Luca Vittuari

The Drygalski Ice Tongue is a tide-influenced floating extension of the David Glacier in East Antarctica, where field observations are constrained by polar logistics. We investigated how horizontal ice motion and vertical tidal response vary along and across the tongue. To extend earlier observations, we analysed a 2025–2026 Global Navigation Satellite System (GNSS) campaign comprising 12 stations and 248 station-days. Kinematic GNSS coordinate series were produced with Bernese 5.4 and subsequently analysed under four stochastic covariance descriptions for horizontal and apparent vertical-rate estimation. Comparisons with a regional Mario Zucchelli Station (MZS)/Terra Nova Bay tide reference characterized station-level up-component response to tidal forcing, and two tide-removal paths were evaluated for apparent vertical-rate analysis. Floating-tongue speeds increased downstream, spanning 663.9–749.4 m/yr, with a clockwise turn toward the southeast and north-to-south differences in fitted campaign rates. The main horizontal patterns persisted across all four tested covariance descriptions. Vertical tidal response was weak in the near-grounding group, spatially heterogeneous in the transition sector, and near-unity downstream. Across the four covariance descriptions, neither tide-removal approach yielded a model-robust apparent vertical-rate field. The campaign resolved spatial patterns of horizontal motion and vertical tidal response, while covariance testing identified the limits of apparent vertical-rate inference from short GNSS records.

Mobile Mapping Systems (MMSs) are increasingly adopted for transport infrastructure inspection, where reliable characterization of global and local geometry is essential. This is critical in GNSS-denied environments such as road tunnels, where trajectory estimation relies on inertial navigation and drift degrades geometric accuracy. This study investigates the influence of acquisition speed and ground control point (GCP) configuration on trajectory accuracy and geometric consistency in a 2.3 km road tunnel. MMS data were acquired at 15, 30, and 50 km/h and processed under progressively denser, uniformly distributed GCP configurations, from unconstrained to fully constrained solutions. Post-processed forward–backward smoothed GNSS/INS trajectories were validated against a topographic control network, and cross-sectional geometry against terrestrial laser scanning (TLS). Without external constraints, planimetric errors at independent check points reached 16.5 m and global deformation reached 0.46% at low speed. Ground control markedly reduced trajectory errors, depending on both control density and acquisition speed. In contrast, locally registered cross-sections showed millimeter-level consistency with TLS in all configurations: median discrepancies of 4–5 mm, with 75–82% of values below 10 mm. Local cross-sectional shape is therefore preserved despite substantial global drift, supporting MMS for shape-based tunnel inspection, while adequate ground control remains necessary for absolute georeferencing.

The tie vectors at co-location sites are required for combining the individual space geodetic solutions and computing the International Terrestrial Reference Frame. Their provision to the combination centres entails a proper and accurate alignment of the tie vector into the global geocentric frame. In order to assess the impact of different alignment strategies on the space geodetic combination, we test four alignment approaches—two similarity transformations and two transformations based on the local values of the deflection of the vertical—and focus on the residuals of the combination. Our results clearly show that the alignment strategy not only controls the orientation on the tie vector in the global frame—determining its global coordinates and components—but also impacts the magnitude of the tie vector’s combination residuals to an extent that can exceed the 1 cm level. In practice, the information carried by the same tie vector is affected by the alignment strategy—a step in the tie vectors’ production that we prove is critical and that can impact the combination of space geodetic techniques. At both sites, the alignment strategy (local topocentric to the global geocentric ITRF frame) that ensures higher consistency (i.e. smaller residuals) in the computation is the similarity transformation that makes use of all the available space geodetic reference points and no information about the DoV.

