© M. Jaud et al., Published by EDP Sciences 2025

1 Introduction

Virtual Field Trips (VFTs) have been developed extensively in recent years, particularly as a consequence of the COVID-19 pandemic, which caused major disruption to classroom teaching and field-based activities. According to the literature (e.g. Bonali et al., 2021; Firomumwe, 2022; Horota et al., 2023; Peace et al., 2021), and quite logically, these VFTs mainly concern disciplines where fieldwork is a required part of the curriculum, in particular geosciences (earth sciences, physical geography, etc.).

The definition of a “Virtual Field Trip" varies over time and depending on the community and is therefore very broad. According to Horota et al. (2023): “a VFT aims to represent the real-world environment through digital data and without the cost of being ‘there’ physically”. And according to Wang et al. (2016): “A ‘virtual field trip’ is an amalgam of Internet and multimedia techniques”.

The first VFTs, often based on the use of Google Earth^®, had limited educational objectives, mainly aiming to “develop a sense of place in the local and regional environment” (Monet and Greene. 2012). However, innovations both in terms of hardware and software solutions in Virtual Reality (VR) over the last few years have contributed to make these VFTs more “immersive” (e.g. Harknett et al., 2022; Horota et al., 2023), i.e. with 3D perception through a head-mounted display rather than conventional 2D user interface. Immersive experiences are likely to provide better perception of the spatial and geometrical properties than in non-immersive VFT (Harknett et al., 2022; Lukacevic et al., 2020).

The literature points to a number of advantages for these VFTs, including the following:

• providing easy access to outcrops inaccessible in the field (e.g. Harknett et al., 2022),

• being more inclusive for students unable to go in the field for financial or health (or other) reasons (e.g. Chiarella and Vurro, 2020; Giles et al., 2020)

• not being dependent on field constraints (weather, transport, difficult/dangerous access, etc. − e.g. Giles et al., 2020)

• potentially facilitating access to new environments or new skills for developing countries (Firomumwe, 2022; Harknett et al., 2022; Omieno et al., 2013)

• economic and ecological considerations (Cliffe, 2017; Harknett et al., 2022).

However, the state of the art (including recent studies) also highlights the limitations of these virtual experiences. Some of the shortcomings to be considered include the following:

• Studies show that from the students’ point of view, VFTs cannot replace field camps because they do not allow them to develop certain technical skills that are useful in the field (Arrowsmith et al., 2005; Cliffe, 2017).

• Due to technical limitations, or limitations in the digital data available, VFTs do not provide an adequate representation of the environment (Wang et al., 2016). Furthermore, Bonali et al. (2021) note that younger users, probably more exposed to hyper real visualisations and gaming, are less impressed by the immersive VR experience.

• As a consequence of the previous point, the VFT has to be adapted to educational and scientific needs; adequate data, appropriate software and technical skills are required. All this has a financial and time cost (Hagge, 2021; Stott et al., 2009).

•  Wright et al. (2023) also described the VFTs as a relatively solitary experience, with less human interactions for the students, whether with peers or teachers, in comparison with field experience.

Based on these findings, a number of studies have looked at the potential of VFTs to replace ground fieldwork and note that students do not expect these virtual experiences to completely substitute for in situ experiences (Cliffe, 2017; De Paz-Álvarez et al., 2022).

Nevertheless, this state of the art also reveals that:

• some of the limitations of VFTs were linked to the quality of the data used in these environments (type of data, resolution, geometric fidelity, etc.);

• the users of these environments are generally not perfectly aware of the methods underlying the creation of these environments;

With the above considerations in mind, we thought it would be interesting to propose a hybrid field camp for Master students in Earth Sciences, combining a naturalistic approach in the field and a VFT. Nevertheless, in order to prevent the students from being solely ’immersed’ in the virtual environment, they also contribute to the creation of this environment, collecting and processing photogrammetric data. Students are familiarised with multi-source and multi-scale data and the use of associated digital tools. They also learn about low-tech technologies, collaborative methods and FAIR data.

2 Field Trip design: context, objectives

The ‘OceanField’ project is a mixed field trip, combining an actual “in situ” fieldwork component and a “digital field camp” component (including geological and numerical field data acquisitions as well as data exploitations, once back indoor). This field trip takes place in the French Alps, where sit major sites for understanding the geology of the Alps, witnessing the birth evolution, and the eventual disappearance of an oceanic domain, followed by continental collision leading to the final Alpine surrection.

