See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/41821985

Monitoring ice-capped active Volcan Villarrica, southern Chile, using

terrestrial photography combined with automatic weather stations and global
positioning systems

Article in Journal of Glaciology · December 2008

DOI: 10.3189/002214308787780076 · Source: OAI

CITATIONS READS

37 173

5 authors, including:

Andrés Rivera Javier Corripio

University of Chile meteoexploration
229 PUBLICATIONS 9,097 CITATIONS 37 PUBLICATIONS 1,590 CITATIONS

SEE PROFILE SEE PROFILE

Ben Brock Jorge Clavero

Northumbria University Amawta Geoconsultores
83 PUBLICATIONS 3,429 CITATIONS 122 PUBLICATIONS 1,619 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Seismo-acoustic study of ice-covered volcanoes View project

Estudio de la dinámica actual de glaciares en alta resolución en la Patagonia Argentina, mediante la combinación de técnicas de sensoreamiento remoto View project

All content following this page was uploaded by Jorge Clavero on 20 March 2015.

The user has requested enhancement of the downloaded file.

920 Journal of Glaciology, Vol. 54, No. 188, 2008

Instruments and Methods

Monitoring ice-capped active Volcán Villarrica, southern Chile,
using terrestrial photography combined with automatic weather
stations and global positioning systems
Andrés RIVERA,1,2,3 Javier G. CORRIPIO,4 Ben BROCK,5 Jorge CLAVERO,6
Jens WENDT1
1
Centro de Estudios Cientı́ficos, Avenida Arturo Prat 514, Casilla 1469, Valdivia, Chile
E-mail: [email protected]
2
Centro de Ingenier ía de la Innovación del CECS, Avenida Arturo Prat 514, Casilla 1469, Valdivia, Chile
3
Departamento de Geografı́a, Universidad de Chile, Marcoleta 250, Santiago, Chile
4
Institut für Geographie, Universitat Innsbruck, Innrain 52, A-6020 Innsbruck, Austria
5
Department of Geography, University of Dundee, Dundee DD1 4HN, UK
6
Servicio Nacional de Geologı́a y Minerı́a, Avenida Santa Marı́a 0104, Casilla 10465, Santiago, Chile

ABSTRACT. Volcán Villarrica (39◦ 25 12 S, 71◦ 56 27 W; 2847 m a.s.l.) is an active ice-capped volcano
located in the Chilean lake district. The surface energy balance and glacier frontal variations have been
monitored for several years, using automatic weather stations and satellite imagery. In recent field
campaigns, surface topography was measured using Javad GPS receivers. Daily changes in snow-, ice-
and tephra-covered areas were recorded using an automatic digital camera installed on a rock outcrop.
In spite of frequently damaging weather conditions, two series of consecutive images were obtained,
in 2006 and 2007. These photographs were georeferenced to a resampled 90 m pixel size SRTM digital
elevation model and the reflectance values normalized according to several geometric and atmospheric
parameters. The resulting daily maps of surface albedo are used as input to a distributed glacier-melt
model during a 12 day mid-summer period. The spatial pattern of cumulative melt is complex and
controlled by the distribution of airfall and wind-blown tephra, with extremely high melt rates occurring
downwind of the crater and exposed ash banks. The camera images are also used to visualize the pattern
of glacier crevassing. The results demonstrate the value of terrestrial photography in understanding the
energy and mass balance of the glacier, including the generation of meltwater, and the potential value
of the technique for monitoring volcanic activity and potential hazards associated with ice–volcano
interactions during eruptive activity.

INTRODUCTION (Fig. 1). Lahars produced by eruptions of Volcán Villarrica in

1948/49, 1963/64 and 1971/72 have resulted in the deaths of
Volcán Villarrica (39◦ 25 12 S, 71◦ 56 27 W; 2847 m a.s.l.; more than 75 people (Stern, 2004), and are considered the
Fig. 1) is a highly active ice-capped volcano, which is charac- main hazard of the volcano (Moreno, 2000). The volcano
terized in historical times mainly by mild strombolian activity is covered by a 30.3 km2 glacier (Rivera and others, 2006),
(González-Ferrán, 1995; Lara, 2004), permanent degassing mainly distributed towards the south and east where the main
and periodic explosions, with the lava lake remaining at glacier basin (Glaciar Pichillancahue–Turbio (17.3 km2 );
a high level (90–180 m below the surface), at least since Fig. 2), composed of partially ash/debris-covered ice, is
1984. It is also very sensitive to magmatic-conduit activ- located, partially infilling a volcanic caldera depression
ity (Calder and others, 2004; Witter and Delmelle, 2004). (Clavero and Moreno, 2004). The energy balance of this
Concentrations of acid gases measured at the summit of the glacier has been monitored since 2003 (Brock and others,
crater have been recognized as a hazard to climbers on 2007), and global positioning system (GPS) as well as radio-
the volcano, who may be exposed to concentrations above echo sounding (RES) measurements were carried out in Janu-
the limits defined by the US National Institute of Occupa- ary 2005 (Rivera and others, 2006).
tional Safety and Health (Witter and Delmelle, 2004). Its One of the most interesting direct effects of volcanic activ-
historical eruptive activity (Petit-Breuihl and Lobato, 1994; ity on the overlying glaciers is the ice cracking or crevassing
Lara, 2004) indicates a low frequency of large explosive observed before or during eruptive events (Klohn, 1963;
eruptions (volcanic explosivity index (VEI) 3–4). More than Fuenzalida-Ponce, 1976; Fuentealba and others, 1984;
50 eruptive events have been documented since 1558 (Petit- González-Ferrán, 1995). Ice cracking has been detected by
Breuihl and Lobato, 1994). The latest most violent eruption seismicity (Métaxian and others, 2003), but here we propose
took place in 1971/72 when lava flows were generated, as observing possible ice cracking by obtaining daily photo-
well as 30–40 km h−1 laharic flows (Naranjo and Moreno, graphs of the ice. This process could be important for ice flow
2004) descending towards Lagos Villarrica and Calafquén and for the hydrological balance of the glacier, as crevasses
Rivera and others: Instruments and methods 921