Geodetic observations play a key role in quantifying glacier dynamics and long-term vertical surface changes in Antarctica, where Global Navigation Satellite System (GNSS) techniques are widely used under challenging environmental and logistical conditions. This study investigates the impact of ambiguity resolution in Precise Point Positioning (PPP) processing for glaciological applications through comparisons with both network-based solutions and earlier PPP implementations relying on float ambiguity estimation. Two complementary datasets collected within the Italian National Antarctic Research Program are analyzed: a short-term kinematic GNSS campaign on the rapidly flowing Drygalski Ice Tongue and a long-term continuous time series from the permanent GNSS station DCRU installed at Station Concordia (Dome C, East Antarctic Plateau). The analysis focuses on the consistency of displacement time series, stochastic noise characteristics and long-term velocity estimates derived from different processing strategies. Results indicate that ambiguity-fixed PPP solutions provide displacement and velocity estimates consistent with network-based and float PPP processing, while generally exhibiting reduced short-term scatter and improved temporal stability. This behaviour is particularly evident for the kinematic observations at Drygalski, where the improved signal-to-noise ratio enhances the resolution of small-amplitude glacier motions associated with ocean-tide forcing. Spectral analysis further confirms the tidal origin of the dominant periodic components, with peaks at approximately 25.6 h and 23.4 h. At Dome C, the long time-span observation record confirms the robustness of three-dimensional velocity estimates over more than a decade, with only minor differences between processing approaches. These findings highlight the overall consistency of glacier motion estimates with respect to the adopted processing strategy and demonstrate the benefits of ambiguity-resolved PPP for improving the interpretation and long-term reliability of Antarctic GNSS time series.

High-resolution Digital Terrain Models (DTMs) are essential for precise terrain analysis, yet their production remains constrained by the high cost and limited coverage of LiDAR surveys. This study introduces a deep learning framework based on a modified Residual Channel Attention Network (RCAN) to super-resolve 10 m DTMs to 1 m resolution. The model was trained and validated on a 568 km² LiDAR-derived dataset using custom elevation-aware loss functions that integrate elevation accuracy (L1), slope gradients, and multi-scale structural components to preserve terrain realism and vertical precision. Performance was evaluated across 257 independent test tiles representing flat, hilly, and mountainous terrains. A balanced loss configuration (α = 0.5, γ = 0.5) achieved the best results, yielding Mean Absolute Error (MAE) as low as 0.83 m and Root Mean Square Error (RMSE) of 1.14–1.15 m, with near-zero bias (−0.04 m). Errors increased moderately in mountainous areas (MAE = 1.29–1.41 m, RMSE = 1.84 m), confirming the greater difficulty of rugged terrain. Overall, the approach demonstrates strong potential for operational applications in geomorphology, hydrology, and landscape monitoring, offering an effective solution for high-resolution DTM generation where LiDAR data are unavailable.

High-resolution Digital Terrain Models (DTMs) are essential for precise terrain analy-sis, yet their production remains constrained by the high cost and limited coverage of LiDAR surveys. This study introduces a deep learning framework based on a modified Residual Channel Attention Network (RCAN) to super-resolve 10 m DTMs to 1 m res-olution. The model was trained and validated on a 568 km² LiDAR-derived dataset us-ing custom elevation-aware loss functions that integrate elevation accuracy (L1), slope gradients, and multi-scale structural components to preserve terrain realism and ver-tical precision. Performance was evaluated across 257 independent test tiles repre-senting flat, hilly, and mountainous terrains. A balanced loss configuration (α = 0.5, γ = 0.5) achieved the best results, yielding Mean Absolute Error (MAE) as low as 0.83 m and Root Mean Square Error (RMSE) of 1.14–1.15 m, with near-zero bias (–0.04 m). Er-rors increased moderately in mountainous areas (MAE = 1.29–1.41 m, RMSE = 1.84 m), confirming the greater difficulty of rugged terrain. Overall, the approach demonstrates strong potential for operational applications in geomorphology, hydrology, and land-scape monitoring, offering an effective solution for high-resolution DTM generation where LiDAR data are unavailable.

