Constraining the magmatic processes that control how magmas differentiate is essential for understanding reservoir dynamics before and during eruptions. Crystallisation and mixing are the two primary processes governing the evolution of magma reservoirs. However, the influence of crystallisation on the physical and chemical mixing of magmas remains poorly constrained, limiting our understanding of textural and chemical evolution of eruptible magma prior to eruptions. Here, we present an experimental study investigating the simultaneous occurrence of crystallisation and dynamic magma mixing using basaltic and dacitic end members at sub-liquidus conditions. Our experiment directly captures the interaction of crystallisation and magma mixing under dynamic conditions, revealing how these processes produce enclave disaggregation, filaments, and compositional gradients in the melt. Our experiment reproduces the interaction of mafic and felsic magmas, and the derived processes of mixing while the crystallisation proceeds. The results indicate that basaltic magmas crystallise rapidly, forming crystal-rich mafic enclaves within a felsic host and producing basaltic andesitic to andesitic melts. Advection promotes stretching and folding, which enhance both chemical exchanges and physical magma mixing effects, leading to enclave disaggregation and the formation of crystal clusters in disequilibrium with the surrounding melt, within a few hours in the investigated experimental setup. We used the parameter sigma 2n (normalized variance) to quantify the mixing efficiency and differential elemental mobilities. This indicates that crystallisation of the mafic magma can promote the evolution of melt compositions that enhances magma mixing efficiency with the more evolved end-member.

Introduction: The most common type of rock present ont he surface of terrestrial planets in the Solar System are volcanic-magmatic rocks, which are constituted by lava flows and fragmented pyroclasts whose texture presents both glassy and crystalline silicate phases.Planet Mars is, besides the Earth, the most diverse terrestrial planet in the Solar System, presenting magmatic rocks with compositions ranging from ultramafic/basaltic to alkaline/trachy-andesitic [1].Understanding the influence that chemical composition and different phases (both crystalline and amorphous) have on the spectral response of volcanic rocks is pivotal to interpret remotely sensed spectra that are commonly used to interpret the geology of terrestrial planets. Thus, we synthesized Martian simulants with two putative Martian compositions, on which we performed cooling experiments together with a rheological, mineralogical and spectral characterization in order to provide reference to interpret the geological processes that have occurred on Mars.Methods:Samples were created by mixing powdered oxides to mimic the composition (Table 1) of volcanic products from Gusev and Gale craters, as they were hypothesized from data belonging to different missions [1].Table 1: Composition of the starting materials accounted GusevGaleSiO247.3352.04TiO20.570.66Al2O311.1915.83FeO19.1210.89MnO0.420.18MgO10.294.44CaO8.176.76Na2O2.767.06K2O0.142.14 The powders were molten at 1450°C to form a silicate melt, which were then quickly cooled to produce glasses. Then, for each of the produced glasses three cooling experiments were conducted in a Gero HTRV 70-250/18, equipped with Anton Paar rheometer [see 2 for reference]. Powdered glass was brought to superliqidus temperature and, after having reached the equilibrium, it was cooled down at a 100°C/h rate to crystallize. Crystallization was monitored through the measurement of viscosity during cooling, so that thre phases were individuated and samples: the point at which crystals start nucleating, the point at which crystal were grown and a final point of advanced crystallization where viscosity was not measurable anymore due to extreme rigidity of the samples (see arrows in Fig. 1).Figure 1: Example of Viscosity path duting cooling for Gale simulant. Arrows indicate stages at which experiments were stopped.The synthesized samples were analyzed using X-ray powder diffraction (XRPD), and quantitative phase analysis (QPA) was performed using the Rietveld profile fitting method with internal standards to determine amorphous content.Spectral analyses: Reflectance spectra were collected at the Cold Surface Spectroscopy (CSS) facility (https://cold-spectro.sshade.eu) located at the Institut de Pla-nétologie et d'Astrophysique (IPAG), Grenoble, France. The instrument used was the SHINE Spec-tro-Gonio Radiometer. The instrument is equipped with a cryogenic simulation chamber, CarboN-IR, to control the temperature of the samples. Spectra were collected in the 1-4.2 μm spectral range at different low temperatures between 105 and 290 K. The sample is cooled to a given temperature and then held for about 5–10 min to wait for thermal equilibrium and then during the measurement time, about 65–80 min depending on some acquisition parameters. We kept each sample for 70–90 min at a given temperature before starting our measurements.Future perspectives:Spectral characterization at low temperature is currently ungoing and will shed light on the influence of temperature on the detection of silicate phases. Preliminary spectral investigations suggest us that for samples with identical bulk chemical composition but different mineralogical assemblage, the spectral response within Visible and Near InfraRed (VNIR) and mid-infrared (MIR) is deeply different. VNIR spectra are influenced by the nucleation of iron-related phases whereas MIR spectra are dependant on Si-bearing phases. The mechanism influencing the shape of spectra results in fact from a complex interaction of the spectral response of different mineral/amorphous phases. In this way, the shape is easily misinterpreted as the features of the different phases which are often no longer recognizable. A thorough analysis helped us to understand which are the features that can be accounted for the interpretation of this kind of igneous materials, but a more comprehensive study on different compositions is needed for a complete assessment of this approach.References:[1] McSween Jr, H. Y. (2015). Petrology on mars. American Mineralogist, 100(11-12), 2380-2395.[2] Vetere, F., Petrelli, M., Perugini, D., Haselbach, S., Morgavi, D., Pisello, A., ... & Holtz, F. (2021). Rheological evolution of eruptible Basaltic-Andesite Magmas under dynamic conditions: The importance of plagioclase growth rates. Journal of Volcanology and Geothermal Research, 420, 107411.Additional Information: This work was carried on thanks to ASI-UniPG agreement 2019-2-HH.0 and in the framework of Trans-National Access research project selected and funded by Europlanet-2024 RI (European Union’s Horizon 2020 RI under grant agreement No. 871149). We also acknowledge the support from MUR in the framework of SUPER-C, ‘Dipartimento di Eccellenza 2023-2027’.

