et al., 2009). Combined, the findings here emphasise the significance of the nonglacial area for control of a deglaciating catchment's sediment yield (e.g., O'Farrell et al., 2009) and infer caution for interpretations of moraines as analogues or indicators of palaeoenvironments, landforming processes, or climate. Notably, these arguments do invoke a degree of speculation, particularly given the brief 2year time period between lidar surveys; however, Barrand et al. (2010) report the mass balance of Midtre Lovénbreen during 2003–2005 to be −0.51 ± 0.02 m w.e. a−1, while the longer-term average (1966–2005) was only slightly less negative (−0.41 ± 0.03 m w.e.a−1) suggesting the rates of change over the forefield during the past 50 years might be expected to be similar to those reported here, lending support to the longer-term perspective of sediment redistribution processes.

7. Conclusion In summary, through the analyses presented for short-term morphological change in the forefield of Midtre Lovénbreen between 2003 and 2005, we demonstrate the viability of using lidar for the detailed modelling of surface and near-surface processes in glacial environments. Furthermore, the results clearly show the need to couple remote sensing methods with in situ field studies to aid in a more complete understanding of processes. Identification of periglacial controls on the sediment redistribution across the forefield reaffirms the importance of considering a complete sediment cascade model from slope to glacier to forefield in understanding changes within deglaciating environments. If Midtre Lovénbreen serves as a site representative of the retreat/thinning rates characterising small glaciers over western Svalbard, this study establishes the importance of determining volumes of subsurface buried ice that may significantly influence predictions of sediment yields from deglaciating High-Arctic catchments, particularly in light of recent glacier recession and the forecast changes in glacier extent under anticipated climatic forcing.

37

Acknowledgements Lidar data were collected by the UK Natural Environment Research Council (NERC) Airborne Research and Survey Facility (ARSF) and its analysis funded through NERC grant NE/B505203/1 (SLICES, P.I.: TM) and NERC studentship NER/S/A/2003/11279 awarded to NEB. GPS base-stations and associated training for the lidar campaign were provided by the NERC Geophysical Equipment Facility. NEB and TM acknowledge the help of Applied Imagery for the use of QT Modeler software. TIF thanks the University of Sheffield (Dept. of Geography Research Support Grant), Gino Watkins Memorial Fund, Dudley Stamp Memorial Trust, and NERC grant NE/G006253/1; AJH acknowledges financial support from the Geological Society, WG Fearnsides Fund, Earth and Space Foundation and the University of Sheffield Knowledge Transfer Fund. In Ny-Ålesund, fieldwork support was provided by Nick Cox and Steve Marshall (NERC Arctic Research Station), Anita Asadullah, Brian Barrett, Edward Hanna, and Fiona Hunter. The comments from Anders Schomacker and an anonymous reviewer are appreciated and the scientific editor, Richard A. Marston, is greatly thanked for his patience and helpful inputs.