The Beyond EPICA Oldest Ice project is a European project that aims to retrieve a continuous ice core up to 1.5 Ma through deep drilling at Little Dome C (LDC), Antarctica. In order to determine the age of the ice at a given depth before the ice is analysed in detail, 1D numerical models are often employed. However, they do not take into account any effects due to horizontal ice flow. We present a 2.5D inverse model that determines the age–depth profile along a flow line from Dome C (DC) to LDC, which is assumed to be stable in time. This means that flow line features such as flow direction and dome location have not changed over the time period considered. The model is constrained by dated radar internal reflecting horizons. Surface velocity measurements are used to determine the flow line and flow tube width, which also allows the model to consider lateral divergence. This new inverse model therefore improves on the methods used by 1D models previously applied to the DC area. This 2.5D model uses a previously developed numerical scheme with the novelty being the inverse methods used to optimise multiple parameters by comparison to radar constraints. By inferring a mechanical ice thickness, the model predicts either the thickness of a basal layer of stagnant ice or a basal melt rate. Results show that the deepest ice at Beyond EPICA Little Dome C (BELDC) originates from around 15 km upstream. The threshold for ice useful for paleoclimatic reconstruction is 20 kyr m⁻¹ (20 000 annual layers per metre in the ice column). The oldest ice that meets this age resolution requirement is 1.12 Ma at BELDC according to the model. Over the LDC area, the thickness of a modelled basal layer is 200–250 m at the base of the ice sheet. Looking at modelled ice particle trajectories, interpretations indicate that this layer could be composed of stagnant ice, disturbed ice or even accreted ice (possibly containing debris). We explore the possibilities, though this is an open question that may only be answered by analysis of the Beyond EPICA ice core itself. Finally, we look at other ice cores that have already been drilled and measured, where the model has been applied. We discuss in detail the thinning in the deepest section of these ice cores, which is less than predicted by the model. This could mean that modelled ages are significantly overestimated in the deepest part of the ice column. Given that the age estimate from the 2.5D model is younger than previous estimates, this work shows the importance of considering the representation of the effects of horizontal flow when modelling the age profile.

Marine habitats are subject to monitoring activities to assess the impact of climate change, as required by EU and national directives. Coralligenous reefs are typical Mediterranean underwater seascapes, which originate from the growth of calcareous bioconstructions, including Lithophyllum stictiforme, a slow-growing calcareous red algae. The impact of temperature variations on its growth has only recently been investigated in the field, although with destructive sampling methods. The MANATEE (Monitoring and mApping of mariNe hAbitat with inTegrated gEomatics technologiEs) project aims to establish a non-destructive monitoring approach by leveraging geomatic techniques and underwater photogrammetry. The project will test three optical-based systems for divers and on different types of uncrewed underwater vehicles – UUVs (a low-cost micro Remotely Operated Vehicle – ROV and an observation-class UUV). This article presents the experimental procedure to verify that the developed underwater photogrammetry-based method can guarantee the detection of millimetre-level changes per year in the algae growth. A statistically significant number of samples were collected, prepared, and fixed on specially designed tiles and nets featuring photogrammetric targets and placed on the seabed. Following metric traceability principles, control elements (calibrated scale bars and colour checkers) were employed to verify geometric and colorimetric results of underwater photogrammetric surveys. The focus here is on the geometric verification, and the analyses presented demonstrate how the sub-millimetre accuracy potential and reproducibility meet the requirements for Lithophyllum stictiforme monitoring.

Understanding the behaviour of domes under both contemporary and historical environmental conditions is essential to facilitate the study of dome-divide dynamics and the interpretation of ice core records. This paper presents nearly 20 years of Global Navigation Satellite System observations at Dome C in East Antarctica, focusing on ice velocity and accumulation rates. The 38 measuring poles established in 1996 for the EPICA Dome C project were surveyed three times in 18 years. The data analysis indicates alterations in ice velocity patterns, including a horizontal velocity shift across the dome and a dynamic summit migration of about 100 m a ⁻¹ . Specifically, increases in velocity on the southeastern slope were counterbalanced by a similar reduction in the northwestern sector. These changes are likely related to variations in accumulation distribution as indicated by snow radar measurements and shifts in the drainage basin of the Byrd Glacier. Furthermore, a 10% alteration in snow accumulation rates at Dome C over the past decade compared with previous centuries was observed, accompanied by an elevation increase of about 3.5 mm a ⁻¹ . The recent findings of the Beyond EPICA – Oldest Ice (BE-OI) project highlight the minimal perturbations of the climate signal on the ice core, attributable to glaciological variability at the dome position.