Crystallisation processes in magmatic plumbing systems strongly influence magma rheology and eruptive styles. The presence of a deformation field in magma storage regions and transport systems can greatly affect crystal nucleation and growth. Here we present new experimental results on the effect of shear strain rates () on texture, crystal zoning patterns, mineral phase proportion (plagioclase, clinopyroxene, olivine, and oxides), and residual glass composition. The experiments, five in total, were carried out using natural trachybasalts under controlled temperature conditions at atmospheric pressure. The main experiment has been initially conducted at 1130 degrees C under a shear strain rate gradient (1 s-1-0 s-1, from the rotating spindle to crucible walls) using a Concentric Cylinder Apparatus. Then, the deformation has been removed and minerals continued to evolve under temperature oscillations (1170-1130 degrees C). The main rationale behind our approach is to demonstrate the impact of deformation on early crystallisation and the subsequent evolution of the system, even after deformation ceases to be effective. In natural settings, such conditions may arise during conduit dynamics, lava flow emplacement and during the development of a shallow magmatic system. Major element analyses and elemental maps were analysed using custom-built unsupervised and supervised machine learning algorithms (e.g. Hierarchical Clustering and Random Forest) to quantify how the area proportions of different chemical zoning patterns vary with inferred . The effect of on nucleation, growth, and mineral phase proportions was quantitatively investigated through shape and crystallographic preferred orientation analysis. The experimental results demonstrate how a small increase in can lead to a significant increase in nucleation rate and thus in crystal number density. While this general relationship has been observed in previous studies (Vona & Romano, 2013; Kolzenburg et al., 2017; Vetere et al., 2017; Mollo et al., 2024), our results demonstrate how these changes directly influence growth competition among different mineral phases, leading to measurable variations in their growth rates, final mineral phase proportions, and residual melt composition. The application of our results to the February-April 2021 eruptive sequence of Mt. Etna reveals that the chemical variability observed in our experiments is in the same range as that observed at Mt. Etna during the studied eruptive sequence. This underlines how different (e.g. different magma ascent rates) can provide a major contribution to the chemical variability of the erupted products.

Abstract Understanding how the glass/crystal ratio influences the spectral response of volcanic rocks is crucial for interpreting planetary remote sensing data. Here, four mafic rocks simulating a possible Martian composition were synthesized with identical bulk chemistry but different mineralogical assemblages, from fully amorphous to ∼70 wt.% crystal content, to investigate how crystal content affects Visible and Near‐Infrared (VNIR) reflectance spectra. Bi‐directional VNIR reflectance was collected at room temperature across a range of incidence (0°, 30°, 60°) and emergence (−70° to +70°) angles. The diagnostic absorptions of the two most abundant phases, pyroxene and glass, are not distinguishable in the spectra and can even reproduce the spectral fingerprint of olivine, absent from the samples; only minor iron oxides (magnetite and hematite, ∼1–7 wt.%) produce clearly identifiable absorptions. Instead, spectral slope emerges as the primary proxy for the glass/crystal ratio: NIR/VNIR slope decreases systematically with increasing crystal content, driven jointly by the incorporation of iron oxides and the loss of the positive‐slope contribution of residual glass. Principal Component Analysis and k‐means clustering independently confirm this control, identifying three clusters that map onto the glass/crystal ratio. This decoupling indicates that the crystal content of a mafic terrain cannot be inferred from pyroxene‐ or glass‐related absorptions alone, but rather from spectral slope and iron‐oxide features. These results experimentally support previous hypotheses linking the spectral diversity of Martian mafic terrains to their degree of crystal content and oxidation state and highlight variable glass abundance as an under‐considered contributor to the spectral interpretation of the Martian surface.

