1
Abstract
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Little is known about the spatial distribution and nature of rock glaciers in the arid environment of the
Bolivian Andes, yet they are potentially important for future water supplies. This paper reports on the
first rock glacier inventory for the region (15° - 22° S), in which 94 rock glaciers have been identified
and mapped using remote sensing approaches through Google Earth, supported by a programme of
field validation. Of the rock glaciers identified, 57% were classified as active (containing ice) and the
remaining as relict (not containing ice). We show that the majority (87%) are orientated in a southerly
aspect (SE, S, and SW). Average rock glacier length and area was 500 m and 0.12 km2, respectively.
From our results we approximate the lower limit of permafrost to be at 4700 m in the Bolivian Andes,
with mean minimum altitude of rock glacier fronts (MAF) estimated to be around 4980 m for active
rock glaciers, and lower for relict rock glaciers. Our results are an important first step towards
assessing the spatial distribution of regional permafrost, and provide information to allow permafrostbased water resources in the Bolivian Andes to be understood against a backdrop of severe ice glacier
recession.
16
1. Introduction
17
Rock glaciers are important elements of the mountain cryosphere, yet currently very little is known
18
about their occurrence in the Bolivian Andes. They are key stores of frozen water in arid mountains
19
and they act as water reservoirs where precipitation is very low (Francou et al., 1999; Brenning, 2005;
20
Azócar & Brenning, 2010; Kellerer-Pirklbauer et al., 2012; Rangecroft et al., 2013). In the dry Andes,
21
glaciers are small and occur infrequently (Esper Angillieri, 2009) because the Equilibrium Line
22
Altitude (ELA), where glacier accumulation balances with ablation (Zemp et al., 2007), exceeds some
23
of the highest peaks (> 6000 m) (Francou et al., 1999), and precipitation is scarce. Currently, there is a
24
pressing need for information on alternative stores of water at high elevations in many regions of the
25
world (Brenning et al., 2007), such as rock glaciers, due to the projected scarcity of water if glacier
26
recession continues in response to climate change (Bradley et al., 2006). Bolivian glaciers have lost
27
roughly half of their volume in the past 60 years (Soruco et al., 2009) and with glaciers being one of
28
the main sources of water for drinking, agriculture and energy generation (Jordan, 2008; Vuille et al.,
29
2008; Chevallier et al., 2011), this impacts negatively on the population. It is estimated that the
30
glaciers of the Cordillera Real supply between 12 - 40 % of potable water for La Paz, depending on
31
seasonal variation (Vergara, 2009; Soruco, 2012). Therefore, with current and projected glacier
32
recession, this contribution to water supplies is expected to reduce (UNFCCC, 2007) and
33
subsequently the importance of rock glacier contributions is likely to increase (Schrott, 1996; Millar
34
& Westfall, 2008).
1
35
Rock glacier inventories have been conducted in various mountainous regions worldwide, with
36
the highest density in the European Alps (e.g. Guglielmin & Smiraglia, 1998; Curtaz et al., 2011;
37
Kellerer-Piklbauer et al., 2012; Krainer & Ribis, 2012; Scotti et al., 2013), and with increasing
38
research in the South American Andes (Brenning, 2005; Esper Angillieri, 2009; Falaschi et al., 2014).
39
Although known to be abundant and very well developed forms of locally important long term water
40
storage in the semi-arid Chilean Andes (Trombotto et al., 1999; Brenning, 2005), their contribution to
41
mountain water supplies is ambiguous and understudied in other parts of the Andes, such as Bolivia.
42
To mitigate the impacts of climate change it is important to understand all inputs to the mountain
43
hydrological cycle, including permafrost features such as rock glaciers.
44
This research does not delve into the debate of rock glacier genesis. Rock glaciers were identified
45
and classify by their activity status, according to their assumed ice content and flow behaviour using
46
morphological and geomorphological criteria from satellite image interpretation (Table 1) and/or
47
ground truthing. We sub-divide them as active/inactive (containing ice) or relict (no ice content)
48
(Barsch, 1996; Baroni et al., 2004). Key features for identification include: surface micro-relief such
49
as ridges and furrows representing decelerating flow (Barsch, 1996); steep well-defined lateral
50
margins; and a steep frontal snout if ice is present (Payne, 1998; Harrison et al., 2008). Frontal slopes,
51
with an angle usually greater than that of repose (~30-35°, depending on the material), indicate ice
52
within the rock glacier mass (Barsch, 1996). Active rock glaciers are known to contain 40 – 60 % ice
53
(Barsch, 1996; Brenning, 2005). Inactive rock glaciers also contain an ice core protected by sediment
54
but are no longer mobile due to melting of most of the upper layers within the frontal slope (Barsch,
55
1996; Scotti et al., 2013). Relict rock glaciers are formerly active features from which the ice has
56
melted. They are characterized by collapse structures on their surface, more subtle surface relief and
57
shallower frontal and lateral slope angles (< 30°, depending on the material) (Barsch, 1996; Scotti et
58
al., 2013) than active features, and regularly have vegetated surfaces (Scotti et al., 2013). Relict rock
59
glaciers are usually found at lower elevations than active rock glaciers (Baroni et al., 2004). Active
60
rock glaciers are typically found between the 0 °C isotherm and the snow line whilst relict rock
61
glaciers are generally developed just below the permafrost level/0 °C isotherm.
2
62
The aims of this study are to map the distribution and summarise the results of this new inventory
63
for the Bolivian Andes. It is an important first step towards addressing water resource issues in the
64
region, and by undertaking a survey of the current permafrost based water resources, their
65
hydrological contribution and importance at local, regional and national scales can be better
66
understood.
