In Ecuador, debris flows are usually related to ongoing volcanic eruptions (e.g. Jones et al., 2015; Mothes et al., 2004; Pistolesi et al., 2014; Sierra et al., 2019). However, they occur on any type of mountains as has been observed in the Alps and the Cordillera Blanca of Peru (e.g. Ames-Márquez & Francou, 1995; Chiarle et al., 2007; Lliboutry et al., 1977; Rebetez et al., 1997). A special type of debris flows, known as lahars, are water-saturated volcanic debris and mud flows generated on the high flanks of a volcano or hill, which flow downslope driven by gravity (Vallance, 2005; Waitt, 2013; Mulas et al., 2019). According to Vallance and Iverson (2015), this phenomenon can occur: i) during a pyroclastic flow-forming eruption due to partial glacial melting or the disruption of a crater lake (primary lahars), or ii) by heavy rainfalls, outbursts of trapped subglacial water, extreme glacial thawing due to weather conditions and seismic events which mobilize preexisting water-saturated deposits (Barriga-López, 2015; Vasconez et al., 2017). The latter one, known as secondary lahars, are much more common and smaller in volume than the former (Jones et al., 2015). In addition, Pierson et al. (1996) described two main simplified sedimentological types of lahars, which can be distinguished according to their solid concentration: i) hyperconcentrated flows, from 20 to 60 vol% of sediments, and ii) debris flows, > 60 vol% of solid particles.

Tropical glaciers can be the water-source for lahar formation related to volcanic or meteorological phenomena as was previously described. Chimborazo (6263 masl, 1.47ºS, 78.81ºW) is a potentially active, ice-capped volcano (Jordan and Hastenrath, 1998; Barba et al., 2008; Cáceres, 2010; Samaniego et al., 2012) located in the Ecuadorian Western Cordillera, 150 km southwest of the capital city, Quito. During the Holocene, it had at least seven eruptive periods between 8000 and 1000 BP (Barba et al., 2008). These events produced limited tephra fall, pyroclastic density currents and lahars. Based on radiocarbon records, Samaniego et al., (2012) proposed an eruptive recurrence of 1000 years, with the last eruption having taken place between the early part of the 5^th century ((AD 420) and the end of the 7^th century (Barba et al., 2008). Based on deposits of these past eruptions, a hazard map of Chimborazo volcano was published in 1990 (Beate et al. , 1990). In addition, three moraine groups were recognized at the high flanks of the volcano (Samaniego et al., 2012). The moraines are at the top of lava flows dated at 30(3 ka (Bablon et al., 2019) and therefore, they were formed during the Last Glacial Maximum (LGM), Younger Dryas and Neo-Glacial periods, respectively (Barba et al., 2008; Samaniego et al., 2012).

On the early morning of 1 December 2015, a secondary lahar occurred on the southeastern flank of Chimborazo without any associated rainfall. The lahar descended down the Yamborumi drainage (Figure 1, Online Resource 1), which is fed by the 13^th glacier tongue and dead ice of Chimborazo at 4912 m asl, causing damage to a section of a pipeline, a bridge, roads and railway infrastructures. The flow also destroyed irrigation water channels and farmlands used for agricultural and livestock purposes. Four more similar events occurred by March 2017 (Online Resource 2), directly threatening downstream towns such as Santa Lucia de Chuquipogyos and Calshi (~1000 inhabitants) and could have also represent a potential hazard for critical infrastructure and larger cities such as San Andrés and Guano (Figure 1).

Figure 1

Map of Chimborazo and Tungurahua volcanoes with location of the 2015-2016 lahar deposits in red. White squares are seismic stations and the yellow circle is the automatic weather station (AWS) at the foot of the 13^th glacier tongue of Chimborazo volcano installed in December 2016. Inset shows the main geodynamic structures, GFZ: Grijalva fracture zone, GSC: Galapagos spreading center, Quito city and Antisana and Chimborazo volcanoes.