We introduce Orange-Volcanoes, an add-on for the open-source Orange Data Mining platform, designed to enhance data-driven workflows in petrology, geochemistry, and volcanology. Orange-Volcanoes extends the core features of Orange by incorporating tools for Compositional Data Analysis (CoDA), geochemical data preprocessing, and thermobarometric estimations.These integrated tools enable users to perform machine learning, statistical evaluations, and predictive modeling on large petro-volcanological datasets while providing intuitive, interactive visualizations. The visual programming framework of the platform fosters collaborative research and ensures accessibility for a wide audience (e.g., scientists, educators, and students) without requiring programming expertise.The combination of advanced machine learning and explainable artificial intelligence techniques, such as feature importance and Shapley additive explanations, supports deeper insights into geochemical variability and improves the interpretation of magmatic processes.We explore the potential of Orange-Volcanoes through various case studies, showcasing applications such as clustering geochemical data and conducting petrological analyses. As the volume of volcanological and geochemical data continues to grow, this tool facilitates the integration of machine learning and data mining into standard scientific practices. The ability to apply diverse statistical and machine learning tools to geochemical data, while interactively visualizing step-by-step results, makes Orange and Orange-Volcanoes valuable assets for managing large multivariate datasets and supporting petrological volcano monitoring. Orange-Volcanoes represents a significant step forward in promoting reproducible, transparent, and collaborative research methodologies.

Orange-Volcanoes is an extension of the open-source Orange data mining platform specifically tailored for geochemical, petrological, and volcanological investigations. Orange-Volcanoes enhances the original platform by incorporating specialized tools to enable interactive data-driven investigations in geochemistry, such as performing Compositional Data Analysis (CoDA). Applying CoDA transformations enables the use of many standard and multivariate statistical methods like principal component analysis, discriminant analysis, and hierarchical clustering on compositional data. In this way, Orange-Volcanoes allows for the application of a wide range of data mining and statistical methods implemented in Orange using geochemical data. Moreover, Orange allows the use of advanced methods in the field of explainable artificial intelligence, such as feature importance and Shapley additive explanations. Also, within Orange-Volcanoes, we demonstrate the flexibility of the Orange platform by developing visual tools that allow for conducting mineral-liquid equilibrium tests and calculating thermo-barometric estimates. The Orange-Volcanoes supports collaborative efforts and reproducibility by offering a visual programming interface that requires no coding experience, making it accessible to a wide range of users, including scientists, educators, and students. We provide a series of case studies, including interactive petrological data exploration and clustering in tephra studies to highlight Orange-Volcanoes’ potential and versatility in volcanological applications. Orange-Volcanoes can be downloaded using pip, and its documentation is available at https://orange3-volcanoes.readthedocs.io/en/latest/.

Silicates are the main constituent of volcanic terrains on terrestrial planets in the Solar system. On Earth, we know that volcanic terrains are constituted by lava flows and fragmented pyroclasts whose texture presents both glassy and crystalline phases. Understanding the influence of glass/crystal ratio on the spectral response of volcanic rocks is therefore focal to interpret remotely sensed spectra that are commonly used to interpret the geology of terrestrial planets. Thus, we performed spectral characterization of lab-made mafic volcanic products which were synthesized in the PVRG labs with the aim to present different degrees of crystallinity.Samples were synthesized by mixing and melting oxides to resemble the composition of a Nakhlite meteorite. First we produced a homogeneous silicate glass (Nglass) which was then used to prepare three more samples by melting at 1500°C and then cooling down slowly (52-56 °C per hour) towards subliquidus temperatures of ca. 1200°C (N12), 1100°C (N11), and 1000°C (N10), respectively. Each sample stayed for 48 hours at the target temperature and was finally quenched in air. SEM and XRPD analyses and Rietveld method quantitative phase analyses were performed to assess type and degree of crystallinity, showing how N11 and N10 present a similar mineralogical assemblage with ca. 30% of glass and crystal species including augite, magnetite, cristobalite and other minor phases similar to the mineralogical composition of a natural Nakhlite. Sample N12 presents less diverse mineralogy (augite and magnetite) and ca. 70% of glass. Bi-directional reflectance spectra was collected at room temperature in the 1-4.2 µm range considering a set of 3 incidence angles (i = 0°; 30°; 60°) and emergence (e) angles between -70° and 70° using the custom-made bidirectional reflectance spectro-goniometers SHINE at the Cold Surface Spectroscopy facility (CSS) of the IPAG laboratory in the frame of the Trans-National Access program, project number 21-EPN-FT1-025, of Europlanet 2024. Spectral analyses show how, in the Visible and Near-Infrared, increasing crystallinity causes slopes of spectra to gradually shifts from positive for glasses towards flat-negative for crystalline material. The spectral features of the single mineral phases are barely distinguishable for N10 and N11, where spectra are flattened probably because of the presence of magnetite, whereas the spectral signature of Fe in augite is distinguishable for N12 located at ~0.9 and ~1.15 µm.Changing observation geometry, reflectance values and spectral slope show important variations while the bands position remains unchanged. We observe important dependence of band and slope in correspondence of low phase (< 30°) angle and high phase angle (> 100°). To identify distinctive features we used principal component analysis (PCA) obtaining four clusters in the PC space which are relatable to the four samples. K-means clustering was used to verify our clustering obtaining a very low level of misclassification, especially regarding Nglass. These results provide further information on the spectral response of synthesized rock samples, especially for what concerns glass-bearing materials, that can be used for modeling of spectral information coming from volcanic rocky bodies in the Solar system.

