1Population growth and the lack of land management in areas potentially affected by eruption at large snow-clad volcanoes, such as Cotopaxi volcano in Ecuador, make the issue of hazard mitigation very challenging. Indeed, risk reduction implies the minimization of interactions between human activities and volcanic hazards. Cotopaxi is one of the most active stratovolcanoes in the Eastern Cordillera of the Ecuadorian Andes, with eruptions of VEI 3 occurring at a frequency of once every 117 years during the last two millennia (Barberi et al., 1995; Hall and Mothes, 1995; Andrade et al., 2005). Vulnerability to volcanic hazards has increased with the growing urbanization of the Inter-Andean valley (fig. 1), especially hazards related to debris flows or lahars (lahar is an Indonesian term for debris- and hyperconcentrated flows that originate on a volcano).
Fig. 1 – Cotopaxi’s drainage systems. Topographic map of Cotopaxi and surroundings. The summits around Cotopaxi represent “old” eroded volcanoes.
Fig. 1 – Le réseau de drainage du Cotopaxi. Carte topographique du Cotopaxi et de ses environs. Les sommets autour du Cotopaxi représentent d’anciens édifices volcaniques, très érodés.
1: town; 2: lahar travel path; 3: border of the Chalupas caldera; 4: contours (in meters) (modified from Barberi et al., 1995).
1 : ville ; 2 : trajectoire des lahars ; 3 : bordure de la caldera de Chalupas ; 4 : courbes de niveau (en mètres) (modifié d’après Barberi et al., 1995).
2Cotopaxi is known to have produced devastating lahars on all of its flanks during each of its historic eruptions (Humboldt, 1810; Sodiro, 1877; Wolf, 1878; Hall et al., 2005). All current information about lahar hazards of Cotopaxi has been derived from the analysis of outcrops found along its northern and southern drainages, principally deposits left by the large Chillos Valley lahar on the northern flanks (Mothes, 1992; Mothes et al., 1998, 2004). Evaluation of volcano hazards along the northern and southern drainages is ongoing. However, this is the first study to identify lahar deposits and evaluate volcano hazards for the eastern flanks of Cotopaxi. Preliminary signs of a possible reactivation of Cotopaxi in 2001 (Troncoso, 2005; Rivero, 2006) prompted a revaluation of the potential threat from the volcano and the publication of a new edition of the Cotopaxi hazard map (Hall et al., 2004 a,b). In addition, mapping and analysis of this area was recently initiated (Ettinger et al., 2005, 2006), dedicated to the study of lahar deposits on the eastern flank of Cotopaxi.
3We present the results from this study of lahars that have flowed down the eastern flank. After we present the setting of the study area and a brief explanation of the methods, we interpret the results of grain-size analysis and spatial distribution of the deposits to discern the effect of topography on flow dynamics for the area between the base of the volcano and the alluvial fan in the Amazon basin. Further, we use these results to help evaluate the potential hazard from future lahars as well as delineate the factors that increase the vulnerability to lahar hazard in this area.
4Cotopaxi (0°38’S, 78°26’ W, 5897 m a.s.l.) is a stratovolcano situated in the Eastern Cordillera or « Cordillera Real » of the Ecuadorian Andes. The nearest cities are Quito (> 1,500,000 inhabitants) at 60 km NNW of Cotopaxi; Latacunga (40,000 inhabitants) at 45 km to the South, and Ambato (250,000 inhabitants) at 70 km to the South. The nearest town on the eastern flank is Puerto Napo (c. 1200 inhabitants), 124 km east of the volcano (fig. 2 A).
Fig. 2 – The eastern drainage of Cotopaxi volcano.
Fig. 2 – Le réseau de drainage du flanc est du volcan Cotopaxi.
A) Lahar deposits and morphological setting from Cotopaxi’s base in the Sierra to Puerto Napo in the Amazon basin. Area 1: proximal, upper deposition zone; Area 2: transition zone; Area 3: distal, lower deposition zone.
A) localisation de tous les dépôts analysés dans les trois zones de morphologie différente de la base du Cotopaxi dans la Cordillère et Puerto Napo dans la plaine amazonienne. Zone 1 : zone proximale du dépôt ; zone 2 : zone de transition ; zone 3 : zone distale du dépôt..