67
68
1.1 Study Area
69
Stretching over 7000 km between Columbia and Patagonia, Bolivia is situated in the central, dry part
70
of the South American Andes (Payne, 1998). Bolivia has a distinctive climate consisting of a wet
71
season (Dec - Feb) and a dry season (May - Aug) (Francou et al., 2003), with aridity increasing
72
towards the south of the country (Fig. 1). This rock glacier inventory encompasses the two Cordillera
73
mountain ranges of Bolivia between 15° S and 22° S (Fig. 1), home to 20 % of the world’s tropical
74
glaciers (Rabatel et al., 2013). The inventory divides the Bolivian Andes into three regions based on
75
location, climate and topography where rock glaciers were found to exist (Fig. 1): i) the Cordillera
76
Real; ii) Sajama; and iii) the Western Cordillera.
77
The first region, ‘Cordillera Real’, is a glaciated fold mountain range (15° - 17° S) situated
78
close to La Paz. With the wettest climate of the Bolivian Andes (Fig. 1), this region contains the
79
highest density of glaciers in the Bolivian Andes and these are currently in recession (Francou et al.,
80
2003; Vergara et al., 2007). The geology of this region forms mainly resistant rocks of the early
81
Paleozoic, producing prominent massifs. The second and third regions, ‘Sajama’ and ‘Western
82
Cordillera’, are based along the Bolivia-Chile border in the mountain chain of the Cordillera
83
Occidental. The Cordillera Occidental is composed of Cretaceous-Tertiary volcanoes surrounded by
84
Quaternary age deposits; all the high peaks on this mountain range are volcanic cones. The second
85
region, Sajama (17° - 18° S), is centred around the isolated ice capped volcanic mountains in the
86
Sajama National Park (Fig. 1). The third region, Western Cordillera, covers the area southwards of
87
Sajama, along the dry, barren mountain range of the Cordillera Occidental (18° - 22° S) (Fig. 1).
88
Almost no glaciers exist in this region (Francou et al., 1999) as rainfall is very low (Fig. 1); annual
89
precipitation is estimated to be less than 250 – 300 mm on the summits (Vuille & Amman, 1997 cited
3
90
in Francou et al., 1999). With its dry climate, this area is the best example of arid high mountains in
91
the inner tropics (Francou et al., 1999).
92
93
2. Methods
94
Unlike glaciers, automated mapping of rock glaciers cannot be easily achieved from remotely sensed
95
data because they exhibit spectrally similar characteristics to their surroundings (Brenning, 2009;
96
Shukla et al., 2010). The optimal approach for generating this inventory was therefore to manually
97
identify rock glacier features (Casassa et al., 2002) (Table 1) within the Google Earth dataset and
98
digitise them. This inventory was achieved using expert photomorphic mapping from Google Earth
99
and a 30 m Global Digital Elevation Model (GDEM) derived from the Advanced Spaceborne Thermal
100
Emission and Reflection radiometer (ASTER) sensor.
101
102
2.1 Remote sensing approach
103
Initial rock glacier mapping and evaluation used freely available, fine spatial resolution remote
104
sensing data, available through the platform of Google Earth (version 7.1.1.18888, Google Inc.) using
105
the criteria identified in Table 1, and supported with field validation work. Satellite data available
106
through Google Earth ranged from 5 to 30 m resolution imagery. Google Earth is highly suited to this
107
type of research as it is a user friendly Geographical Information System (GIS) tool that facilitates
108
exploration of satellite data without the need for the user to turn to bespoke software to undertake
109
basic spatial queries, and negating the need for purchasing expensive satellite imagery. This is
110
especially pertinent for a project focused on developing an operational solution for use by a
111
developing world non-governmental organisation (e.g. by the Bolivian NGO 'Agua Sustentable').
112
Additionally, remote sensing data of the type available through Google Earth are known to be
113
particularly useful in large-scale geomorphological surveys (Slaymaker, 2001; Shukla et al., 2010)
114
and invaluable in environments where field access is difficult and/or limited (Kääb et al., 2005).
115
Since its development data available through the portal of Google Earth have been applied
116
widely across a range of research areas (Butler, 2006; Nourbakhsh et al., 2006; Ballagh et al., 2007;
117
Chang et al., 2009; Sheppard & Cizek, 2009; Ballagh et al., 2011; Yu & Gong, 2012). The imagery
4
118
has also been used to complement other sources of satellite imagery, such as aerial imagery or satellite
119
imagery for geomorphological mapping (e.g. Brenning, 2005; Morén et al., 2011). For research and/or
120
projects in developing countries with limited budgets, Google Earth is a useful source of data that
121
would otherwise be costly to obtain.
122
However, it should be noted that there are some uncertainties to consider with this approach,
123
specifically regarding the acquisition of remotely sensed data over mountainous terrain as
124
unfavourable meteorological conditions, topographic distortions and geometric uncertainties (e.g.
125
differences in scale, horizontal displacements and shadows; Buchroithner, 1995) can affect the data.
