Very high-resolution terrain surveys of the Chã das Caldeiras lava fields (Fogo Island, Cape Verde)

Fogo in the Cape Verde archipelago off Western Africa is one of the most prominent and active ocean island volcanoes on Earth, posing an important hazard to both local populations and at a regional level. The last eruption took place between 23 November 2014 and 8 February 2015 in the Chã das Caldeiras area at an elevation close to 1,800 m above sea 15 level The eruptive episode gave origin to extensive lava flows that almost fully destroyed the settlements of Bangaeira, Portela and Ilhéu de Losna. In December 2016 a survey of the Chã das Caldeiras area was conducted using a fixed-wing unmanned aerial vehicle and RTK GNSS, with the objective of improving the mapping accuracy derived from satellite platforms. The main result is an ultra-high resolution 3D point cloud with a Root Mean Square Error of 0.08 m in X, 0.11 m in Y and 0.12 m in Z, which provides unprecedented accuracy. The survey covers an area of 23. 9 km2 and used 2909 20 calibrated images with an average ground sampling distance of 7.2 cm. A digital surface model and an orthomosaic with 25 cm resolution are provided, together with elevation contours with an equidistance of 50 cm and a 3D texture mesh for visualization purposes. The delineation of the 2014-15 lava flows shows an area of 4.53 km2 by lava, which is smaller but more accurate than the previous estimates from 4.8 to 4.97 km2. The difference in the calculated area, when compared to previously reported values, is due to a more detailed mapping of flow geometry and the exclusion of the areas corresponding 25 to kīpukas. Our study provides an ultra high-resolution dataset of the areas affected by Fogo’s latest eruption – crucial for local planning – and provides a case study to determine the advantages of ultra high-resolution UAV surveys in disasterprone areas. The dataset is available for download at http://doi.org/10.5281/zenodo.4035038 (Vieira et al., 2020).

like Cape Verde, where volcanic eruptions are likely to have disproportionate impacts, on account of their more limited resources and geographical isolation (Komorowski et al., 2016). Accordingly, solid and realistic volcanic hazard assessments in such areas, more than in any other settings, greatly benefit from very-high resolution datasets from which detailed 35 volcanological, geophysical, and environmental parameters can be inferred. In particular, very high-resolution digital terrain datasets of recently erupted lava-flow fields, may provide rapidly produced topographical datasets that can also be used to plan mitigation and reconstruction strategies, as well as the opening of vital new communication infrastructures. The usefulness of such datasets is greatly enhanced when these datasets are freely available to governmental agencies, decisionmaking bodies and the scientific community alike. In line with this vision, in this paper we report and make available a 40 recently acquired very high-resolution digital surface model and orthomosaic of the lava flow-flow field created during the 2014-15 eruption of Fogo volcano in Cape Verde.
Fogo in the Cape Verde Archipelago off Western Africa is one of the most prominent and active ocean island volcanoes on Earth, posing an important hazard to both local populations and at a regional level Heleno da Silva et al., 1999;Ramalho et al., 2015;Eisele et al., 2015;Jenkins et al. 2017). Crucially, Fogo is the site of recurring volcanic activity, 45 with a record of at least 27 historical eruptions since the island was discovered in the mid-fifteenth century, yielding a mean recurrence interval between eruptions of approximately 19.8 years, with individual intervals ranging from 1 to 94 years (Ribeiro, 1954;Torres et al., 1998;Day et al., 1999;Mata et al., 2017). The latest events occurred in 1995 and in 2014-15, both of which extruded extensive lava flow fields at the Chã das Caldeiras, a summit depression lying at an elevation of approximately 1800 m above mean sea level. Effectively, the settlements of Bangaeira, Portela and Ilhéu de Losna located in 50 Chã das Caldeiras were almost fully destroyed in the 2014-15 eruption (Fig. 1). With a population of about 1200 inhabitants prior to the eruption, the local economy was based in agriculture (mainly wine and fruit orchards), grazing, and tourism.
Fortunately, there were no casualties.