B) LAHARZ simulation results for four different volumes plotted on a 40 m shaded relief map representation of a DEM. Figures along river path indicate number of study outcrop.
B : résultats des simulations effectuées par LAHARZ avec quatre volumes différents sur un MNT ombré de 40 m. Les nombres, le long de la rivière, désignent les affleurements étudiés.
5With its conical shape, its large glacier cap and its relief of 2000 to 3000 m above the metamorphic basement, Cotopaxi dominates the landscape of the Inter-Andean valley. The basal diameter of the volcano is c. 22 km. It is drained by numerous incised valleys with steep banks, often several tens of meters deep. Glaciers cover about 14 km2 of the summit, have thicknesses that range between 30 and 120 m and have an estimated total volume of 0.5 km3 (Ramirez et al., 2004). On the eastern flank, the glacier extends from an altitude of about 4600 m to the summit exposed to winds carrying humidity from the Amazonian plain to the Altiplano, whereas on the western flank, the glacier has almost completely receded to the crater rim at ca 5800 m. Moraine materials and sparse vegetation dominate the landscape in areas of intermediate elevation (4600 – 4300 m altitude), on the eastern flank. In contrast, areas below 4300 m are covered by predominately dense vegetation comprised of high grasses and small bushes (“páramo”-type).
6Descriptions of historic activity at Cotopaxi (Humboldt, 1810; Wolf, 1878), suggest all significant lahars were triggered by eruptions (Aguilera et al., 2004). Since 1532, Cotopaxi has produced 13 episodes of activity, of which the most noteworthy were those of 1742-1744, 1768 and 1877. During these eruptive periods, observers noted ash, scoria and pumice falls, pyroclastic flows, lava flows and lahars. During the 1877 eruption, three lahars followed the Ríos Pita and San Pedro on the northern flank, affecting the villages of Sangolqui, San Rafael and Tumbaco (Sodiro, 1877; Wolf, 1878). Mothes et al. (2004) estimate peak discharges for lahars of the 1877 eruption to be > 50,000 m3/s for the northern drainage about 25 km from the crater and 55,000 m3/s for the Río Cutuchi drainage to the South. The contributions to the Cutuchi drainage from three individual channels include discharges of 6,000 m3/s, 12,000 m3/s, and 18,000 m3/s.
7The consequences of Cotopaxi’s lahars are known principally for their impact on cities situated on the North, Northwest, Southwest and South sides of the volcano (d’Ercole, 1991, 1996). These areas are currently the most populated sectors near the volcano and have remained so since before European colonization in the XVIIth century. Local lahar outcrops are easily accessible and several studies have shown that the volcanic threat to those cities today is very high (d’Ercole, 1991, 1996; Mothes, 1992; Hall et al., 2004 a,b; Aguilera et al., 2004; Samaniego, 2004). Between the interval 1742-1744 and 1877, eleven significant lahars descended the southern flank within the Río Cutuchi basin, affecting the villages of Latacunga (fig. 1), Salcedo and Baños. The 1877 lahars also had a major impact on proximal regions of Cotopaxi, leading to the deaths of more than 1000 people, the loss of many farm animals and the destruction of infrastructure (d’Ercole, 1996). For the eastern basin of the Tambo-Tamboyacu-Napo river system, there are few details reported except the accounts of Sodiro (1877); that even at Puerto Napo c. 124 km from the volcano, there were native fatalities and destruction of houses and boats by a lahar generated by the 1877 eruption. Since 1877 the population of this area has increased tenfold.
8Identifying key facts of the volcano morphology and human settlement for the eastern flanks of Cotopaxi are critical to assess volcano hazards. Glaciers extend down to 4600 m and provide a significant source of melt water in the event of eruption. The gradient varies from the upper parts of the volcano (27° to 35°) to its base (8° to 15°). In this sector, the c. 200,000-year-old Chalupas caldera dominates the Altiplano. The Tambo and Tamboyacu Rivers originate from the glacier and traverse this large and nearly flat area (fig. 3).
Fig. 3 – Cotopaxi seen from the eastern side within the Chalupas caldera. Note the blocky field in the foreground that represents lahar deposits of the 1877 eruption. The hills in the background comprise the rim of the Chalupas caldera (photography S. Ettinger).
