Content uploaded by Cristina García-Hernández
Author content
All content in this area was uploaded by Cristina García-Hernández on Mar 25, 2024
Content may be subject to copyright.
1
REFORESTATION AND LAND USE CHANGE AS DRIVERS FOR A
DECREASE OF AVALANCHE DAMAGE IN MID-LATITUDE MOUNTAINS
(NW SPAIN)
Cristina García-Hernández1, Jesús Ruiz-Fernández1, Covadonga Sánchez-Posada2,
Susana Pereira3, Marc Oliva3, Gonçalo Vieira3
1 Department of Geography, University of Oviedo, C/ Amparo Pedregal s/n, 33011, Oviedo, Spain.
2 Department of Statistics, Operational Research and Mathematics Didactics, University of Oviedo, Calvo Sotelo s/n,
33007, Oviedo, Spain.
3 Centre for Geographical Studies, Institute of Geography and Spatial Planning, Universidade de Lisboa, R. Branca
Edmée Marques, 1600-276 Lisboa, Portugal.
Corresponding author
Cristina García-Hernández, cristingar@hotmail.com
Department of Geography, University of Oviedo
C/ Amparo Pedregal s/n, Oviedo, 33011, Spain
Tel: +34 985104405
2
HIGHLIGHTS
• Evolution of snow avalanche damage is studied over a 215-year period.
• A synthetic damage index is created, mapping its variations over time and space.
• The evolution of damage is closely related to human-induced changes in vegetation
cover.
• Deforestation explains the concentration of damages in the late 19th century.
• Changes in land use management explain the decrease of the damage, mainly since
the 1960s.
3
ABSTRACT
Natural conditions that explain the triggering of snow avalanches are becoming better-
known, but our understanding of how socio-environmental changes can influence the
occurrence of damaging avalanches is still limited. This study analyses the evolution of
snow avalanche damage in the Asturian Massif (NW Spain) between 1800 and 2015,
paying special attention to changes in land-use and land-cover patterns. A damage index
has been performed using historical sources, photointerpretation and fieldwork-based
data, which were introduced in a GIS and processed by means of statistical analysis.
Mapping allowed connecting spatiotemporal variations of damage and changes in
human-environment interactions. The total number of victims was 342 (192 dead and
150 injured). Results show stability in the number of avalanches during the study
period, but a progressive decrease in the damage per avalanche. Changes in land use
explain the evolution of damage and its spatial/temporal behaviour. The role played by
vegetation cover is at the root of this process: damage was the highest during the late
19th and early 20th centuries, when a massive deforestation process affected the
protective forest. This deforestation was the result of demographic growth and intensive
grazing, disentailment laws and emerging coal mining. Since the mid-20th century, the
transformation of a traditional land-management system based on overexploitation into
a system based on land marginalization and reforestation, together with the decline of
deforestation due to industrial and legal causes, resulted in the decrease of avalanches
that affected settlements (mostly those released below the potential timberline). The
decrease of damage has been sharper in the western sector of the Asturian Massif,
where oak deforestation was very intense in the past and where lithology allows for a
more successful ecological succession at present. Taking into account that reforestation
can be observed in mountain environments of developed countries worldwide, and
considering present initiatives conducted to counteract its negative cultural effects by
means of grazing and clearing operations, planning is imperative, and this research
provides useful information for environmental management policies and risk mitigation
in avalanche prone areas.
Keywords: Snow avalanches; Deforestation; Reforestation; Land-use change; Damage
index; Cantabrian Mountains.
4
1. INTRODUCTION
Snow avalanches are a major slope process in mountain environments, posing a
significant threat in the coolest and highest regions of the world (Schneebeli et al.,
1997; Johannesson and Arnalds, 2001; Fuchs et al., 2004; Keiler et al., 2005; Podolskiy
et al., 2014). In the mountain environments of the Iberian Peninsula, snow avalanches
are frequent; the Pyrenees, Cantabrian Mountains and Central Range are periodically
affected by this kind of events (Mases and Vilaplana, 1991; Furdada et al., 1995; Julián
and Chueca, 1999; Muntán et al., 2009; Fernández-Cañadas et al., 2015). In the
Cantabrian Mountains only a few studies have addressed this issue (Castañón, 1984;
Wozniak and Marquínez, 2004; Santos et al., 2010; Vada et al., 2012; García-
Hernández et al., 2014; Serrano et al., 2016). This poor understanding of avalanches and
their triggering factors is particularly significant in the case of the Asturian Massif, the
highest and largest sector of Cantabrian Mountains, where this research focuses.
The consequences of hazards are usually evaluated taking into account the degree of
loss or the level of damage derived from a specific hazard (UNISDR, 2009) that can be
estimated in terms of monetary units or social values, accounting the number of dead
and injured people (Alexander, 1996), affected buildings (Merz et al., 2004) or
infrastructures (Geertsema et al., 2009), or using a synthetic index to quantify personal
injuries, damages to property and disruptions of economic activities (Fekete, 2009;
Pereira et al., 2016). Only a few studies have specifically quantified the effects of
changes in population, exposure of people and assets, and vulnerability as determinants
of loss (IPCC, 2012). In addition, since exposure and vulnerability vary over time and
space depending on economic, social, geographic, demographic, cultural, institutional,
governance and environmental factors (IPCC, 2012), if we want to understand the
dynamics of natural hazards, we must analyse how human and natural systems interact
in the long-term (Mileti and Myers, 1997). Measuring the degree of influence that
human-induced modifications have on damage ratio of events may provide a better
prediction on future effects of these modifications on the natural systems. Moreover, not
only avalanche risk, but exposure to that risk, is variable in time due to transformations
affecting the human system and its links to the natural system, such as changes in land
use and social vulnerability of inhabitants (Fuchs et al., 2005; Keiler et al., 2005). In
this sense, it is important to emphasize the importance that induced changes on
5
vegetation cover, as a consequence of land use changes, could have on the increase of
the avalanche risk (Gubler and Rychetnik, 1991).
The development of natural hazard databases is decisive for risk management because
they allow identifying the relationships between the occurrence of dangerous natural
phenomena and the existence of exposed and vulnerable population. Such databases
also enable a quick access to data that can be analysed using the most recent advances in
geostatistical and modelling techniques providing avenues for risk mitigation. Historical
documents have been increasingly employed in natural hazard research (Pfister et al.,
2006; Raška et al., 2014; Pereira et al., 2016) and several research studies have stressed
the potential use of documentary data sources to reinforce the knowledge on snow
avalanche events, illustrating the strong interaction between economic and social
development and snow avalanche activity (Irwin and Owens, 2004; Hétu et al., 2011;
Podolskiy et al., 2014; Hétu et al., 2015). Since the beginning of the 21st century,
policies for disaster risk reduction have shifted from structural mitigation measures to a
more comprehensive and integrated approach in which mapping is an essential tool for
better understanding natural disaster dynamics (EEA, 2010). Geographical location and
extent of hazards serve the purposes of associating reported impacts with socio-
economic and environmental contextual parameters (Peduzzi et al., 2005). Thus, hazard
mapping is essential in order to assess the causes and consequences of natural disasters,
and it provides a solid basis for rational land-use planning (Kienholz, 1978).
The purposes of this paper are to: i) Map and analyse the spatial and temporal
distribution of snow avalanches, as well as its level of damage in the Asturian Massif
from 1800 to 2015; ii) Examine temporal variability of the damage as a result of
changes on the interactions between human and natural systems; iii) Identify the key
factors explaining the occurrence of damaging avalanches or influencing its
consequences; iv) Assess, in particular, how induced changes in vegetation cover from
changes in land uses, could have influenced the variation of damage over time.
2. STUDY AREA
The Asturian Massif (Fig. 1) makes up the western area of the Cantabrian Mountains,
located in the NW of the Iberian Peninsula. This massif is essentially composed of
sedimentary and metamorphic rocks from the Palaeozoic Era, folded and/or fractured by
6
two orogenies, the Hercynian and the Alpine (Julivert, 1983; Martínez-Catalán et al.,
1992). The lithology is very mixed, showing a predominance of siliceous materials
(mainly sandstone, slate and quartzite) in the western part of the massif, and an
alternation of calcareous and siliceous sedimentary rocks in the central and eastern
parts. The Picos de Europa, one of the main limestone massifs in the world (Ruiz-
Fernández et al., 2016), is located in the latter part.
The topography of the Asturian Massif shows a substantial north-south dissymmetry,
with considerably steeper and sharper descents in the northern slope. Its highest areas
generally exceed 2,000 metres, and the highest altitudes are reached in the Picos de
Europa (Torrecerredo, 2,648 metres). Also, this massif contains a wide surface
developing above 1,500 metres, as well as numerous upland areas between 1,000 and
1,500 metres. Therefore, this mountain area is characterised by the existence of large
altimetric differences, very enclosed valleys and rather steep hillsides.