126
127
2.2 Digitising landform characteristics
128
Planimetric landform characteristics extracted and recorded for each rock glacier (Fig. 2) included
129
geographic coordinates; elevation (minimum and maximum); length (parallel to flow); width
130
(perpendicular to flow); aspect (main aspect of flow); and notes on surface texture and features. For
131
each rock glacier the Minimum Altitude at the Front (MAF) of the rock glacier was determined by
132
locating the elevation of the lowest point on the rock glacier snout where it meets the slope
133
underneath on satellite imagery. The maximum elevation at the head of the rock glacier (MaxE) was
134
similarly recorded; however, it is acknowledged that defining the upper boundary of a rock glacier is
135
difficult (Krainer & Ribis, 2012). Consistent judgement was made on where the upper boundary of the
136
rock glacier meets the input accumulation zone above it. Feature measurements were made using the
137
ruler tool on Google Earth with the length (parallel to the direction of flow) and width (perpendicular
138
to the direction of flow) being used (Fig. 2). Aspect was divided into 8 classes between N and NW
139
and manually defined along the main direction of the flow of the rock glacier (Fig. 2). Rock glacier
140
[surface] area (km2) of the digitised polygons was calculated through Google Earth Pro.
141
Although rock glaciers are known to be difficult to identify due to their composition of rock
142
debris supplied from surrounding rock areas (Shukla et al., 2010), morphological analysis can help
143
distinguish them from similar features such as protalus lobes, debris-covered glaciers, rock slope
144
failures and rock avalanches (Hamilton & Whalley, 1995; Seligman, 2009; Jarman et al., 2013). The
145
ratio of length to width measurements distinguishes rock glaciers from other periglacial features
5
146
(Harrison et al., 2008) such as protalus lobes and protalus ramparts. A length:width ratio greater than
147
1 represents a tongue shaped rock glacier (Guglielmin & Smiraglia, 1998; Harrison et al., 2008) and a
148
ratio of less than 1 implies a lobate rock glacier or a protalus lobe or protalus rampart. Although other
149
permafrost features were identified (e.g. detachment failures), only active and relict rock glaciers
150
meeting the identification criteria in Table 1 and were analysed.
151
In existing rock glacier inventories, vegetation has been used as a measure for rock glacier
152
activity, however here we However not using vegetation as an identifying characteristic has not
153
changed the number of rock glaciers in the inventory because they still meet the other criteria (mainly
154
the surface morphology of ridges and furrows). Furthermore, in Bolivia the environment is arid and
155
vegetation is very sparse and therefore it was not a critical factor in the classification of rock glacier
156
activity in this region.
157
Mean rock glacier parameters were used to compare between regions. Statistical analysis on
158
rock glacier characteristics were conducted in SPSS (version 19, IBM) using ANOVA, General
159
Linear Models and Tukey post hoc tests to address regional differences in rock glacier parameters
160
according to their activity. All statistical significance was tested at the 0.05 significance level. ArcGIS
161
(version 10.1, ESRI) and R (version 3.0.0, R Project for Statistical Computing) were used to assess
162
the relationship between mountain slope aspect and rock glacier orientation.
163
164
2.3 Field validation
165
To assess the reliability of the photomorphic mapping and subsequent rock glacier type classification,
166
field surveys were conducted during July - August 2011 and 2012 across the Bolivian Andes to
167
provide ground control for information extracted remotely. Field surveys included validation of rock
168
glacier identification and activity classification through feature observations and frontal slope angle
169
measurements using an inclinometer. Rock glaciers visited in the field programme are identified in
170
Tables S1 – S6 in the supplementary data file. Analysis of satellite data in Google Earth assisted in the
171
identification of appropriate field sites for validation and further research and for assessing the
172
accessibility of particular rock glaciers from settlements and roads.
173
6
174
3. Results
175
A total of 94 rock glaciers were identified in the Bolivian Andes (Table 2; Fig. 3). Active rock
176
glaciers accounted for 54 of these (57 %) and the remaining were classified as relict (43 %). 89 % of
177
the rock glaciers were classified as tongue shaped. 87% were developed with a southward flow
178
direction. 90 % of the rock glaciers were situated between 4700 and 5200 m altitude with 6 %
179
occurring above 5200 m and 4 % below 4700 m. The calculated mean MAF for active rock glaciers
180
was 4983 m (± 30m) and 4870 m (± 30m) for relict rock glaciers (Table 2). Rock glacier
181
characteristics and features were analysed at a national and a regional level. Tables S1 – S6 in the
182
supplementary data file detail the information for each rock glacier recorded in the inventory. Figure 4
183
shows in situ photographs and the relating Google Earth imagery for example rock glaciers that were
184
visited in the field.
185
186
3.1 Rock glacier elevation and range
187
Rock glaciers occurred within an elevation range of 4475 m to 5345 m in the Bolivian Andes (Fig. 5;
188
Fig. 6a and 6b). Active rock glaciers occurred at higher elevations than relict rock glaciers.
189
Nationally, the mean MAF for all active rock glaciers was 4985 m (± 30 m) (Table 2), with a range of
190
4709 to 5345 m (Fig. 5; Fig. 6a), while the most frequent elevation band was 4900 - 5000 m (28 %).
191
Over half of the active rock glaciers (52 %) were situated in the elevation bands 4900 - 5100 m (Fig.
192
5). The highest active rock glacier was located in the Sajama region [B124a]. The mean MAF for
193
relict rock glaciers was roughly 100 m lower, at 4870 m (± 30 m) (Table 2), with a range of 4475 to
194
5202 m (Fig. 5; Fig. 6b). The highest relict rock glacier was also situated in the Sajama region
195
[B118b]. Over a third of the relict rock glaciers were situated in the elevation band 4800 - 4900 m (35
196
%), with 58 % of them found between 4800 and 5000 m (Fig. 6b). The box plots of Figure 5 show
197
that the Western Cordillera had the largest spread of both active and relict rock glacier elevations. In
198
contrast, the Cordillera Real had the smallest range of active rock glacier MAFs (Fig. 5).