Fig. 3 – Cotopaxi vu depuis la caldera de Chalupas à l’est. Remarquez le champ de blocs au premier plan qui appartiennent aux dépôts de lahars de l’éruption de 1877. Les collines à l’arrière-plan constituent le rebord de la caldera de Chalupas (photo S. Ettinger).
9In this manner, the eastern drainage system differs from the other landscape morphologies around the volcano. Hills forming the rim of the old Chalupas caldera present obstacles to normal river flow; the Tambo and the Tamboyacu rivers have cut through these hills and their deeply incised channels indicate zones of high energy flow. Nevertheless, these rivers flow across a low-lying floodplain area (average slope of 1.5°) within the Chalupas caldera; this site of the confluence of the two river channels (zone 1 of fig. 2) is a zone of deposition. After the confluence, flow continues as the ‘Río Verdeyacu’ through the Eastern Cordillera down to the Amazon basin with a marked increase in slope (zone 2 of fig. 2). The difference in altitude from the origin at the glacier (c. 4600 m) down to Puerto Napo (440 m) is 4160 m, yielding a mean slope value of 8.9% over the 124 km distance (fig. 4). When the river reaches the Amazon basin, it is called the ‘Río Jatunyacu’, which in the local Quechua language means ‘steep water’. The Jatunyacu river becomes the Napo river, a major tributary of the Amazon, a few kilometers downstream from Puerto Napo. The break-in-slope leads to the formation of a large depositional fan at the foot of the eastern Cordillera (zone 3 of fig. 2A).
Fig. 4 – Longitudinal profile from Cotopaxi volcano to Puerto Napo. Longitudinal profile of the Tambo-Tamboyacu river system from Cotopaxi’s eastern glaciers (its source) downwards to the Amazon plain, showing the succession of two deposition zones separated by an erosion zone.
Fig. 4 – Profil longitudinal depuis le volcan Cotopaxi jusqu’à Puerto Napo. Profil longitudinal des rivières Tambo-Tamboyacu depuis leur source dans les glaciers du Cotopaxi jusqu’à l’aval dans la plaine amazonienne, montrant la succession de deux zones de dépôts séparées par une zone d’érosion.
10The Altiplano on the eastern side of the volcano (as part of the regional Park of Cotopaxi volcano) is not populated, with the exception of two small intermittently occupied “haciendas” located on the borders of old lahar terraces. The first intensively populated area on this river system is located on the piedmont that opens up to the Amazon basin near Puerto Napo about 100 km on the East. Human settlements along the Río Jatunyacu are relatively limited; with a combined local population of c. 4000 people. This area has been developed quite recently. In the 1950’s Puerto Napo was built long after the last eruption of Cotopaxi in 1877. Unfortunately, the majority of the population has absolutely no awareness of the potential threat posed by a volcano located 100 km upstream.
11Mapping and sampling of deposits included areas of the proximal eastern flank of Cotopaxi volcano, along the drainage systems of the Tambo and Tamboyacu rivers (zone 1) and of the distal area along the Jatunyacu river, downstream to Puerto Napo (zone 3). In both areas, the study’s aim was to locate the most recent lahar deposits. A total of ten cross-sections were measured on the Tambo-Tamboyacu drainages where evidence of the 1877 lahar passage could be discerned, either by upper limits of deposits, erosion lines or vegetation boundaries. Cross-sections were measured using a precision hand level, a 50-m tape and a handheld GPS. Areas of maximum lahar inundation were calculated from these measurements. Grain-size analyses were conducted in the field with a grid system (fig. 5).
Fig. 5 – Lahar deposit section in the upper zone of river Tambo. This 3.34-m-thick lahar deposit is found over a horizontal distance of about 200 m. Note the 1 m grid and person for scale. The massive clast-supported deposits originate from non-cohesive debris flows.
Fig. 5 – Dépôt de lahar vu en coupe dans la zone supérieure de la rivière Tambo. Ce dépôt de lahar affleure sur environ 200 m de distance horizontale avec une épaisseur de 3,34 m. Notez la grille de 1 m et la personne comme échelle. Les dépôts massifs à support clastique sont issus des coulées de débris non cohésives.
12The size of the grid was either 1 x 1 m or 0.5 x 0.5 m (determined by the layer’s thickness) and the diameter of clasts under each grid node was measured. Representative layers of each of the 16 deposits were sampled, sometimes twice at the base and at the top of the outcrop. Nineteen samples were analyzed in the laboratory using a sedigraph and nine sieves (from 128 mm to 0.038 mm).