The northern slope of the Asturian Massif is characterised by an oceanic climate, with
mild average temperatures (between 10ºC and 14ºC), and heavy rains (between 1,400
and 1,500 mm annually in general, and even more than 2,000 mm annually in the
highest areas) spread evenly throughout the year (Muñoz-Jiménez, 1982). Conversely,
the southern slope has a transitional oceanic-Mediterranean climate, characterised by
annual average temperatures below 10ºC, rains exceeding 1,000 mm in the highest areas
only, and summer dryness for one or two months. This moderate and mostly humid
climate allows the development of Atlantic deciduous forests where the natural tree line
is placed at 1600 m a.s.l. on average (Ruiz-Fernández et al., 2008), 1450/1650 m a.s.l.
in the northern slopes and 1700/1850 m a.s.l. in the southern aspect (Muñoz-Sobrino et
al., 2012).
Human presence in the mountainous areas of the Asturian Massif has been traditionally
connected with livestock farming (Suárez et al., 2005), which has contributed to the
mobility of human groups between the higher and lower areas of the massif, and
between its northern and southern slopes through mountain passes. Numerous
infrastructures can be seen in these areas, the construction of which experienced a
substantial increase from 1850 on. Historically, the mountainous villages of the
Asturian Massif have been isolated and densely populated, which enhanced the
7
potential damages of avalanches to people and property. Although recent
socioeconomic changes have led to a lower demographic pressure in most of the highest
inhabited lands, there has been a parallel improvement of the road network which has
brought more visitors to the highest areas of the massif, where the avalanche hazard is
more noticeable. Therefore, in this massif, the potential for damage to recreation and
transport during the cold season is remarkable and, even though it has caused large
socio-economic damages, the history of snow avalanche accidents has not been
reconstructed.
Fig. 1
3. MATERIALS AND METHODS
3.1. Obtaining the data
The compilation of the data series of avalanches that occurred between 1800 and 2015
has been carried out examining historical sources, such as newspapers and church
documents. For the former, we have analysed 32 newspapers at the regional and
national level covering the period under analysis. From these, 891 issues were manually
surveyed focusing on the news published between the months of November and April,
which reflects the snow season. The sources were the hardcopy microfilms from the
Central Library of Asturias and the National Library of Spain. Further 3208 issues have
been examined through virtual newspaper libraries using the Spanish key-words
“avalancha” (avalanche) and “alud de nieve” (snow mass movement). The procedure
resulted in the identification of 273 news related with snow avalanches in the Asturian
Massif. The wide criteria used to select the news aimed at guaranteeing a wide spatial
distribution and to avoid gaps. Moreover, we have examined the records of church
documents (records of deceased people) from 62 mountain parishes across the Asturian
Massif. These documents, written between 1800 and 1960, allowed identifying other 18
damaging events which had not been reported in newspapers, resulting in a total of 291
events.
The paths of 126 of these avalanches were reconstructed in the field through
geomorphological evidence and testimonies of residents and witnesses. The criterion for
8
selecting these 126 avalanches was good data availability. The analysis of present
vegetation formations colonising the starting areas was conducted by means of photo-
interpretation and fieldwork (as a way to verify the findings of the photointerpretation).
Four categories of vegetation cover were assessed: forest, shrubland, grassland, sparse
vegetation. We have also used aerial photography, ancient photographs and prints to
support field observations.
3.2. Conducting the geostatistical analysis
Data on avalanche location, characteristics and consequent damages were stored in a
database management system in order to facilitate the subsequent statistical analysis.
All 291 historical avalanches were introduced in a GIS with a point in the geometric
centroid of the depletion area. The analysis was based on a high resolution Digital
Elevation Model (MDT05 of the Spanish National Geographic Institute, with a 5m grid
size). The boundaries of the 126 events with well-documented avalanche path were
drawn as polygons (depletion, transportation and accumulation areas). For each
avalanche, data on place and date of occurrence, together with information about people
and affected structures, were also compiled. Besides, regarding the 126 avalanches with
a well-documented avalanche path, we also collected and analysed data referring to the
starting point (aspect and release angle θ), average angle over the run of the avalanche
(α), and distance to the impact point with people or structures.
For data processing and figures, we have used the free software environment R.
Statistical analysis has been performed through classical descriptive statistics and
several techniques of robust statistics: median, 25th and 75th percentiles as measures of
central tendency; interquartile range as a measure of spread; winsorised means; Tukey’s
method instead of classic method of mean +/- 3*standard deviation. These types of
techniques prevent results from being affected by outliers. Tukey’s method (Tukey,
1977) has been used for outlier detection, characterizing extreme values of the
distribution as “high” (Eq. 1) or “far high” (Eq.2):
High = 75th percentile (P75) + 1.5*Intercuartile Range (IQR)……………………Eq. 1
Far-high = P75+3*IQR…………………………………………………………….Eq.2
9
We made use of statistical inference to detect statistically significant differences
between groups. Parametric techniques (the Student’s t-test, and non-parametric
methods as the Wilcoxon rank-sum test) have been employed as well.
3.3. Calculating a damage index
To assess the evolution of the damage resulting from avalanche activity, a damage index
has been performed. These kinds of indices are indirect numerical surrogates of
phenomena (Fekete, 2009) that allow for quantifying numerically the damage. To
calculate the index we have considered eight dimensions of damage that have been
incorporated into two components: the personal and the material damage (PD and MD).
These two components have been synthesised on a total damage (TD) index, depending
on equation 3:
TD= 0.8*PD+0.2*MD………………………………………………………………Eq. 3
PD and MD depend on the weight of their related dimensions (Table 1). These
dimensions have been selected and weighted using as reference framework specific
criteria used by reinsurers and institutions such as the Munich Reinsurance Company
and the Centre for Research on the Epidemiology of Disasters (Below et al., 2009).
When any of the conditions exposed in table 1 is fulfilled, the avalanche is considered
to be damaging.
Table 1
3.4. Spatial and temporal analysis
The study period (1800-2015) has been divided in three well-defined stages in the
territorial dynamics of the Asturian Massif: 1800-1899 (rural), 1900-1959 (mostly rural
with early urban industrialization) and 1960-2015 (rural abandonment with gradual
tourism increase) (García-Fernández, 1980; Ortega-Valcárcel, 1989; Fernández-García,
2006; Rodriguez-Pascual, 2006). These stages are used to explain damage index shifts
over time, establishing connections with socio-economic, demographic and
environmental changes. The territory was divided into three sectors (Western, Central
and Eastern), in order to analyse those causes of index evolution related to the different
10
territorial dynamics across the massif. Moreover, we established four typologies of land
use affecting the avalanche areas in the moment of the events (settlements, agro-silvo-
pastoral, road or railroad, recreational).
3.5. Connecting the damage index with social factors
In order to eliminate the effect that population oscillations could have on the evolution
of the index, we examined the evolution of the number of fatalities (parameter with the
greatest weight in the calculation of the TD index) per 100,000 inhabitants, compared
with the evolution of the number of avalanches. For the population analysis we used
data since 1870s for five mountain municipalities regularly affected by avalanches
(Somiedo, Cangas del Narcea, Lena, Villamanín and Cabrales) (Fig. 1). These
municipalities summed 59% of the avalanches, 50% of the victims and 57% of TD, and
its economic and demographic dynamics are uniform. This population sample was used
as a measure of the inhabitants exposed to avalanche hazard.
Finally, the main processes considered in the discussion have been included in a
timeline (Fig. 17), to connect them with the evolution of the TD index. A base-1 log
scale has been used in order to represent those processes in the same scale, enabling the
evaluation of their evolution with regard to their own highest and lowest values over
time. Sources used to perform this timeline have been the Statistical Yearbook 1858-
2015, National Forest Inventory 1966-2007, Mining Statistics of Spain 1861-2010,
Asturian Society of Economic and Industrial Studies and Spanish Statistical Office.
4. RESULTS
4.1. Evolution of the number of the damaging avalanches
A total of 291 damaging avalanches were documented for the period 1800 to 2015. No
decreasing trend is detected in the number of avalanches by stages (Fig. 2a). In fact, the
stage 1960-2015 recorded the highest percentage of years with damaging avalanches
(48%) as well as the highest percentage of years recording more than five events (13%)
(Fig. 2b).