References Arctic Climate Impact Assessment (ACIA), 2004. Impacts of a Warming Arctic: Arctic Climate Impact Assessment. Cambridge University Press, Cambridge, UK. 144 pp. Baltsavias, E., Favey, E., Bauder, A., Bosch, H., Pateraki, M., 2001. Digital surface modelling by airborne laser scanning and digital photogrammetry for glacier monitoring. Photogrammetric Record 17 (98), 243–273. Barrand, N.E., 2007. Changes in volume and extent of NW Svalbard glaciers using airborne lidar and digital photogrammetry. Ph.D. Thesis, Dept. of Geography, Swansea University, Swansea, Wales. Barrand, N.E., Murray, T., James, T.D., Barr, S.L., Mills, J.P., 2009. Optimizing photogrammetric DEMs for glacier volume change assessment using laserscanning derived ground control points. Journal of Glaciology 55 (189), 103–116. Barrand, N.E., James, T.D., Murray, T., 2010. Spatiotemporal variability in elevation changes of two high-Arctic valley glaciers. Journal of Glaciology 56 (199), 771–780. Bennett, M.R., Huddart, D., Glasser, N.F., Hambrey, M.J., 2000. Resedimentation of debris on an ice-cored lateral moraine in the high-Arctic (Kongsvegen, Svalbard). Geomorphology 35, 21–40. Beylich, A.A., Gintz, D., 2004. Effects of high-magnitude/low-frequency fluvial events generated by intense snowmelt or heavy rainfall in Arctic periglacial environments in northern Swedish Lapland and northern Siberia. Geografiska Annaler 86A (1), 11–29. Beylich, A.A., Laute, K., Liermann, S., Hansen, L., Burki, V., Vatne, G., Fredin, O., Gintz, D., Berthling, I., 2009. Subrecent sediment dynamics and sediment budget of the braided sandur system at Sandane, Erdalen (Nordfjord, western Norway). Norsk Geografisk Tidsskrift 63 (2), 123–131. Björnsson, H., Gjessing, Y., Hamran, S.-E., Hagen, J.O., Liestøl, O., Pálsson, F., Erlingsson, B., 1996. The thermal regime of sub-polar glaciers mapped by multi-frequency radioecho sounding. Journal of Glaciology 42 (140), 23–32. Bruland, O., Marechal, D., Sand, K., Killingtveit, A., 2001. Energy and water balance studies of a snow cover during snowmelt period at a high arctic site. Theoretical and Applied Climatology 70 (1–4), 53–63. Charlton, M.E., Large, A.R.G., Fuller, I.C., 2003. Application of airborne lidar in river environments: the River Coquet, Northumberland, UK. Earth Surface Processes and Landforms 28, 299–306. Driscoll, F.G., 1980. Wastage of the Klutlan ice-cored moraines, Yukon Territory, Canada. Quaternary Research 14, 31–49. Dyke, A.S., Savelle, J.M., 2000. Major end moraines of Younger Dryas age on Wollaston Peninsula, Victoria Island, Canadian Arctic: implications for palaeoclimate and for formation of hummocky moriane. Canadian Journal of Earth Science 37 (4), 601–619. Etienne, S., Mercier, D., Voldoire, O., 2008. Temporal scales and deglaciation rhythms in a polar glacier margin, Baronbreen, Svalbard. Norsk Geografisk Tidsskrift 62, 102–114. Etzelmüller, B., 2000. Quantification of thermo-erosion in proglacial areas — examples from Svalbard. Zeitschrift für Geomorphologie 44, 343–361. Etzelmüller, B., Hagen, J.O., 2005. Glacier-permafrost interaction in Arctic and alpine mountain environments with examples from southern Norway and Svalbard. In: Harris, C., Murton, J. (Eds.), Cryospheric Systems: Glaciers and Permafrost. Geological Society Special Publication, London, pp. 11–27. Etzelmüller, B., Vatne, G., Ødegård, R.S., Sollid, J.L., 1993. Mass balance changes of surface slope, crevasse and flow pattern of Erikbreen, northern Spitsbergen: an application of a geographical information system. Polar Research 12, 131–146. Etzelmüller, B., Ødegård, R.S., Vatne, G., Mysterud, R.S., Tonning, T., Sollid, J.L., 2000. Glacial characteristics and sediment transfer system of Longyearbreen and Larsbreen, western Spitsbergen. Norsk Geografisk Tidsskrift 54, 157–168. Everest, J., Bradwell, T., 2003. Buried glacier ice in southern Iceland and its wider significance. Geomorphology 52, 347–358.