13A preliminary hazard map was drawn based on georeferenced data of lahar deposits and local topography; the latter obtained from 1:50 000 scale topographic maps. In addition, numerical simulations were constructed using the statistics-based software LaharZ (Iverson et al., 1998) on a 40 m digital elevation model (DEM) to delineate potential inundation areas from future lahars. We chose volumes for the simulations that bracketed estimated volumes of the 1877 lahar by Mothes et al. (2004) and Barberi et al. (1992). Four specified lahar volumes were used for the simulations: 1 billion m3, 316 million m3, 100 million m3 and 31,6 million m3. LaharZ calculated thousands of cross-sections to yield a total planimetric area for each volume that represents a lahar-inundation hazard zone. Zones having smaller planimetric areas were plotted over larger ones to produce a “nested” set of inundation hazard zones (fig. 2 B).
14Grain-size distributions vary from massive coarse-grained debris flow deposits that we group into zone 1 in the Tambo-Tamboyacu area (fig. 2 A) to fine-grained hyperconcentrated streamflow deposits that we group into zone 3 in the lower Jatunyacu area (fig. 2 A and fig. 6).
Fig. 6 – Average grain-size distribution in the top layer of the Tambo – Tamboyacu and Jatunyacu lahar deposits. Note the increase in fines (silt, clay) in the Jatunyacu basin.
Fig. 6 – Distribution moyenne de la granulométrie dans les niveaux supérieurs des dépôts du Tambo – Tamboyacu et Jatunyacu. Remarquez l’augmentation du pourcentage des fines (limons, argile) dans le bassin du Jatunyacu.
15Debris flow deposits in the zone 1 area exhibit two different sets of characteristics: 1) normal superposition of deposits with channel incision exposing pre-existing older lahar sequences; 2) inset terraces of progressively younger age accreted to the walls of pre-existing channels. The Tambo-Tamboyacu lahar terraces are poorly sorted, mainly normally graded, and have a weak fabric. Samples analyzed were taken from either massive outcrops, without any notable stratigraphy, or multi-layer outcrops representing multiple deposition phases (figs. 5 and 7).
Fig. 7 – Multi-layer stratigraphy of outcrop n° 7 in the Tamboyacu drainage. An example for a multi-layer stratigraphy deposit presenting normal grading from the base to the top. Probably there have been several flow pulses within one event.
Fig. 7 – Stratigraphie à couches multiples du dépôt n° 7 dans le drainage du Tamboyacu. Exemple d’une stratigraphie à plusieurs couches qui présentent un granoclassement normal de la base vers le haut. Probablement, la concentration sédimentaire du lahar devait varier durant l’écoulement, en bouffées successives.
16The multi-layer stratigraphy is marked by a contact indicating an abrupt change in the succession of the layers; an upper, normally graded sequence with a coarse base suggests a new flow pulse. Blocks > 50 cm in diameter at the base of the proximal deposits are ubiquitous. Deposit thicknesses in the Tambo-Tamboyacu varied from 1.64 m to 4.60 m reflecting differences in topography and channel structure. The deposits thinned in large valleys and thickened in smaller gullies or at confluences. The Tambo-Tamboyacu deposits and the 1877 deposits on the northern and southern flank of the volcano fall into the category of ‘non-cohesive’ or ‘clay-poor’ in which the fine component (silt and clay-size particles) is generally < 6 weight percent; gravel fractions range from 46 to 85 weight percent, and sand fractions are 10-43 weight percent. Clay content from hand samples are minor to insignificant.
17Channels are quite large within the Tambo-Tamboyacu drainage: measured cross-section areas often ranged between 1,500 and 2,300 m2. Flow characteristics described by Mothes et al. (2004) for the 1877 lahars on the northern flank reveal velocities up to 20 m/s (based on run-up calculations), and peak discharges in the order of 40,000 m3/s within 20 km of the crater. Studies on the southern side of Cotopaxi, showed the discharge reached about 50,000 m3/s and the lahar had an average velocity of 15 m.s-1 (Mothes et al., 2004). Comparison of morphology and bedrock characteristics at the foot of the volcano, show the eastern Altiplano is similar to the northern environment. Our observations suggest the eastern lahar velocities reached up to 18 m/s and discharges equal to or higher than the lahars that descended the northern drainages. Future field measurements will help better identify run-ups, estimate flow velocities, and calculate potential discharges.