11
Fig. 2
4.2. Avalanche-related damages evolution
The total number of victims was 342 (192 dead, 150 injured), with the maximum
concentration reached during the stage 1800-1899 (100 dead, 62 injured) and
dramatically declining in the last stage 1960-2015 (20 dead, 50 injured) (Fig. 3a).
Between the two first stages there is no substantial difference in the survival rate, which
remains at around 38%. The remarkable change is observed during the last stage, when
the survival rate reaches 66% (Fig. 3b).
Fig. 3
Regarding the number of victims, a decreasing trend over time is observed (Fig. 4). The
western sector of the Asturian Massif encompasses 37% of the total victims, the eastern
36% and the central 27%. The distribution of these percentages vary across the three
stages, with a clear decline in the western sector (49-38-7%) and increases in the central
(29-23-43%) and especially in eastern areas (22-39-50%).
Fig. 4
A progressive decrease of the index of average damage per avalanche is detected.
According to Student’s t-test (p-values <0.0001, in both cases), the differences in the
averages are significant between the first and second and, also, between the second and
third stages (Fig. 5a y 5.b). These differences have been confirmed by the Wilcoxon
rank-sum test (p-values <0.0001 and 0.0007), which also shows a staggered decrease by
stages of the TD index, as observed in the figure 2a.
Fig. 5
12
Using the criteria of Tukey to analyse the annual aggregate of avalanches events, the
number of deaths, number of victims, number of victims per 100,000 inhabitants and
number of deaths per 100,000 inhabitants, we defined two types of outlier years: years
far-high and high (Table 2). Outlier years concerning fatalities and victims are
concentrated in the first stage, except for 1910. From 1910 onwards no atypical far-high
year is recorded, with the exception of 2015, which recorded a large number of
avalanches.
Table 2
The normalised values of the number of avalanches per municipalities area are larger in
the central sector of the Asturian Massif, with 0.085 events/km2 (0.040 in the eastern
and 0.020 in the western parts). The western sector concentrates 39% of the index, 30%
the eastern and 31% the central. Spatial distribution of the number of avalanches and
personal and material index values according to the total index values per municipality
and stage are presented in Fig. 6a. A general decline of the TD is observed between the
first stage (index value 1109) (Fig. 6b) and the third (210.4) (Fig. 6d), though the
decrease is more pronounced in the western sector: it includes 46% of the index in the
first stage but only 14% in the third, in the central part this trend is more moderated (35
vs. 28%), while in the east is reversed (19 vs. 58%). As for the percentage distribution
of PD and MD, the PD is significantly higher in the last stage, namely in the center-east
sectors (Fig. 6d).
Fig. 6
4.3. Influence of environmental drivers and land-use on avalanche hazard
Fig. 5a shows that the index of average damage per avalanche decreases significantly
towards recent times, but this pattern is different if we consider the land use of the
affected area at the time of the incident. TD index distribution according to the land use
13
of the affected area during the three stages is presented in Fig. 7. This distribution
shows an inverted "V" with further damage during the intermediate stage, or without
significant differences between stages in the case of road and railroad typology.
Fig. 7
Half of the avalanches has affected communication routes (Fig. 8a). However, if we
look at the evolution of the phenomena over time, we observe that the settlements have
been the most affected land use during the first stage, while the road and railroad
communication networks were the most disturbed after 1900, when settlements started
to be less affected. During the third stage, between 1960 and 2015, the recreational
category includes 19% of the damages generated by avalanches. The settlements are the
spaces that concentrate more PD (68% of the index value throughout the study period)
(Fig. 8b). This fact is more evident in the first stage, in which settlements recorded more
than 80% of damage, subsequently descending until 21% of the damages produced
during the last stage. Avalanches affecting recreational experienced the opposite trend,
since they only appear in the last stage representing 35% of damage and the agro-silvo-
pastoral category recorded its most remarkable peak in the stage of 1900-1959.
Fig. 8
The settlements concentrated the highest percentage of personal victims (49%) and
fatalities (59%) (Fig. 9a, 9b) as well as the highest number of victims and fatalities per
avalanche, and the lowest survival rate (Table 3). The recreational land use type
concentrates less personal victims and fatalities, with the lowest number of victims and
fatalities per avalanche, and the highest survival rate (Table 3).
Fig. 9
14
Table 3
Using the criterion of Tukey to the distribution of the TD index, 36 extreme avalanches
were detected. Settlements stand out as the space where 83% of these avalanches have
impacted. Spatial distribution of the number of avalanches according to the land use of
the affected areas by municipality and stage are represented in Fig. 10. Avalanches
affecting the roads and railroads are the majority throughout the entire study period and
for most of the territory, especially in the central part of the massif where they constitute
70% of the total (compared to 30-40% in the other sectors) (Fig. 10a). The avalanches
affecting settlements represent 22% of those occurred in the eastern sector and 34% of
the western sector, where this type of damaged space represents the higher percentage;
however, 85% of avalanches in settlements in the western sector occurred in the first
stage (77% in the central sector and 61% in the eastern) (Fig. 10b). Avalanches
affecting the agro-silvo-pastoral space have been more numerous in the eastern and
western areas (28% and 22% of the total), while in the central part represent only 4%
(Fig. 10a). Finally, avalanches affecting recreational areas, that only appear in the last
stage, mainly occurred in the eastern part, where they represent 17% of the events
(compared to 3-4% in the other areas) (Fig. 10d).
Fig. 10
4.4. Avalanche damage and population variation
The evolution of the TD index is compared to population in Fig. 11. Variations of the
TD index follow, with a slight gap, the same pattern of population: between 1874 and
1888 the population rose 18% and between 1888 and 1895 the index rose 20%; between
1918 and 1931 the population grew by 27% and between 1931 and 1935 the index rose
by 27%. Similarly, when the population stabilizes and begins to fall, the index
stagnates: between 1900s and 1960s the population grew by 60% and the index by 66%,
whereas between 1960s and 2015 the population declined 66% and the index increased
only by 7%.
15
Fig. 11
The evolution of the number of fatalities per 100,000 inhabitants, compared with the
number of avalanches, shows that prior to 1870 the occurrence of fatalities was higher
than the avalanches (Fig. 12). The growth followed a similar rate from 1870 until the
1960s, when the avalanche occurrences began to grow at a higher rate than the number
of fatalities, which slowed down. This pattern is remarkable between 2000 and 2010,
when the difference became more significant. While the growth rate of fatalities showed
decreasing trends throughout the three stages (49.7%; 49.0%; 14.2%), the growth rate of
damaging avalanches show increasing trends (99.0%; 121.7%; 212.7%), which suggests
that the number of documented damaging avalanches is not reduced but, those
avalanches, gradually, caused fewer fatalities.
Fig. 12
4.5 Spatial distribution of the avalanches and their consequences for damage
The upper level of the dense forest in the Asturian Massif is located, on average, at 1600
m a.s.l. Regarding the 126 avalanches with a well-documented avalanche path, 4% of
avalanches were triggered and deposited above the timberline (AT), 33% were triggered
above the timberline and deposited below the timberline (AT-BT) and 63% were
triggered and deposited in the forest environment, below the timberline (BT). Up to
88% of AT avalanches took place from 1960 to 2015, while 70% of AT-BT and 75% of
BT occurred before 1960. AT concentrated 1% of the total damage, while AT-BT and
BT included 46% and 53%, respectively.
The average distance travelled until impacting the target was 932 m, with a maximum
of 2,831 m and a minimum of 96 m. There was a significant relationship (according to
the Student’s t-test) between distance and damage: the average TD of the avalanches
travelling distances greater than P75 (>1,375 m) was 18.9; in the group that travelled
medium distances, between P25 and P75 (476-1,375 m) was 9.9, and in the smallest
16
ones, travelling distances lower than P25 (<476 m) was 3.2 (Fig. 13). Large avalanches
occurred mainly during the first stage (41%), showing the lowest proportion in the latter
stage (24%), while the rest of avalanches were distributed evenly over time. Up to 58%
of the damage in settlements was due to BT avalanches (28% to BT avalanches that
started between 1,400 and 1,600 m) and the impact of those small to medium
avalanches releasing below the timberline was maximum (Table 4).
Fig. 13
Table 4
The vegetation colonizing nowadays the release areas of the 126 avalanches with a
well-documented avalanche path is shrubland in 34% of the cases, grassland in 25% and
forest in 21% of the study cases. Another 20% of release areas present sparse
vegetation. 50% of the forested release areas have not been functional since 1900 and
up to 75% since the mid-1940s (Fig. 14a). Those avalanches which release area is now
forested, are the oldest avalanches and they have the highest average damage rate (19)
(Fig. 14b). The release areas colonized by sparse vegetation have been functional
mostly during last stage (Fig. 14a) and show the lowest average damage rate (5) (Fig.