55
The latest eruption started on the 23rd November 2014 and lasted until the 8th February 2015, with magma being erupted from a 700 m-long NE-SW trending eruptive fissure located on ¬the SE flank of the previous 1995 crater row, on the SW flank of Pico do Fogo Mata et al., 2017). Reportedly, the eruption started with vigorous fire-fountain activity, which quickly evolved to a more explosive strombolian style, forming a crater row roughly parallel to the 1995 60 fissure. Later, the eruption was characterized by simultaneous or alternating hawaiian, strombolian and vulcanian eruptive styles (from the different craters of the fissure) lasting for several days, and by an almost constant emission of lava flows from the lowermost terminus of the vent (Mata et al., 2017). These formed two initial thick 'a'ā flow lobes: the first advanced towards the southwest and eventually stalled after 1.7 km, at the foot of the caldera wall; the second progressed intermittently 3 km to the northeast, towards the village of Portela, razing a large portion of this settlement (Mata et al. 2017;65 Jenkins et al., 2017). During the later stages of the eruption, however, this flow lobe was reactivated, producing more fluid https://doi.org/10.5194/essd-2020-289 'a'ā and pāhoehoe breakouts to the west and north, the latter of which destroyed most of what was left of the Portela settlement and descended to the village of Bangaeira, causing widespread destruction there (Mata et al. 2017;Jenkins et al., 2017). The resulting lava flow field affected an area of ci. 4.8 km2, with extruded volumes estimated at ~45 × 106 m3 (Bagnardi et al., 2016;Richter et al., 2016;Cappello et al. 2016). 70 Remote sensing techniques have been used by several authors to study the Fogo eruption of -15. Capello et al. (2016 used the HOTSAT satellite thermal monitoring system for the analysis of MODIS and SEVIRI data for the location of the hotspot, lava thermal flux, and effusion rate estimation. To forecast lava flow hazards during the 2014-15 Fogo eruption they used the MAGFLOW model. Validation of numerical simulations was done using Landsat 8 OLI and EO-1 ALI images and field observations. Bargnardi et al. (2016) used very high-resolution tri-stereo optical imagery acquired by the Pleiades-1 75 satellite constellation and generated a 1 m resolution digital elevation model (DEM) of the Fogo Volcano. From the Pleiades-1 post-eruption topography they subtracted the heights from a pre-eruption DEM, that was obtained using spaceborne synthetic aperture radar (SAR) data from the TanDEM-X mission. To measure the subsidence of the lava flow field, they used Sentinel-1 for interferometry. Height differences between the post-eruptive Pleiades-1 DEM and the preeruptive topography from TanDEM-X show a lava volume of 45.83 ± 0.02 × 106 m3, emplaced over an area of 4.8 km2 at a 80 mean rate of 6.8 m3/s. Richter et al. (2016) did lava flow simulations based on field topographic mapping and satellite remote sensing analysis. They produced a topographic model of the 2014-15 lava flows from combined Terrestrial Laser Scanner (TLS) and photogrammetric data. The pre-eruptive DEM used was 5 m/pixel and was generated from the contours based on photogrammetric data. They estimated a lava flow volume of 43.7 ± 5.2 x 106 m3. TerraSAR-X imagery was used to access the lava flow model performance. The authors point out the need of having up-to-date topographic information 85 because lava flow hazards change as result of topography modifications.
More recent, Bignami et al. (2020) used a combined method of 21 images from Sentinel-1, COSMO-SkyMed, Landsat 8, and Earth-Observing-1 missions from November 2014 to January 2015, to retrieve lava flow patterns. They applied an automatic change detection technique for estimating the lava field and its temporal evolution, combining the SAR intensity and the interferometric SAR coherence. Results showed a SW-NE oriented dyke, located inside Chã das Caldeiras, SW of 90 the Pico do Fogo, as reported by Gonzalez et al. (2015). The area coverage of the lava flow obtained by visual analysis (L8 and EO-1) was estimated at 4.97 km2 as in Cappello et al. (2016), very close to the 4.8 km2 estimated by Bagnardi et al. (2016), and the 4.85 km2 estimated using Terrestrial Laser Scanner (TLS) combined with structure from motion data by Richter et al. (2016).