This work investigates using Brillouin light scattering (BLS) spectroscopy the relationship between chemistry and longitudinal (or pressure) sound-wave propagation on a large dataset of alkaline to sub-alkaline silicate glasses. Results show that the frequency fB of the Brillouin peak decreases with the silica content and the Silica vs. Calcia- Ferrous oxide- Magnesia (SCFM) parameter, while it increases with the degree of polymerization expressed by the ratio of nonbridging oxygens to tetrahedral cations (NBO/T). It is possible to infer that the values of both fB and the real part of the longitudinal elastic modulus M' are tightly related to the content of divalent cations (M2+) participating in the silicate network. Our findings suggest that alkaline earth metals and Fe2+ linearly speed up the longitudinal acoustic waves in silicate glasses. This might open a new window on the possibility of using the BLS technique for rapid and accurate determinations of physical and chemical properties of natural glasses present on Earth and other planetary bodies.

Magmatism and volcanism are key processes that shaped the surface of rocky planets in the Solar System, and understanding how magmatism occurred on these planets is essential to reconstruct their geological histories. It is very likely that volcanic terrains on planetary surfaces are covered by products that have a porphyric, aphanitic or hyalocrystalline texture, in which crystalline phases represent only a fraction of the rocks that are mainly composed by amorphous materials such as glasses.Remote investigation of planetary surfaces combines geomorphological analysis with spectral data. Regarding the latter, the interpretation of spectral information from planetary terrains needs to be based on the comparison with reference material.Up to now, libraries and repositories are mostly report information about so-called planetary analogues consisting of crystalline materials rather than amorphous products. This happens because of the lack of spectral features of amorphous products. However, a the presence of amorphous material together with crystalline phases deeply influence the spectral response of the latter [1].For this reason, the Petro-Volcanology Research Group at the University of Perugia started to build up a spectral database of volcanic rocks, both natural and synthetic, to understand how and how much are crystalline and amorphous products influencing each other in determining the spectral response of a rock/terrain. In this work we report the spectral characterization of various products within two ranges: Visible and Near Infrared (VNIR) and Mid-Infrared (MIR).The accounted products consist of:- Synthetic silicate glasses with wide range of chemical composition [2,3]- Synthetic planetary analogues consisting of rocky materials containing both amorphous and crystalline products [4]- Natural volcanic rocksThe overall analysis of the spectral response shows that, within the MIR, it is relatively easy to link the wavelength position of well-known spectral features (Christiansen feature, Reststrahlen Bands, Transparency features) to the silica content of the product itself, no matter which crystalline phases are present within the rock. This finding is of pivotal importance since the determination of silica content is fundamental for a characterization of volcanic products.On the other hand, investigations and analyses on VNIR spectra show that the interpretation within such is more complicated. Indeed, a small presence of Iron-bearing crystals can deeply influence the VNIR spectrum. However, some constrains can be build using empirical parameters and machine-learning approaches.All the spectra produced and published by our research group is or will be made available to everyone, open-source, within the platform hosted by ASI-SSDC (www.ssdc.asi.it/rockspectra/), making it possible for scientist from all-over the world to integrate such data in other researches or to directly compare planetary spectra to laboratory data [4].[1] Horgan, B. H., Cloutis, E. A., Mann, P., & Bell III, J. F. (2014). Near-infrared spectra of ferrous mineral mixtures and methods for their identification in planetary surface spectra. Icarus, 234, 132-154.[2] Pisello, A., Ferrari, M., De Angelis, S., Vetere, F. P., Porreca, M., Stefani, S., & Perugini, D. (2022). Reflectance of silicate glasses in the mid-infrared region (MIR): Implications for planetary research. Icarus, 388, 115222.[3] Pisello, A., De Angelis, S., Ferrari, M., Porreca, M., Vetere, F. P., Behrens, H., ... & Perugini, D. (2022). Visible and near-InfraRed (VNIR) reflectance of silicate glasses: Characterization of a featureless spectrum and implications for planetary geology. Icarus, 374, 114801.[4] Pisello, A., Zinzi, A., Bisolfati, M., Porreca, M., & Perugini, D. (2022). A new spectral database for silicate glasses: a fundamental resource to interpret characteristics of volcanic terrains on planetary bodies (No. EPSC2022-539). Copernicus Meetings.