In practical terms, the OceanField project can be summarised into three main components (Fig. 1):

• geological fieldwork and remote sensing in the field, with an introduction to the practical implementation of different Structure-from-Motion (SfM) photogrammetry approaches,

• remote sensing in the classroom, with presentations of the different types of data, vectors, products, etc. and training in photogrammetric processing using the datasets collected in the field, Geographical Information Systems (GIS) training, etc.

• virtual field camp, combining non-immersive and immersive tools.

The work presented here was carried out as part of the 2023 field trip, involving ten first year master students of the ‘Marine and Coastal Sciences’, with a ‘Marine Geosciences’ major, at the University of Brest (based in the European Institute for Marine Studies; IUEM). They are accompanied by three instructors, two geologists and an expert in remote sensing. The field camp takes place as soon as the school year begins in early September in order to minimise weather related risks (snow). Students are to mobilise on the knowledge and know-hows acquired during the previous years (undergraduate studies), since all of them formerly graduated in geology. In-class teaching eventually complements the field training.

The educational objectives of this hybrid field camp are:

• to encourage interdisciplinary understanding of earth sciences by combining traditional geological fieldwork with remote sensing techniques and GIS analysis;

• to train students with practical skills in data collection, processing, analysis, and visualisation, essential for conducting research and solving real-world geological and environmental problems;

• to develop students’ spatial thinking abilities and problem-solving skills both in the field and through the application of GIS techniques and spatial analysis methods to geological and environmental datasets;

• to introduce students to innovative technologies such as SfM photogrammetry, remote sensing, and immersive virtual field experiences, preparing them to utilise these tools in their future academic and professional missions;

• to encourage critical thinking and interpretation of geological and geospatial data, enabling students to draw meaningful conclusions from multi-source and multi-scale datasets.

Fig. 1 Timeline of the OceanField project.

3 Geological fieldwork and study areas

The fieldwork is part of the fundamental curriculum for our Master students majoring in Geology. It enables them to acquire and/or improve their field geological skills, such as: petrology, stratigraphy and structural analyses at different scales, oceanic lithosphere identification, relative chronology and outcrops sketching and mapping. In turn they improve their ability to observe, describe and interpret the geology of an oceanic domain.

In order to facilitate the students understanding of the geodynamic evolution in the Franco-Italian Alps, (from the Jurassic extension, to the Cenozoic shortening − collision and the eventual the present-day active deformation of the outer Alpine arc), the field camp focuses on six key study sites, visited in September 2023 (Fig. 2).

Fig. 2 a) Location of the field trip in the French Alps. b) Location of the various field camp study sites on the geological map (BRGM©). c) and d) Students carrying out geological fieldwork.

3.1 Ornon

This area is located on the Eastern flank of the Taillefer Massif in the external massifs of the Alps (Dauphinois zone). Students are there confronted with the relics of the rifting: the continental extension that precedes oceanisation. This site features Hercynian continental basement rocks crosscut by a subvertical tectonic accident (the so-called Ornon normal fault), overlaid by early Jurassic sediments deposited within the newborn marine basin. This sedimentary episode witnessed an important seismic activity, accompanying the fault functioning and recorded it by “fossilising” the fall of metric to hectometric Trias-age sedimentary and magmatic rock blocks (so-called olistoliths), dropped down the cliff, into the still unconsolidated Jurassic sediments. Finally, all of these rocks and structures were extensively deformed during the latest W-E Alpine compression episode. All these are fully documented in Barféty et al. (1979) or Lemoine et al. (2000). At this ∼1 × 1 km large outcrop, students are to: (i) identify the petrological diversity (dolomites, basaltic rocks, mudstone, and gneiss); (ii) recognize the fault surface, define its present-day geometry and use tectonic markers (such as tectonic striations) in order to determine its past motion; (iii) perceive the “abnormal” location and orientation of the olistoliths within the Jurassic sediments, and use their Triassic ages, in order to narrow down the timing of the early rifting.