Several papers have been published on the latest eruption of the Fogo volcano, e.g. focusing on the variation of land surface 95 temperatures during the eruption , on lava geochemistry and small-scale mantle heterogeneity (Mata et al., 2017), mineralogy and geochemistry of incrustations (Silva et al., 2019), conduit dynamics and surface deformation (Gonzalez et al. 2015), on lava flow mapping and volume estimates (Bargnardi et al. 2016;Bignami et al., 2020), and lava flow hazards (Richter et al., 2016;Cappello et al. 2016;Jenkins et al., 2017). Portugal), an extensive survey using a survey-grade unmanned aerial vehicle (UAV) was conducted in December 2016 with the purpose of generating a very high resolution digital elevation model (DEM) and orthomosaic of the lava field to be used as baseline data for assessment of the eruption impacts, support to geological mapping and studies of the lava flow field, as well as for modelling lava flow dynamics. A study commissioned by the United Nations Development Program in Cape Verde stresses that an improvement in the assessment of hazards on the island of Fogo can only be achieved from a detailed 105 analysis and the modelling of the lava flow . The data presented here is the result of that campaign and is the most detailed and updated survey of the area, with an ultra-high resolution digital surface model (DSM) that is essentially a DEM in most of the area due to the overall lack of vegetation and scarce number of buildings, and a digital orthomosaic.

Application of UAVs to volcanic areas 110
Digital elevation models and the dissemination of Geographical Information Systems have changed the way the terrain is characterized, analysed, monitored and modelled, especially since the 1990's. DEMs have been produced from dense collections of topographical points, manned aircraft photogrammetry, digitizing of topographic maps (Stevens et al., 1999), satellite remote sensing (Baldi et al., 2002;Kerle, 2002;Diefenbach et al., 2013), light detection and ranging -LiDAR (Mouginis-Mark and Garbeil, 2005;Mazzarini et al., 2007;Favalli et al., 2009;Fornaciai et al., 2010), radar interferometry 115 (InSAR) (Rowland et al., 1999;Poland, 2014). Since the mid-2010's, with the technological developments and decreasing cost of unmanned aerial vehicles, accompanied by the development of photogrammetry software and computing power, a revolution took place. Very accurate and high quality DEMs became increasing available, leading to the possibility to easily achieve centimetric to decimetric resolution, even in large areas. The number of UAV-based surveys have been increasing steadily but many of them stay stored in the producers or client computers and are not of open-access. 120 UAVs have had various applications, such as for wildlife recognition (e.g. Christiansen et al., 2014;Wang et al.,  rough terrains of difficult access, such as the situation of volcanic eruptions, make UAVs a powerful tool to acquire real-time or near-real-time data on processes that often cannot be observed by naked eye (Di Felice et al 2018). Hence UAVs are increasingly used in situations of risk where it is essential to make the rapid terrain recognition following natural disasters (Gomez and Purdie 2016). Software and computing power have evolved very fast in last decade and the application of 135 advanced photogrammetry algorithms involving image matching and structure from motion started to allow facilitated production of high-quality digital surface models and orthophoto maps. The collection of accurate GNSS ground control points allows to generate these products with centimetre resolution and accuracy (Favalli et al 2018). The recent development of RTK UAVs allows for even faster workflows in the terrain and to produce highly accurate models.
Photogrammetry techniques have been widely applied to study detailed changes in the morphology and structures of 140 volcanos, like Mount St. Helens, the Colima and the Merapi (Major et al., 2009;Walter et al., 2013;Salzer et al., 2016).
Optical and thermal cameras transported on UAVs have been used for identifying meter to sub-meter topography changes and for the detection of thermal anomalies (Nakano et al., 2014;Thiele et al., 2017;Nakano et al., 2014;Müller et  for collecting immediate and real-time aerial data in volcanic environments during and after an eruption is provided by Jordan (2019), highlighting its advantages for mapping, sample collection, thermal imaging, magnetic surveys, slope stability studies, and also as platforms for carrying outgassing measurement sensors.

