Measuring Lava Flow Parameters

Chapter 1: Introduction

1.1 Introductory Remarks and Greater Context

Lava flows are frequently observed at volcanoes across the full array of tectonic environments and magma compositions (Arnold et al., 2019). Mapping, measuring, and modelling these volcanic hazards are a priority for scientists and volcano observatories to understand the advancing pathways and rate of lava flows. Extensive observations and records of lava effusion rates, extent, shape, and volume can disclose changes in the magma composition or conduit behaviour and dimensions (Stasiuk et al., 1993) (Naranjo et al., 2016). Surface area, length, depth, volume and effusion rate of historical lava flow activity can all be combined to reconstruct previous eruptions to test the success of lava flow models. Retrospective analysis of lava flow morphology can better our understanding of such hazards to develop or improve upon proposed models used to produce lava flow hazard probability maps as insights into the lava emplacement process improves our forecasting abilities (Barberi et al., 1993).

Satellite remote sensing is crucial to the volcanic activity mapping framework because it allows researchers to monitor dangerous or inaccessible areas (Boccardo et al., 2015). Cloud and ash cover limitations of optical satellite imagery when studying volcanic environments has pushed synthetic aperture radar monitoring approaches the head of lava flow mapping. Lava flow surface area and extent are measured from the Synthetic Aperture Radar (SAR) imagery, with depth and volume calculated from a topographic/planimetric approach from DEM construction during post-processing (Ebmeier et al., 2012). Estimations of lava volumes, calculated from average depth and extent, are used to determine effusion rate (Naranjo et al., 2016). Finally, using the Q-LAVHA model with both Manhattan and Euclidean iterations morphological information from the mapped lava flows will be inputted to create a lava hazard probability map to highlight the areas at greatest risk of inundation.

1.2 Project Aims

Motivations for studying lava flows are both practical and scientific. Governing bodies and agencies with scientist pursue more effective predictions of hazards associated with volcanic activity, including lava flows (Fink, 1990). The principal end-user requirement of lava flow mapping is to assess the extent of new lava emplacement on the ground and track the flow evolution as well as analyse potential damage and risk to infrastructure (Boccardo et al., 2015). Active sensor remote sensing provides a possibility where lava flow mapping is no longer limited by cloud and ash cover, daylight hours and long orbit times from optical satellites that decrease revisit time (Boccardo et al., 2015).

Aims of this dissertation also include accurately measuring lava flow parameters, including extent, surface area, depth, volume and effusion rate. These characteristics are essential to investigate the wider behaviour of the El Reventador volcanic system, specifically the magma conduit system cycles. Finally, this dissertation aims to assess the future probability of lava flow inundation in the El Reventador landscape from volcanic lava flow modelling software. This can be compared to prior lava flows from this past eruptive phase and period to predict and analyse the future hazards that El Reventador poses to the surrounding landscape and infrastructure.

1.3 Literature Review

Risks posed from lava flows are considerably less than other primary impacts of volcanic eruptions, such as lahars, pyroclastic flows and poisonous gas emissions. But as a hazard, they still prove to be destructive as an advancing lava flow cannot be contained and will engulf or burn anything in its flow path (Chester, 1993). Only in exceptional circumstances do deaths occur as a result of lava flows because of their slow-moving nature (Riley, 2020). Their hazard risk ultimately depends on whether the distance they extent interferes with infrastructure, property or communities.