38

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39

Finsterwalder, R., 1954. Photogrammetry and glacier research with special reference to glacier retreat in the eastern Alps. Journal of Glaciology 2 (15), 306–315. Førland, E.J., Hanssen-Bauer, I., Nordli, P.Ø., 1997. Climate Statistics and Longterm Series of Temperature and Precipitation at Svalbard and Jan Mayen. DNMI Klima Report 21/97, Oslo. Geist, T., Elvehøy, H., Jackson, M., Stötter, J., 2005. Investigations on intra-annual elevation changes using multi-temporal airborne laser scanning data: case study Engabreen, Norway. Annals of Glaciology 42, 195–201. Gibas, J., Rachlewicz, G., Szczucinski, W., 2005. Application of DC resistivity soundings and geomorphological surveys in studies of modern Arctic glacier marginal zones, Petuniabukta, Spitsbergen. Polish Polar Research 26 (4), 239–258. Glasser, N.F., Hambrey, M.J., 2001. Styles of sedimentation beneath Svalbard valley glaciers under changing dynamic and thermal regimes. Journal of the Geological Society 158, 697–707. Graham, D.J., Bennett, M.R., Glasser, N.F., Hambrey, M.J., Huddart, D., Midgley, N.G., 2007. ‘A test of the englacial thrusting hypothesis of “hummocky” moraine formation: case studies from the northwest Highlands, Scotland’: comments. Boreas 36 (1), 103–107. Gurnell, A.M., Warburton, J., Clark, M.J., 1988. A comparison of the sediment yield characteristics of two adjacent glacier basins, Val d'Herens, Switzerland. IAHS Publication 174, 431–441. Gurnell, A.M., Hodson, A., Clark, M.J., Bogen, J., Hagen, J.O., Tranter, M., 1994. Water and sediment discharge from glacier basins: an arctic and alpine comparison. IAHS Publication 224, 325–334. Hagen, J.O., Liestøl, O., Roland, E., Jørgensen, T., 1993. Glacier Atlas of Svalbard and Jan Mayen. Meddelelser, 129. Norsk Polarinstitutt, Olso. Hallet, B., Prestrud, S., 1986. Dynamics of periglacial sorted circles in western Spitsbergen. Quaternary Research 26, 81–99. Hambrey, M.J., 1984. Sedimentary processes and buried ice phenomena in the proglacial areas of Spitsbergen glaciers. Journal of Glaciology 30, 116–119. Hambrey, M.J., Bennett, M.R., Dowdeswell, J.A., Glasser, N.F., Huddart, D., 1999. Debris entrainment and transfer in polythermal valley glaciers. Journal of Glaciology 45 (149), 69–86. Hambrey, M.J., Murray, T., Glasser, N.F., Hubbard, A., Hubbard, B., Stuart, G., Hansen, S., Kohler, J., 2005. Structure and changing dynamics of a polythermal valley glacier on a centennial timescale: Midre Lovénbreen, Svalbard. Journal of Geophysical Research 110 (F01006). doi:10.1029/2004JF000128 19 pp. Han, H., Ding, Y., Liu, S., 2006. A simple model to estimate ice ablation under a thick debris layer. Journal of Glaciology 52 (179), 528–536. Hansen, S., 2003. From surge-type to non-surge-type glacier behaviour: midre Lovénbreen, Svalbard. Annals of Glaciology 36, 97–102. Hanssen-Bauer, I., Førland, E.J., 1998. Long-term trends in precipitation and temperature in the Norwegian Arctic: can they be explained by changes in atmospheric circulation patterns? Climate Research 10 (2), 143–153. Hjelle, A., 1993. Geology of Svalbard. Norsk Polarinstitutt Handbook, 7. Hodgkins, R., Cooper, R., Wadham, J., Tranter, M., 2003. Suspended sediment fluxes in a high-Arctic glacierised catchment: implications for fluvial sediment storage. Sedimentary Geology 162 (1–2), 105–117. Hodson, A.J., Ferguson, R.I., 1999. Fluvial suspended sediment transport from cold and warm-based glaciers in Svalbard. Earth Surface Processes and Landforms 24 (11), 957–974. Hodson, A., Kohler, J., Brinkhaus, M., 2005. Multi-year water and surface energy budget of a high-latitude polythermal glacier: evidence for overwinter water storage in a dynamic subglacial reservoir. Annals of Glaciology 42, 42–46. Hoelzle, M., 1993. DC Resistivity Soundings 1992 in North-western Svalbard Arbeitsheft Nr. 13, Versuchsanstalt für Wasserbau. ETH, Zürich, Switzerland. 37 pp. Hopkinson, C., Demuth, M.N., 2006. Using airborne lidar to assess the influence of glacier downwasting on water resources in the Canadian Rocky Mountains. Canadian Journal of Remote Sensing 32 (2), 212–222. Hopkinson, C., Barlow, J., Demuth, M., Pomeroy, J., in press. Mapping changing temperature patterns over a glacial moraine using oblique thermal imagery and LiDAR. Canadian Journal of Remote Sensing. Hugenholtz, C.H., Moorman, B.J., Barlow, J., Wainstein, P.A., 2008. Large-scale moraine deformation at the Athabaska Glacier, Jasper National Park, Alberta, Canada. Landslides 5 (3), 251–260. doi:10.1007/s10346-008-0116-5. Irvine-Fynn, T.D.L., 2008. Modelling runoff from the maritime Arctic cryosphere: water storage and routing at Midtre Lovénbreen. Ph.D. Thesis, Dept. of Geography, University of Sheffield, Sheffield, UK, 384 pp. Irvine-Fynn, T.D.L., Moorman, B.J., Sjogren, D.B., Walter, F.S.A., Willis, I.C., Hodson, A.J., Williams, J.L.M., Mumford, P.N., 2005. Cryological processes implied in Arctic proglacial stream sediment dynamics using principal components analysis and regression. In: Harris, C., Murton, J. (Eds.), Cryospheric Systems: Glaciers and Permafrost. Geological Society Special Publication, London, pp. 83–98. James, T.D., Murray, T., Barrand, N.E., Barr, S.L., 2006. Extracting photogrammetric ground control from LiDAR DEMs for change detection. The Photogrammetric Record 21 (116), 312–328. Johnson, P.G., 1971. Ice cored moraine formation and degradation, Donjek Glacier, Yukon Territory, Canada. Geografiska Annaler 53A (3–4), 198–202. Johnson, P.G., 1992. Stagnant glacier ice, St. Elias Mountains, Yukon. Geografiska Annaler 74A (1), 13–19. King, L., Volk, M., 1994. Glaziologie und glazialgeomorphologie des Liefde- und Bockfjordgebieties NW-Spitzbergen. Zeitschrift für Geomorphologie Supplement 97, 145–160. Kohler, J., Nordli, P.Ø., Brandt, O., Isaksson, E., Pohjola, V., Martma, T., Aas, H.F., 2003. Svalbard Temperature and Precipitation, Late 19th Century to the Present, ACIA Final Report.