199
Regionally, larger differences were observed between active and relict rock glacier elevations
200
(MAFs), although no significant difference was found (ANOVA: F-value = 5.756, df within group =
201
1, between groups= 88, p = 0.138). It can be assumed that was due to lack of difference in MAFs in
7
202
the Cordillera Real (Fig. 8). Active rock glaciers in Sajama were on average at 175 m higher
203
elevations than relict ones. Equally, active rock glaciers in the Western Cordillera were 173 m higher
204
than relict ones in the same region (Table 3). Rock glaciers were found at significantly different
205
elevations (Linear Model: F-value: 3.697, df within groups = 2, between groups = 88) in Sajama
206
compared to the Western Cordillera (p=0.007) and also the Cordillera Real (p=0.001), however the
207
Cordillera Real and Western Cordillera were not significantly different (p=0.949).
208
Tukey post hoc testing (ANOVA: F-value: 8.161, df within groups = 5, between groups = 88,
209
p = <0.001) showed that active rock glacier MAFs in Sajama were statistically at higher elevations
210
than in the other regions (Fig. 7): Cordillera Real (p = 0.002) and Western Cordillera (p = 0.018).
211
Sajama’s relict rock glacier MAFs were also found at significantly higher elevations than relict MAFs
212
in the Western Cordillera (p=0.048).
213
214
3.2 Aspect
215
Results from this inventory suggest that south-facing slopes are the most suitable for rock glacier
216
development and formation. 87 % of rock glaciers in the inventory have developed on south-facing
217
aspects (SE, S and SW) (Fig. 8) with SE being the most predominant aspect (34 %). 88 % of the rock
218
glaciers classified as active had a main direction of flow as southerly; 86 % of the relict rock glaciers
219
had these same south-facing aspects. Figure 8a shows that southerly aspects allowed for rock glaciers
220
to exist at lower elevations. Although rock glaciers of all sizes occurred on slopes with southerly
221
aspects, such slopes contained the largest rock glaciers in the country (Fig. 8b). Figures 9a and 9b
222
show that the dominance of southerly aspects in rock glaciers is independent to any bias of available
223
mountain slope aspects. The spread of slope aspects was relatively uniform across the Bolivian Andes
224
(Fig. 9a), but there was a clear clustering of active rock glaciers with southerly-facing slopes in this
225
inventory (Fig. 9b).
226
227
3.3 Rock glacier morphology
228
89 % of the rock glaciers in the study area were found to be tongue shaped and this proportion does
229
not differ greatly between active and relict forms (91 % and 87 % respectively). Each region had at
8
230
least one rock glacier of length greater than 1 km, but overall 93 % of the rock glaciers had a length of
231
less than 1 km, with an overall average length of 500 m. On average, Sajama had the longest active (
232
= 753 m ± 30 m) and relict (
= 519 m ± 30 m) rock glaciers in Bolivia (Fig. 10a) and the
233
Cordillera Real had the smallest active (
234
glaciers (Table 3; Fig. 10a). However, there was no significant effect of region on rock glacier length
235
(ANOVA: F-value = 5.18, df within groups = 2, between groups = 88, p=0.162).
= 380 m ± 30 m) and relict (
= 378 m ± 30 m) rock
236
In total, it is estimated that rock glaciers cover 11 km2 of the Bolivian Andes (Table 4). Rock
237
glacier area varies between 0.006 and 0.789 km2 in the Bolivian Andes, a similar range of to those in
238
the Chilean Andes (Brenning, 2005, p.234). Mean rock glacier area was 0.12 km2, with a median area
239
of 0.08 km2 (Table 4). In total, 56 % of the rock glaciers were smaller than 0.1 km2. The largest rock
240
glacier areas were found in Sajama, with active rock glacier areas averaging 0.2 km2 (Table 3). No
241
correlation was observed between rock glacier length and elevation (Linear model: r=0.07, df=92,
242
p=0.501) (Fig. 10b) and between rock glacier area and elevation (r=0.01, df=92, p=0.929). Similarly,
243
no significant correlation was seen between latitude and both rock glacier length (r=0.056, df=92,
244
p=0.591) and area (r=0.052, df=92, p=0.616). Bolivian rock glaciers are found to be of similar size
245
or slightly larger than rock glaciers found elsewhere (e.g. Colorado, Austria) (Table 4), however their
246
abundance in Bolivia is much lower.
247
248
4. Discussion
249
This national rock glacier inventory for Bolivia has identified a number of active (and fewer relict)
250
rock glaciers in the high arid mountains in the region. Along the Cordillera Occidental, rock glaciers
251
are clustered around a small number of isolated mountain peaks (e.g. Sajama 18° S, 68° W). This is
252
recognised as 'island permafrost', known to be common across other parts of the Central Andes
253
(Travassos et al., 2008). The MAF for rock glaciers is often considered a good approximation of the
254
lower limit of discontinuous permafrost (Scotti et al., 2013). No active rock glaciers are found lower
255
than 4700 m, implying that this is the lower limit of permafrost for the Bolivian Andes. Although
256
slightly higher, this is in line with research from the Argentina Andes at 22° 30’ S where the lower
257
limit of active rock glaciers was observed at 4500 m (Corte et al., 1982 cited in Esper Angillieri,
9
258
2009). Rock glaciers were found at higher elevations than of those in the Chilean Andes; Brenning
259
(2005) identified rock glaciers at 3000 m and higher, whereas in the Bolivian Andes no rock glaciers
260
were found lower than 4400 m. Rock glacier size appears to be similar to those in recent European
261
rock glacier inventories (Krainer & Ribis, 2012; Kellerer-Pirklbauer et al., 2012), yet their frequency
262
is not as high, and thus their total surface area is not as great as those of the European Alps (Table 4).