To predict and manage the hazard posed by volcanic lava flows, it first must be understood how these flows advance. Lava flows shape, size and surface features differ greatly depending on a variety of controls; lava effusion rate, vent geometry, underlying slope, topography barriers and flow rheology (Walker et al., 1973) (Wadge, 1981) (Fink and Griffiths, 1992) (Hulme, 1974) (Gregg and Fink, 2000) (Dietterich and Cashman, 2014) (Diettrich et al., 2015) (Rumpf et al., 2018) (Sparks et al., 1976) (Fink and Griffiths, 1998) (Kerr et al., 2006). Studies conducted on lava flow rheology andesitic volcanoes, like that of Mt Kilauea, Hawaii and basaltic volcanoes such as Mt Etna, Italy, show lava behaves as a non-Newtonian fluid, comparable to a model of a Bingham plastic material (Hulme, 1974) (Park and Iverson, 1984). Yield stress determines flow dimensions for non-Newtonian fluids (Hulme, 1974). At a given flow rate the thickening of the flow is controlled by yield strength, and therefore, a high rate of flow brings slower cooling rates. Thus, allows lava to create longer flow lengths (Park and Iverson, 1984). Depending on the silica content of the volcanic magma, lava flow characteristics will differ. El Reventador is an andesitic volcano, with silica concentrations between 53 and 59% (Arnold et al., 2019). Lava flows with greater silica contents construct flows with lesser extents and lower velocities (Dragoni and Tallarico, 1994). Basaltic (also referred to as mafic) lava flows have lower concentrations of silica, and consequently, lower viscosities allowing them to flow greater distances (Kauahikaua et al., 2003). Contrastingly, andesitic flows are high in silica, giving them high viscosities and lower velocities limiting their final distance from the vent.

Previous work by Aguilera et al. (1988) and Barberi et al. (1988) outlined the volcano-tectonic and petrological evolution of all Ecuadorian volcanoes. The previous history of volcanic activity in Ecuador notes that the volcanic axis at the Eastern Cordillera Real only became active 5 million years ago, with the Western Cordillera Real axis active approximately 1 -1.5 million years ago (Barberi et al., 1998). Within both Reals, andesites are the most prevalent. Lava deposits at El Reventador’s current cone are either andesites or basaltic andesites (Naranjo et al., 2016) (Samaniego et al., 2008). The present cone at El Reventador has been developing throughout the last 20ka in the caldera from a prior collapse of the eastern flank (Arnold et al., 2017). The geologic history of Reventador shows its construction can be marked into three principal phases. Starting with the basal volcanic complex when this stratovolcano began its construction 350,000 – 400,000 years ago. After, the basal volcanic complex collapse 30,000 years ago Reventador moved into its second phase of construction, the Volcán Paleo Reventador. While growing inside the caldera remains from the previous collapse, the eastern flank collapsed after 10,000 years (20,000 years ago), creating a vast debris avalanche that still scars the topography of the landscape at El Reventador today (Samaniego et al., 2005). The final phase of El Reventador’s construction is still ongoing, and the present cone undergoes frequent alterations from volcanic activity (Samaniego et al., 2005).

El Reventador’s location on the subduction plate boundary between the Nazca and Pacific plate has prompted the first studies of the volcano to focus on the source region of the magmas between mantle wedge vs. slab melting, and the space-time subduction control on lava compositions (Ridolfi et al., 2008). However, in the most recent decades, a shift in research theme has meant scientists have been particularly interested in modelling the magmatic chamber feeding system of El Reventador.

Sub-Plinian eruptions characterized by low volumes and lower magma discharges rates which in turn develop unsteady, oscillating eruptive columns. They have a lesser intensity to Plinian eruptions but dynamics are similar (Cioni et al., 2015). Whereas strombolian eruptions entail more moderate frequent eruptions with expanding gases driving the ejection of smaller lava flows. Both eruption types have occurred within the entire eruptive period beginning from 2002 to the present. The 2002 eruption was considered to be sub-Plinian, but strombolian eruptions were mainly responsible for volcanic activity previously to create the present cone, and after 2002 volcanic eruptions at El Reventador are commonly strombolian in nature (Aguilera et al., 1988).