This study highlights the rheological variation of magmatic systems during the early crystallization stage, which undergoes very different shear stress. Etna basaltic glass, made from natural rock powder, was used as a starting material. Nine shear rate-controlled experiments were conducted at 1150 degrees C (below liquidus temperature and undercooling degree Delta T similar to 40 degrees C) with shear rates (gamma(center dot)) of 0.1, 1 and 10 s(-1) using wide-gap concentric cylinder viscometry. Three additional experiments were conducted without spinning the melts (no shear, gamma(center dot)= 0 s(-1)). Run-products were collected after 3, 6 and 9 h. The experiment with the highest shear rate (gamma(center dot)=10s(-1)) showed a brittle failure when viscosity reached the value of 2.90 log (Pas) and after ca. 1150s. The measured viscosity matches the shear stress at 7244 Pa, corresponding to the brittle failure in our partly crystallised system at these conditions. After 9 h, the response of the partly crystallised Etna basalt to different deformation rates results in decreasing viscosity from 4.89 to 3.83 and 2.90 (log Pa s) as the gamma(center dot) increases from 0.1 to 1 and 10 s(-1), respectively. The main outcome of this study relates to the nucleation and growth of minerals with shear deformation. The deformation-free (gamma(center dot) = 0 s(-1)) runs show the presence of only two phases: glass and Fe-oxides (Fe-ox) with only a few vol% (1-3) of oxides crystals after 3, 6 and 9 h. The deformation-bearing (gamma(center dot) = 0.1, 1 and 10 s(-1)) runs show different scenarios: after 3 h, we observed only Fe-ox for a gamma(center dot) of 0.1 s(-1) (similar to deformation-free ones). As the shear rate increases to 1 s(-1) and 10 s(-1), solid phases after 3 h experiments are Fe-ox, plagioclase and clinopyroxene. Crystal growth rate depends on the applied shear rate: the highest rate is 1.1 x 10(-6) cm/s and was measured for plagioclase after a 3 h experiment for gamma(center dot) = 10 s(-1). At gamma(center dot) = 0.1 s(-1), the plagioclase growth rate decreases to 2.70 and 1.35 x 10(-6) cm/s as experimental time increases from 6 to 9 h, respectively. The vast range of shear stress and the systematic data obtained in this study are fundamental to deciphering crystallization dynamics suffered by magmas in volcanic reservoirs, dikes, conduits and lavas.

Rationale Silicate glasses are significant components of volcanic products, but spectral libraries typically provide references of crystalline materials rather than amorphous ones, even if blurred/featureless spectra are observed on planetary terrains. Thus, the Petro-Volcanology Research Group (PVRG) started an investigation of the spectral response of silicate glasses developing a database, to distribute spectral data applying the FAIR (Findable, Accessible, Interoperable, Reusable) principles ​[1]​. We have analyzed these spectra using PCA and, in particular, we focused on MIR range, in which information about silicates’ arrangement in the material is detectable ​[2].Fig. 1: a) TAS diagram of all products. Circles:sub-alkaline , triangles: alkaline. b) spectra of the entire dataset MethodologyThe investigated datasets consist of MIR spectra of 20 powdered silicate glasses (

The magma accumulation phase preceding caldera-forming eruptions is crucial to identify the signs of an impending catastrophic event. While extensively studied in silicic systems, mafic volcanoes present unique challenges. The high eruptive temperatures of these magmas imply short storage in the cold upper crust and, thus, short periods of unrest preceding eruption, which could challenge our capacity to mitigate the impact of an imminent event. We analyse crystals in erupted magma to reconstruct the lead-up phase to the last caldera-forming eruption of Colli Albani, an active volcano near Rome (Italy). Here, we show that the caldera-forming eruption was preceded by effusive to mildly explosive eruptions fed by multiple crustal magma reservoirs. Following a pause in volcanic activity, magma rose directly from the mantle and accumulated to form a reservoir of a few tens of cubic kilometres. The ascent of one of these pulses ultimately triggered the last caldera-forming eruption of Colli Albani. Our results provide a new framework to identify the processes leading to caldera-forming eruptions in mafic volcanic systems.