PACIFIC
OCEAN

ARGENTINA

Fig. 1. General location map of Volcán Villarrica.

are the main pathway for meltwater to enter the glacier’s en- (Figs 2 and 3). The AWS recorded incoming and reflected
and subglacial drainage system (Fountain and Walder, 1998; shortwave radiation, net all-wave radiation, air temperature,
Fountain and others, 2005). air humidity and wind speed at 2 m above the surface with
A second direct effect of volcanic activity is the deposition hourly mean values recorded on a Campbell CR10 data
of pyroclastic materials on top of the glaciers, changing logger. Albedo was measured using a Kipp & Zonen CM6B
the albedo and affecting the energy balance by providing albedometer sensitive to radiation in the range 0.3–2.8 μm,
insulation from atmospheric heat and insolation where with the sensors mounted in a surface-parallel plane. Details
particles are large or the cover is continuous (Adhikary and of the collected variables and technical details of the sensors
others, 2002; Kirkbride and Dugmore, 2003). Studies of at the AWS are described by Brock and others (2007).
tephra thermal properties and glacier melt rates at Volcán
Villarrica (Brock and others, 2007) identified a very low Oblique photography
thermal conductivity of 0.35 W m−1 K−1 in the lapilli tephra For measuring albedo, surface changes and tephra cover,
which blankets most of the ablation zone. Furthermore, a an automatic camera (Canon EOS 300D) was installed at
critical thickness of just 5 mm of tephra cover was found the upper part of the rim surrounding the main volcano
to reduce the melt rate of the buried ice compared with a edifice (upper part of the caldera, Fig. 2), from where daily
bare surface. Consequently, volcanically produced materials photographs of the glacier were obtained. The camera has
appear to have a large positive impact on the mass balance a 6.3 megapixel complementary metal oxide semiconductor
of glaciers on Volcán Villarrica, due to an extensive mantle (CMOS) sensor and recorded the images on a 2 GB flash-
of insulating tephra in the lower ablation zones. In this study, card memory. It was fitted with a high-quality, fixed 24 mm
quantitative assessment of the total area of tephra cover focal length lens to minimize optical distortions. This camera
and tephra-free ice slopes and snow albedo has been made sensor has some sensitivity in the near-infrared spectrum, at
through terrestrial photography. least to 1000 nm, and beyond that point if the infrared filter is
removed (experimental tests by the authors). The camera was
inserted into a Pelikan sealed box where it was controlled by
AIMS AND METHODS a Canon timer. The system was powered by 12 V batteries
The main aim of this paper is to analyse albedo variation which were fed by a solar panel installed nearby.
and its impact on melt on Glaciar Pichillancahue–Turbio Conventional photography is a powerful medium for
of Volcán Villarrica by means of terrestrial photography. collecting and storing information. If this information can
Previous measurements of the glacier areal variations (Rivera be located precisely in space, then photography becomes
and others, 2006) have been updated until 2007, to allow the a powerful tool for quantitative analysis. Here we use a
assessment of recent ice retreat on the volcano’s glaciers. tool for georeferencing oblique photography developed by
Corripio (2004), using a single image and a digital elevation
Meteorological data model (DEM). The accuracy of the technique depends on the
The meteorological data were obtained by an automatic accuracy of the DEM and on the quality of the photographic
weather station (AWS) located during the summer of 2003/04 image, especially the degree of distortion and aberration
on the surface of the glacier at 1933 m a.s.l. (near the produced by the lens. The technique does not produce
equilibrium-line altitude, ∼2000 m a.s.l.) (Rivera and others, elevation data and requires an existing DEM. What the
2006). During the winter the AWS was moved to a location tool does is to locate the geographical position of every
on a rock outcrop at 1925 m a.s.l., next to the fixed camera pixel in a photographic image. It is therefore useful in
922 Rivera and others: Instruments and methods