14b).
Fig. 14
With regards to the aspect of the release areas, the most frequent aspect was the
southern, especially SW (28%), with few events in northern aspects (Fig. 15a). The
most damaging avalanches occurred in eastern-exposed slopes (average rate of 15%),
while the least damaging took place in western aspects (average rate of 8%).
Avalanches in southern aspects concentrate 61% of the index (Fig. 15b).
17
Fig. 15
If we split the avalanches in three groups depending on the distribution of TD (high,
medium and low damage), we find that the damage increases as the angle α decreases
(Fig. 16a). The average angle α in the high damage group was 29.4° (31.4° in the
medium damage group and 33.4° in the low damage group). Up to 75% of avalanches
of the high damage group occurred in slopes below 32°. This trend is attenuated for
angles θ: the average angle θ in the high damage group was 33.6°, and in the case of the
medium damage and low damage groups was 35.7° (Fig. 16b). Both the angles α
(average angle 29.5°) as θ (average angle 32.7°) are lower in the settlements than in the
rest of affected spaces (Fig. 16c; 16d).
Fig. 16
5. DISCUSSION
5.1. Temporal evolution of the damaging avalanches
While in mountains such as the Alps, the number of damaging avalanches has decreased
over time due to significant mitigation efforts (Schneebeli et al., 1997; SLF, 2000;
Fuchs et al., 2004), however, in the Asturian Massif, no decrease is observed per decade
for the late 20th and early 21st centuries. This may be due to the fact that natural
avalanche activity has not significantly varied over time, as it is the case of the Alps
(Schneebeli et al., 1997; Laternser and Schneebeli, 2002). However, contrary to the
Alps, in the Asturian Massif, there has been an absence of mitigation measures.
Although the number of damaging avalanches in the Asturian Massif has not decreased,
a significant and progressive decrease is detected, both of the TD index and the damage
index per avalanche (Fig. 5A). The decrease of the damages caused by snow avalanches
over recent decades is a worldwide trend (United Nations, 2002). However, the
distribution of the avalanche damage index, considering the land use, shows an increase
18
between the first and the second stage (Fig.7). Therefore, the progressive decline
observed in the global patern is only explained by the change detected in land uses:
while in the first stage 55% of damaging avalanches occurred in settlements (the land
use in which the average damage was the highest), in the second stage this type only
accounted for 15% of the events (Fig.8A), prompting the other categories to gain
relative importance, causing the decline of the TD index.
Settlements concentrated most of the victims caused by avalanches (with the lowest
survival rate), the most extreme events and most of the TD index, namely during the
stage 1800-1899 (which produced almost half of the total victims). Specifically, the last
third of the 19th century recorded the highest number of extreme avalanches with the
highest number of injured and deaths. This exceptionally damaging period corresponds
to the last decades of the Little Ice Age, an interval between 1570 and 1900 when
temperatures fell significantly below the 1961-1990 mean in many mountain regions of
the Northern Hemisphere (Matthews and Briffa, 2005). During the second half of the
19th century, there are widespread evidence in historical Iberian sources of colder
temperatures together with intensification of climate variability and extreme events
(Álvarez-Vazquez, 1986; Barriendos and Martín-Vide, 1998).
The stage 1960-2015 (in which the number of avalanches that affected settlements was
substantially lower), concentrated only 20% of the victims and there was a significant
increase of the survival rate (66%). Even taking into account the possibility that the
survival rate of the past has been affected by a sub-register, as it happens in other
studies on natural hazards inferred from historical documents (Podolskiy et al., 2014), in
this study, avalanches in settlements show the lowest survival rate recorded in
comparison to the rest of land uses. This explains the high damages recorded in
southern aspects since, in the Asturian Massif, during the Middle Ages settlements used
to be placed close to the land suitable for cultivation, preferably with orientations S, SE
and SW (García-Fernández, 1980). Therefore, the evolution of damage in settlements
determines the long-term decrease of the damages from the first stage to the last.
5.2. Changing spatial patterns of avalanche damage
The damage decline over time has been widespread throughout the Asturian Massif,
though the decrease has been more pronounced in the western sector. In turn, in all
19
areas most of the avalanches affecting settlements occurred during the first stage (1800-
1899). Personal and material damage started increasing (not in general but in relative
terms, compared to the other sectors) in the eastern side of the massif between 1900 and
1959, and increased further during the third stage, when the central area of the massif
remained stable. During the third stage, the recreational typology increased
significantly, especially in the eastern sector. The Picos de Europa National Park
(containing the Montaña de Covadonga area, the oldest protected environment in Spain)
is one of the most appreciated areas by mountaineers, especially since the 1960s (Suarez
et al., 2005). This region includes a large part of the eastern sector of the Massif and
concentrates the highest peaks and steeper slopes of the Cantabrian Mountains, with the
highest snow accumulation rates (Muñoz-Jiménez, 1982). These conditions determine a
high natural avalanche activity in a highly frequented area by mountaineers, which
explains the increased occurrence of avalanches related with recreational activities in
this sector since the 1960s. This kind of avalanches usually causes more personal
injuries than material damages, which explains why in the eastern sector the former has
been higher than material damage during the last stage. On the other hand, avalanches
affecting mountain climbers explain the concentration of AT avalanches, as well as the
concentration of triggering areas in which vegetation is sparse, that have been active in
the last stage (1960-2015).
In the central sector of the massif, there are three of the most transitated mountain
passes and three of the four ski resorts existing in the Asturian Massif. This enhances
the potential of damage, mainly through avalanches affecting roads, although an over-
representation of damage may exist in this sector, due to its greater socio-economic
interest (Raška et al., 2014). By contrast, the western sector does not have neither the
transport infrastructure density of the central sector nor the tourist attraction of the
eastern sector. Also, almost the totality of the damage caused by avalanches belonging
to the agro-silvo-pastoral category occurred during the first two stages and affected to
peripheral environments of the massif where economy was completely based on the
primary sector until mid-20th century (García-Fernández, 1980). Other research
examining similar study periods in regions of Australia (Irwin and Owens, 2004), North
America (Hétu et al., 2011) and Eurasia (Podolskiy et al., 2014), show the same
progressive reduction in the number of snow avalanches affecting settlements and work
20
environments, and an increase of avalanches related with recreational activities during
the second half of the 20th century.
5.3. Historical and geographical background for avalanches until 1960
The decrease in the number of avalanches affecting settlements is key for the evolution
of the general damages. This decrease could have been partly driven by population
fluctuations, since the Asturian mountains have been subject to a depopulation process
since the mid-20th century (Fernández-García, 2006; Rodríguez-Pascual, 2006). The TD
index responds with a slight delay to the demographic performance. However, the fact
that the number of fatalities decreases regardless of population decrease (Fig.12), even
when we do not dispose of data about current floating population (what would reinforce
our conclusion at this regard), suggests that other historical factors (not directly but
indirectly related to population dynamics) have played a more critical role.
Interestingly, 63% of the recorded avalanches occurred below the potential timberline.
This must be framed with the fact that the dense forest stabilizes snow cover and limits
avalanches (Gubler and Rychetnik, 1991; Newesely et al., 2000; Feistl et al., 2014).
This type of avalanches was responsible for 53% of total damage between 1800 and
2015, although only 8% occurred after 1960. Consequently, changes in vegetation cover
must be taken into account in order to interpret this spatio-temporal pattern. Until the
mid-20th century, grazing was very intense in the Cantabrian Mountains (García-
Fernández, 1980; Rodriguez-Pascual, 2006). Pastures induced a lowering of the
timberline and developed even in steep slopes close to villages, forest covered only
about a quarter of the potential forest area and tree-felling was frequent even in the
remaining forest (García-Fernández, 1980). Besides, since the early 18th century the
northern mountains of the Iberian Peninsula showed population growth peaking in late-
19th century (Anglada et al., 1980). In this context of rising demographic pressure, the
disentailment process was initiated by a number of liberal laws from 1812, promoting
the sale of public forests to private sector that was free for tree-felling, both for grazing,
agriculture, and for industrial purposes (Martí, 2003). This promoted the most
widespread deforestation in Spanish history, which extended until the end of the 19th
century (Bauer, 1980). Although large areas were excluded from sale in the northern
slopes of the massif, privatization in mountainous areas was more intense since 1860s,
21
because the Catalogue of Public Utility Woodland enacted strict requirements to
exclude from sale public mountain forests, requirements remaining until the late 19th
century (Fernández-García, 2006). Also, the onset of mining activities in the region
since the 1850s demanded large quantities of native wood, exceeding 200,000 m³ per
year in the 1910s (Manuel et al., 2003) and causing extended deforestation (Pérez-
González, 1980).