The study area and the volcanic activity of 2014-15
The island of Fogo is one of ten islands of Cape Verde, an archipelago located off the west African coast, about 600 km from 170 Senegal. The Cape Verdes are regarded as the type-example of a volcanic archipelago formed in a stationary plate environment relatively to its hotspot, which probably explains the arcuate distribution of its islands (Burke & Wilson 1972;Lodge and Helffrich, 2006;Ramalho et al., 2010aRamalho et al., , 2010bRamalho et al., , 2010cRamalho, 2011). In more detail, this arcuate geometry is defined by two island chains: a "northern", from São Nicolau to Santo Antão, and a "eastern-to-southern", from Sal to Brava. There is no evident hotspot track but there is a morphological suggestion of an age progression in the "eastern-to-175 southern chain", from east (oldest islands) to west (youngest islands), which is supported by ages of the oldest exposed lithologies (see Ramalho, 2011). Fogo is located close to the southern terminus of this latter chain and is the only island in the archipelago with historical (i.e. last 500 years) eruptions (Bebiano, 1932;Ribeiro, 1954;Machado, 1965;Day et al., 1999;Faria and Fonseca, 2014).
Fogo is a large ocean island volcano showing a conical shape with a diameter of about 30 km (at sea level) and rising to an 180 elevation of 2829 m, approximately 7 km above the surrounding seafloor. Structurally, the island is a compound volcano, featuring a "somma-vesuvio" association, with a younger stratovolcano -Pico do Fogorising from the central depression -Chã das Caldeirasof an older collapsed volcano, sometimes referred as Monte Amarelo (Ribeiro, 1954;Day et al., 1999). This depression, however, is open to the east, being limited in the remaining three sides by a horseshoe shaped steep rock wall, over 1,000 m high, called Bordeira. This morphology, in turn, is either interpreted as a gravitational collapse headwall 185 Paris et al., 2011) or as volcanic caldera walls, whose eastern portion later experienced a gravitational flank failure (Torres et al., 1998;Brum da Silveira et al., 1997a, 1997bMadeira et al., 2008). Notwithstanding the different interpretations for the origins of this summit depression, it is clear that the opening to the east resulted from a massive flank failure of the flank of the volcano. Effectively, marine geophysical surveys undertaken off Fogo revealed the presence of voluminous submarine debris avalanche deposits extending offshore into the channel between Fogo and Santiago, and into 190 the seafloor south and north of these islands, thus attesting the occurrence of this collapse (Le Bas et al., 2007;Masson et al., 2008;Barrett et al., 2019b). Moreover, field evidence attesting to the impact of a megatsunami triggered by Fogo's flank failure has been documented in the neighbouring islands of Santiago (Paris et al., 2011(Paris et al., , 2018Ramalho et al., 2015) and Maio (Madeira et al., 2020), confirming the catastrophic nature of the collapse and suggesting a 65-84 ka age for this event.