Fig. 2. Glacier variations, 1961–2007, at Volcán Villarrica. The yellow dot shows the summer AWS location. The star shows the location of the
camera, the GPS base station and the AWS in winter. Blue is snow or ice cover, and brown shows areas of permanently tephra-covered ice.

mapping land cover and assessing changes in surface cover. effect of radiation reflected from the surrounding slopes.
The technique follows standard procedures for perspective The final result is a map of normalized reflectance values,
views in computer graphics and photogrammetry (Slama or relative albedo. The atmospheric transmittance was
and others, 1980; Fiume, 1989; Foley and others, 1990). It calculated using the radiative-transfer model MODTRAN
creates a virtual photograph of the DEM that can be scaled (Berk and others, 1989) and general knowledge of the local
to the resolution of the photographic image to establish a atmospheric conditions from the AWS. The photographic
mapping function between pixels in the photograph and image was mapped to a resampled DEM from the Shuttle
gridcell points. This allows the exact position of pixels Radar Topography Mission (SRTM, acquired by JPL (Jet
to be located on the oblique image. The georeferencing Propulsion Laboratory, Pasadena, CA, USA)/NASA in 2000).
process consists of a viewing transformation applied to the It is important to note that the resampling procedure to
DEM, in which the coordinates of every gridcell are firstly a 10 m resolution DEM cannot increase the resolution of
translated to refer them to a coordinate system with the origin the original DEM, at 90 m, but allows the extraction of
at the camera position. A transformation is then applied more information from the photography within the known
according to the viewing direction and focal length of the spatial limits of the original DEM gridcell. The DEM was not
camera. This results in a three-dimensional set of points contemporary with the photographic image acquisition. This
corresponding to the cells in the DEM as seen from the may introduce errors into the results, as the slope and aspect
point of view of the camera. Finally, the resulting viewing may be slightly different from one date to another, and the
transformation is projected into a two-dimensional viewing snow accumulation may change. It is beyond our financial
’window’, corresponding to the area of the film and scaled capability to acquire a high-resolution DEM for every
proportionally. The process is explained in detail by Corripio campaign, but we are working on the design of some terrain
(2004); it has been coded in Interactive Data Language (IDL) measurements that allow the precise evaluation of the errors
under the Creative Commons license and is freely available incurred by using slightly old DEMs, such as simultaneous
from the authors. measurements of albedo at different points and comparison
Using this technique the evolution of the snow cover was with those derived from georeferenced photography.
mapped accurately during the acquisition periods. Once Unfortunately this project was a constant battle against the
the photograph is georeferenced, the reflectance values are elements. Despite testing the kit under laboratory conditions,
normalized according to the viewing geometry, the angle of the camera case was twice flooded during severe weather
incidence of the sun’s rays on the slope, the ratio of direct conditions, which also prevented collection of data on one
to diffuse radiation, the atmospheric transmittance between occasion. The camera set-up was finally destroyed by a
the pixel location and the position of the camera, and the lightning strike in March 2007.
Rivera and others: Instruments and methods 923

Fig. 3. (a) Location of GPS measurements and the AWS in winter and summer. (b) Locations of flags installed on the glacier surface which
were measured using GPS whilst a photograph was acquired for each site.

GPS survey mounted on bedrock was occupied to provide geodetic

quality measurements for differential positioning at the
Several kinematic and static GPS surveys were conducted centimetre level. Thereby, baseline lengths never exceeded
to map the extent of the glacier and to georeference tie a few kilometres. GPS data were post-processed using
points for photogrammetric purposes (Fig. 3). Javad’s dual- the Waypoints GrafNav software package, version 7.70,
frequency GPS receivers and antennas were used exclusively where precise ephemeris and clock information provided by
at a sampling rate of at least 2 s and an elevation cut- the International GNSS (global navigation satellite systems)
off angle of 10◦ . At all times a nearby reference station Service (IGS, final products) was incorporated. In January
924 Rivera and others: Instruments and methods

Table 1. Aerial photographs and satellite images (MSS: multispectral scanner; ASTER: Advanced Spaceborne Thermal Emission and Reflection
Radiometer)

OEA∗ Landsat SAF∗ SAF∗ GeoTec∗ ASTER† ASTER† ASTER†

MSS†

Date 13 Dec 1961 8 Feb 1976 7 Jan 1983 15 Dec 1987 30 Jan 1988 13 Feb 2003 2 Feb 2005 24 Feb 2007
Scale/resolution (m) 1 : 50 000 57 1 : 30 000 1 : 50 000 1 : 70 000 15 15 15

∗ Aerial photographs. † Satellite images.