263
Optimum rock glacier conditions are found in areas with high elevation, cold temperatures
264
and low precipitation (Baroni et al., 2004): conditions characteristic of the Bolivian Andes. The key
265
controls on rock glacier characteristics and formation are thought to be climatic conditions, glacial
266
history and rock supply (Johnson et al., 2007; Guglielmin & Smiraglia, 1998). On a regional scale,
267
rock glacier distribution is climatically controlled through the parameters of precipitation and
268
temperature; the latter is dependent on elevation and aspect. The limited distribution of rock glaciers
269
along the Bolivian Andes despite the appropriate climatic conditions, demonstrates the small scale
270
variability in rock glacier distribution, with other factors such as rock supply, lithology, glacial
271
history, competition with glaciers and geothermal fluxes (Brenning, 2005; Johnson et al., 2007)
272
ultimately determining their location. Unfortunately, high quality lithological data were not available
273
for this research, but would future enhance the understanding of controls on rock glacier distributions.
274
Even though temperature is a key control on rock glacier activity, no correlation was observed
275
between rock glacier length and elevation (r=0.07) or rock glacier area and elevation (r=0.01), also
276
suggesting the importance of other factors such as debris supply. However, the inventory did show
277
that aspect is an important parameter for rock glacier formation (Fig. 9b). In the Southern Hemisphere
278
south-facing slopes are more likely to host glaciers and rock glaciers (Esper Angillieri, 2010) in
279
association with reduced insolation and increased accumulation of snow, ice and rock debris (Esper
280
Angillieri, 2009). This inventory also suggests that southerly aspects, with their reduced insolation,
281
allow rock glaciers to exist at lower elevations than at other aspects (Figure 8a).
282
Rock glaciers develop between the 0 °C isotherm and the regional snow line (e.g. Payne,
283
1998; Milana & Maturano, 1999), and it follows that the elevation at the front of active rock glaciers,
284
the MAF, corresponds closely to the 0 °C isotherm. This inventory implies that the present day 0 °C
285
isotherm lies at 4900 – 5000 m. Relict features interpreted as former rock glaciers reflect increased
10
286
elevation of the 0 °C isotherm, associated with warming, or changes in debris supply (Esper
287
Angillieri, 2010). The mean 100 m difference found here between active and relict rock glaciers
288
(Table 2) represent this upward shift in the isotherm over time. The modern 0 °C isotherm drops in
289
elevation from north to south along the Western Cordillera (Payne, 1998) but the ELA increases with
290
aridity, leading to a bigger niche for rock glaciers. This is demonstrated by the larger range of rock
291
glacier MAFs in this region (Fig. 5). However, with rising regional temperatures there will be smaller
292
niches for active rock glaciers to occupy as the 0 °C isotherm moves closer to, or above, mountain
293
summits. Therefore a warming climate will decrease the number of active rock glaciers and increase
294
the number of relict rock glaciers (Krainer & Ribis, 2012). Equally, it can also be hypothesised that
295
rock glaciers may become more frequent in the Cordillera Real with glacier recession through
296
increased debris supply to glacier surfaces producing glacial derived rock glaciers.
297
In the wetter climate of the Cordillera Real, conditions favour [ice] glaciers. This smaller
298
niche for rock glaciers was confirmed by our analyses, which determined that the Cordillera Real had
299
the smallest elevation range of active rock glaciers (4740 – 5050 m) of our three study regions (Fig.
300
5). The drier Western Cordillera had a larger total area of rock glaciers, a surface area of 2.88 km2 for
301
active rock glaciers (Table 4), and the largest range of elevations of all regions (4500 – 5255 m).
302
From the remotely sensed imagery we observed that rock glaciers were also more noticeably
303
developed in the Western Cordillera. These results and observations suggest that conditions in the
304
Western Cordillera region are the most conducive for active rock glaciers in the Bolivian Andes.
305
Given the lack of glaciers in the Western Cordillera, this rock glacier inventory has shown
306
that rock glaciers in that region could be considered as hydrologically important to mountain
307
communities (18° - 22° S). Similarly, in Sajama where ice is also limited, it can be assumed that rock
308
glaciers may be important for local mountain communities and the Sajama National Park. However,
309
glaciers dominate in Cordillera Real, with a recent estimated coverage of 185.5 km2 (Ramirez et al.,
310
2012), while rock glaciers are less abundant and smaller in this region (Table 3). Here, we estimated
311
that rock glaciers cover 1.07 km2 and hence we argue that rock glaciers are not an important source of
312
water (in comparison to glaciers) for the La Paz region. However, unlike glacier recession, the ice of
313
rock glaciers is protected from small thermal changes by the insulating rock layer, resulting in a
11
314
longer lag time in their response to climate change, responding on the decadal time scale (Haeberli et
315
al., 2006). Therefore, rock glaciers are likely to become increasingly important in mountain regions
316
for water reservoirs given climate warming (Schrott, 1996; Millar & Westfall, 2008). Hence there is a
317
need to understand current controls on rock glacier development and explore the impact of climate
318
change on rock glaciers.