Pico do Fogo, currently the highest point in the island (2829 m in elevation), is a large and roughly symmetrical strato-cone 195 that grew on top of the collapse scar, partially infilling this feature (Ribeiro, 1954;Torres et al., 1997;Brum da Silveira et al., 1997a, 1997bDay et al., 1999). The sheer volume of this volcanic edifice is a testimony to the vigorous eruptive activity https://doi.org/10.5194/essd-2020-289  , however this seems to be contradicted by the stratigraphic sequence at the summit, which exhibits mostly altered volcanic successions. Effectively, historical records suggest that all historic eruptions were extruded from 200 adventitious vents localised at the base and lower flanks of Pico do Fogo, or at Chã das Caldeiras and the eastern flank of the island, in the periphery of this strato-cone (Ribeiro, 1954;Torres et al., 1997;Brum da Silveira et al., 1997a, 1997b. This is the case of the 1951, 1995, and 2014-15 eruptions, which vents were located in the NW, SW and S flanks of Pico do Fogo, close to its base at Chã das Caldeiras. Chã das Caldeiras (Fig. 2) is thus a lava-infilled high-altitude summit depression, which resulted from the gradual 205 accumulation and ponding of lava flows (and pyroclasts) erupted from Pico do Fogo and its adventitious/satellite cones, against the vertical walls of Bordeira. Morphologically, Chã can be divided in two large semi-circular sectors: a southern, larger, with approximately 3 km of radius, and with an elevation of 1780 m, and a northern, with a shorter radius of approximately 1 km, and with a mean elevation of 1650 m. These two sectors, which are roughly separated by the prominent Monte Amarelo spur, have been interpreted as two coalescent volcanic calderas by Torres et al. (1997), Brum da Silveira et 210 al. (1997a, 1997b, and Madeira et al. (2008). Chã das Caldeiras is a generally flat landscape, punctuated by a few volcanic cones and extensively covered by 'a'ā and pāhoehoe lava flows and ash and lapilli deposits, which make it very irregular in detail and a challenging terrain for mapping. In particular, the extensive 'a'ā lava flow lobes of the 2014-15, 1995 and 1951 eruptions covered large portions of Chã, resulting in large swaths of virtually inaccessible rocky surfaces, given their extreme roughness. Hummocky landscapes also exist, generally corresponding to older 'a'ā lava flow fields with scattered 215 large rafted blocks of spatter sequences in its surface (resulting from the gravitational collapse of strombolian cones and subsequent transport by lava flows), which are now partially buried under a blanket of lapilli and ash that smoothed the surface. A good example of such surfaces can be found to the east and particular to the west of the Monte Beco cone, being genetically associated to this vent. The foot and slopes of Pico do Fogo, in contrast, are extensively covered by a thick blanket of lapilli and ash, conferring a very smooth and uniform conical surface. Despite this cover, fanned leveed 220 channelled morphologies can also be recognized at the foot of Pico do Fogo, corresponding to buried lava flow fans and alluvial fans. Overall, vegetation is scarce and is mostly confined to the surfaces of talus accumulated at the foot of Bordeira, where a thin soil exists, or to some scattered vineyards along some ash-covered slopes.  Human settlement at Chã das Caldeiras started towards the end of the 19th century (Ribeiro, 1954). Chã, as a high-altitude depression, is cooler and more humid than the rest of the island, with frequent fog condensation and occasional frosts, 230 providing ideal conditions for the planting of orchards and vineyards. Attracted by the prospect of a more prosperous agriculture, people gradually settled Chã, mostly in the vicinities of Monte Amarelo, where some water springs and enhanced but ephemeral flow (from the larger canyons draining Bordeira) allowed easier access to water. Here they established the settlements of Portela, Boca Fonte, and later Bangaeira, which slowly and gradually grew in size until the 1995 eruption, when Boca Fonte was all but destroyed and the main access road to these settlements was blocked by the 235 advancing flows (the 1951 eruption, although of a higher magnitude, had a lower impact in these settlements; Jenkins et al., 2016). After the 1995 eruption, however, the prospect of an additional income provided by a burgeoning wine industry and the rapidly growing flow of tourists that came to see the volcano, fuelled the rapid growth of Portela and Bangaeira, with population reaching as much as ~1500 resident inhabitants by 2014 Jenkins et al., 2016). The 2014/2015 eruption, in contrast, had a profound impact in these villages, given that the advancing lava flows either razed or 240 https://doi.org/10.5194/essd-2020-289 buried up to 90% of the existing buildings, as well as covered large swaths of the adjacent agricultural land. Gradually, however, reconstruction is taking place, either through new constructions over the recent lava flows, or by the painstaking reclamation of lava-buried but structurally intact buildings.