2005 the local reference station was linked to the SIRGAS estimated from atmospheric humidity and temperature plus
(Geocentric Reference System for the Americas) which is the terrain parameters and ground/snow emissivity according
regional realization of the International Terrestrial Reference to geological characteristics if data are available. Tests on
Frame. complicated surfaces in the Andes, such as penitentes, testify
to the ability of the model to reproduce snow surface
Satellite imagery temperature (Corripio and Purves, 2005). Application to a
In order to update the glacier variations at the volcano, watershed in the Alps for the estimation of meltwater runoff
several satellite images and aerial photographs were used gave values within 6% of measured runoff for a runoff gauge
(Table 1). All the satellite images were geolocated and with a 10% accuracy (unpublished data).
orthorectified using the regular IGM (Instituto Geográfico
Militar) cartography and available DEMs (e.g. SRTM and
airborne synthetic aperture radar (AirSAR)) (Rivera and RESULTS AND DISCUSSION
others, 2006). Once the satellite images were orthorectified, The frontal and areal glacier variations of Glaciar Pichillan-
classification procedures based on spectral band ratios were cahue–Turbio have been updated up to 2007 (Table 2). Both
applied to account for the glacier extent and snow/ice/debris the main arms of the glacier have continued retreating at
classification (Paul and others, 2002). Aerial photographs similar rates, as indicated by frontal length changes measured
were stereoscopically analysed and the resulting information in recent decades (Casassa and others, 2004). These glacier
was transferred with a zoom transfer scope (ZTS) to the tongues are totally debris-covered and only at steep flanks
regular cartography, with ice fronts being digitally compared is bare ice visible, due to the backwasting ablation process
to the satellite (Benson and Follet, 1986). All glacier limits (Fig. 2). In spite of the insulation provided by the ash and
were analysed using GIS (geographical information system) debris covering the ice, the glacier has lost significant area
commercial software, such as IDRISI 32 for Windows, in recent decades, much more than other debris-free glaciers
ArcInfo version 8.0.1 and PCI Geomatica, allowing an located on top of active volcanoes (Rivera and others, 2006).
accurate estimation of areas and frontal changes (Fig. 2). The present extent of Villarrica’s glaciers is shown in Figure 2,
and recent variations are summarized in Table 2.
Energy-balance and melt modelling In order to monitor the glacier at a daily resolution, we
We ran a modelling experiment to assess the effect of tephra used oblique terrestrial photographs that were georeferenced
deposition on the glaciated surface of the volcano and its to a DEM with the help of accurate ground-control points
influence on mass balance and runoff. Using meteorological (GCPs) measured on the glacier surface, as shown in Figure 4.
data for 12 days in January 2007 and the corresponding This tool can be applied to precisely locate snow features on
albedo variations derived from the photographic images, a the surface of the glaciers in areas that are very difficult to
distributed energy-balance model was applied to the eastern access. The eastern upper side of the volcano is constantly
slopes of the Volcán Villarrica. Thus, we could assess the swept by ash falls and toxic gases from the crater fumaroles.
effect of diminished albedo due to volcanic ash deposition This made direct surveying impossible without appropriate
on the mass balance and meltwater production from the protective clothing and breathing masks. Thus remote sens-
volcano glacier. In January 2007 there were no available ing is a more convenient alternative, as shown in Figure 5.
wind-speed data from the AWS. These data were derived The figure shows the changing positions of crevasses on the
from the NOAA (US National Oceanic and Atmospheric upper section of the volcano. The left image is from 25 March
Administration) archived 1◦ resolution aviation (AVN) model 2006 and the right image is from 14 January 2007. The deriv-
outputs for the gridcell corresponding to Volcán Villarrica. ation of flow rates and surface variations using this approach
The data were chosen at a 750 hPa pressure level, which is currently being studied. This preliminary demonstration
corresponds to the middle height of the volcano’s cone. is given to illustrate the potential of the technique for high-
The model applied (SnowDEM: Snow Distributed Energy- temporal-resolution surface monitoring in hazardous or in-
balance Model) is explained in detail by Corripio (2003a). It accessible environments.
is a distributed, multilayered energy-balance model, which The pattern of tephra dispersion is clearly visible on the
takes into consideration radiative fluxes, heat interchange georeferenced image of the volcano of 25 December 2005
with the atmosphere, evaporation or sublimation and heat (Fig. 6). The conical shape of the georeferenced image is
flux due to precipitation. Shortwave radiation is evaluated due to the field of view of the camera. There is a thick tephra
according to a detailed parametric radiative transfer model layer around the crater rim and a band of darker snow to
of the atmosphere plus terrain effects (Corripio, 2003b, the southeast following the prevailing west and northwest
2004; Strasser and others, 2004). Longwave radiation is winds. The amount and area of tephra deposition depends
Rivera and others: Instruments and methods 925

Fig. 4. Image of Volcán Villarrica and superimposed perspective projection of the DEM. Every red dot corresponds to a gridcell in the DEM.
Circles in the insets show the location of precise ground-control points. In the terrain these were marked by flags visible in the original
high-resolution photograph.