Introduction Silicates are the main constituent of volcanic terrains on terrestrial planets in the Solar system. On Earth, we know that volcanic terrains are constituted by lava flows and fragmented pyroclasts whose texture presents both glassy and crystalline phases. Understanding the influence of glass/crystal ratio on the spectral response of volcanic rocks is therefore focal to interpret remotely sensed spectra that are commonly used to interpret the geology of terrestrial planets. Thus, we performed spectral characterization of lab-made mafic volcanic products which were synthesized in the petro-vulcanology research group (PVRG) labs to present different degrees of crystallinity.Samples preparation and characterizationSamples were created by mixing powdered oxides to mimic the composition of a Nakhlite meteorite. The powder was melted at 1450°C to form a silicate melt, which was then quickly cooled to produce a glass and divided into four sub-samples. As observable in Fig. 1, one sub-sample was rapidly cooled in air to form Nglass, while the other three were slowly cooled (52-60°C per hour) to target temperatures of approximately 1200°C (N12), 1100°C (N11), and 1000°C (N10).Figure 1: Cooling ramps for the syntheses of products with different crystallization degreeThe synthesized samples were analyzed using X-ray powder diffraction (XRPD), and quantitative phase analysis (QPA) was performed using the Rietveld profile fitting method with internal standards to determine amorphous content. Crystallized samples (N10, N11, and N12) showed peaks corresponding to various crystalline phases including Augite, Cristobalite, Magnetite, and Hematite. Quantitative analyses are reported in as reported in Figure 2.Figure 2: Mineralogical assemblage of the produced samplesReflectance analysisSpectroscopic characterization was performed with SHINE (SpectropHotometer with variable INcidence and Emergence) at the Cold Surface Spectroscopy facility (CSS) of IPAG laboratory. This setup allows the collection of spectra in the visible (VIS) and near-infrared (NIR) ranges (0.35–4.5 µm) across a broad range of illumination (0–30°) and emergence angles (e=0-70°) at room temperature. Both illumination and emergence angles are zenithal angles, measured from the normal direction to the surface. Approximately 2 g of powdered samples (grain size < 80 µm). All acquired spectra are shown in Figure 3 where they are divided in four subgraphs representing the four samples N10, N11, N12 and Nglass.Figure 3: VNIR spectra of four samples at different acquisition geometries.Principal component analyses (PCA)Our set of spectra was analyzed using PCA and K-Means clustering to visualize differences and similarities between the spectra from an unsupervised machine learning perspective. Principal component analysis was then performed without normalization, retaining the first two PCs, which encompass ~99.82% of the total variance (Figure 3a). The elbow method was used to determine the optimal number of clusters, resulting in 3 clusters (Figure 4). This led to a well-defined classification with only 3 misclassified samples. The mean spectra from each cluster were visibly similar to the spectra of each class, with low standard deviation for Nglass and N12, and higher for the cluster comprising N10 and N11. The resulting PCs are explainable from the loadings, showing that the first PC mainly represents the shape of a Nglass spectrum, while the second PC considers the height of the shoulder before ~800 nm and the dip after that point..Figure 4: Clustering of spectra after PCA analysesFuture perspectivesWe have observed how, for samples with identical chemical composition and different mineralogical assemblage, the spectral response within Visible and Near InfraRed (VNIR) can be substantially different. The mechanism influencing the shape of spectra results in fact from a complex interaction of the spectral response of different mineral/amorphous phases. In this way, the shape is easily misinterpreted as the features of the different phases which are often no longer recognizable. A PCA analyses helped us to understand which are the features that can be accounted for the interpretation of this kind of igneous materials, but a more comprehensive study on different compositions is needed for a complete assessment of this approach.AcknowledgementsThis work was carried on thanks to ASI-UniPG agreement 2019-2-HH.0 and in the framework of Trans-National Access research project selected and funded by Europlanet-2024 RI (European Union’s Horizon 2020 RI under grant agreement No. 871149)

In planetary science, visible (Vis) and near-infrared (NIR) reflectance spectra allow deciphering the chemical/mineralogical composition of celestial bodies’ surfaces by comparison between remotely acquired data and laboratory references. This paper presents the design of an automated test rig named Exoland Simulator equipped with two reflectance spectrometers covering the 0.38–2.2 µm range. It is designed to collect data of natural/synthetic rocks and minerals prepared in the laboratory that simulate the composition of planetary surfaces. The structure of the test rig is conceived as a Cartesian robot to automatize the acquisition. The test rig is also tested by simulating some project trajectories, and results are presented in terms of its ability to reproduce the programmed trajectories. Furthermore, preliminary spectral data are shown to demonstrate how the soil analogs’ spectra could allow an accurate remote identification of materials, enabling the creation of libraries to study the effect of multiple chemical–physical component variations on individual spectral bands. Despite the primary scope of Exoland, it can be advantageously used also for tribological purposes, to correlate the wear behavior of soils and materials with their composition by also analyzing the wear scars.