This dissertation will build upon two key pieces (Naranjo et al., 2016 and Arnold et al, 2019) of literature that focused on lava flows at El Reventador between 2002 and 2016. Both these pieces of literature focused on first identifying the lava flows that were emplaced over the study period, an aim that will be carried forward over a new period (September 2016 to December 2020) within this dissertation. The eruptive period at El Reventador since 2002 has been persistent and ongoing with lava flows consistently observed within the constraints of this period. This eruptive period has been split into different phases (A-E) by Naranjo et al. (2016) and Arnold et al. (2019) to distinguish between major eruptions within the period and changes to volcanic activity characteristics. A sixth phase (F) within this eruptive period is going to propose during this dissertation that will date from the end of the Arnold et al. (2019) research, in August 2016, to the end of 2020.

Lava flows recorded within this eruptive period all emerge from within the central summit lava dome, and no visible vents or fissures have been identified on the remainder of the structure. Flows within phases A-E, but especially in the earliest phases, were clearly confined by what is left of the 2002 eruption crater, and thus, forced down the north or south flanks of the volcano (Arnold et al., 2019). Although, as new lava effusion has begun to infill the crater depression, this pattern in initial flow bearings is becoming less distinctive with greater numbers of flows pushing down the east flank towards the end of eruptive Phase E.

Volcanic behaviour observed over Phases A-E lies between a constant low rate of effusion and a more discrete explosive flow (Arnold et al., 2019). This has prompted the theory of the magmatic feeding system for El Reventador to be proposed as a conduit within the magma reservoir. The conduit is of critical importance in controlling lava effusion rates at El Reventador. Patterns of high frequencies of small lava flow events observed at El Reventador falls between a constant low rate of extrusion volcanic behaviour and an infrequent but explosive volcanic behaviour are typical of other volcanic systems with a conduit feeding system (Costa et al., 2007) (Arnold et al., 2019). The temporary magma storage in the conduit found in the upper-mid crust roughly 8km deep (Arnold et al., 2019). The geometry of the conduit is believed to be in north south orientation that aligns with the presence of the north and south vents on the summit crater (Arnold et al., 2017). Within the conduit, magma ascends from the mantle to the conduit storage until a certain threshold is reached. After the threshold has been exceeded, volcanic activity at El Reventador increases, and thus lava effusion rate increases until the conduit is emptied below the threshold. Volcanic activity remains slow until the conduit threshold has been exceeded once more.

1.4 Study Site and Eruptive History

El Reventador, a dominantly andesitic stratovolcano, is located on the Ecuadorian Sub-Andean uplift rear arc zone in the eastern Andes of Ecuador within a remote area of the Parque Nacional Cayambe-Coca (Ridolfi et al., 2008). It is one of 55 in Ecuador that are part of the Northern Volcanic Zone (NVZ) and is an important strato-volcano that comprise the andesitic volcanic row of the Eastern Cordillera Real chain of mountains (Hall et al., 2008). Two volcano arcs in Ecuador, Eastern and Western Cordilleras, are cut apart by the InterAndean Valley (Hall et al., 2017). Typical of the Eastern Cordillera, El Reventador is a highly active andesitic volcano, hence its name translates as ‘The Exploder’. It is built on a 50km dense continental crust in a low compressional tract between the Cordillera Real and Amazonian rainforest foreland (Barberi et al., 1988) (Guillier et al., 2001).

Though 90km east of Quito and major population centres in Ecuador, there is still a pressing threat from El Reventador lava flows to critical oil pipeline and state highway infrastructure roughly 8km away from the summit (Hall et al., 2004). Although flows have been limited to 3km in extent away from the summit, the cumulative extrusive volume of lava from 2002 to the present would be more than enough to inundate the highway and pipeline.

The volcanic complex is comprised of two structures; first, the old caldera that has suffered two sectorial collapses that have left a large landslide escarpment; and second, the current cone that has grown within the amphitheatre left by these landslides (Hall et al., 2004). Andesitic lava flow activity is dominant at El Reventador and is concentrated around the 1,300m tall strato-cone with a 4km wide horse-shoe shaped caldera (Naranjo et al., 2016).