3.2 Villard Notre Dame

This sector is located less than 6 km East from the Ornon fault and dominated by the presence of a ∼6 × 7 km massive granitic body (Rochail Granite), corresponding to a Hercynian basement horst, distinct from the Taillefer. As in Ornon this crystalline formation has been individualised by the Jurassic rifting, and syntectonic muddy sediments filled the basin. In this zone, the late alpine compression was more sustained than in Ornon, and most of the structures are folded, and partly to totally reversed, with originally ∼45° stip normal faults (from the rifting) turned into sub-horizontal inverse thrusts (granitic basement overlapping and deforming the latter Jurassic sediments). Here the students are to recognize all the types of rocks and formations discovered in the more preserved Ornon area, and apprehend how the late compression folded and shortened the whole rifted basin during the Cenozoic.

3.3 Saint Crépin

Saint Crepin is located East from the Pennic front (outer-inner Alpes limit) and therefore lies within the Briançonnais zone. This ∼1 × 1 km outcrop displays a small anticline valley, where Cretaceous calcshists cover the reduced deposit of Upper Jurassic limestones (Ammonitico Rosso facies, absent from the Dauphinois zone) in direct discordant contact with Triassic dolomites. These new facies, together with the absence of lower Jurassic sediments (massively present in Ornon and at the Rochail), should lead the students to a better understanding of the Tethys ocean paleogeography, where high points (horsts), formed during the rifting, were punctually located above the sea level (forming islands) and were, therefore, not subject to sediment deposition, but on opposite, could be affected by continental erosion, leading to stratigraphic hiatus. Here the students are to identify the rocks (including the new facies), apprehend the anticline geometry, locate and characterise the discordance and figure out that this part of the oceanic domain was necessarily distant from the Dauphinois zone, considering the different sedimentary histories.

3.4 Béraudes

This study area is located ∼40 km further north within the Briançonnais zone, on the Western rim of the Haute Clarée valley. This zone features another distorted relic of the rifted and subsequently-tectonized margin. Triassic sediments (sandstone and dolomites) covering carboniferous layers are cross-cut by two antithetic normal faults inherited from the rifting phase and subsequently deformed by the E-W Cenozoic Alpine shortening to form the so-called Cerces syncline. Here the students can remobilise the know-how (petrology, structural geology, etc.) gathered at the three previous studies and take advantage of the peculiar topography, in order to get an actual 3D view of such a structure and figure out its polyphased geological history.

3.5 Chenaillet

This area located 30 km south of the Clarée Valley, close to Montgenèvre, is a major example of a preserved Tethysian ophiolitic sequence, including serpentinized peridotites, albitites, isotropic gabbros, dolerites, basalts (pillow lavas), sediments and hydrothermal deposits which are typical constituents of an ocean lithosphere section. The actual geodynamic significance and structure of the section exhumed here remain debated (see and references therein), and is beyond the point of the fieldclass. At the Chenaillet, students are mostly expected to train their igneous petrological skills, get a broad kilometric vision of the geometry of the complex and understand the petrologic control on the topography.

3.6 Saint Veran

In this zone, sited ∼80 km south-East of the Chenaillet, takes place within the so-called Liguro-piemontais domains. the geology of the sector is dominated by green schists detritical metasediments (shiny schists), packing masses of tens to hundreds of metres wide of Paleocene-Eocene blue schists metamorphosed oceanic rocks (serpentinites, metagabbros, metabasalts, see Tricart and Schwartz, 2006 for details). In this area, students are to recognize petrological and map the different units, get a broad 3D view of the overall geometry of the complex and point out the structural relationship between the blocks and the surrounding metasediments. The petrological recognition of metamorphic rocks containing minerals, typical of the blueschist facies (e.g. glaucophane), with evidence of protolites likely to be similar to the rocks observed in the preserved ophiolitic sequence of the Chenaillet, should lead the students to identify here, a witness of the final stage of the Tethys ocean floor history, once subducted back into the mantle, during the ocean closure and later exhumed as a metamorphic ophiolite during the Alpine surrection.

4 Multi-source remote sensing data

The fieldwork is paired with a classroom module on remote sensing and Geographical Information System (GIS), in which students acquire both theoretical knowledge and practical skills (remote sensing data handling, software tools, etc.) on multi source remote sensing data acquisition/processing/visualising and interpreting. Since the OceanField project was set up, this module has been adapted to give greater prominence to data relating to OceanField. A multi-source and multi-scale approach will enrich the virtual field camp and familiarise students with the variety of remote sensing data, the different platforms and the concepts of spatial coverage, resolution and scale (Fig. 3).

Fig. 3 Combining the multi-source and multi-scale data used in OceanField.