IntroductionRemotely sensed hyperspectral data provide essential information on the composition of rockson planetary surfaces. On Mars, these data types are provided by the CRISM instrument(Compact Reconnaissance Imaging Spectrometer for Mars) [1], a hyperspectral camera thatoperated onboard the MRO (Mars Reconnaissance Orbiter) probe that collected more than 10Tb of data over more than 14 years of operation. CRISM covers a spectral range going from362 to 3920 nm, with a spectral resolution of 6.55 nm/channel and a spatial resolution of 18.4m/px, from 300 km altitude.The most advanced CRISM data products are the MTRDRs (Map-Projected Target ReducedData Records) [2]. These data are re-projected onto the Martian surface and are cleaned fromthe so-called ”bad bands”, noisy stripes. The two main CRISM MTRDR subproducts are thehyperspectral datacube and the spectral parameter datacube. The first contains the detectedreflectance spectra, while the second is composed of 60 different spectral parameters, as definedin [3], usually to produce RGB maps emphasizing specific minerals within the scene.The analysis of such products is often made by selecting Regions of Interest (ROI) from whichto extract the spectra. Recently, other more global and general methods are beginning to beused, such those that involves the application of machine learning algorithms to analyze and/ormap the planetary surface spectra (i.e., [4]), in order to obtain more general and exhaustiveresults.Regarding the model architecture, undercomplete autoencoders represent a fundamental conceptin the domain of unsupervised learning using neural networks. At its core, an autoencoderis a type of artificial neural network trained to reconstruct its input data at the output layer.It consists of two main components: an encoder and a decoder. The encoder compresses theinput data into a lower-dimensional representation, while the decoder attempts to reconstructthe original input from this compressed representation. An undercomplete autoencoder is characterizedby having a bottleneck in the network architecture, where the dimensionality of eachlayer, from the first one to the encoded representation layer is lower than the previous layerdimension. This architecture force the model to capture the most relevant information stored inthe input data, thus facilitating effective feature extraction. The train process of a undercompleteautoencoder is no different than the training of a classical Fast Forward Neural Network,so involving the minimization of a reconstruction loss function evaluated between the inputdata and the reconstructed output.For this work, we analyzed as a case study the CRISM named FRT00003E12in Nili Fossae. The choice of this specific scene was motivated by the morphological and spectralrichness that it offers and by the mole of literature [5,6,7] available for comparison.MethodsThis work explores the potentials of Undercomplete Autoencoders to perform:The dimensionality reduction of a CRISM MTRDR product; the clustering of similar spectra regions and subsequent region spectra analysis; the generation of specific nonlinear combination of spectral parameters to enhance specificmineralogies In detail, we trained the networks using the spectral parameters datacube after eliminatingthe pure reflectance parameters and the immediately reflectance-relatable parameters, thosebeing R770, RBR, RPEAK1, R440, IRR1, R1330, IRR2 , IRR3,R530, R600, R1080, R1506, R2529 and R3920 [3]. We obtained a total number of 46 spectralparameter for each pixel image.Then, we checked the pixel values for each spectral parameter, setting to zero each value belowzero. This was done because less than zero valued pixels do not have a physical meaning insidethe spectral parameter image. Finally, each feature were normalized with the following: xnorm = (x-xmean)/σ.The subsequent step consisted of model selection by hyperparameter optimization. To achievethis goal, we trained 60 random different models with various, random sampled hyperparameterssets on a subset of the original datacube (20%) for 50 epochs.The best model was then chosen by selecting the trial that, in 50 epochs, gave us the lowestloss function value, where the chosen loss function was the Huber Loss, chosen as encompassboth the properties of the L1 and MSE losses ensuring so a better overall performance.The optimizer we used for both the random search and the training process is Adam, andthe batchsize for both the final training and the hyperparameter optimization was set to theentire image/subset. Then we trained the optimal model for 100 epochs and we extracted thelatent space for each pixel. After this we scaled the resulting pixel values setting to zero all thepixels below the median value (constraining so between the 50th and 100th percentiles) and wedivided each pixels in different classes. Each class derives from the combination of latent spaceneurons in that pixel. For example, the classes 00000 and 11111 represents all the pixels thatdoes not activate any latent space neuron and all that activates all the latent space neurons,respectively (see figure 3 top left and bottom right). Finally, from each class, we extracted themean spectra and its standard deviation.Table 1 and figure 1 reports the final model resulting from the random search hyperparametersoptimization process. Also, Figures 2 and 3 shows its application to the reflectance-deficient CRISM spectral parameter datacube. Figures 2 and 3 highlight that each pixel in the CRISMdatacube activates different neurons (Fig 2) and neuron combinations (Fig 3) in the latent space.Figure 4 reports the superimposition of the different unique neuron combinations reported infigure 3. In figure 4 b, each color represent a class of activated neurons and permits a moredirect comparison between the model’s extracted information and the morphological/albedofeatures of the CRISM scenario (a). Figure 5 reports the mean spectra extracted for eachneuron combination that could be used for further investigations or for the comparison withlaboratory investigations.Overall, we highlighted that undercomplete autoencoders could be a possible reliable instrumentfor the analyses of hyperspectral datacube, from simple spectra extraction by clustering to thegeneration of synthetic spectral parameters. table 1L. Rate Enc. N. Dec. N. W. Decay Enc. Dim. Act.func. 0.05 25,15 45 0.0001 4 SiLU figure 1 figure 2 figure 3figure 4figure 5[1] Murchie2007,E05S03[2]Seelos2016,LPSCXXXXVII[3]Viviano-Beck2014,119(6):1403-1431[4]Baschetti2024,XIXCongressoNazionalediScienzePlanetarie[5]Mustard2009,doi:10.1029/2009JE003349.[6]Elhmann2008,Science,322,1228-1832[7]Mustard2008,https://doi.org/10.1038/nature07097https://doi.org/10.1038/nature07097