However, more recent eruption reports from ground monitoring surveys have identified the emergence of several new vents for lava flow activity. Cone has an elongated shape towards the East, having slopes of 34°. Due to constant volcanic activity and eruption’s construction and destruction around the cone, the height of El Reventador changes daily. The last estimations of the cone summit height were 3,570m (Almeida et al., 2019) (Arnold et al., 2017). Typically, the volcano behaves with effusive types of bimodal activity that generates lava and explosive flows between small and moderate magnitudes. Flows contain ballistic blocks and are accompanied by 500-6,000m ash columns.

The explosive and effusive behaviour patterns make El Reventador an ideal location to study lava flows over this current eruptive period from 2008 to the present, with other notable recent eruptions in 2002, 2004 and 2007 (Smithsonian, 2017). The volcanic activity of this volcano is not well documented due to the remote location, poor accessibility and weather. But estimated to have 16 eruptions since the 16th century, with these eruptions being characterised by lava flows, pyroclastic flows, mudflows and ash falls. For this dissertation, only lava flows are being studied.

Despite the remoteness of El Reventador, it is still crucial to monitor current activity and model the possible future lava flow hazard risk because as recently as the 2002 eruption lava flows, pyroclastic flows and lahars inundated and damaged roads and a nearby oil pipeline. The 2002 sub-Plinian eruption of El Reventador forced significant economic and social impacts on Ecuador from damage to major nearby oil pipelines and state highways from pyroclastic flows, as well as provoking the closure of Quito schools and airports due to the ash fallout (Ridolfi et al., 2008). It is worth mentioning that during recent 21st-century eruptions lava flows never reached the eastern edge of the Caldera. However, if flows do extent to this point, the front of the flows could collapse and produce avalanches of hot material that would destroy infrastructure in the Quijos-Coca river valley.

El Reventador is an inaccessible location for most because of its remote location in the middle of a national park, poor weather and frequent eruptive activity. Ground monitoring is present at El Reventador, but research typically relies on remote sensing as field measurement can only be taken from a small selection of locations within the caldera (Naranjo et al., 2016). Optical satellite imagery, aerial photographs and radar are tools that have been used previously, but the radar was the chosen method to conduct the research presented in this dissertation.

1.5 Lava Flow Modelling

Volcanic activity, especially lava flows, is controlled by the thermo-rheological properties of the magma, and subsequent lava once above the surface (Hess, 1980). Owing to the complex nature the of magma and lava production process and a lack of understanding and empirical data relating to the physiochemical and thermo-rheological properties of the lava material, models that have been produced at present do not fully embrace each parameter that influences lava flow events (Costa and Macedonio, 2005)

Deterministic models account for the thermo-rheological properties of lava flows and can be used to predict lava flow paths and run-out. Attempting to solve transport equations, which involve mass conservation and momentum/energy balances, can be very difficult for a 3D model to complete, thus more simplified models tend to be used in assessing lava flow hazard – using cellular automata (CA) and 2D models (Wadge et al., 1994). Probabilistic models are based on the assumption that the topography of the volcanic environment is the most important factor that controls the pathways that lava flows take (Macedonio et al., 1990). They use the principle that paths can only propagate downwards. Probabilistic models do not need to solve physical equations relating to thermo-rheological characteristics of the lava flows and thus computation times are fast. They have been used to evaluate eruptions of Mt. Etna, Mt. Cameroon and Mt. Nyiragongo in the recent past (Barberi et al., 1992a, 1993; Bonne et al., 2008; Costa and Macedonio, 2005; Favalli et al., 2006, 2009a; b, 2012; Macedonio et al., 1990).