319
320
5. Conclusions
321
This study used remotely sensed data contained within the Google Earth platform to identify and
322
assess rock glacier distribution across the Bolivian Andes. Rock glacier characteristics were recorded
323
and analysed and the total number and distribution of rock glaciers was established. 94 rock glaciers
324
have been identified, covering an estimated 11 km2, of which 57 % were classified as active and the
325
remaining as relict. We found that the majority (87 %) of Bolivian rock glaciers are orientated in a
326
southerly aspect (SE, S and SW), suggesting the importance of reduced solar input for rock glacier
327
development. From our results we assess the lower limit of permafrost to be at 4700 m in the Bolivian
328
Andes, an estimation which is in agreement with those of surrounding regions. Regional differences
329
in rock glacier elevation, distribution and size were found between the Cordillera Real, Sajama and
330
Western Cordillera and result from variations in climate (specifically aridity), rock supply and direct
331
competition with ice glaciers. The Western Cordillera has the largest niche zone for rock glacier
332
formation and development, which has led to some of the largest and most well developed rock
333
glaciers in the Bolivian Andes. Rock glaciers in Bolivia are important sources of water in the Western
334
Cordillera, and may increase in importance in the Cordillera Real region with ice glacier recession;
335
however, their longer-term future given rising temperatures and the disappearing 0 °C isotherm
336
remains unknown. This is important because there is growing evidence of severe ice glacier recession
337
in the region and regional water supplies will become increasingly dependent on other permafrost
338
stores in the future.
339
340
Acknowledgements
12
341
342
343
344
345
This research was funded by a NERC Case Studentship (NE/H018875/1). Oxfam and Agua
Sustentable provided valuable help and support in the field in Bolivia. Comments from James Duffy
(re: analysis) and Andrew Suggitt (re: manuscript) and anonymous reviewers notably improved the
manuscript. The authors would like to thank Helen Jones at Exeter University for her help with the
figures.
346
347
348
349
References
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
Esper Angillieri MY. 2009. A preliminary inventory of rock glaciers at 30°S latitude, Cordillera Frontal of San
Juan, Argentina. Quaternary International 195: 151 – 157.
Esper Angillieri MY. 2010. Application of frequency ratio and logistic regression to active rock glacier
occurrence in the Andes of San Juan, Argentina. Geomorphology 114: 396 – 405.
Azócar GF & Brenning A. 2010. Hydrological and Geomorphological Significance of Rock Glaciers in the Dry
Andes, Chile (27° - 33°S). Permafrost and Periglacial Processes 21: 42 – 53.
Ballagh LM, Parsons MA & Swick R. 2007. Visualising cryopsheric images in a virtual environment: present
challenges and future implications. Polar Record 43(4): 305 – 310.
Ballagh LM, Raup BH, Duerr RE, Khalsa SJS, Helm C, Fowler D, Gupte A. 2011. Representing scientific data
sets in KML: Methods and challenges. Computers and Geosciences 37(1): 57 – 64.
Baroni C, Carton A & Seppi R. 2004. Distribution and behaviour of rock glaciers in the Adamello-Presanella
Massif (Italian Alps). Permafrost and Periglacial processes 15: 243 – 259.
Barsch D. 1996 . Rockglaciers – indicator for the present and former Geoecology in high mountain
environments. Springer, Berlin.
Bradley RS, Vuille M, Diaz HF, Vergara W. 2006. Threats to Water Supplies in the Tropical Andes. Science
312: 1755 – 1756.
Brenning A. 2005. Geomorphological, hydrological and climatic significance of rock glaciers in the Andes of
Central Chile (33-35°S). Permafrost and Periglacial Processes 16: 231 – 240.
Brenning A. 2009. Benchmarking classifiers to optimally integrate terrain analysis and multispectral remote
sensing in automatic rock glacier detection. Remote Sensing of Environment 113: 239 – 247.
Brenning A, Grasser M. & Friend DA. 2007. Statistical estimation and generalized additive modeling of rock
glacier distribution in the San Juan Mountains, Colorado, United States. Journal of Geophysical Research
112(F2), DOI: 10.1029/2006JF000528
Buchroithner MF. 1995. Problems of mountain hazard mapping using spaceborne remote sensing techniques.
Advances in Space Research 15(11): 57 – 66.
Bulter D, 2006. Virtual globes: The web-wide world. Nature 439: 776 – 778.
Casassa G, Smith K, Rivera A, Araos J, Schnirch M & Schneider C. 2002. Inventory of glaciers in isla Riesco,
Patagonia, Chile, based on aerial photography and satellite imagery. Annals of Glaciology 34: 373-378.
13
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
Chang AY, Parrales ME, Jimenez J, Sobieszczyk ME, Hammer SM, Copenhaver DJ & Kulkarni RP. 2009.
Combining Google Earth and GIS mapping technologies in a dengue surveillance system for developing
countries. International Journal of Health Geographics 8(49). DOI:10.1186/1476-072X-8-49
Chevallier P, Pouyaud B, Suarez W & Condom T. 2011. Climate change threats to environment in the tropical
Andes: glaciers and water resources. Regional Environmental Change 11(1): 179 – 187.
Curtaz M, Vagliasindi M, Letey S, Morra di Cella U & Pogliotti P. 2011. A new rock glaciers inventory in the
North-Western Alps. Geophysical Research Abstracts 13.
Falaschi D, Castro M, Masiokas M, Tadono T & Ahumada A L. 2014. Rock glacier inventory of the Valles
Calchaquíes region (~25°S), Salta, Argentina, derived from ALOS data. Permafrost and Periglacial Processes
25(1): 69 – 75.
Francou B, Fabre D, Pouyaud B, Jomelli V & Arnaud Y. 1999. Symptoms of degradation in a tropical rock
glacier, Bolivian Andes. Permafrost and Periglacial Processes 10: 91 – 100.