4.1 Satellite and aerial imagery

DINAMIS — “Dispositif Institutionnel National d’Accès Mutualisé en Imagerie Satellitaire”, the French national facility for institutional procurement of Very High Resolution satellite imagery — enables us to program Pleiades stereo-imagery acquisitions above OceanField study areas. The Pléiades satellites 1A and 1B, from Airbus DS and CNES (French Spatial Agency), were launched in 2011 and 2012 and provide very high-resolution optical imagery (0.5 m resolution) with a swath of 20 km. The stereo-pairs were collected on 2022/07/26, 2022/07/31, 2022/08/13 and 2022/09/22.

We also use data from IGN^© (the French National Geographic Institute) databases, available online: https://geoservices.ign.fr/. For the relief data, we downloaded georeferenced raw point clouds from the LiDAR HD^® program (IGN, 2023), which aims to provide 3D mapping of the French territory. LiDAR HD 3D point clouds are derived from aerial LiDAR acquisition with a density of at least 10 pulses per m² and with a maximum mean square error (RMSE) of 50 cm for planimetry and 10 cm for altimetry. In the Alps, the data was collected in 2021. For texture, we downloaded some images from the BD ORTHO^®, a collection of aerial orthophotographs with a resolution of 20 cm. These images were collected in 2021 (for Isère department) or in 2022 (for Hautes Alpes department).

4.2 In situ photogrammetry

The in situ photogrammetry data includes data collected during the field trip and for which the students participated in the acquisition (Fig. 4). Where regulations, weather conditions and site frequentation allowed, Unmanned Aerial Vehicle (UAV) photogrammetry data was acquired using a DJI Mini 3 pro®, equipped with a 12 Mpix camera (focal length: 7 mm). As this UAV is ultra-light (249 g), it implies less operational restrictions and requires no training. The UAV can therefore be flown by students without the need for a licence (Fig. 4a), as long as flight regulations are respected. As this UAV is not equipped with Real Time Kinematic Global Navigation Satellite System (RTK GNSS) positioning, to guarantee the geometric quality of the reconstruction, we used ground control points (GCPs) (Eltner et al., 2016; Tonkin and Midgley, 2016). The students contribute to establish the GCP network and to measure the GCP with RTK GNSS.

Where drone flights were impossible or not appropriate, the students got a terrestrial photogrammetry training. Terrestrial photogrammetry has the advantage of being straightforward to implement, inexpensive, highly adaptable to weather conditions, highly transportable and largely free of regulatory constraints. It is in turn a particularly well-suited method for the training of non-specialist students. Students carried out a terrestrial photogrammetry technique using Real Time Kinematic (RTK) positioning as proposed by Jaud et al. (2020) and illustrated in Figure 4b. An RTK GNSS antenna (in this study, a SparkFun RTK Facet^®, connected to the Centipède collaborative network of bases − https://centipede.fr/) is mechanically held by a wooden frame at a constant distance from the camera, along a vertical axis. This allows to subsequently determine the position of each image with centimetre accuracy, avoiding the use of GCP or scale bars.

On other sites, the students also carried out collaborative photogrammetry surveys, using their personal smartphones, in order to photograph geological objects and subsequently gather all the photos together for processing (Fig. 4c). As the geotag accuracy of the smartphone photos was not sufficient to ensure a geometrically compliant reconstruction (Bessin et al., 2023; Jaud et al., 2020), scale bars were placed in the environment in order to refine the self-calibration of the cameras and in turn improve the quality of the reconstruction.

Apart from the collaborative smartphone surveys, for which they were all mobilised at the same time, students were split into 2–4 people groups, focusing either on geology work or photogrammetry acquisitions. The turnover in operations allowed all the students to experiment with the different approaches to in situ photogrammetry.

Fig. 4 Involvement of students in in situ photogrammetry surveys in Béraudes (a,b) and St Véran (c) study areas, using UAV photogrammetry (a), terrestrial photogrammetry method with RTK positioning (b) and smartphone collaborative photogrammetry (c).

5 Data processing, analysis and sharing

5.1 Data processing to create 3D models

Pleiades stereo-imagery was processed using Agisoft Metashape Pro software (v.1.6.4), as described in Lastilla et al. (2020) and Bessin et al. (2023). The raw satellite images were not made available to the students. Instead, they were provided with Digital Elevation Models (DEMs − 1 m resolution) and orthoimages (0.5 m resolution) generated previously by the scientific team.

