EVALUACIÓN DE TOXICIDAD EN SISTEMAS TERRESTRES 36

a key role in the ablation of debris-covered glaciers (Benn et al., 2012; Immerzeel et al., 2014a; Pellicciotti et al., 2015; Steiner et al., 2015; Buri et al., 2016b; Thompson et al., 2016). Understanding of key processes occurring in supraglacial ponds has advanced conceptually to include conduit-collapse formation (Kirkbride, 1993; Sakai et al., 2000), subaqueous and waterline melting (Sakai et al., 2000; Röhl, 2006), calving (Benn et al., 2001; Sakai et al., 2009), and englacial filling and drainage (Gulley and Benn, 2007). Most process observations have been made on individual features (Benn et al., 2001; Röhl, 2008; Xin et al., 2011).

However, few studies have attempted to quantify the energy exchanges associated with supraglacial ponds, or their effects on glacier ablation. Xin et al. (2011) identified two kinetic types of melt: (1) winds may force currents to drive thermo-erosion and notch development near the lake surface and (2) free convection, due to the density/temperature relationship of water, may drive pond circulation and therefore promote melt along the entire water/ice interface. Sakai et al. (2000) and Röhl (2008) each adapted empirical relationships from iceberg melt observations to examine subaqueous and waterline melting of ice cliffs, while Luthje and Pedersen (2006) adapted a method based on free convection to study basal melting of ponds on the Greenland ice sheet. To date, no effort has been made to compare these algorithms or to apply a melt model based on physical principles to supraglacial ponds on debris-covered glaciers. Perhaps more importantly, the ablation effect of such ponds at the glacier scale is reported for few locations, where it is poorly constrained (Sakai et al., 2000; Röhl, 2008).

Based on the current understanding of pond-related processes and associated mass loss, research is needed to:

• revisit the Sakai et al. (2000) modelling approach with modern instrumentation and numerical modelling approaches

• compare the subaqueous melt modelled using the Sakai et al. (2000); Röhl (2008); Luthje and Pedersen (2006) models and investigate a free-convection algorithm

• assess the surface and subaqueous energy balances of several ponds

• determine the total surface energy balance of all ponds across several glaciers to determine their influence ablation at the glacier scale

2.4

Site Description: Langtang Valley, Nepal

The Langtang Valley was selected as a study site based on accessibility, environmental data availability, potential for collaboration, and its heritage of previous scientific studies. Located

60 km north of Kathmandu and just three days’ walk from a roadhead, Langtang Valley is exceptionally accessible for a glacierised catchment in the HKKH (Figure 1). The upper Langtang basin borders the Tibetan Autonomous Republic, China to the north and measures

350 km2 (Figure 2.6a and b). Elevation ranges from 3650 m.a.s.l. at Langtang village to

7234 m.a.s.l. at the peak of Langtang Lirung; they are located only 4.5 km apart, highlighting the extremely steep topography in the basin. The basin has a mean slope of 23° as calculated from a 90 m DEM.

Local climate is primarily influenced by the South Asian monsoon, with the majority of precipitation occurring concurrently with the warmest temperatures, from mid-June to September (Immerzeel et al., 2011). This period also contains the highest daily maximum and minimum temperatures experienced at the site, and consequently the highest rates of ablation occur during these months (Ragettli et al., 2015). Occasional precipitation events occur in the post-monsoon (October to November) and in the much colder winter (December to February). The pre-monsoon (March to mid-June) is characterised by rising temperatures, which are responsible for melting much of the annual snowpack deposited during the post-monsoon and winter months; occasional precipitation events also occur during the pre-monsoon (Collier and Immerzeel, 2015). The rugged topography drives spatial meteorological variability through relief, illumination, and wind channelisation.

Air temperature has been thoroughly studied throughout Langtang Valley, both through

in-situobservation (Immerzeel et al., 2014b; Heynen et al., 2016) and numerical modelling of

atmospheric dynamics (Collier and Immerzeel, 2015), with seasonal lapse rates established over a 3-year record (Heynen et al., 2016). On-glacier lapse rates for debris-covered glaciers are often much steeper than for non-glacier terrain (Brock et al., 2010; Shaw et al., 2016; Steiner and Pellicciotti, 2016), and a network of temperature-loggers was installed on Lirung Glacier to monitor temperature across the glacier, which identified local and general seasonal temperature lapse rates along the debris surface (Steiner and Pellicciotti, 2016).

