The 2014-2015 Holuhraun eruption extruded >1 km 3 of lava in a barren region of the Icelandic highlands. Due to its large volume and the abundance of data for this eruption, Holuhraun is an ideal site to investigate fissure-fed eruption products for comparison with other large lava flowsfields on Earth and other planetary bodies. To characterize lava morphologies associated with this event, we used 0.01-0.5 m/pixel image data, acquired by aerial surveys and small Unoccupied Aircraft Systems

more »

... S) to create a 1:800-scale facies map that was ground-truthed using field observations during the summers of 2015-2019. Each facies region exhibits similar attributes in the remote sensing data, including albedo, surface texture, and geomorphology. However, at our mapping scale of 1:800, the facies typically include mixtures of lava types. Results show that transitional lava types (e.g., rubbly pāhoehoe, slabby pāhoehoe, and spiny pāhoehoe) dominate the 2014-2015 Holuhraun lava flow-field (83.82 km 2 ), rather than the traditional end-members of ʻaʻā and classical pāhoehoe. At 1:800-scale, we distinguish the following eight facies (with the percentage of total flow-field area shown in parentheses): rubbly (57.35%), spiny (25.96%), undifferentiated rubbly-spiny (9.59%), shelly (5.58%), pāhoehoe J o u r n a l P r e -p r o o f Journal Pre-proof 1953). Other so-called "transitional" lava types include spiny pāhoehoe, slabby pāhoehoe, rubbly pāhoehoe, and platy to platy-ridged lava, and are volumetrically significant components of many flow-fields (see Harris et al. (2016) and references therein). Rubbly pāhoehoe may even be the dominant eruption product of most mafic fissure-fed eruptions in Iceland and elsewhere (e.g., Guilbaud et al., 2005; Keszthelyi et al., 2000; Keszthelyi et al., 2004; Thordarson and Höskuldsson, 2008; Voigt and Hamilton, 2018) . J o u r n a l P r e -p r o o f Journal Pre-proof 3 This study systematically examines the 2014-2015 Holuhraun lava flow-field to identify and describe its constituent facies based on their albedo, surface texture, and shape. Additionally, we aim to elucidate the processes that led to the development of different lava surface textures during this eruption. The term "facies" is used here in a volcanological context to define domains with lava surfaces that exhibit a similar physical appearance and characteristics implying formation under a distinct set of emplacement conditions. In this sense, a single geologic unit (i.e., a flow-field) is likely to include multiple facies that may relate to changes in one or all of the following parameters: magma composition, eruption parameters, lava transport history, and modes of emplacement. Since we use the term facies to refer to lava surfaces, we distinguish between facies and lava types, in that facies provide information about the cumulative emplacement history recorded within a specific domain, whereas lava types reflect the mode of emplacement at a particular stage of the eruption. Facies and lava types can be equivalent, if specific textural domains correspond to a single lava flow type (e.g., pāhoehoe), or they can include multiple lava flow types that cannot be uniquely separated at a given mapping scale. The goal of this study is therefore to investigate the surface of the 2014-2015 Holuhraun lava flow-field to identify the different lava facies and their relationships to one another, as well as their constituent lava types and characteristics. In turn, this information informs our understanding of lava emplacement processes associated with large fissure-fed lava flow-fields, and helps us to determine what information can and cannot be gleaned about eruption processes by considering its final products, especially using visual and topographic remote sensing data. To achieve our goals, we qualitatively describe the morphological characteristics of the Holuhraun lava flow-field surface using remote sensing observations to generate a 1:800-scale facies map. This map is freely accessible as a geodatabase via the University of Arizona Campus Repository (Voigt and Hamilton, 2021). Facies identified on the basis of remote sensing observations were also investigated in the field to determine their constituent lava types and finer scale characteristics. This study therefore provides fundamental new insights into facies-based approaches to examining lava flow-field evolution, thereby contributing more broadly to the characterization of other large fissure-fed eruption products. The Bárðarbunga- Veidivötn volcanic system and the 2014-2015 Holuhraun eruption J o u r n a l P r e -p r o o f Journal Pre-proof 1.21 km 3 (Bonny et al., 2018) to 1.36 ± 0.07 km 3 (Dirscherl and Rossi, 2018). The 2014-2015 Holuhraun eruption extruded olivine tholeiite magma of homogenous composition (Halldórsson et al., 2018), with a mean output rate of approximately 80 m 3 /s (Bonny et al., 2018). Data and Methods a. Regional aerial datasets To evaluate the geomorphological characteristics of the 2014-2015 Holuhraun lava flowfield and its surroundings, we examined several image and topographic datasets at different pixel scales. Regional aerial data products include images and Digital Terrain Models (DTMs) acquired by two different organizations. First, pre-eruption air photos were captured by J o u r n a l P r e -p r o o f Journal Pre-proof 5 Loftmyndir ehf. on August 23 rd , 2003 and August 12 th , 2013 and used to generate a 5 m/pixel stereo-derived DTM and a corresponding 0.5 m/pixel orthomosaic. Loftmyndir ehf. also captured post-eruption air photos over the Holuhraun lava flow-field on August 30 th , 2015, and used these data to construct a stereo-derived DTM and orthomosaic with the same pixel scales as the preeruption datasets. Then, as part of the Iceland subglacial volcanoes interdisciplinary early warning system (IsViews), run by the Ludwig Maximilian University, Munich, the region was imaged on September 6 th , 2015 using the UltraCam-Xp camera to produce a 0.5 m/pixel stereoderived DTM and 0.2 m/pixel orthomosaic (Jaenicke et al., 2014; Münzer et al., 2016) . Fig. 2 shows the locations of the remote sensing data and supporting field photographs presented in this study. b. Small Unoccupied Aircraft Systems (sUAS) data Regional aerial survey data were complemented by higher-resolution data collected using sUAS during the summers of 2015-2019 (Bonnefoy et al. (2019); Scheidt and Hamilton (2019) ). Although contextual images and videos were acquired using multi-rotor systems, including several DJI Phantom 3 Pro and Phantom 4 Pro sUAS, the main inputs used in this study were acquired using a fixed-wing Trimble UX5-HP. These data include: a continuous 20 cm/pixel DTM and 5 cm/pixel orthomosaic over the vent region, and two 5 cm/pixel DTM blocks, with corresponding 1.3 m/pixel orthomosaics, that were acquired on either side of Baugur (see the blue and red outlined boxes in Fig. 2 , respectively). All datasets and associated pixel scale are summarized in Table 1 . c. Facies mapping Geological maps typically delimit lithostratigraphic units, whereas facies maps further subdivide geological units into regions of similar attributes, which in this study includes visual, textural, and geomorphological attributes identified using remote sensing data. However, facies are not directly equal to lava types because lava types co-occur locally and are not uniquely resolved at our digitizing scale of 1:800. Our facies are therefore expected to be mixed-classes in terms of their constituent lava types, making ground-truthing essential for correctly linking remote sensing facies to specific lava types present within the lava flow-field. J o u r n a l P r e -p r o o f Journal Pre-proof J o u r n a l P r e -p r o o f Journal Pre-proof