The datasets from the IGN databases do not require any special processing. Nevertheless, to facilitate their use in the GIS and immersive VR software (presenting later), the Lidar point clouds were interpolated onto a regular grid (at 40, 50 or 60 cm resolution depending on the sites and the surfaces considered) in order to convert them into raster DEMs.

Agisoft Metashape Pro software (v.2.0) was also used for SfM processing of UAV and terrestrial photogrammetric datasets, following the workflow presented in Figure 5a. For this part, the students worked in pairs or trios (Fig. 5b). After a brief introduction to the method, they process the data themselves, relying on tutorials (on paper or video) and asking the teacher if they had any problems or questions. Due to time constraints, not all the datasets were processed by the students (the biggest ones were processed by the supervisors). Computers powerful enough for this type of processing and equipped with appropriate software were made available to them for this task. The processing protocol is of course adapted according to the dataset and the approach used (GCP, RTK positioning of the camera, geotag of Smartphone photos and scale bars). In certain areas with vertical surfaces or overhangs, processing was not continued beyond the generation of the dense point cloud (Fig. 5c) so as not to lose the 3D information during rasterization (2.5D).

Each group of students has to submit a report in which they present their work (theoretical principles of photogrammetry, procedures, quality of the results) so that we can ensure that they have properly assimilated the methods.

Fig. 5 a) SfM photogrammetry workflow. b) Trio of students processing data in Agisoft Metashape for the reconstruction of the Col d’Ornon study area. c) Example of the 3D point cloud obtained by photogrammetric reconstruction using collaborative smartphone images in St Véran study area.

5.2 Data analysis using different tools

From the collecting data and resulting 3D models, we have designed several pedagogical activities to extend the OceanField in situ fieldtrip with a VFT (with immersive and not immersive activities). The fact that the students have already investigated the geological context (regionally and site by site) during the field course enables them to be more effective in the geological analysis of the digital data. Consequently, the VFT sessions also devote a large part to the analysis of remote sensing data itself: comparison of different types of data over the same area, impact of spatial resolution, suitability of this resolution and of the acquisition geometry (aerial or terrestrial) for the scientific question, etc. Three main software tools are used for these VFTs:

• QGIS, a well-known free and open source Geographic Information System, for viewing, modifying and analysing geospatial data.

• CloudCompare, a free and open source software for viewing, editing and processing 3D data (mainly clouds or meshes), offering a wide range of measurement and analysis functions (distance calculations, statistics, estimation of geometric properties, projection, registration, segmentation, etc.).

• VRExplorer, an immersive and interactive VR platform (developed for teaching purposes by VR2Planets: https://www.vr2planets.com/education/vrexplorer/), in which several scenes can be prepared and selected.

QGIS and CloudCompare being non-immersive tools, used in a “classic” way on a computer screen, we won’t go into further detail about their use. VRExplorer proposes a multi-user environment (with avatars), which enables collaborative learning in immersive environments. Students had a short step by step presentation of the VR environment functionalities. A set of mapping and measurement tools are proposed: measuring distances, positioning or delineating geological features or landforms via shapefile edition, measuring an azimuth or dip and dip direction, etc. The students then worked independently in two-person groups (Figs. 6a and b). What the ‘immersed’ person sees (with the virtual reality headset) is transcribed onto a computer screen. So, while one of the students is immersed, his/her partner can guide or assist him/her, in particular by referring to the information gathered in the field or by controlling the measurements performed in a virtual environment (Fig. 6c).

QGIS and the VRExplorer immersive environment are used in parallel. These two tools are ideally suited to combine multi-scale data: Pléiades satellite imagery, IGN^© HD databases and in situ photogrammetry data. In addition, most of the measurements taken in VRExplorer can be exported in GeoJSON format and then imported into QGIS.

CloudCompare was used in particular with in situ photogrammetric models, exported as dense 3D point clouds, on study areas with sub-vertical outcrops or overhangs. The students had to carry out a number of actions: colouring the point cloud in true colours or according to altitude, extracting profiles, creating lists of points, measuring distances, calculating dips and dip directions (Fig. 6d), segmenting the cloud manually (Fig. 6e), etc. In this way, they familiarise themselves with as many of the software’s functions as possible. At the end of these activities, when relevant, they were asked to compare the measurements taken in the digital environment with their own measurements taken during the field trip (Fig. 6d), and to consider any differences and possible sources of error or uncertainty (both in the digital environment and in the field).