UAV Surveying 245
The field campaign was conducted about 20 months after the end of the eruption of 2014-15, when the lava flows had already cooled substantially, but with the occupation of the few houses existing at the Chã das Caldeiras, being forbidden and still hazardous, mainly due to gas emissions. Hence, the field team stayed at the village of São Filipe and travelled daily to the survey area. The expertise of the team on local conditions, geology and logistics and cooperation with the Cape Verde authorities, greatly facilitated the success of the mission. 250 The main challenges to overcome were i. the weather, which in December frequently shows high winds and low visibility (clouds) in the Chã das Caldeiras, ii. finding good landing sites for the UAV, iii. coping with the 1000 m high vertical rock wall of the Bordeira and with its possible influence on the positioning and communications system of the UAV, and iv.
collecting enough high-quality ground control points.
The survey of the Chã das Caldeiras area was conducted from 12 to 16 December 2016 with a field team of 4 members. Two 255 members focussed on conducting the UAV flights and the other two on the collecting ground control points.
Communications among team members were done using VHF radios.
The weather during the campaign was excellent with clear skies and no wind in the first days, but deteriorated towards the end of the week, with clouds entering the survey area and affecting initially the illumination conditions and even limiting the flights in the last two days. This has affected the quality of the orthomosaic, which shows illumination artefacts in the 260 northern part of the Chã das Caldeiras.
The survey was conducted using a professional survey-grade fixed-wing UAV SenseFly eBee classic, which has a structure of expanded polypropylene (EPP) foam, with carbon and composite parts. It has a 96 cm wingspan and under 0.7 kg take-off weight, which when disassembled allows for lightweight packing. This model allows for flights with wind speeds up to 45 km/h, flight durations of up to 50 min and a radio link distance up to 3 km. Two cameras were used: a Canon G9X 16MP in 265 the initial flights, which had a critical failure, and a backup Canon IXUS 12 MP which was used subsequently.
Take-off with the eBee is performed by hand, which facilitates selecting the location, but landing is done normally in fully automatic mode needing several tens of meters of approach area, and a smooth landing surface in order not to damage the EPP UAV body. This was a significant limitation to the survey, since the area of the Chã das Caldeiras is mostly covered by very rough lava surfaces, with scarce smooth ash and lapilli cover sites, which are normally far apart. Given these 270 constraints, five sites allowing for good landing conditions were selected: (i) 14.93477º N, 24.35407º W, (ii) 14.928092º N, https://doi.org/10.5194/essd-2020-289  N, 24.356605º W, (iv) 14.9334999º N, 24.3722831º W and, (v) 14.962204º N, 24.3690256º W.
The survey consisted of 20 flights that do not show the ideal spatial setup nor homogenous illumination conditions in the resulting aerial photos, but it was the best solution given the logistical constraints. This was due to the following problems: 275 sparse location of the take-off and landing sites, changes in wind-speed affecting power consumption, unexpected cloud advection and low visibility during some days, duration of daylight, fast changing shadowing effect from the Bordeira rock wall and by Pico do Fogo, battery limitations (due to heat, high risk of damaging the UAV in case of a need to crash land over lava flow, we decided not to conduct flight of over 30 min), long-distances to move between landing sites, these resulting on the need for constant on-site modifications of the original flight planning. The average flight elevation above the 280 ground was 190 m, resulting on an average ground sampling distance of 6 cm, over 2,900 aerial photos and a total surveyed area of 24 km2 (Fig. 3).

Ground control points
Coordinates of ground control points (GCPs) were measured at markers distributed in the field prior to the survey and at easy identifiable points, such as large boulders and building edges. The measurements were obtained using a Leica Viva (GS08) GNSS with base stations located at known coordinate sites (Monte Beco and Monte Amarelo) and a rover for surveying in RTK mode during the field surveys in December 2018. Extra GCPs were collected in February 2017 in small boulders 290 selected in the preliminary orthophoto mosaic, in order to improve georeferencing quality. The accuracy of the GCP coordinates is of about 3 cm.  principles of stereoscopic photogrammetry, in which a 3D structure results from a series of 2D image overlays taken in motion to an object. The geometry of the images, position and orientation of the camera are calculated automatically without the need for a priori indication. Decisions are made simultaneously based on an iterative adjustment of package images, and the structures are automatically extracted from several overlapping images and the points are searched from image to image allowing the estimation of the coordinates of the objects. The 3D point clouds are generated in an image-space coordinate 305 system and are transformed into an absolute real-world object-space coordinate system. The transformation is achieved by using 3D functions based on a relatively small number of GCPs (Westoby et al. 2012;Smith et al. 2016).