Other authors have linked deforestation as a result of increased population pressure and
industrial activities, with increasing avalanches in Eurasia (Podolskiy et al., 2014) and
North America (Hétu et al., 2015). One of the most heavily deforested areas in the
Asturian Massif was the Western sector, which concentrated the largest oak forests. Oak
was the less excluded species with successive disentailment laws, and it had a great
interest for mining roof support (Sendín, 1996; Manuel et al., 2003). In Asturias, oak
forests owned by the state were almost entirely privatized during this period (Martí,
2003). Sessile oak (Quercus petraea), which grows in the highest areas of the massif,
had been traditionally protected by the inhabitants of the highest lands, who considered
their presence an effective tool against avalanches (Lastra, 2002). In fact, dense oak
forests meet the requirements to be considered protective forest as they are irregular and
pluri-specific forests in which young and old specimen coexist (Gubler and Rychetnik,
1991), allowing the proliferation of saplings in lower stratum as they are open forests.
Hence, tree-felling in the forests of the Western sector, especially in the great oak forest,
explains concentration of damages there, especially during the late 19th century and,
once tree-felling and grazing started declining, the decrease of damage was remarkable.
In addition, we must take into account that oaks tend to preferentially be on south-
facing slopes, which are generally the most heavily affected by human action on the
Cantabrian Mountain (García et al., 2005), a fact that fits with a higher concentration of
damage in south orientations.
5.4 Recent environmental changes driving avalanche dynamics
From the mid-20th century, the Spanish State conducted massive reforestation efforts
with fast growing forest species, and the use of native trees from mountainous areas of
the massif decreased, being replaced by species such as Eucalyptus globulus and Pinus
pinaster (Morales, 1982). At the same time, mining activities started to decline (Pérez-
González, 1980), so the demand for native wood (and therefore also deforestation)
22
decreased markedly. Oak forests were reforested as well, multiplying their surfaces by
four (Manuel et al., 2003).
An important factor contributing to reduce avalanche damage, especially in settlements,
was the abandonment of the traditional management system and subsequent land-use
changes in mountain environments. Since the mid-20th century, there has been a
regeneration of natural vegetation due to the decrease of the importance of farming
activities; this is a process that can be observed in mountain environments of many
developed countries of Europe, America and Asia (Baudry and Bunce, 1991; Rudel et
al., 2000; Piussi, 2000; Müller and Zeller, 2002; Gellrich, 2007; Cramer et al., 2008). In
the Iberian Peninsula, this process has been mainly examined in the Pyrenees, where
there was a progressive land marginalization because of the decrease of stocking rate,
and grazing focused on flat environments located closer to the villages (Lasanta and
Ruiz, 1990). This led to a vegetation colonization process which last stage is the passive
reforestation of formerly grazed areas (Lasanta et al., 2000). In the Asturian Massif, this
phenomenon has been equally intense (Zúazua et al., 1985; Cascos, 2011; Álvarez-
Martínez et al., 2013). Here, there was an abandonment of the most distant farm lands
and pastures (Rodríguez-Pascual, 2006) that resulted in the recovery of the forest. In the
first decade of 21st century, the main mountain municipalities of the massif had
increased their forested area by 15% (on average) with respect to the 1960s (Bengoa,
2011). In this period, there has been a significant expansion and densification of the
forest at higher elevations (García de Celis et al., 2004).
Apart from the lower demographic pressure, another factor to be taken into account is
the progressive specialisation in bovine breeding, which intensified after 1940
(Fernández-García, 2006). Since then, most of the livestock consisted of foreign cattle
breeds for milking, unfit for grazing steep areas (Ortega-Valcárcel, 1989). Shepherds
almost disappeared together with goat flocks, which were expelled from many areas
because they were considered “enemies of the forest” (Rodríguez-Pascual, 2006). The
decrease of goat grazing in steeper slopes, where they usually grazed exerting an
important biomass control (Osoro et al., 2000), has promoted the extension of protective
forest (Gubler and Rychetnik, 1991; Bebi et al., 2001). These social processes are
displayed in a timeline and related with the evolution of the damage index (Fig. 17).
23
Fig. 17
5.5 Vegetation recovery and interactions between social and natural systems
Once grazing is abandoned and deforestation decreases, the avalanche activity depends
mainly on the resilience of the vegetation (Germain et al., 2005). In this sense, land-use
intensity and time-lag following the abandonment will be of crucial importance, as well
as the frequency and magnitude of avalanches, since they can constrain the growth of
trees (Tasser et al., 2007). In the 126 avalanches with a well-documented avalanche
path, the damage increased as the average slope gradient was lower; the most damaging
avalanches (specifically those affecting settlements, which are in areas with lower slope
angles) were released on moderated slopes, in which avalanches are not frequent,
allowing for a more successful succession of vegetation. This is confirmed by the fact
that the forests currently cover the majority of the most severely affected areas in the
past, namely before 1950.
Another important aspect to consider is the lithology. While the limestones of the
eastern sector of the massif do not allow for the development of deep soils, the bedrock
of the western sector composed of quartzite, sandstone and slates favours the formation
of well-developed soils and enhances a faster re-vegetation process. Spontaneous
reforestation processes have been especially active between 1,400 and 1,600 m a.s.l., an
elevation belt intensely used for grazing activities (García-Fernández, 1980; Ruiz-
Fernández et al., 2008) where several avalanches occurred causing significant damage
to settlements.
Other factors, besides forest expansion, could explain the decrease of damage since the
braking effect of the forest on avalanche propagation is very efficient in the case of
small-to medium-sized avalanches released below the timberline (Teich et al., 2012),
though it is limited for large avalanches starting above it (Bartelt and Stöckli, 2001;
Margreth, 2004; Teich et al., 2012; Feistl et al., 2014). Such other factors may be the
abandonment or under occupation of settlements affected by these avalanches.
Therefore, the decrease in damage caused by large avalanches that started above the
timberline cannot be explained only by forest growth.
24
Apart from the protective effect produced by the presence of the forest, revegetation
may also has negative effects, since it increases the potential for fires (Pausas and
Keekey, 2009), decreases the landscape heterogeneity and even the biodiversity
(Hunziker, 1995; Lasanta et al., 2011), although this last point is highly debated at the
moment (Blanco-Fontao et al., 2011). Considering all of the above, at present, efforts
are being made to counteract this process (Lasanta, 2011). Some authors have noted that
grazing activities can diminish the protective capacity of the forest (Gubler and
Rychetnik, 1991), although in recent years grazed forests have been also used to stop
avalanches (Bebi et al., 2001; Mayer and Söckli, 2005). In addition, frequent human-
induced fires conducted in order to control revegetation can cause wildfires, which in
turn can reactivate avalanches near settlements and roads (Santos et al., 2010).
Therefore, well-managed and controlled grazing can be compatible with the
maintenance of the protective effect of the forest and can be a useful tool to prevent
both fires and avalanches.
5. CONCLUSIONS
From 1800 to 2015, the number of damaging avalanches in the Asturian massif has not
decreased, but there has been a progressive decrease in damage. The decline of damage
per avalanche is explained by changes in land use of affected areas from settlements to
recreational uses and transportation networks, in which damages are less severe. This
phenomenon explains the spatial behaviour of damages.
The transformation of a traditional land-management system based on overexploitation
into a system based on land abandonment and marginalization, has been the main factor
influencing the evolution of damage over the last fifty years, by means of passive
reforestation. Before this, massive felling and clearing had led forests to lose their
function as protective structures against avalanches; this process was particularly
intense during the late 19th and early 20th centuries because of demographic growth,
disentailment laws and emerging coal mining. This explains the rise of damage during
these decades, namely because of avalanches affecting settlements, which released
mostly below the potential timberline. Since the mid-20th century, population decline
and reduction of livestock activity, together with a sharp drop in demand of native
wood, allowed the recovery of vegetation, which experienced the consecutive stages of
25
succession resulting in dense, protective forests. Forest succession was more effective in
less steep areas, which were also the areas where the most damaging avalanches,
affecting settlements, occurred.
The deceleration of a more intense deforestation, in parallel to a more successful
process of natural vegetation recovery, favoured by the existence of deeper soils,
resulted in a sharper and earlier decline of damage in the Western sector of the Asturian
Massif. Since the 1960s, the damage per avalanche started an evident decrease in the
whole territory, but the Eastern sector gained relative importance by absorbing the
majority of damage caused by avalanches related to recreation and sports in mountain
environments, and the relative importance of the Central sector remained stable since it
contained most of the infrastructures.