Fig. 6 a) and b) Immersion in the OceanField data using VRExplorer software. In groups of two, while one student works in the virtual environment, their partner assists them by controlling from the screen or guiding them through the space. After a while, the roles are swapped. c) Example of measurement of dip direction and dip (in degrees) carried out in the VRExplorer virtual environment, at the Béraudes study area.

5.3 Data sharing and re-use

To ensure that students are aware of all the stages in the data lifecycle, we also address the issue of data sharing and reuse. For example, during exercises in QGIS and VRExplorer comparing different data sources, students are encouraged to take particular interest in metadata.

With a goal of sharing and re-use, the 3D models produced as part of the OceanField project are deposited on an online platform, with Creative Commons licences (Fig. 7). We choose the V3Geo platform (https://v3geo.com/), purpose-built for sharing virtual 3D models of high quality within the geoscience community (Buckley et al., 2021). This platform introduces students to the value of metadata through practical examples. When data is deposited, basic information must be completed (model’s name, author, funding source, country/region, “geological tags”), which is then used to search for models (Fig. 7b).

In addition, due to time constraints, not all the data sets were analysed during the VFT. So, making the data available on a user-friendly web platform means that students can re-access the 3D models that we didn’t have time to detail during the VFT in class. Students are particularly keen to see the models created from the surveys in which they were involved.

Fig. 7 a) Map interface showing the 3D models shared in V3Geo platform in the frame of the OceanField project. b) Example of model page on V3Geo (St Veran study area), showing metadata, model description, tags and image carousel serving as a hyperlink to launch the 3D model viewer. c) Example of V3Geo web viewer (Chenaillet study area) for 3D interactive navigation.

6 Discussion

6.1 Practical constraints for the combination of real and virtual fieldworks

The aim of combining “real” and virtual field camps is to optimise the synergy between the two approaches. There are, however, certain constraints. For example, for geology and remote sensing teachers, it has been necessary to adapt the content of the various teaching modules. In addition, as the students work in VFTs on their own data, this requires a significant amount of time each year for the supervisors to prepare the digital datasets.

During the field camp, this combined approach requires greater involvement from the students to be able to manage both the geology-oriented activities and the photogrammetric survey activities. Furthermore, including photogrammetric surveys in the field camp adds a number of constraints: transporting equipment, weather conditions, legislation on drone flights, etc. Indeed, these photogrammetric surveys are even more sensitive to weather conditions than traditional geological fieldwork. In order to adapt to these field constraints, alternative methods can be used (e.g. kite photogrammetry if the wind is too strong for the drone) or low-cost methods can be carried out to spare expensive instruments in case of breakage. The decision to choose a given technique is the result of a collective discussion involving the students (although guided by the teachers), in order to find the best compromise between the geological problem to be addressed, the remote sensing resources available and the constraints in the field. Experiencing alternative techniques teaches students to adapt to changing field conditions and constraints and enhances their versatility. In this way, students also gain insight into the decision-making processes (finding a trade-off between cost-effectiveness, feasibility, safety, and data quality).

6.2 Benefits for the student’s curriculum

With the OceanField project, students increase their skills in field geology, by developing their ability to observe, describe and interpret the geology of an emblematic domain to understand rifting, oceanisation and collision processes. With the possibility of enriching the catalogue of virtual sites available from year to year, this offers a degree of flexibility for adapting the sites visited in person during the field camp according to the physical condition of the students in the group, the weather conditions, etc. The VFT complements the time-limited fieldwork with digital models and immersive environments, but it also enables the students to learn about tools that are now nearly indispensable, such as GIS, or to familiarise themselves with tools that are currently expanding, such as immersive VR environments and low-tech techniques.

The great originality of OceanField is that it involves students in the acquisition and processing of photogrammetry data, and then in the creation of their virtual environment. In this way, they develop their ability to define data acquisition plans. They acquire additional technical skills for generating 3D models and manipulating and analysing this 3D data, by involving them in the various stages of the data lifecycle (Fig. 8). This approach gives them a better understanding of data (scale, spatial coverage, resolution, data availability, feasibility of the survey, etc.). Being confronted with multi-source and multi-scale data and with various analysis tools encourages them to think about the right match between the scientific need and the technical constraints and the most relevant method to answer a given scientific question.