Kohler, J., James, T.D., Murray, T., Nuth, C., Brandt, O., Barrand, N.E., Aas, H.F., Luckman, A., 2007. Acceleration in thinning rate on western Svalbard glaciers. Geophysical Research Letters 34 (L18502). doi:10.1029/2007GL030681 5 pp. Kozarski, S., 1982. The genetic variety of ice cores in the marginal forms of some Spitsbergen glaciers, Hornsund region. Acta Universitatis Wratislaviensis 525, 153–164. Krabill, W., Abdalati, W., Frederick, E.B., Manizade, S., Martin, C.F., Sontag, J., Swift, R.N., Thomas, R.H., Wright, W., Yungel, J., 2002. Aircraft laser altimetry measurement of elevation changes of the Greenland ice sheet: technique and accuracy assessment. Journal of Geodynamics 34 (3–4), 357–376. Krüger, J., Kjær, K.H., 2000. De-icing progression of ice-cored moraines in a humid, subpolar climate, Kötlujökull, Iceland. The Holocene 10 (6), 737–747. Laffy, D., Mercier, D., 2002. Global change and paraglacial morphodynamic modification in Svalbard. International Journal of Remote Sensing 23 (21), 4743–4760. Lefauconnier, B., Hagen, J.O., Ørbæk, J.B., Melvold, K., Isaksson, E., 1999. Glacier balance trends in the Kongsfjorden area, western Spitsbergen, Svalbard, in relation to the climate. Polar Research 18 (2), 307–313. Liestøl, O., 1976. Pingos, springs and permafrost in Spitsbergen. Norsk Polarinstitutt Årbok 1975, 7–29. Lønne, I., Lyså, A., 2005. Deglaciation dynamics following the Little Ice Age on Svalbard: implications for shaping of landscapes at high latitudes. Geomorphology 72, 300–319. Lougeay, R., 1974. Detection of buried glacial and ground ice with thermal infrared remote sensing. In: Santeford, H.S., Smith, J.L. (Eds.), Advanced Concepts and Techniques in the Study of Snow and Ice Resources. NAS, Washington DC, pp. 487–494. Lukas, S., 2005. A test of the englacial thrusting hypothesis of ‘hummocky’ moraine formation: case studies from the north-west Highlands, Scotland. Boreas 34 (3), 287–307. Lukas, S., 2007. ‘A test of the englacial thrusting hypothesis of “hummocky” moraine formation: case studies from the northwest Highlands, Scotland’: reply to comments. Boreas 36 (1), 108–113. Lukas, S., Nicholson, L.I., Ross, F.H., Humlum, O., 2005. Formation, meltout processes and landscape alteration of High-Arctic ice-cored moraines — examples from Nordenskiöld Land, central Spitsbergen. Polar Geography 29 (3), 157–187. Lyså, A., Lønne, I., 2001. Moraine development at a small High-Arctic valley glacier: Rieperbreen, Svalbard. Journal of Quaternary Science 16 (6), 519–529. Mattson, L.E., Gardner, J.S., 1991. Mass wasting on valley-side ice-cored moraines, Boundary Glacier, Alberta, Canada. Geografiska Annaler 73A (3–4), 123–128. Mattson, L.E., Gardner, J.S., Young, G.J., 1993. Ablation on debris covered glaciers: an example from the Rakhiot Glacier, Punjab, Himalaya. IAHS Publication 218, 289–296. McKean, J., Roering, J., 2004. Objective landslide detection and surface morphology mapping using high-resolution airborne laser altimetry. Geomorphology 57, 331–351. McKenzie, G.D., 1969. Observations on a collapsing kame terrace in Glacier Bay National Monument, southeastern Alaska. Journal of Glaciology 8 (54), 413–425. Mercier, D., Laffly, D., 2005. Actual paraglacial progradation of the coastal zone in the Kongsfjorden area, western Spitsbergen (Svalbard). In: Harris, C., Murton, J. (Eds.), Cryospheric Systems: Glaciers and Permafrost. Geological Society Special Publication, London, pp. 111–117. Moreau, M., Mercier, D., Laffly, D., Roussel, E., 2008. Impacts of recent paraglacial dynamics on plant colonization: a case study on Midtre Lovénbreen forefield, Spitsbergen (79°N). Geomorphology 95, 48–60. Nakawo, M., Young, G.J., 1982. Estimate of glacier ablation under a debris layer from surface temperature and meteorological variables. Journal of Glaciology 28 (98), 29–34. Nicholson, L., Benn, D.I., 2006. Calculating ice melt beneath a debris layer using meteorological data. Journal of Glaciology 52 (178), 463–470. Nordli, P.Ø., Kohler, J., 2003. The Early 20th Century Warming: Daily Observations at Green Harbour, Grønfjorden, Spitsbergen. DNMI Klima Report 12/03. Nuth, C., Kohler, J., Aas, H.F., Brandt, O., Hagen, J.O., 2007. Glacier geometry and elevation changes on Svalbard (1936-90): a baseline dataset. Annals of Glaciology 46, 106–116. O'Farrell, C.R., Heimsath, A.M., Lawson, D.E., Jorgensen, L.M., Evenson, E.B., Larson, G., Denner, J., 2009. Quantifying periglacial erosion: insights on a glacial sediment budget, Matanuska Glacier, Alaska. Earth Surface Processes and Landforms 34 (15), 2008–2022. Østrem, G., 1959. Ice melting under a thin layer of moraine, and the existence of ice cores in moraine ridges. Geografiska Annaler 41, 228–320. Østrem, G., 1965. Problems of dating ice-cored moraines. Geografiska Annaler 47A (1), 1–38. Otto, J.-C., Schrott, L., Jaboyedoff, M., Dikau, R., 2009. Quantifying sediment storage in and high alpine valley (Turtmanntal, Switzerland). Earth Surface Processes and Landforms 34 (13), 1726–1742. Pearce, J.T., Pazzaglia, F.J., Evenson, E.B., Lawson, D.E., Alley, R.B., Germanoski, D., Denner, J.D., 2003. Bedload component of glacially discharged sediment: insights from the Matanuska Glacier, Alaska. Geology 31 (1), 7–10. Pickard, J., 1983. Surface lowering of ice-cored moraine by wandering lakes. Journal of Glaciology 29 (102), 338–342. Rippin, D., Willis, I., Arnold, N., Hodson, A., Moore, J., Kohler, J., Björnsson, H., 2003. Changes in geometry and subglacial drainage of Midre Lovenbreen, Svalbard, determined from digital elevation models. Earth Surface Processes and Landforms 28 (3), 273–298. Ross, A.B., 1976. A form and process study of a terminal ice-cored moraine. M.A. Thesis, University of Ottawa, Ottawa, 178 pp. Rutter, N., Hodson, A., Irvine-Fynn, T., Kristensen, M., submitted for publication. The hydrology and hydrochemistry of deglaciation in a high-Arctic catchment, Svalbard. Journal of Hydrology.