on the intensity of fumarolic activity and the concurrent the visible area of the volcano was derived from the photo-
winds. It would be possible to model wind flow across the graphic images and incorporated into the energy-balance
volcano, but it is far more difficult to predict the spatial model. The albedo derived from the images reveals increas-
and temporal variability of fumarolic activity. It is therefore ing values toward the fringes of the visible area (the northern
very difficult to anticipate the ash dispersion pattern. The and southern slopes). We believe this is a realistic result, as
only solution is to observe it at relatively high temporal those slopes are away from the prevailing winds and suffer
resolution and incorporate the results in any model of the less tephra pollution by volcanic fallout. In fact, a visual in-
glacier surface. spection of the northern slopes while skiing Villarrica a week
Here we present the results for a set of 11 images during earlier showed thick ash layers on the uppermost section
a clear-sky period from 5 to 16 January 2007. The albedo of of the volcano near the crater and clean, metamorphosed

Table 2. Frontal and areal changes of Glaciar Pichillancahue–Turbio. Data from 1961–2003 are taken from Casassa and others (2004); those
from 2003–07 are from the present work

Period Length change Annual length change Area change

m m a−1 km2

Glaciar Turbio (−39◦ 25 , −71◦ 53 ) 1961–83 −407 −19 −0.02
1983–87 −43 −9 −0.08
1987–98 −200 −20 −0.35
1998–2003 −679 −135 −0.43
2003–05 stable −0.14
2005–07 −60 −30 −0.05
1961–2007 −1389 −30 −1.07

Glaciar Pichillancahue (−39◦ 26 , −71◦ 52 ) 1961–83 −288 −14 −0.93
1983–87 −51 −11 −0.11
1987–98 −414 −41 −1.05
1998–2003 −52 −10 −0.51
2003–05 −136 −68 −0.24
2005–07 −95 −48 −0.20
1961–2007 −1036 −23 −3.03
926 Rivera and others: Instruments and methods

Fig. 5. Detail of two images showing the changing positions of crevasses on the upper section of the volcano: (a) 25 March 2006; (b) 14 January
2007. Areas not visible from the standpoint of the camera are shown as black.

granular snow below that point to about 1800 m a.s.l. On the surface, with insulation and melt reduction overriding
the upper section snow was fine-grained, interspersed with the influence of albedo lowering (Brock and others, 2007).
small wind deposits of highly broken precipitation particles In our analysis we assume the critical debris thickness to
(Colbeck and others, 1990); on the lower section it consisted correspond to an albedo of 11% (based on the broadband
of larger snow grains with high water content. albedo of andesitic basaltic tephra, ASTER spectral library,
Previous work suggests that the critical ablation threshold http://speclib.jpl.nasa.gov/), below which snowmelt is re-
is passed as soon as the tephra forms a continuous layer at duced relative to bare-snow surfaces. Spurious high values

Fig. 6. Pattern of tephra dispersion on Volcán Villarrica, 25 December 2005. The inset shows the original image. The conical shape of the
georeferenced image is due to the field of view of the camera, located at the lower right-hand corner. There is a thick tephra layer around
the crater rim and a band of darker snow following the prevailing west and northwest winds. Areas of the georeferenced image not visible
from the camera standpoint are shown as white.
Rivera and others: Instruments and methods 927

Fig. 7. Map of cumulative ablation from 5 to 16 January 2007, modelled with albedo derived from daily photographs of the glacier (see
Fig. 6). The spatial variations are due to the pattern of tephra dispersion, which is more concentrated leeward of the prevailing west and
northwest winds. Axes units are 10 m.

may be due to errors in the precise boundary between tephra- differences in ablation of several centimetres over a few
covered and bare-snow areas. These errors are of the order tens of metres. On the vertical transect we can observe
of two pixels of the original DEM resolution at the borders of a clear anomaly, as ablation increases with elevation and
the image, or ±180 m. Areas where the viewing angle is very reaches a local maximum near the crater (distance 0–500 m),
shallow have been masked, but sub-pixel variation in slope where ash depositions are more intense. It then follows an
may add to the error. It is unlikely that the precision can irregular but decreasing trend downslope, as ash gets more
be increased without the use of photogrammetric cameras
and a very up-to-date DEM. The resulting ablation map
(Fig. 7) shows high spatial variability. This variability may
be enhanced by post-depositional reworking of the tephra
Vertical transect
layers, and facilitated by a positive feedback, as surface Along-contour 2500m transect
particles will tend to aggregate while melting on concave
surfaces (Drewry, 1972), which are initially caused by
differential melt. Of particular note are areas of locally
enhanced melting downwind of exposed ash banks in the
lower third of the image, and accelerated melting over large
areas downwind of the crater, which are exposed to almost
continuous airfall tephra deposition. These irregularities are
also discernible on any transect along and across the eastern
slopes of the volcano, as shown in Figure 8. This figure
shows the modelled differences in accumulated ablation for (km)
a transect along the 2500 m elevation contour and along
a transect from the summit crater towards the camera. Fig. 8. Profiles of cumulative ablation along the 2500 m elevation
Along the elevation isopleth the pattern is approximately contour (black) and across a vertical profile from the crater down to
symmetrical, with minimum values towards the slopes that the southeast (grey). The black transect follows the 2500 m contour
are less subject to ash deposition and a maximum on the from the south anticlockwise to the northeast. The origin of the
southeast slopes. The horizontal gradient is rather large, with vertical transect (grey) is the summit crater.
928 Rivera and others: Instruments and methods