The effects of the revegetation process are currently being highly discussed by the
scientific community. The results of this study highlight one of its possible positive
consequences: the reduction of avalanche damage. Nevertheless, taking into account
other effects of this process (which may increase the potential for fires and has been
widely considered negative from the cultural point of view), planning is imperative.
Plans must be carried out to preserve protective forests, but these actions can be
combined with controlled grazing in forests assessing potential release areas, so we can
get all the advantages of those clearing actions without sacrificing the existence of the
protective effects of the forest against snow avalanches.
ACKNOWLEDGEMENTS
This research was supported by the FPU program of Ministerio de Educación, Cultura y
Deporte under doctoral fellowship of Cristina García-Hernández [grant number MECD-
15-FPU14/01279]. Susana Pereira was funded by the project FORLAND –
Hydrogeomorphologic risk in Portugal: driving forces and application for land use
planning [grant number PTDC/ATPGEO/1660/2014], Fundação para a Ciência e a
Tecnologia (FCT).
REFERENCES
Alexander, D. (1996). The health effects of earthquakes in the mid-1990s. Disasters, 20,
231-6.
26
Álvarez-Martínez, J., Gómez-Villar, A., Lasanta, T. (2013). The use of goats grazing to
restore pastures invaded by shrubs and avoid desertification: a preliminary case
study in the Spanish Cantabrian Mountains. Land Degradation & Development,
27: 3-13.
Alvarez-Vazquez, J.A. (1986). Drought and rainy periods in the province of Zamora in
the 17th, 18th and 19th centuries. In: Lopez-Vera, F. (Ed.), Quaternary Climate in
Western Mediterranean, Universidad Autonoma, Madrid, pp. 221-235.
Anglada, S., Balcells, E., Creus, J., García-Ruiz, J. M., Martí, C.E., Puigdefábregas, J.
(1980). La vida rural en la montaña española: orientaciones para su promoción.
Consejo Superior de Investigaciones Científicas, Universidad de Zaragoza, Jaca.
Barriendos, M., Martin-Vide, J. (1998). Secular climatic oscillations as indicated by
catastrophic floods in the Spanish Mediterranean coastal area (14th–19th
centuries). Climatic Change, 38 (4), 473-491.
Bartelt, P., Stöckli, V. (2001). The influence of tree and branch fracture, overturning
and debris entrainment on snow avalanche flow. Annals of Glaciology, 32, 209-
216.
Bauer, E. (1980). Los montes de España en la historia. Ministerio de Agricultura.
Madrid.
Baudry, J., Bunce, R.G.H. (1991). Land abandonment and its role in conservation.
Options Méditerranéennes, série A: Séminaires Méditerranéennes nº 15.
CIHEAM, Zaragoza.
Bebi, P., Kienast, F.,Schönenberger, W. (2001). Assessing structures in mountain
forests as a basis for investigating the forests’ dynamics and protective function.
Forest Ecology and Management, 145 (1), 3-14.
Below R., Wirtz A., Guha-Sapir D. (2009). Disaster Category Classification and Peril
Terminology for Operational Purposes. Common accord. Centre for Research on
the Epidemiology of Disasters and Munich Reinsurance Company. CRED
Working Paper, 264. CRED: Brussels, Belgium.
Bengoa, J. (2011). Cambios en el paisaje en la Cordillera Cantábrica (Asturias,
Cantabria y Castilla y León) en los últimos 40 años: comparativa de superficies
arboladas a partir del mapa forestal. In: Ezquerra, F.J., Rey, E. (Coords.), La
evolución del paisaje vegetal y el uso del fuego en la Cordillera Cantábrica, Juan
de Castilla y León, Valladolid, 185-198.
Blanco-Fontao, B., Quevedo, M., Obeso, J. R. (2011). Abandonment of traditional uses
in mountain areas: typological thinking versus hard data in the Cantabrian
Mountains (NW Spain). Biodiversity and Conservation, 20(5), 1133-1140.
Cascos, C. (2011). La simplificación y la uniformidad crecientes en los paisajes
ganaderos de la montaña húmeda de Castilla y León: el ejemplo de Babia y Luna.
Polígonos. Revista de Geografía, 21, 149-178.
Castañón, J.C. (1984). Sobre el modelado originado por los aludes de nieve en el Prau
del Albo (Alto Huerna, Asturias). Ería, 6, 106-112.
Cramer, V.A., Hobbs, R.J., Standish, R.J. (2008). What's new about old fields? Land
abandonment and ecosystem assembly. Trends in Ecology & Evolution, 23 (2),
104-112.
European Environment Agency: Mapping the impacts of natural hazards and
technological accidents in Europe. An overview of the last decade (2010), EEA,
Copenhagen.
Feistl, T., Bebi, P., Teich, M., Bühler, Y., Christen, M., Thuro, K., Bartelt, P. (2014).
Observations and modeling of the braking effect of forests on small and medium
avalanches. Journal of Glaciology, 60 (219), 124-138.
27
Fekete, A. (2009). Validation of a social vulnerability index in context to river-floods in
Germany. Natural Hazards and Earth System Sciences, 9 (2), 393-403.
Fernández-Cañadas, J.A., Palomo, M., Pantoja, L. (2015). Delimitación espacial de las
zonas probables de salida de aludes en el macizo de Peñalara mediante el uso de
SIG, Espacio Tiempo y Forma. Serie VI, Geografía, 6-7, 73-94.
Fernández-García, F. (2006). La quiebra del modelo histórico de organización de la
Montaña Cantábrica. In: Delgado-Viñas, C. (Ed.). La Montaña Cantábrica, una
montaña viva; Cátedra Cantabria, Parlamento de Cantabria y Universidad de
Cantabria; Santander, pp. 79-98.
Fuchs, S., Bründl, M., Stötter, J. (2004). Development of avalanche risk between 1950
and 2000 in the municipality of Davos, Switzerland. Natural Hazards and Earth
System Science, 4 (2), 263-275.
Fuchs, S., Keiler, M., Zischg, A., Bründl, M. (2005). The long term development of
avalanche risk in settlements considering the temporal variability of damage
potential. Natural Hazards and Earth System Science, 5, 893-901.
Furdada, G., Martí, G., Oller, P., García, C., Mases, M., Vilaplana, J.M. (1995).
Avalanche mapping and related G.I.S. applications in the Catalan Pyrenees.
Surveys in Geophysics, 16, 681-693.
García, D., Quevedo, M., Obeso, J. R., Abajo, A. (2005). Fragmentation patterns and
protection of montane forest in the Cantabrian range (NW Spain). Forest Ecology
and Management, 208(1), 29-43.
García de Celis, A.J., Guerra Velasco, J.C., Martínez Fernández, L.C. (2004). Los
abedulares de la Omaña Alta (León): notas sobre dinámica vegetal y cambios en
el aprovechamiento de los montes. Boletín de la Asociación de Geógrafos
Españoles, 38, 245-260.
García-Fernández, J. (1980). Sociedad y organización tradicional del espacio en
Asturias, Silverio Cañada, Gijón.
García-Hernández, C., Ruiz-Fernández, J., Sánchez de Posada, C., Poblete, M.A.
(2014). El impacto del episodio avalanchoso de 1888 en el Macizo Asturiano, a
través de la prensa. In: Gómez-Ortiz, A., Salvador, F., Oliva, M., Salvá, M. (Eds.),
Avances, métodos y técnicas en el estudio del periglaciarismo, Universidade de
Barcelona, Barcelona, 55-64.
Geertsema, M., Schwab, J.W., Blais-Stevens, A., Sakals, M.E. (2009). Landslides
impacting linear infrastructure in west central British Columbia. Natural Hazards,
48 (1), 59-72.
Gellrich, M., Baur, P., Koch, B., Zimmermann, N.E., (2007). Agricultural land
abandonment and natural forest re-growth in the Swiss mountains: a spatially
explicit economic analysis. Agriculture Ecosystems and Environment, 118, 93-
108.
Germain, D., Filion, L., Hétu, B. (2005). Snow avalanche activity after fire and logging
disturbances, northern Gaspe Peninsula, Quebec, Canada. Canadian Journal of
Earth Sciences, 42, 2103-2116.
Gubler, H., Rychetnik, J. (1991). Effects of forests near the timberline on avalanche
formation. In: Snow, Hidrology and Forest in High Alpine Areas, Proceedings of
the Vienna Symposium, IAHS, 205, 19-37.
Hétu, B., Brown, K., Germain, D. (2011). Les victimes d’avalanche au Québec entre
1825 et 2009. The Canadian Geographer/Le Géographe Canadien, 55 (3), 273-
287.