In summary, deterministic models are more realistic representations of the physical conditions that lead to lava flows, but they often include simplifications of physical aspects due to a lack of understanding of the complexity involved in lava flow hazards and unavailability of accurate thermo-rheological data for Reventador lava without conducting ground fieldwork. Probabilistic models do not require physical assumptions as the major role in determining the spatial extent of the lava flow hazard is the topography. They are a more efficient model to use and yet the results, at least for predicting the spatial extent of lava flows, compare well with empirical data from the field and the deterministic model results from previous studies (Felpeto et al., 2001; Kauahikaua et al., 1995; Wadge et al., 1994). It is for this reason that this study intends to use the Quantum-Lava Hazard Assessment (Q-LAVHA) model plugin on QGIS to estimate the probability of lava flow inundation (Mossoux et al., 2016). Q-LAVHA is predominantly a probabilistic model making it highly efficient and excludes the requirement for large quantities of thermo-rheological data. However, it combines certain elements of deterministic models so the properties of andesitic lava are not ignored.

1.6 Research Questions and Objectives

Research Question 1 (RQ1): Using Synthetic Aperture Radar (SAR) imagery, is it possible to identify and map previous lava flows from the El Reventador volcano via Multi-Temporal Coherence (MTC) interferometry between September 2016 and December 2020?

Successful mitigation of volcanic hazards depends upon the detection of reoccurring volcanic activity (Hall et al., 2004). The practice of mapping lava flows for inaccessible volcanos has not always been possible from optical satellite imagery. The primary objective of this dissertation is the demonstrate the usefulness of SAR imagery to detect lava flows at El Reventador.

Research Question 2 (RQ2): Are there significant differences in lava flow behaviour, spatial patterns, parameters or effusion rates within the same eruptive period by contrasting Phase F (novel flows mapped in this dissertation) with Phase A-E (flows previously identified and mapped by Naranjo et al. (2016) and Arnold et al. (2019))?

Referencing previous lava flow events at El Reventador identified by Naranjo et al. (2016) and Arnold et al. (2019) can aid in our understanding of the characteristics and frequency of lava flows observed in this environment. Historical information of volcanic activity is used to better understand the theories behind Reventador’s conduit magma feeding system and predict volcanic hazards at an andesitic stratovolcano. This research objective seeks to improve upon a vacuum in global andesitic lava flow research, as basaltic volcano research provides the majority of global research into such volcanic activity. Flows at El Reventador are andesitic and mapping their spatial patterns and parameters aim to improve upon the less comprehensive existing knowledge of andesitic lava extrusion behaviours.

Research Question 3 (RQ3): How can the probability of future lava flow hazards at El Reventador be modelled based on the existing topography using the Q-LAVHA freeware plugin to simulate lava flow inundation?

Accurate lava flow inundation models are fundamental information for scientists and governing bodies or agencies to respond to imminent or persisting lava flow hazards (Mossoux et al., 2016). A better understanding of likely lava flow behaviour informs key actors responsible for land use decisions and vulnerability response plans, even when the flows being addressed are only expected to damage property and not expected to result in the loss of human life (Herault et al., 2009). Numerical simulations models that predict size and extent, like that of Q-LAVHA, can explore various eruptive situations inundation areas (Del Negro et al., 2007).

With the use of the model, Q-LAVHA, that uses topography as the greatest control on flow pathways, lava flow inundation will be simulated from volcanic vents on a Digital Evolution Model (DEM) (Mossoux et al., 2016). Combines existing probabilistic and deterministic models to calculate the probability of a lava flow spatial propagation and terminal extent from the vent source (Mossouz et al., 2016). Topography is the primary constraint on the lateral and lineal spatial propagation of lava. Euclidean model option simulates lava flow until the maximum extent input parameter is reached as the crow flies from the source vent. Whereas, the Manhattan model option simulates lava flow until the maximum extent is reached where the lava reaches that travel as a line. The final output of the model iterations will publish a hazard probability map that demonstrates which areas of the landscape are most at risk from lava flow inundation at El Reventador. This hazard map can then be used by organisations and communities to better predict and prepare for the potential future risk lava flows will pose to nearby infrastructure.