Fig. 9. Map of ablation differences between the observed ash-covered volcano and a hypothetically clean snow surface with albedo
parameterized according to Brock and others (2000).

dispersed with distance from the source. There are two peaks glacier variation in melt rates, other than that due to aspect
∼5 km from the crater, associated with older ashes that are and shading, is not incorporated and, in particular, the high
resurfacing after the overlying snow has disappeared. In these spatial variability in melt rates immediately below the crater
areas the tephra is probably thick enough to insulate the and in the vicinity of exposed ash banks is missed (Fig. 9).
underlying ice and reduce melting. Brock and others (2007) This demonstrates that using photographic-derived albedo
measured a mean daily melt rate of 46 mm w.e. at three provides an improved depiction of the actual spatial variation
stakes set on snow with mean albedo 0.51 in the vicinity of surface albedo.
of the AWS, under similar meteorological conditions in the The differences in cumulative ablation between models
second half of January 2004. The measured melt rate (12 day accounting or not for tephra deposition reach peak values of
cumulative melt = 506 mm) corresponds well with modelled >40 cm in some spots, but these peak values are probably
values using photographic-derived albedo in January 2007 an overestimation. The errors in this computation are due to
over large areas of snow with relatively light tephra cover (1) sub-pixel slope and albedo variation; (2) local areas of
(Figs 7 and 8). very shallow viewing angle where the average DEM slope is
The effects of variable deposition, and redistribution of steeper and (3) projection of cones and mounts into adjacent
fallen tephra, on spatial patterns of melt are illustrated by grids due to almost parallel viewing angles. Therefore only
comparison with modelled melt rates when these spatial vari- the areas where the viewing angle of the camera is >60◦
ations are neglected. Figure 9 shows the differences in mod- are reliable. This results in differences in ablation of ∼20–
elled cumulative melt rates for the same period, replacing 30 cm between model runs considering ash cover or clean
photographic-derived albedo with the Brock and others snow. While at the end of the accumulation season the
(2000) ageing-curve albedo parameterization. This param- surface of the glacier is relatively smooth, at the end of the
eterization calculates albedo as a function of temperature ablation season it presents many concavities a few metres
since the most recent snowfall, and assumes albedo decay in depth and a few tens of metres in diameter (Fig. 10). The
is caused by snow metamorphism and the build-up of negative values are areas where the photographic-derived
impurities over time, but does not account for spatial vari- albedo is higher than the parameterized albedo. This is likely
ation in snow impurity content. While the model with to happen in small concave areas which get more shading
parameterized albedo is able to account for some of the from the sun. These small concavities are not registered
along-glacier (vertical) variation in melt rates associated in the DEM; a more precise estimation of their albedo
with slower snow metamorphism at higher elevations, cross- would require a higher-resolution DEM. The irregular snow
Rivera and others: Instruments and methods 929

Fig. 10. Photograph of Volcán Villarrica in spring, showing an irregular snow surface, which is likely to be a combination of ice flow and
crevassing together with differential ablation.

surface is likely to be due to a combination of effects from as a result increase the surface melt rate. The maximum
glacier flow and crevasse formation together with differential values were found near ash/debris hummocks where winds
ablation. blow material onto the snow. However, large areas of
The increase in ablation of snow due to tephra impurities the glacier experience high ablation (low albedo) in the
was estimated by Brock and others (2007) to be 8% at areas located downstream to the east and southeast of
the location of the AWS. The same authors recorded rates the crater, where airfall tephras are frequently deposited.
of melting similar to those produced here for the areas Another interesting result obtained here is the possibility of
with higher ash depositions. Ablation certainly seems higher detection of crevassing at the flanks of the volcanic cone.
further up the slope, as the albedo is lower. This can Changing patterns of crevasses could indicate subglacial
also be seen from the longitudinal transect (Fig. 8). The volcanic activity, as well as glacial ice dynamics. Further
average modelled increase in ablation for the pixels visible research will be needed to extract surface velocities using,
from the camera viewpoint is 13.6% in this study. It is for example, feature tracking in the daily photographs. The
questionable whether this increased ablation and related frontal variations of the glacier have also been updated,
meltwater production might affect the glacier dynamics, confirming previous receding trends, in spite of the thick
given the high porosity of the ground. It is, however, likely insulating debris which mantles most of the glaciers’ ablation
that it has an impact on the groundwater recharge and on zones. Continued glacier retreat in spite of the insulating
runoff further down the valley, with additional implications tephra is likely to be due to a combination of reduced
for the ecological system (e.g. Hauer and others, 1997). A accumulation due to decreasing precipitation trends in the
network of piezometers together with runoff gauges would region (Bown and Rivera, 2007), rapid melt at exposed ice
be the ideal instrumentation to detect this effect on water cliffs along crevasses and possibly enhanced basal melting in
levels, and hopefully will be incorporated in future research areas of high geothermal heating (Rivera and others, 2006).
programs in the region. The role of ash deposition on top of the snow surfaces
has been discussed here, but more detailed analysis of the
mass balance of the glacier needs to be done in order
CONCLUSIONS to understand the specific driving factors for the retreating
A number of photographs were obtained from a fixed camera trend, especially how the decrease in precipitation observed
installed at Volcán Villarrica. These photographs have been in the Chilean lake district (Bown and Rivera, 2007) is
used to incorporate daily spatial albedo variations into a affecting accumulation on the glacier.
glacier-melt model based on AWS data, SRTM elevation
data and GPS measurements on the ground. The modelled
ablation shows significantly higher values than previously ACKNOWLEDGEMENTS
measured on the volcano (Brock and others, 2007) in areas This work was sponsored by Fondo Nacional de Ciencia
directly downwind from ash sources, i.e. the crater and y Tecnologı́a, Chile, (FONDECYT 1040515 and 7050177)
exposed tephra banks. These ashes reduce albedo and and Centro de Estudios Cientı́ficos (CECS), Chile. CECS is
930 Rivera and others: Instruments and methods