Francou B, Vuille M, Wagnon P, Mendoza J & Sicart JE. 2003. Tropical climate change recorded by a glacier in
the central Andes during the last decades of the twentieth century: Chacaltaya, Bolivia, 16°S. Journal of
Geophysical Research 108: 1 – 12.
Guglielmin M & Smiraglia C. 1998. The rock glacier inventory of the Italian Alps. Permafrost – Seventh
International Conference (Proceedings), Yellowknife (Canada), Collection Nordicana. No. 55: 375 – 382.
Haeberli W, Hallet B, Arenson L, Elconin R, Humlum O, Kääb A, Kaufmann V, Ladanyi B, Matsuoka N,
Springman S & Mühll DV. 2006. Permafrost creep and rock glacier dynamics. Permafrost and Periglacial
Processes 17: 189 – 214.
Hamilton SJ & Whalley B. 1995. Rock glacier nomenclature: A re-assessment. Geomorphology 14(1): 73 – 80.
Harrison S, Whalley B & Anderson E. 2008. Relict rock glaciers and protalus lobes in the British Isles:
implications for Late Pleistocene mountain geomorphology and palaeoclimate. Journal of Quaternary Science
23(3): 287 – 304.
Jarman D, Wilson P and Harrison S. 2013. Are there any relict rock glaciers in the British mountains? Journal
of Quaternary Science 28: 131–143. DOI: 10.1002/jqs.2574.
Johnson BG, Thackray GD & Van Kirk R. 2007. The effect of topography, latitude, and lithology on rock
glacier distribution in the Lemhi Range, central Idaho, USA. Geomorphology 91: 38 – 50.
Jordan E. 2008. Glaciers of South America: Glaciers of Bolivia. Satellite Image Atlas of Glaciers of the World.
United States Geological Survey 18th March 2008. Retrieved 12 June, 2013, from
http://pubs.usgs.gov/pp/p1386i/bolivia/intro.html
Kääb A, Huggel C, Fischer L, Guex S, Paul F, Roer I, Salzmann N, Schlaefli S, Schmutz K, Schneider D,
Strozzi T & Weidmann Y. 2005. Remote sensing of glacier- and permafrost-related hazards in high mountains:
An overview. Natural Hazards and Earth System Sciences 5: 527-554.
Kääb A & Weber M. 2004. Development of transverse ridges on rock glaciers: field measurements and
laboratory experiments. Permafrost and Periglacial Processes 15(4): 379 – 391.
Kellerer-Pirklbauer A, Lieb GK & Kleinferchner H. 2012. A new rock glacier inventory of the Eastern
European Alps. Austrian journal of Earth Sciences 105(2): 78 – 93.
Krainer K & Ribis M. 2012. A rock glacier inventory of the Tyrolean Alps (Austria). Austrian Journal of Earth
Sciences 105(2): 32 – 47.
Milana JP & Maturano A. 1999. Application of Radio Echo Sounding at the arid Andes of Argentina: the Agua
Negra Glacier. Global and Planetary Change 22: 179 – 191.
14
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
Millar CI & Westfall RD. 2008. Rock glaciers and related periglacial landforms in the Sierra Nevada, CA, USA;
inventory, distribution and climatic relationships. Quaternary International 188: 90 – 104.
Morén B, Heyman J & Stroeven AP. 2011. Glacial geomorphology of the central Tibetan Plateau. Journal of
Maps 7(1): 115 – 125. DOI:10.4113/jom.2011.1161.
Nourbakhsh I, Sargent R, Wright A, Cramer K, McClendon B & Jones M. 2006. Mapping disaster zones.
Nature 439: 787 -788.
Payne D. 1998. Climatic implications of rock glaciers in the arid Western Cordillera of the Centra Andes.
Glacial Geology & Geomorphology rp03.
Rabatel A, Francou B, Soruco A, Gomez J, Cáceres B, Cebellos JL, Basantes R, Vuille M et al. 2013. Current
state of glaciers in the tropical Andes: a multi-century perspective on glacier evolution and climate change. The
Cyrosphere 7: 81 – 102.
Ramirez E, Ribeiro R, Machaca A, Fuertes R, Simões J. 2012. Glacier Inventory of the Cordillera Real –
Bolivia using high resolution satellite images ALOS and CBERS-2B. Geophysical Research Abstracts, EGU
General Assembly 2012, vol 14.
Rangecroft S, Harrison S, Anderson K, Magrath J, Castel AP, Pacheco P. 2013. Climate change and water
resources in arid mountains: an example from the Bolivian Andes. AMBIO 42 (7): 852 – 863.
Scotti R, Brardinoni F, Alberti S, Frattini P, Crosta GB. 2013. A regional inventory of rock glaciers and protalus
ramparts in the central Italian Alps. Geomorphology 186: 136 – 149.
Schrott L. 1996. Some geomorphological-hydrological aspects of rock glaciers in the Andes (San Juan,
Argentina). Zeitschrift für Geomorphologie 104: 161 – 173.
Seligman ZM. 2009. Rock glacier distribution, activity, and movement, Northern Absaroka and Beartooth
Ranges, MT, USA. MSc Thesis, University of Colorado, Boulder, Colorado, USA.
Sheppard SRJ & Cizek P. 2009. The ethics of Google Earth: Crossing thresholds from spatial data to landscape
visualisation. Journal of Environmental Management 90: 2102 – 2117.