18Lahar deposits along the Jatunyacu river (zone 3 of fig. 2A), beginning at c. 118 km from their source, are hyperconcentrated flow deposits (‘Jatunyacu facies’). These deposits extend 50 to 100 m laterally from the riverbed into the rainforest, have a mean thickness of 2 m and represent the main lahar facies deposited in the distal area. Although we did not find organic material for 14C dating, stratigraphic position, lithologic characteristics, limited soil formation and the lack of visible alteration of clasts or matrix suggest the deposits were left by the 1877 lahar. The Jatunyacu facies shows slight imbrication, well-developed clast fabrics, bedding structures and relatively good sorting, but is not graded. Well-preserved terraces along the Jatunyacu river are sparse, a result of the river’s high erosive capacity during rainy periods. The deposits in the Jatunyacu area show thicknesses of 2.60 m at about 110 km from the volcano, 2.20 at about 114 km and 1.20 m at about 120 km distance from Cotopaxi. Slope values in the Jatunyacu drainage are < 0.4%, i.e. ideal for a depositional environment.
19Boulders > 1 m in diameter are deposited mostly within 13 km of the edifice indicating the hazard from coarse-grained, non-cohesive debris-flows is very high within the proximal area (zone 1 of fig. 2 A). The hazard remains high as far as 120 km downstream near Puerto Napo where eastern flank lahars have left deposits c. 1 m thick. Because of the abrupt, decreasing slope at the base of the eastern Cordillera and the shallow entrenchment of the Jatunyacu river downstream, the area of distal flooding may be substantial depending on the volume of a lahar and its bulking value. The bridge at Puerto Napo could act as a dam for future lahars, thus enhancing the potential for flooding of areas at higher elevations upstream. The slight meandering of the riverbed may favor over-spilling of arriving flood waves at channel bends. The stratigraphy and characteristics of deposits analyzed in the Jatunyacu basin suggest emplacement by progressive flooding (channel overflow) of the riverbank instead of the rapid passing of a large lahar flood wave.
20No one knows the size of a lahar that will be generated during the next eruption, but deposits left by lahars that have flowed across the eastern flanks in the past help constrain volume estimates of future lahars. In many cases, the extent of inundation areas generated by LAHARZ simulations agree with the extent of mapped lahar deposits. For example, such agreement suggests that inundation from large lahars may cover the entire width of the Jatunyacu floodplain. Simulations also help when assessing worst-case scenarios. Our LAHARZ runs, having volumes > 500 million m3, represent low-probability, high-impact events. However, in some places simulations do not agree with the mapped extent of lahar deposits. For example, one simulation indicates that lahars 316 million m3 may not reach Puerto Napo (fig. 2B). However, there are historical accounts of the 1877 lahar affecting human settlements at Puerto Napo (Sodiro, 1877). Possible reasons for the difference between the simulations and the observed accounts are: 1) the DEM does not represent the current topography adequately; 2) volume estimates from deposits are incorrect; 3) future lahars may traverse post-1877 lahar topography differently than lahars flowing over pre-1877 lahar topography; 4) the model does not account for the wave of river water pushed ahead of the lahar; or 5) LAHARZ does not simulate flows that change from debris flows to hyperconcentrated flows adequately. Regardless, the goal of the simulations is not to re-enact a particular event, but to depict a range of possible future lahar inundation areas as part of an ongoing assessment of volcano hazards for this area.
21The construction of villages along Jatunyacu river has taken place over the past 50 years, especially the past several years, and several are built directly on 1877 lahar deposits. Since many houses are constructed on historical lahar deposits and the majority of the population cultivates fruit and vegetable crops along the river or even sometimes on islands within the riverbed, even a medium-intensity lahar would rapidly destroy the paths connecting the villages, the houses located nearest to the river, and the plantations that represent the main food supply for the local population. These crops are often the only guarantee of economic income for the local population. In addition, the vulnerability of this area increases with schools situated less than 50 m from the river channel, the construction of poor quality houses, the lack of electricity or radio equipment to alert the people of an eruption, the lack of infrastructure to support evacuation and the native language ‘Quechua’, which may create difficulties in communication and comprehension between local inhabitants and many national or foreign experts, who frequently speak only Spanish.