Hétu, B., Fortin, G., Brown, K. (2015). Climat hivernal, aménagement du territoire et
dynamique des avalanches au Québec méridional: une analyse à partir des
28
accidents connus depuis 1825. Canadian Journal of Earth Sciences, 52 (5), 307-
321.
Hunziker, M. (1995). The reforestation in abandoned agricultural lands: perception and
aesthetical assessment by local and tourists. Landscape and Urban Planning, 31,
399-410.
IPCC. (2012). Managing the Risks of Extreme Events and Disasters to Advance Climate
Change Adaptation, Special Report of the Intergovernemental Painel on Climate
Change. Field, C.B., Barros, V., Stocker, T.F., Dahe, Q. (Eds.). Cambridge
University Press, Cambridge.
Irwin, D., Owens, I. (2004). A history of avalanche accidents in Aotearoa, New
Zealand. In: Proceedings of the International Snow Science Workshop. Jackson
Hole, WY, pp. 484-491.
Johannesson T., Arnalds, T. (2001). Accidents and economic damage due to snow
avalanche and landslide in Iceland. Jökull, 50, 81-94.
Julian, A., Chueca, J. (1999). Cartografía de zonas probables de aludes en el valle de
Ordesa (Pirineo aragonés). Geographicalia, 37, 73-86.
Kienholz, H. (1978). Maps of geomorphology and natural hazards of Grindelwald,
Switzerland: scale 1: 10,000. Arctic and Alpine Research, 10 (2), 169-184.
Keiler, M., Zischg, A., Fuchs, S., Hama, M., Stötter, J. (2005). Avalanche related
damage potential – changes of persons and mobile values since the mid-twentieth
century, case study Galtür. Natural Hazards and Earth System Science, 5, 49-58.
Lasanta, T. (1990). Diversidad de usos e integración espacial en la gestión tradicional
del territorio en las montañas de Europa occidental. In: García-Ruiz, J.M., (Ed.),
Geoecología de las áreas de montaña, Geoforma, Logroño, pp. 235-266.
Lasanta, T., Ruiz, P. (1990). Especialización productiva y desarticulación espacial en la
gestión reciente del territorio en las montañas de Europa occidental. In: García-
Ruiz, J.M., (Ed.), Geoecología de las áreas de montaña, Geoforma, Logroño, pp.
267-295.
Lasanta T., Vicente-Serrano S.M., Arnáez J. (2011). La revegetación en áreas de
montaña. ¿Dejar hacer o intervenir en el territorio? Geographicalia, 59-60, 199-
211.
Lastra, J.J. (2002). Bosques naturales de Asturias. Servicio de Publicaciones de la
Universidad de Oviedo. Oviedo.
Laternser, M., Schneebeli, M. (2002). Temporal trend and spatial distribution of
avalanche activity during the last 50 years in Switzerland. Natural Hazards, 27,
201-230.
Manuel, C.M., Díaz, P., Gil, L.A. (2003). Tercer inventario forestal nacional 1997-
2006: La transformación histórica del paisaje forestal en Asturias. Ministerio de
Medio Ambiente, Dirección General de Conservación de la Naturaleza,
Subdirección General de Montes, Madrid.
Margreth, S. (2004). Die Wirkung des Waldes bei Lawinen. Schutzwald und
Naturgefahren. Forum für Wissen 2004, Birmensdorf, pp. 21–26.
Martí, F. (2003). La desamortización española. Ediciones Rialp. Madrid.
Mases, M., Vilaplana, J.M. (1991). Zonas de aludes en la Vallferrera: clasificación y
riesgo geomorfológico. Pirineos, 138, 39-52.
Matthews, J. A., Briffa, K. R. (2005). The ‘Little Ice Age’: Re‐evaluation of an
evolving concept. Geografiska Annaler: Series A, Physical Geography, 87 (1),
17-36.
29
Mayer, A. C., Stöckli, V. (2005). Long-Term Impact of Cattle Grazing on Subalpine
Forest Development and Efficiency of Snow Avalanche Protection. Arctic,
Antarctic, and Alpine Research, 37 (4), 521-526.
Merz, B., Kreibich, H., Thieken, A., Schmidtke, R. (2004). Estimation uncertainty of
direct monetary flood damage to buildings. Natural Hazards and Earth System
Science, 4 (1), 153-163.
Mileti, D., Myers, M.F. (1997). A bolder course for disaster reduction: imagining a
sustainable future, Revista Geofisica, 47, 41-58.
Morales, G. (1982). La explotación de los recursos forestales en Asturias. Ería, 3, 35-
54.
Müller, D., Zeller, M. (2002). Land use dynamics in the central highlands of Vietnam: a
spatial model combining village survey data with satellite imagery interpretation.
Agricultural Economics, 27, 333-354.
Muntán, E., García, C., Oller, P., Martí, G., García, A., Gutiérrez, E. (2009).
Reconstructing snow avalanches in the Southeastern Pyrenees. Natural Hazards
and Earth System Sciences, 9, 1599-1612.
Muñoz-Jiménez, J. (1982). Geografía Física. El relieve, el clima y las aguas. In: Quirós,
F. (Ed.), Geografía de Asturias I, Ayalga, Oviedo.
Muñoz-Sobrino, C., Suarez-Perez, F.J., Nava-Fernandez, H. S., Fernandez-Casado,
M.A., Gomez-Orellana, L., Rodriguez-Guitian, M. A., Fernandez-Prieto, J. A.,
Ramil-Rego, P. (2012). Environmental changes in the westernmost Cantabrian
Range during the postglacial period: the Pena Velosa (Muniellos, Asturias) pollen
record. In Environmental Changes and Human Interaction along the Western
Atlantic Edge (pp. 79-94). Associação Portuguesa para o Estudo do Quaternário
Coimbra.
Newesely, C., Tasser, E., Spadinger, P., Cernusca, A. (2000). Effects of land-use
changes on snow gliding processes in alpine ecosystems. Basic and Applied
Ecology, 1 (1), 61-67.
Ortega-Valcárcel, J. (1989). La economía de montaña, una economía de equilibrio.
Ería, 19, 115-128.
Osoro, K., Vassallo, J.M., Celaya, M.R., Martínez, A. (2000). Resultados de la
interacción vegetación por manejo animal en dos comunidades vegetales naturales
de la Cordillera Cantábrica. Investigación agraria. Producción y sanidad
animales, 15, 137-158.
Pausas, J.G., Keeley, J.E. (2009). A burning story: the role of fire in the history of life.
BioScience, 59 (7), 593-601.
Peduzzi, P., Dao, H., Herold, C. (2005). Mapping disastrous natural hazards using
global datasets. Natural Hazards, 35 (2), 265-289.
Pérez-González, R. (1980). Industria, población y desarrollo urbano en la cuenca central
hullera asturiana. PhD Thesis, University of Oviedo, Oviedo.
Pereira, S., Ramos, A.M., Zêzere, J.L., Trigo, R.M., Vaquero, J.M. (2016). Spatial
impact and triggering conditions of the exceptional hydro-geomorphological event
of December 1909 in Iberia. Natural Hazards and Earth System Sciences, 16 (2),
371-390.
Pfister, C., Weingartner, R., Luterbacher, J. (2006). Hydrological winter droughts over
the last 450 years in the Upper Rhine basin: a methodological approach.
Hydrological Sciences Journal, 51, 966-985.
Piussi P. (2000). Expansion of European mountain forests. In: Price, M.F., Butt, N.,
(Eds.), Forests in Sustainable Mountain Development: A State of Knowledge
Report for 2000. CABI Publishing, Wallingford, UK, pp. 19-25.
30
Podolskiy, E. A., Izumi, K., Suchkov, V. E., Eckert, N. (2014). Physical and societal
statistics for a century of snow-avalanche hazards on Sakhalin and the Kuril
Islands (1910–2010). Journal of Glaciology, 60 (221), 409-430.
Raška, P., Zábranský, V., Dubišar, J., Kadlec, A., Hrbáčová, A., Strnad, T. (2014).
Documentary proxies and interdisciplinary research on historic geomorphologic
hazards: a discussion of the current state from a central European perspective.
Natural hazards, 70, 705-732.
Rodríguez-Pascual, M. (2006). Evolución de la ganadería y el paisaje en la zona
meridional de la Cordillera Cantábrica. In: Delgado-Viñas, C. (Ed.), La Montaña
Cantábrica, una montaña viva. Universidad de Cantabria, Santander, pp.197-205.
Rudel, T.K., Perez-Lugo, M., Zichal, H., 2000. When fields revert to forest:
development and spontaneous reforestation in post-war Puerto Rico. Professional
Geographer, 52, 386-397.