funded in part by the Millennium Science Initiative and the In De Jong, C., D.N. Collins and R. Ranzi, eds. Climate and hy-
Centers of Excellence Base Financing Program of Comisión drology of mountain areas. Chichester, J. Wiley and Sons, 15–27.
Nacional de Investigación Cientı́fica y Tecnológica de Chile Drewry, D.J. 1972. A quantitative assessment of dirt-cone dynamics.
(CONICYT). Centro de Ingenier ía de la Innovación del CECS J. Glaciol., 11(63), 431–446.
is funded by CONICYT and the Gobierno Regional de Fiume, E.L. 1989. Mathematical structure of raster graphics. Boston,
MA, Academic Press.
Los Rios. M. Rodriguez, F. Bown, A. Uribe, G. Gazitúa,
Foley, J.D., A. van Dam, S.K. Feiner and J.F. Hughes. 1990. Computer
Corporación Nacional Forestal, Chile, (CONAF), Servicio graphics: principles and practice. Second edition in C . Boston,
Nacional de Geologı́a y Minerı́a, Chile, (SERNAGEOMIN), MA, Addison-Wesley.
C. Acuña and R. Koschitzki helped with data collection, Fountain, A.G. and J.S. Walder. 1998. Water flow through temperate
preparation of some of the figures, field campaigns and the glaciers. Rev. Geophys., 36(3), 299–328.
installation of the equipment. The paper was written while Fountain, A.G., R.W. Jacobel, R. Schlichting and P. Jansson. 2005.
one of the authors (J.G.C.) was funded by a grant from the Fractures as the main pathways of water flow in temperate
Austrian Science Foundation (FWF project No. M952-N10). glaciers. Nature, 433(7026), 618–621.
We are grateful to the scientific editor, M.T. Gudmundson, Fuentealba, G., A.P. Riffo, R.H. Moreno and P. Acevedo. 1984.
and to L. Nicholson and M. Nolan for helpful comments. La erupción del Volcán Villarrica. Temuco, Universidad de la
Frontera.
Fuenzalida-Ponce, R. 1976. The Hudson Volcano. In González-
Ferrán, O., ed. Proceedings of the Symposium on Andean and
REFERENCES Antarctic Volcanology Problems, Santiago, Chile, September
1974. Naples, Stabilimento Tipografico Francesco Giannini &
Adhikary, S., Y. Yamaguchi and K. Ogawa. 2002. Estimation of snow Figli, 78–87.
ablation under a dust layer covering a wide range of albedo. González-Ferrán, O. 1995. Volcanes de Chile. Santiago, Instituto
Hydrol. Process., 16(14), 2853–2865. Geográfico Militar.
Benson, C.S. and A.B. Follett. 1986. Application of photogrammetry Hauer, F.R. and 8 others. 1997. Assessment of climate change
to the study of volcano–glacier interactions on Mt. Wrangell, and freshwater ecosystems of the Rocky Mountains, USA and
Alaska. Photogramm. Eng. Remote Sens., 52(6), 813–827. Canada. Hydrol. Process., 11(8), 903–924.
Berk, A., L.S. Bernstein, and D.C. Robertson. 1989. MODTRAN: Kirkbride, M.P. and A.J. Dugmore. 2003. Glaciological response
a moderate resolution model for LOWTRAN 7. Hanscom, to distal tephra fallout from the 1947 eruption of Hekla, south
MA, Air Force Geophysics Laboratory. (Technical Report Iceland. J. Glaciol., 49(166), 420–428.
GL-TR-89-0122.) Klohn, E. 1963. The February 1961 eruption of Calbuco volcano.
Bown, F. and A. Rivera. 2007. Climate changes and recent glacier Bull. Seismol. Soc. Am., 53(6), 1435–1436.
behaviour in the Chilean lake district. Global Planet. Change, Lara, L.E. 2004. Overview of Villarrica volcano. In Lara, L.E. and
49(1–4), 79–86. J. Clavero, eds. Villarrica Volcano (39.5◦ S), Southern Andes,
Brock, B.W., I.C. Willis and M.J. Sharp. 2000. Measurement and Chile. Santiago, Servicio Nacional de Geologı́a y Minerı́a,
parameterization of albedo variations at Haut Glacier d’Arolla, 5–12. (Boletı́n 61.)