During fieldwork carried out in May 2016, the bulk deposits of four lahars that occurred on 1 and 6 December 2015, 9 March and 20 April 2016, were mapped. This information allowed estimating peak discharge, mean speed, cumulative volume, grain size distribution and componentry portions. Additionally, we used meteorological (weather time-series) and volcanological (seismic and thermal images) data, acquired by the official monitoring institutions.

Meteorological temporal series show that temperature, incident shortwave radiation and wind speed reached a peak in 2015. Strong winds trigger turbulence in the atmosphere which at glacier altitudes produces sublimation of the ice/snow (Favier et al., 2004). All these factors together can cause a pronounced increase in the glacier-melting rates of Antisana and also Chimborazo volcano. In addition, during 2015, the local and regional climatology was disturbed by the ocean-atmospheric phenomenon known as El Niño (ENSO) (Barnard et al., 2017). According to Francou et al., (2004), this phenomenon causes an increase in the energy available to melt ice/snow, and a decrease in the cloud cover and therefore decrease of precipitation and RH in the Andean region. That implies higher penetration of ISR and an increase in temperature as is observed in the time-series (Figure 4). Additionally, the monthly evolution of the weather variables in 2015 show sustained increases in air temperature at Antisana with peaks of 1.7ºC and 2.3ºC in September and December, respectively.

Moreover, tephra accumulation on glaciers could produce albedo reduction because more solar energy is absorbed, inducing melting (Delgado-Granados et al. , 2015). Schotterer et al., (2003) reported a remarkable change in Chimborazo’s glacier related to ashfall from Tungurahua in 2000. The ashfall resulted in extreme melting and the formation of penitents and intra-glacier ponds at 28m depth beneath the summit. Since then, several more of Tungurahua’s eruptions have affected Chimborazo’s glacier (Figure 5a), including the November 2015 eruptive phase.

Despite the fact that seismic activity is not well-known at Chimborazo volcano, based on the available data, it is possible to infer that during the period from December 2015 to July 2016 a slight increase in internal activity took place at Chimborazo volcano (Figure 6).

In addition, the residents along the riversides of the Yamborumi ravine observed buoyant ice during the first lahar emplacement on 1 December 2015, while the “Cenit Mountain Club” reported superficial water-ponds on the 13^th Chimborazo glacier tongue at the end of December 2015 (Figure 7a). Additionally, during the overflight performed on 20 February 2016 some kettle holes were observed on Chimborazo’s moraine deposits (Figure 7b). Therefore, we conclude that an intense periglacial melting process was ongoing at the time of the lahars occurrence on Chimborazo volcano.

Although the exact melting mechanism remains unknown, based on field observations, meteorological and volcanological evidences, we suggest that: i) the anomalous increase of energy fluxes from the atmosphere linked to ENSO phenomenon, ii) the decrease of glacier albedo due to ash accumulation from Tungurahua volcano and iii) a slight increase in internal activity at Chimborazo shown by more seismic events and thermal anomalies, could all have led to the rapid formation of superficial and intra-glacier water ponds. These, in turn, subsequently burst out incorporating moraines, providing the water and sediment sources to form secondary lahars without associated rainfall nor an eruption.

Similar trigger mechanisms have been suggested for debris flows generated at the Cordillera Blanca in Perú (Lliboutry et al. , 1977b) and more recently at the Himalaya mountain range in India (Nie et al. , 2021; Biswas, 2021).

Importantly, because the published hazard map from 1990 only regards hazards related to eruptive events at Chimborazo, the occurrence of these secondary, eruption-unrelated lahars prompted the elaboration of a new hazard map for Yamborumi ravine (Vasconez et al. , 2016).

Finally, it seems that similar lahar events have occurred in the past near to the Yamborumi drainage, as shown in Figure 2c. Perhaps these phenomena are recurrent in other drainages of the volcano, but they were not reported because no populated areas or infrastructure were affected. In order to prevent future damage to infrastructure and populated areas, further research should focus on the source processes and/or the location of potential ponds that could burst out at Chimborazo volcano and at other Ecuadorian volcanoes.

Figure 7 a) Molten dead ice at the 13th glacier tongue of Chimborazo volcano (ph. Jaime Romero) b) kettle holes detected on Chimborazo’s dead ice which are covered by moraine deposits.