Shukla A, Gupta RP & Arora MK. 2010. Delineation of debris-covered glacier boundaries using
optical and thermal remote sensing data. Remote Sensing Letters 1(1): 11 — 17.
Slaymaker O. 2001. The role of remote sensing in geomorphology and terrain analysis in the Canadian
Cordillera. International Journal of Applied Earth Observation and Geoinformation 3(1):11 – 17.
Soruco AS. 2012. Medio siglo de fluctuaciones glaciares en la Cordillera Real y sus efectos hidrológicos en la
ciudad de La Paz. IRD. ISBN 978-99954-62-0 [Spanish]
Soruco A, Vincent C, Francou B, Gonzalez JF. 2009. Glacier decline between 1963 and 2006 in the Cordillera
Real, Bolivia. Geophysical Research Letters 36(6). DOI:10.1029/2008GL036238.
Travassos JM, Stael GC & Liaudat DT. 2008. A GPR survey on degraded permafrost at Mesón San Juan,
Mendonza, Argentina. Revista Brasileria de Geoffsica 26(4): 519 – 526.
Trombotto D, Buk E & Hernández J. 1999. Rock glaciers in the Southern Central Andes (approx. 33°–34°S),
Cordillera Frontal, Mendoza, Argentina. Bamberger Geographische Schriften 19: 145–173.
UNFCCC. 2007. Climate Change: Impacts, vulnerabilities and adaptation in developing countries. United
Nations framework convention on climate change. Retrieved 12 June, 2013, from
http://unfccc.int/resource/docs/publications/impacts.pdf
Vergara W. 2009. Assessing the potential consequences of climate destabilization in Latin America. Latin
America and Caribbean region sustainable development working paper 32. The World Bank, Latin America and
the Caribbean Region Sustainable Development Department.
15
511
512
513
514
515
516
517
518
519
520
521
522
523
524
Vergara W, Kondo H, Pérez EP, Pérez JMM, Rueda VM, Arango MCM, Murcia JFR, Roldán GJA, Palcios E.
2007. Visualizing future climate in Latin America: Results from the application of the Earth Simulator. Latin
America and Caribbean Region Sustainable Development Working Paper 30. The World Bank, Latin America
and the Caribbean Region Sustainable Development Department.
Vuille M, Francou B, Wagnon P, Juen I, Kaser G, Mark BG, Bradley RS. 2008. Climate change and tropical
Andean glaciers: Past, present and future. Earth-Science Reviews 89: 79 – 96.
Yu L & Gong P. 2012. Google Earth as a virtual globe tool for Earth science applications at the global scale:
progress and perspectives. International Journal of Remote Sensing 33(12): 3966-3986.
Zemp M, Hoelzle M & Haeberli W. 2007. Distributed modelling of the regional climatic equilibrium line
altitude of glaciers in the European Alps. Global and Planetary Change 56(1-2): 83 – 100.
16
Characteristic
features used for
identification
Frontal ramp
Active
Relict
Notes
Steep (typically ≥30°)
Gently sloping
(typically <30°)
Rock glacier
body
Swollen body,
indicating ice
presence
Signs of flow: e.g.
ridges and furrows
A flattened body,
result of ice
disappearing
Less defined flow
lines
Frontal slope indicating the presence of ice
(Baroni et al., 2004, p. 251). An angle
greater than the angle of repose indicates
the presence of ice; however angle of
repose depends on lithology.
Possible swollen body indicating the
presence of ice (Baroni et al., 2004, p. 251)
Surface texture
Well-developed longitudinal and
transversal ridges and furrows which act as
signs of flow (Kääb & Weber, 2004, p.
379)
Table 1: Characteristics used to identify rock glaciers (frontal ramp, body and surface texture). Further
discrimination into active and relict types can be achieved by image based or field based survey of detailed
characteristics.
Active
Relict
Number (%)
of features
54 (57)
40 (43)
MAF*
(m)
4983
4870
MaxE**
(m)
5162
5009
Length
(m)
552
432
Width
(m)
239
253
Area
(km2)
0.132
0.108
Aspect
SE
S
Table 2: Key mean characteristics for active and relict rock glaciers.*Minimum altitude at Front. **Maximum
Elevation of rock glacier.
Area of Activity
region Status
(km2)
Cordillera ~7050 Active
Real
Relict
~4000 Active
Sajama
Relict
~25000 Active
Western
Cordillera
Relict
36050
Total
Region
Number
of
features
16
10
15
16
23
14
94
Mean
MAF*
(m)
4906
4891
5110
4934
4954
4781
Mean
MaxE**
(m)
5033
5036
5307
5073
5158
4915
Mean
feature
length (m)
380
378
753
519
541
373
Mean
feature
width (m)
212
150
242
252
253
328
Mean
area
(km2)
0.07
0.08
0.20
0.13
0.13
0.09
Table 3: Regional rock glacier characteristics. *Minimum altitude at Front. **Maximum elevation of rock
glacier.
17
Modal
aspect
SW
S
SE
S
SE
SW
Number of
rock glaciers
Average rock
glacier area
(km2)
Median rock
glaciers area
(km2)
Total rock glacier
surface area
(km2)
Reference
Bolivia
94
0.12
0.08
11
-
San Juan
mountains,
Colorado
756
0.064
0.051
70
Brenning et
al., 2007
Tyrolean Alps,
Austria
3145
0.088
-
167.2
Krainer &
Ribis, 2012
Eastern European
Alps
347
0.061
-
21.3
KellererPirklbauer et
al. 2012
Table 4: Results on rock glacier size (length and area) from this inventory compared to other peer-reviewed
rock glacier inventories.
18