T.D.L. Irvine-Fynn et al. / Geomorphology 125 (2011) 27–39 Schiefer, E., Gilbert, R., 2007. Reconstructing morphometric change in a proglacial landscape using historical aerial photography and automated DEM generation. Geomorphology 88, 167–178. Schomacker, A., 2008. What controls dead-ice melting under different climate conditions? A discussion. Earth-Science Reviews 90, 103–113. Schomacker, A., Kjær, K.H., 2007. Origin and melt-out of multiple generations of icecored moraines at Brúarjökull, Iceland. Boreas 36, 411–425. Schomacker, A., Kjær, K.H., 2008. Quantification of dead-ice melting in ice-cored moraines at the high-Arctic glacier Holmströmbreen, Svalbard. Boreas 37, 211–225. Serreze, M.C., Walsh, J.E., Chapin, F.S., Osterkamp, T., Dyurgerov, M., Romanovsky, V., Oechel, W.C., Morison, J., Zhang, T., Barry, R.G., 2000. Observational evidence of recent change in the northern high-latitude environment. Climatic Change 46, 159–207. Sletten, K., Lyså, A., Lønne, I., 2001. Formation and disintegration of a high-arctic icecored moraine complex, Scott Turnerbreen, Svalbard. Boreas 30, 272–284. Small, R.J., 1987. Moraine sediment budgets. In: Gurnell, A.M., Clark, M.J. (Eds.), Glacio-Fluvial Sediment Transfer: An Alpine Perspective. Wiley and Sons, Chichester, pp. 165–198. Smith, M.J., Rose, J., Booth, S., 2006. Geomorphological mapping of glacial landforms from remotely sensed data: an evaluation of the principal data sources and an assessment of their quality. Geomorphology 76 (1–2), 148–165.