Students are, in turns, better able to assess the potential and limitations of each approach and how they complement each other. Finally, they become more aware of the quality of observations by remote sensing, the possible sources of error and the consequences for the uncertainties that can generate in geological interpretations. Taking a critical step back is a real asset in a context where virtual environments are becoming more and more visually realistic, although this is not necessarily a guarantee of accuracy.

Fig. 8 Student involvement at the different stages of the data lifecycle in the OceanField project.

6.3 Student’s feedback

Among the methods and tools implemented as part of this project, only one student (out of ten) had already used QGIS. This approach was therefore generally new to them. In the field, they were enthusiastic about the idea of familiarising themselves with photogrammetry methods. And, after the field camp, they were eager to process and analyse the data they had collected themselves.

At the end of the project, we asked them to answer a basic survey to gather their feedback (which are presented below, in Fig. 9 and Tab. 1). All the students testified that they had a better perception of the distances and dimensions of objects in real life than in digital data (Fig. 9a). Some mentioned that this may be due to a lack of familiarity with digital data or to the fact that perception is less direct in these digital environments as it requires the use of measurement tools. A majority (70%) considers that the combination of the real field and VFT approaches provides a better understanding of the area’s geology (Fig. 9b). For the students, going back to the digital multi-scale data, weeks or months after the real field trip, is therefore a valuable contribution.

Software tools that allow data to be viewed from different angles (CloudCompare and VRExplorer) are preferred for data exploration (Fig. 9c), while QGIS and CloudCompare seem the most appropriate for geological measurements and interpretation (Fig. 9d). They also mention QGIS as an appropriate tool for synthesising information about the area. For most students, CloudCompare is more intuitive than immersive VR, where using joysticks (and more generally moving around in VR) is less familiar than using a mouse on a screen. The immersion in VR was seen as fun and useful for familiarising themselves with an area, but not essential for taking measurements. Several students admitted that they found it difficult to point accurately in VR. This may be partly due to the fact that immersive VR software currently only offers the option of importing 2.5D data (DEM + texture), but not 3D data, which results in a loss of information and a blurred effect on quasi-vertical or overhanging walls. 3D data handling should be possible with the next version of the software.

The survey also included open-ended questions in which the students were invited to express their views on the strengths and weaknesses of real and virtual field trips. Table 1 compiles their responses. Compared to our state of the art of the advantages and limitations of virtual terrains (immersive or not), our students also point out that these virtual terrains offer the possibility of accessing certain outcrops (or parts of outcrops) that are not easily accessible on the field (for example, as in Harknett et al., 2022) and that are only accessible in certain weather conditions (for example see Giles et al., 2020). As Cliffe (2017), our students also pointed out that VFTs did not allow them to develop certain technical skills (requiring, for example, breaking a rock, using their sense of smell, using a magnifying glass, etc.) which are fundamental to a detailed understanding of the geological object. They also noted that although fun compared to a ’classic’ course, VFTs did not allow for the same degree of human interaction as real field trips in an outstanding learning environment.

These answers also show a good understanding of the hybrid approach (in situ and digital) and of the multi-source and multi-scale approach. In their view, exploring the site ‘for real’ beforehand makes it easier to analyse and interpret the data in the VFT. They feel that these digital tools are also useful for producing more accurate and more complete deliverables (more exhaustive analysis of the area, more quantitative data) than diagrams drawn up in the field.

In a satisfaction survey carried out at the end of their Master 1 year, 90% of students said that field and digital geology methods were complementary, and 85% that the use of photogrammetry, GIS and 3D tools was suited to the teaching objectives of the course. Overall, the students were very satisfied with their exploration of a wide range of techniques and software, and felt that some of these methods will be useful to them in the future, since 100% of them would recommend this type of hybrid experience to other geosciences students.

Fig. 9 Student responses to the feedback survey (the group is made up of 10 students. Some questions could have multiple answers.)