Switzerland. J. Glaciol., 46(155), 675–688. Métaxian, J.-P., S. Araujo, M. Mora and P. Lesage. 2003. Seismicity
Brock, B., A. Rivera, G. Casassa, F. Bown and C. Acuña. related to the glacier of Cotopaxi Volcano, Ecuador. Geophys.
2007. The surface energy balance of an active ice-covered Res. Lett., 30(9), 1483. (10.1029/2002GL016773.)
volcano: Villarrica Volcano, southern Chile. Ann. Glaciol., 45, Moreno, H. 2000. Mapa de Peligros Volcánicos del volcán
104–114. Villarrica. (Scale 1 : 75 000.) Santiago, Servicio Nacional de
Calder, E.S. and 6 others. 2004. Combined thermal and seismic Geologı́a y Minerı́a. (Documento de trabajo 17.)
analysis of the Villarrica volcano lava lake, Chile. Rev. Geol. Naranjo, J.A. and H. Moreno. 2004. Laharic debris-flows from
Chile, 31(2), 259–272. Villarrica Volcano. In Lara, L.E. and J. Clavero, eds. Villarrica
Casassa, G., C. Acuña, R. Zamora, E. Schliermann and A. Rivera. Volcano (39.5◦ S), Southern Andes, Chile. Santiago, Servicio
2004. Ice thickness and glacier retreat at Villarrica Volcano. Nacional de Geologı́a y Minerı́a, 28–45. (Boletı́n 61.)
In Lara, L.E. and J. Clavero, eds. Villarrica Volcano (39.5◦ S), Paul, F., A. Kääb, M. Maisch, T. Kellenberger and W. Haeberli.
Southern Andes, Chile. Santiago, Servicio Nacional de Geologı́a 2002. The new remote-sensing-derived Swiss glacier inventory.
y Minerı́a, 53–60. (Boletı́n 61.) I. Methods. Ann. Glaciol., 34, 355–361.
Clavero, J. and H. Moreno. 2004. Evolution of Villarrica Volcano. Petit-Breuihl, M.E. and J. Lobato. 1994. Análisis comparativo de la
In Lara, L.E. and J. Clavero, eds. Villarrica Volcano (39.5◦), cronologı́a eruptiva histórica de los volcanes Llaima y Villarrica
Southern Andes, Chile. Santiago, Servicio Nacional de Geologı́a (38◦ –39◦ ). In Proceedings of VII Congreso Geológico Chileno,
y Minerı́a, 17–27. (Boletı́n 61.) Concepción, Chile, Vol. 1, 366–370.
Colbeck, S.C. and 7 others. 1990. The international classification for Rivera, A. and 8 others. 2006. Ice volumetric changes on active
seasonal snow on the ground. Wallingford, Oxon, International volcanoes in southern Chile. Ann. Glaciol., 43, 111–122.
Association of Scientific Hydrology. International Commission Slama, C.C., C. Theurer and S.W. Henriksen, eds. 1980. Manual
on Snow and Ice. of photogrammetry. Fourth edition. Falls Church, VA, American
Corripio, J. 2003a. Modeling the energy balance of high altitude Society of Photogrammetry and Remote Sensing.
glacierised basins in the Central Andes. (PhD thesis, University Stern, C.R. 2004. Active Andean volcanism: its geologic and
of Edinburgh.) tectonic setting. Rev. Geol. Chile, 31(2), 161–206.
Corripio, J.G., 2003b. Vectorial algebra algorithms for calculating Strasser, U., J. Corripio, F. Pellicciotti, P. Burlando, B. Brock
terrain parameters from DEMs and solar radiation modelling in and M. Funk. 2004. Spatial and temporal variability of me-
mountainous terrain. Int. J. Geogr. Inf. Sci., 17(1), 1–23. teorological variables at Haut Glacier d’Arolla (Switzerland)
Corripio, J. 2004. Snow surface albedo estimation using terrestrial during the ablation season 2001: measurements and simulations.
photography. Int. J. Remote Sensing, 25(24), 5705–5729. J. Geophys. Res., 109(D3), D03103. (10.1029/2003JD003973.)
Corripio, J. and R. Purves. 2005. Surface energy balance of high Witter, J.B. and P. Delmelle. 2004. Acid gas hazards in the crater of
altitude glaciers in the central Andes: the effect of snow penitents. Villarrica Volcano (Chile). Rev. Geol. Chile, 31(2), 273–277.

MS received 30 April 2008 and accepted in revised form 6 June 2008

View publication stats