39

Svendsen, J.I., Mangerud, J., 1997. Holocene glacial and climatic variations on Spitsbergen, Svalbard. The Holocene 7 (1), 45–57. Syvitski, J.P.M., 2002. Sediment discharge variability in Arctic rivers: implications for a warmer future. Polar Research 21 (2), 323–330. Szponar, A., 1975. Ablation of ice–moraine ridges and its morphological effects, with glaciers of the Hornsund region as example. Acta Universitatis Wratislaviensis 251, 101–125. Watson, R.A., 1980. Landform development on moraines of the Klutlan Glacier, Yukon Territory, Canada. Quaternary Research 14 (1), 50–59. Werner, A., 1993. Holocene moraine chronology, Spitsbergen, Svalbard: lichenometric evidence for multiple Neoglacial advances in the Arctic. The Holocene 3, 128–137. Woodward, J., Murray, T., McCaig, A., 2002. Formation and reorientation of structure in the surge-type glacier Kongsvegen, Svalbard. Journal of Quaternary Science 17 (3), 201–209. Ziaja, W., 2001. Glacier recession in Sørkappland and central Nordenskiöldland, Sptisbergen, Svalbard, during the 20th century. Arctic Antarctic and Alpine Research 33 (1), 36–41. Ziaja, W., 2005. Response of the Nordenskiöld Land (Spitsbergen) glaciers Grumantbreen, Håbergbreen and Dryadbreen to the climate warming after the Little Ice Age. Annals of Glaciology 42, 189–194.