22We identified two levels (medium and high) of lahar hazard based upon field studies to construct a lahar hazard map. We also constructed a map of vulnerability based on human settlements, agricultural settings and infrastructure. Overlaying these two maps established criteria to create a risk map. For many reasons, most houses are constructed along the riverbanks and the risk map shows clearly that a great number of villages will be affected by a future lahar of Cotopaxi volcano (fig. 8).
Fig. 8 – Hazard, vulnerability and risk map for the Jatunyacu river area. Preliminary maps based on field analysis and topographic map information. The risk map is created by overlaying vulnerability and hazard analysis data.
Fig. 8 – Carte des aléas, de la vulnérabilité et des risques pour une zone le long la rivière Jatunyacu. Les cartes sont fondées sur les enquêtes de terrain et l’information issue des cartes topographiques. La carte de risque a été créée en croisant les données de l’analyse de la vulnérabilité avec celle des aléas.
A) Hazard, 1: present-day Jatunyacu river bed; 2: area of low hazard; 3: area of high hazard.
B) Vulnerability, 1: primary (villages); 2: secondary (agricultural settings); 3: path; 4: road; 5: bridge.
C) Risk (degree of vulnerability and degree of threat, between low risk and high risk).
A) Aléa, 1 : lit actuel de la rivière ; 2 : zone d’aléa faible ; 3 : zone d’aléa fort.
B) Vulnérabilité, 1 : importante (villages) ; 2 : secondaire (cultures) ; 3 : chemin ; 4 : route ; 5 : pont.
C) Risque : degré de vulnérabilité et degré de la menace définissant un risque faible ou fort.
23This study describes the first analysis of lahar deposits of the Eastern flank of Cotopaxi volcano. The location of the deposits and their sedimentologic characteristics permitted us to reconstruct flow behavior and evolution from proximal to distal areas. Multi-layer stratigraphy suggests that multiple lahars or flow pulses moved downstream. The observation of a texture change with increasing distance from the volcano indicates an evolution from a coarse, non-cohesive debris flow at its base to a fine-grained hyperconcentrated streamflow in the distal areas within the Amazon basin. We were able to distinguish three main zones defined in terms of flow dynamics: an upper deposition zone in the flat-lying Chalupas caldera area, principally at the confluence of the Tambo and the Tamboyacu rivers; a transition zone corresponding to the descent in the steep Verdeyacu river valley from the Sierra down to the Amazon floodplain; and a lower deposition zone of finer grained material in the Jatunyacu river area near the margin of the Amazon basin. This study of the Eastern flanks of Cotopaxi shows that lahars originating from eruptions of Cotopaxi volcano have left deposits in the proximal and distal areas and indicate that future lahars will likely have a major impact on populated areas even 120 km downstream in the Amazon basin.
24A preliminary hazard map was constructed based upon georeferenced data of lahar deposits and local topography obtained from 1:50 000 scale maps. Numerical simulations with the LaharZ software showed a reasonable correlation between hand drawn maps and simulations of large lahar volumes in the Tambo-Tamboyacu drainage and the Jatunyacu area. The results of the simulations help to assess hazard from future inundation for areas where the paucity of deposits and/or difficult access limits the collection of data in the field. These data sets were used to extrapolate areas of potential inundation by future lahars. In addition, a preliminary vulnerability map has also been created of villages, critical infrastructure, and agricultural areas observed during fieldwork and on the topographic maps. These two data sets were overlain to obtain a preliminary risk map. The risk map depicts four different risk categories resulting from two classification types (high or low) for the vulnerability and the hazard levels (fig. 8).
25The recent development of the region’s human settlement since the 1950s and the lack of socio-economic data available encourage further work in this area including the development of a GIS database and the analysis of the risk perception by the population. This combination of social and economic information with continued fieldwork and analysis would provide the basis for updating of the risk map accounting for differing levels of local development. A critical problem arising from Cotopaxi lahars is a similar issue often encountered in the field of volcanic risk analysis: the low recurrence rate of lahars mandates that we educate the people at risk regularly, show them how to recognize this danger and suggest ways that they can protect themselves.