The models that we have developed have used 2909 calibrated images with an average ground sampling distance of 7.17 cm and a total area covered of 23.89 km2. The camera optimization resulted in a 0.35% difference between the initial and optimized internal camera parameters. The images showed a median of 49521 key points per image and a median of 22632 310 matches per calibrated image. The large number of flights, large area and different illumination conditions led us to do separate processing and georeferencing of flights, with iterative project merging until the final model was obtained. In this process, we have used a total of 37 3D and 3 2D GCPs measured in the terrain. In order to improve the matching, 696 manual tie points were included, especially in areas covered by pyroclasts (lapilli and ash).
The point cloud was processed using full image scale, matching of image pairs using the aerial grid/corridor model and 315 geometrically verified matching using automatic advanced key points extraction. The advanced camera calibration was done using the alternative method, internal parameters optimization (all), external parameters optimization (all) and no automatic rematch. The point cloud densification was done using multiscale and half-image size, with an optimal point density and a minimum number of 3 matches. This option was selected after intensive testing with 4 and 5 matches, which generated large gaps in the point clouds, in areas which were well-resolved with 3 matches. The point cloud still shows sectors with no data 320 in homogeneous fine ash and lapilli covers, but those areas are small as discussed below and always outside the recent lava fields (Fig. 5). Since the target of the 3D survey are the lava fields, we guarantee that those sectors are well represented with many accurate points in the cloud.  The final digital surface model (DSM) and orthomosaic are presented with a resolution of 25 cm/pixel, which allow for 330 maintaining the root mean square error (RMSE) well below pixel size. Noise filtering and sharp surface smoothing were applied for the DSM, with interpolation using inverse distance weighting.

Point cloud
The point cloud model shows a RMSE of 0.08 m in X, 0.11 m in Y and 0.12 m in Z, evaluated using 13 independent check 335 points (Table 1). Even after including several hundreds of manual tie points, it was not possible to obtain a good quality point cloud all over the survey area. Areas of homogeneous ash and pyroclasts covers lack 3D data, but they do not impact the overall mapping of the area, as is discussed below. Preprint. Discussion started: 12 November 2020 c Author(s) 2020. CC BY 4.0 License.

Digital Surface Model
The DSM of the Chã das Caldeiras with 25 cm resolution shows unprecedented topographic detail and allows for excellent visualization and quantification of the terrain morphometry. The iterative improvement of the point cloud by a 345 detailed visual analysis of the DSM shaded relief model followed by adding almost 700 manual tie points in the model and reprocessing, allowed to reach a very high-quality result (Fig. 6). The final model is the result of a resampling of the first model obtained with 10 cm resolution. In order to guarantee the quality of its use, the DSM was divided in three quality zones (Fig. 7). The zonation is available in the dataset as a shapefile that can be used to mask the DSM depending on user needs. The high resolution (10 cm) shaded 355 relief model derived from the DSM, as well as the 50 cm equidistance contours interpolated from the DSM were used for the systematic visual analysis and for manual delineation of the areas with errors in the DSM.

365
The high-quality zones cover 96.8% of the entire survey and coincide with areas of rough surfaces with numerous automatic and manual tie points, where the morphology is very accurate, and the point cloud model has high resolution. -Pāhoehoe lava flows show a much smoother and homogeneous surface, but they have frequent fractures and lineaments.
They occupy generally small sectors of the orthomosaic and are bound by very rough a'ā lavas, facilitating point matching .