6.4 Future evolution of the class

The field class enhanced the students’ understanding of the proper use of photogrammetric tools, while working with the DEM models deepened and extended their grasp of the geologic observations. As the feedback from the students was positive regarding both the pedagogical and technical aspects, the principle of this hybrid field camp is in turn, intended to be reproduced and further developed during the coming years, by collecting data on new sites of interest, for example, in the Barles aera (Alpes de Haute-Provence, France, field class scheduled for October 2024). In the long term, this will expand the catalogue of available virtual field sites, complementing the limited-time observation work done in the field. It will also provide teachers with a rich library of visual resources for use in more traditional, classroom-based lessons. We are also aiming to provide a database of technical resources in the form of video courses and tutorials. These video resources (https://www.youtube.com/@IUEM-P2I_pole-image/videos) are also intended to be enriched and updated over time, in line with technical developments and the needs expressed by students.

7 Conclusion

With this hybrid approach combining real and virtual field trips, the fieldwork provides knowledge and know-how in geology (mapping, analysing the landscape and the geodynamical processes, identifying rocks and geological features) and in geomatics, as well as technical skills for the acquisition and processing of in situ remote sensing data (RTK GNSS measurements, photogrammetric acquisition and processing). The digital approach that follows, based on QGIS, CloudCompare and immersive VR with VRExplorer, is then enriched by a consideration of the quality of the data and its relevance to scientific questions. Students are thus involved in the various stages of the data life cycle.

This project creates synergy between different disciplines and prepares students for research careers and professional practice in earth sciences by providing them with the necessary theoretical knowledge, practical skills, and technological experience. The overall impression of the students was positive, and all of them suggested that this approach should be continued for future classes. Furthermore, the virtual environments created as part of this project have the potential to be reused in other contexts and with other audiences (outreach, lifelong learning, partnerships with territorial managers, etc.).

Acknowlegdments

Many thanks to Mélina Boulic, Titouan Brousse, Elyna Clément, Margaux Clochon, Maïwen Collas, Mina Gautier, Younic Jarry, Manon Mabo, Corentine Ribbe and Clara Saint-Jean, 1^st year ‘Geosciences Marines’ Master students at the University of Brest in 2023-2024 for participating in this study.

We are also grateful to Jean-Marie Gilliot, lecturer at IMT Atlantique, for his proofreading and expert advice.

This teaching fieldwork, OceanField, benefits from the financial support of ISblue project, the Interdisciplinary graduate school for the blue planet (grant number ANR-17-EURE-0015, co-funded by a grant from the French government under the program “Investissements d’Avenir” embedded in France 2030).

This work was conducted in collaboration with the ImmerSea LAB platform (supported by ISblue and the Dassault Systèmes Foundation) for the development of real and virtual collaborative practices.

It benefited also from a number of interactions with VR2Planets, the company that develops the VRExplorer immersive virtual reality software for teaching. We would like to thank them for their attentiveness and patience.

References

All Tables

All Figures

	Fig. 1 Timeline of the OceanField project.
In the text

	Fig. 2 a) Location of the field trip in the French Alps. b) Location of the various field camp study sites on the geological map (BRGM©). c) and d) Students carrying out geological fieldwork.
In the text

	Fig. 3 Combining the multi-source and multi-scale data used in OceanField.
In the text

	Fig. 4 Involvement of students in in situ photogrammetry surveys in Béraudes (a,b) and St Véran (c) study areas, using UAV photogrammetry (a), terrestrial photogrammetry method with RTK positioning (b) and smartphone collaborative photogrammetry (c).
In the text

	Fig. 5 a) SfM photogrammetry workflow. b) Trio of students processing data in Agisoft Metashape for the reconstruction of the Col d’Ornon study area. c) Example of the 3D point cloud obtained by photogrammetric reconstruction using collaborative smartphone images in St Véran study area.
In the text

Fig. 6 a) and b) Immersion in the OceanField data using VRExplorer software. In groups of two, while one student works in the virtual environment, their partner assists them by controlling from the screen or guiding them through the space. After a while, the roles are swapped. c) Example of measurement of dip direction and dip (in degrees) carried out in the VRExplorer virtual environment, at the Béraudes study area.

In the text

	Fig. 7 a) Map interface showing the 3D models shared in V3Geo platform in the frame of the OceanField project. b) Example of model page on V3Geo (St Veran study area), showing metadata, model description, tags and image carousel serving as a hyperlink to launch the 3D model viewer. c) Example of V3Geo web viewer (Chenaillet study area) for 3D interactive navigation.
In the text

	Fig. 8 Student involvement at the different stages of the data lifecycle in the OceanField project.
In the text

	Fig. 9 Student responses to the feedback survey (the group is made up of 10 students. Some questions could have multiple answers.)
In the text