28% of the glacier area in Langtang Valley is mantled by heterogeneous rock debris, primarily covering the tongues of five valley glaciers (Figure 2.6a). The debris-covered glacier tongues are characterised by extremely variable surface topography, with large depressions occasionally filled by ponds or punctuated abruptly by bare-ice cliffs. The debris mantle varies in thickness from 0.1 to at least 2.5 m (Ragettli et al., 2015), composed of grains ranging in size from sand to large boulders. Lirung Glacier has been the site of numerous field studies of supraglacial ponds (Sakai et al., 2000; Bhatt et al., 2007; Takeuchi et al., 2012; Miles et al., 2016) in spite of its small size and advanced decay (Immerzeel et al., 2014a). The much larger Langtang, Langshisa, and Shalbachum Glaciers have strongly

2.4 Site Description: Langtang Valley, Nepal 31

Fig. 2.6 (a) Geographic context of the study area. (b) The study area, showing the upper Langtang basin’s principal debris-covered glaciers: 1-Lirung, 2-Shalbachum, 3-Langshisa, 4-Ghanna, 5-Langtang. Backdrop is 6S-corrected Landsat TM false-colour composite from 16 June 2009.

negative surface mass balances and show significant supraglacial ponded areas (Pellicciotti et al., 2015).

The five study glaciers sharply differ in size, debris cover and hypsometry. In terms

of size, they range from 1.3 km2 (Ghanna) to 52.8 km2 (Langtang), with debris mantling

22-40% of total glacier area. The glaciers also vary in their altitudinal extents, with terminus elevations ranging from 4025 m.a.s.l. (Lirung) to 4718 m.a.s.l. (Ghanna). All five glaciers are rapidly losing mass in response to climate change. Ghanna Glacier is retreating from its terminal moraines, with Lirung and Langshisa Glaciers also retreating to a lesser degree, while Shalbachum and Langtang are downwasting with a nearly stable terminus (Ragettli et al., 2016). Field observations have noted the pronounced disconnect between the debris- covered tongue and clean-ice upper portion for Lirung Glacier, a process which has recently been noted for Shalbachum Glacier as well. Langshisa and Langtang Glacier have both lost connectivity with minor tributaries since the 1970s (Pellicciotti et al., 2015; Ragettli et al., 2016).

Of the five glaciers, Lirung Glacier has received the most scientific attention. The glacier

is the most accessible in the valley, and its relatively small debris-covered area (1.03 km2)

has been conducted sporadically at Lirung Glacier for several decades, including repeat surveys (Aoki and Asahi, 1998; Immerzeel et al., 2014a), and investigation of pheonomena such as ice cliffs (Sakai et al., 1998, 2002; Steiner et al., 2015; Buri et al., 2016b) and supraglacial ponds (Sakai et al., 2000; Takeuchi et al., 2012).

Fig. 2.7 A view of the debris-covered surface of the terminus area of Lirung Glacier. Ice cliffs, hummocky terrain, and the outwash plain and proglacial lake are typical features for debris-covered glaciers, as is the vegetation growing on less dynamically-active parts of the debris surface.

The glacier’s recent retreat in response to climate change has been documented by these studies. Downwasting led to development of a terminal lake in the 1990s which has a naturally-incised outlet and poses minimal hazard. The glacier’s terminus quickly retreated from contact with this lake and continues to retreat at a more gradual rate (Ragettli et al., 2016). More recently, the debris-covered tongue disconnected from its source icefalls, and the icefall termini have experienced pronounced retreat in recent years.

Over the past several years, an extensive set of glaciological and meteorological observa- tions has been collected at Lirung Glacier and its surroundings (e.g. Immerzeel et al., 2014b; Shea et al., 2015b; Steiner et al., 2015; Kraaijenbrink et al., 2016; Steiner and Pellicciotti, 2016). These datasets provide an excellent glaciological context for the study of pond-related processes at this site.

Chapter 3

Spatial and temporal variability of

supraglacial ponds in the Langtang

Valley

3.1

Executive Summary

Supraglacial ponds play a key role in transferring atmospheric energy to the ice of debris- covered glaciers, partially overcoming the ablation-reducing effects of debris cover, but the spatial and temporal distribution of these features is largely unknown, so their overall effect is poorly constrained. 172 Landsat TM/ETM+ scenes are analysed covering the period 1999-2013 to identify thawed supraglacial ponds for the debris-covered tongues of five glaciers in the Langtang Valley of Nepal. I apply an advanced atmospheric correction routine (LandCor/6S) and improve upon previous band-ratio and image morphological techniques to identify ponds, then characterise their spatial, seasonal, and interannual patterns of ponding. Pond cover exhibits high variability between glaciers (May-October means of 0.06-1.69% of debris-covered glacier area), with ponds most frequent in zones of low surface gradient and velocity. The ponds show a pronounced seasonality, appearing rapidly in the pre-monsoon as snow melts, peaking in cover in the monsoon at ~2% of debris-covered area, then declining in the post-monsoon as ponds drain or freeze. Ponds at the study site are highly recurrent and persistent, with 40.5% of pond locations apparent in multiple years. For the whole investigation period, Langtang Glacier shows an increase in April-October total pond cover, while all glaciers show a decline in August-October pond cover for 2009-2013.