-Infrastructure, such as non-paved roads and houses show numerous matches and provide very accurate results. Cultivated areas occupy small sectors of the surveyed area, but the small holes dug to cultivate vines, as well as other small trees, are also very well represented in the DSM (Figs. 9-C and 9-D).   The medium-quality zones are sectors dominated by ash and lapilli, where sporadic 3D errors occur and occupy 0.6% of the survey. These areas, which in general can be used for visualization purposes and even for quantification, but with special care. Most errors in these zones are very small (dm scale) and can be smoothed by resampling, for example to 1-2 m resolution (Fig. 10).

400
The low-quality zones correspond to patches where the point cloud was poorly resolved, having numerous artefacts in the DSM (Fig. 11). These areas cannot be used for quantification purposes and their visualization shows errors, which sometimes are significant. The low-quality zones only occupy 2.6% of the survey area. These cases occur in very smooth surfaces of ash and lapilli or in sectors where a small number of overlapping aerial photos exists. They are located mainly in the base of some slopes, concave areas and also in the top of Monte Beco, due to the lack of photo overlapping.

Orthophoto mosaic
The digital orthophoto mosaic that shows a resolution of 25 cm (Fig. 12), but the high accuracy of the survey allowed to make a mosaic with a resolution of 10 cm. This product may be delivered upon request. The mosaic shows an overall high 415 graphic quality, with few problems relating to shadow effects close to the Bordeira wall in the south of the Chã das Caldeiras, and with varying illumination conditions in the lava flows of the northwest part of the survey, where stripping occurs. The sectors with medium quality in the DSM do not affect the overall quality of the ortho mosaic, but the interpolation may result in geometrical inaccuracies in the orthophoto mosaic in the areas of low quality in the DSM.

3D models for visualization
A 3D texture mesh (.fbx) was produced for visualization purposes, allowing for the accurate visualization of the whole 425 surveyed area (Fig. 13).  The lava flow field of the 2014-15 eruption was digitized manually using the orthomosaic and DSM. Unfortunately, our survey missed a small area of the lava flow with 0.007 km2 in the northwest sector of Chã das Caldeiras, close to Monte Amarelo and that sector had to be digitised using very high-resolution Google Earth imagery. The accuracy of the present survey allowed to calculate a new area for the 2014-15 lava flow field, which is 4.53 km2, a number smaller than the areas calculated by other authors using coarser resolution imagery, that varied from 4.8 to 4.97 km2 (5.8% to 8.9%). This 435 discrepancy can be explained by the higher spatial resolution of our dataset that allows more accurate delineations, identifying in addition several kīpukas, i.e. small 'islands' (interior elevations surrounded by lava during the 2014-15 eruption) and also to a spatial variation effect (Chen, 1999) that results from the computation of the same areas in products with different spatial resolutions.

Conclusions
The 23.9 km 2 very high-resolution digital surface model and orthophoto mosaic of the Chã das Caldeiras lava fields developed from UAV surveys of December 2016, show unprecedented detail and accuracy (resolution = 25 cm and RMSE = 0.103 m). 96.8% of the survey area has provided a very high-quality DSM, which due to the scarce vegetation and built areas 460 may be used as a DEM. The areas with moderate problems occupy 0.6% of the survey, with only 2.6% of the area showing poor-quality. The sectors with problems in the point cloud and DSM are those associated to very homogeneous ash and lapilli deposits. These areas can be easily masked out of the DSM by using the shapefiles made available in the dataset. The rough surface 'a'ā lavas and the smooth pāhoehoe flows are very accurately determined, as well as the volcanic cones. The resulting DSM and orthomosaic constitute base datasets of unprecedent detail of high-value for geological research and for 465 lava flow modelling with a high potential for applications in risk mitigation. These products allow delineating accurately the borders between different surfaces (lava types and other classes) and perceiving sub-meter surface features, which is less accurate or not achievable at all at meter scale, over an area of several square kilometres. These morphometrics features include pressure ridges, tumuli, flow channels, levées, dragged blocks and remains of human structures, among other smaller features. 470 Finally, we consider that these highly detailed products can play a relevant role in the assessment of volcanic hazards and related research and whose importance is surely excelled by becoming open access.