Ruiz-Fernández, J., González-Trueba, J.J., Poblete, M.A. (2008). La montaña media
cabraliega. In: Ruiz-Flaño, P., Serrano, E., Poblete, M.A., Ruiz-Fernández, J.
(Eds.): De Castilla al mar. La naturaleza del paisaje en la Montaña Cantábrica.
Santander, Asociación de Geógrafos Españoles – Universidad de Valladolid –
Universidad de Oviedo, pp. 99-124.
Ruiz-Fernández, J., Oliva, M., Cruces, A., Lopes, V., Freitas, M.C., Andrade, C.,
García-Hernández, C., López-Sáez, J.A., Geraldes, M. (2016). Environmental
evolution in the Picos de Europa (Cantabrian Mountains, SW Europe) since the
Last Glaciation. Quaternary Science Reviews, 138, 87-104.
Santos-González, J., Redondo-Vega, J.M., Gómez-Villar, A., González-Gutiérrez, R.B.
(2010). Los aludes de nieve en el Alto Sil (Oeste de la Cordillera Cantábrica,
España). Cuadernos de investigación geográfica, 36, 7-26.
Schneebeli, M., Laternser, M., Ammann, W. (1997). Destructive snow avalanches and
climate change in the Swiss Alps. Eclogae geologicae Helvetiae 90, 457-461
Sendín, M.Á. (1996). Asturias en el Segundo Inventario Forestal. Ería, 39-40, 131-150.
Serrano, E., Gómez-Lende, M., Pisabarro, A. (2016). Nieve y riesgo de aludes en la
montaña cantábrica: el alud de Cardaño de Arriba, Alto Carrión (Palencia).
Polígonos. Revista de Geografía, (28), 239-264.
SLF (2000). Der Lawinenwinter 1999. Eidenössisches Institut für Schnee-und
Lawinenforschung, Davos.
Suárez, F., Herrán, M., Ruiz-Fernández, J. (2005). La adaptación del hombre a la
montaña. El paisaje de Cabrales (Picos de Europa). Ería, 68, 373-389.
Tasser, E., Walde, J., Tappeiner, U., Teutsch, A., Noggler, W. (2007). Land-use
changes and natural reforestation in the Eastern Central Alps. Agriculture,
Ecosystems & Environment, 118 (1), 115-129.
Tukey J.W. (1977). Exploratory Data Analysis. Addison-Wesley, London-Amsterdam-
Don Mills-Ontario-Sydney.
Teich, M., Bartelt, P., Grêt-Regamey, A., Bebi, P. (2012). Snow avalanches in forested
terrain: Influence of forest parameters, topography, and avalanche characteristics
on runout distance. Arctic, Antarctic, and Alpine Research, 44 (4), 509-519.
UNISDR (2009). Terminology on disaster risk reduction. UN, Geneva, Switzerland.
United Nations (2002). Living with Risk, A Global Review of Disaster Reduction
Initiatives, Preliminary Version, United Nations, Geneva, Switzerland.
Vada, J.A., Frochoso, M., Vilaplana, J.M. (2012). Evaluación y cartografía del riesgo de
aludes en el camino PR-PNPE 21 de acceso a la Vega de Urriellu, Picos de
Europa (Noroeste de España). Cuaternario y Geomorfología, 26 (1-2), 29-47.
31
Wozniak, E., Marquínez, J. (2004). Evaluación de la susceptibilidad por aludes a escala
regional: el caso de Asturias. In: Benito, G., Díez-Herrero, A. (Eds.), Riesgos
Naturales y Antrópicos en Geomorfología. Sociedad Española de Geomorfología
– Centro de Ciencias Medioambientales (CSIC), Madrid, pp. 509-518.
Zuazúa, M.T., Luis, E., Navascúes, E. (1985). Primeras etapas de la sucesión del
pastizal en campos de cultivo abandonados en las tierras altas de León. Pastos, 15,
9-19.
32
1. Figures
Fig. 1. Location of the study area and the sites mentioned along the text.
Fig. 2. (A) Percentage of years, avalanches and damage (TD index), per stage; (B)
Percentage of years with zero, one to five and more than five damaging avalanches, per
stage.
33
Fig. 3. (A) Number of personal victims (killed and injured) caused by damaging
avalanches per stage; (B) ratio of survivals, per stage.
Fig. 4. Number of personal victims (killed and injured) per affected municipality and
period and number of avalanches per municipality.
34
Fig. 5. (A) Distribution of TD index per avalache, per stage; (B) Confidence intervals at
the 95% confidence level.
35
Fig. 6. Number of avalanches and personal and material index values according to the
total index values per municipality and period; (A) complete period 1800-2015; (B)
1800-1899; (C) 1900-1959; (D) 1960-2015.
36
Fig. 7. TD index distribution according to the land use of the affected area, per stage.
Fig. 8. (A) Percentage of avalanches and (B) percentage of total damage index
according to the land use of the affected areas and periods.
37
Fig. 9. (A) Percentage of personal victims (killed and injured) and (B) percentage of
fatalities according to the land use of the affected areas.
38
Fig. 10. Number of avalanches according to the land use of the affected areas by
municipality and stage; (A) complete period 1800-2015; (B) 1800-1899; (C) 1900-
1959; (D) 1960-2015.
39
Fig. 11. Evolution of the TD cumulative index compared to population in selected
Asturian Massif municipalities from 1874 to 2014.
Fig. 12. Evolution of the cumulative number of avalanches compared to the number of
killed per 100,000 inhabitants.
40
Fig. 13. Distribution of TD index according to the travelled distance until impacting the
target of the 126 avalanches with a well-documented avalanche path.
Fig. 14. According to the present physiognomy of the vegetation types colonising the
starting areas of the 126 avalanches with a well-documented avalanche path: (A)
Distribution of years with damaging avalanches, and (B) distribution of TD index.
41
Fig. 15. According to the aspect of the starting areas of the 126 avalanches with a well-
documented avalanche path: (A) Percentage of avalanches, and (B) percentage of TD
cumulative index.
Fig. 16. Cumulative distribution of α (A) and θ (B) angles, depending on the damage,
and cumulative distribution of α (C) and θ (D) angles, depending on the land use of the
126 avalanches with a well-documented avalanche path.
42
Figure 17. Main factors affecting the evolution of the cumulative index of avalanche
damage in the Asturian Massif from 1860 to 2015. This timeline (A) can be related with
a conceptual model of the two types of slope profile corresponding to the
overexploitation and deforestation period (previously to 1960) (B), and to the land
abandonment and revegetation period (after 1960) (C), respectively.
43
2. Tables
Table 1. PD and MD dimensions and corresponding assigned weight used to obtain TD
index.
Dimensions of damage
Assigned weight
Dead
4
80%
PD
Injured
2
Homeless
1
Home
4
20% MD
Farm Building
3
Train
2
Infrastructure
2
Head of cattle
1
Road vehicle
1
Road/railroad outage
1
Forest disturbance
1
Table 2. Outlier years concerning the annual aggregate of avalanche events, fatalities,
personal victims (injured and killed), fatalities per 100,000 inhabitants, and personal
victims per 100,000 inhabitants.
Year
Nr of
avalanches
Nr of
victims
Nr of victims/
100.000 inh.
Nr of
fatalities
Nr of fatalities/
100.000 inh.
1857
-
20
13.1
-
4.6
1871
-
27
17.3
19
12.2
1888
67
60
33.3
28
15.6
1895
19
24
13.3
18
10.0
1910
-
28
13.2
17
8.0
1935
-
17
6.6
8
-
1945
14
16
-
8
-
2005
17
-
-
-
-
2015
25
-
-
-
-
Table 3. Personal victims and fatalities per avalanche, and ratio of survival according to
the land-use of the affected area.
Land-use
Personal victims
Fatalities
Ratio of survivals
Settlement
2.3
1.5
32%
Agro-silvo-pastoral
1.7
0.9
49%
Recreational
1.6
0.5
67%
Road/rairoad
0.4
0.2
56%
44
Table 4. Percentage of damage per land-use, according to the altitude of the starting area
(determining whether the avalanches start above or below the potential timberline) and
according to the size of avalanches (based on their travelled distance).
% Damage
% Damage
Large sized
avalanches
Small-medium
sized avalanches
All sizes
avalanches
>1600
m a.s.l.
≤1600
m a.s.l.
>1600
m a.s.l.
≤1600
m a.s.l.
1400-
1600 m a.s.l.
Other
Settlements
35%
4%
6%
55%
28%
72%
Road/railroad
56%
3%
10%
31%
4%
96%
Agro-silvo-pastoral
64%
0%
12%
24%
13%
87%
Recreational
31%
0%
49%
20%
0%
100%