The 2018 LERZ eruption of Kilauea featured a wide range of eruptive styles. In particular, Fissure 17 (F17) displayed activity ranging from Hawaiian fountaining in the eastern part of the fissure to Strombolian explosions in the western part. Lava erupted from F17-West was highly viscous and contained magmatic enclaves. Magmatic enclaves have previously been observed in many other volcanic systems (e.g. Vulcano Island, IT and Sete Cidades Volcano, PT), where they have been attributed to injection of mafic magma into an evolved magma chamber, resulting in viscous fingering, quenching, and break-off into fragments. The F17 enclaves differ from previous studies in that the chemical compositions of the enclave and host magmas are very similar, and that the enclaves have a limited spatial distribution and lack signs of viscous behavior and quenching, pointing to a different formation mechanism than inferred for other volcanic systems.In order to test a different formation hypothesis, we conducted fractal analysis of the size distribution of 84 individual enclaves from F17-West lavas. Our results, including a fractal dimension of fragmentation Df of 2.585, indicate that the F17 enclaves likely formed by brittle fragmentation. Since the enclave and host magmas were at temperatures far above the glass transition during the magma hybridization, high strain rates have to be invoked to explain the brittle fragmentation. This may have caused the enclave magma to transition into solid-state behavior, allowing it to break off into fragments that were subsequently picked up by the host magma and carried to the free surface.The enclaves from F17-West therefore offer a unique insight into the diversity of processes that characterizes the shallow parts of volcanic systems, as well as the importance of strain rates in modulating the rheological behavior of magmas.

AbstractSilicates are the main constituent of volcanic terrains on terrestrial planets and other rocky bodies in the solar system [1]. Typically, these volcanic terrains are constituted by fragmented pyroclasts whose texture is often afanitic or porphyric rather than holocrystalline: this means that the fraction of crystalline material is less relevant than the fraction of amorphous, or glassy, material. Thus, it is of paramount importance to take into account amorphous silicate phases to explore the influence of glass/crystal ratio on the spectral response of volcanic rocks, to better interpret available and future remotely sensed spectra from past and future missions [2, 3]. Here we report the results of a study concerning mafic volcanic products which were synthesized in order to present different degrees of crystallinity: three basaltic melts were cooled at different rates to obtain different textures, from totally amorphous to crystalline. Finally, they were analysed by means of emissivity in the thermal-IR range at different temperatures. Samples preparationSamples were produced by melting two natural mafic rocks: a basalt (low-alkali mafic rock from Snake River Plain, USA) and a shoshonite (high-alkali mafic from Vulcano island, Italy) following a two-steps routine of crushing and melting [4, 5]. A third silicate melt was produced by mixing and melting oxides to resemble the composition of Nakhlite meteorite [6]. The three melts were cooled in three different ways (Fig. 1) to obtain nine different samples. In order to obtain pure, crystal-free glasses, melts were directly quenched from super-liquidus temperature (Black line Fig.1), whereas other, identical, melts were cooled down slowly (52-56 °C per hour) and then quenched at subliquidus temperatures of ca. 1100°C (red line Fig. 1) and 1000°C (yellow line Fig.1). Following this approach, we obtained 9 rocky samples from three melts, so that each sample was differing in both chemical composition and crystallinity. Figure 1: The three cooling ramps used for the syntheses of the material. Samples analysisSamples were analysed using SEM, and spectroscopically characterized under different conditions: thermal-IR data have been acquired at the Planetary Spectroscopy Laboratory of the German Aerospace Center in Berlin (DLR), collecting the emitted thermal radiation for samples at different temperatures (150°C, 300°C, 450°C, 600°C; spectral range 5-16 μm) [7]. SEM imaging showed successful different degree of crystallinity for the three steps, which results in a different spectral response, visible in Figure 1.By observing shape of spectra, crystalline Shoshonite and Basalt show similar shape to their relative amorphous but for a shoulder at 8.2-8.3 μm, whereas Nakhlite shows substantially different shapes for the three crystallinity steps (Fig. 2).For what concerns the shift of the spectra, crystal-bearing products seem to show similar features at slightly lower wavelengths for Shoshonite and Basalt, whereas this trend is inverted for Nakhlite, probably due to different phases nucleating in different melts. Higher emissivity temperatures seem to homogenize the spectral response of samples with same chemical composition and different textural properties. These results provides further information on the spectral response of synthetised rock samples [2], that can be used for modeling of spectral information coming from rocky bodies in the Solar system Figure 2: Spectra resulting from emissivity measurements performed on nine different products produced from three initial compositions (Nakhlite:N, Shoshonite:sho and Basalt:B). Emissivity was measured with samples at four different temperatures, two of which are here shown (150 and 600°C). References[1] Namur, O. and Charlier, B. (2017). Silicate mineralogy at the surface of mercury.Nature Geoscience, 10(1):9.[2] Pisello, A., Vetere, F. P., Bisolfati, M., Maturilli, A., Morgavi, D., Pauselli, C., ... & Perugini, D. (2019). Retrieving magma composition from TIR spectra: implications for terrestrial planets investigations. Scientific reports, 9(1), 1-13.[3] Di Genova, D., Hess, K.-U., Chevrel, M. O., and Dingwell, D. B. (2016). Models for the estimation of fe3+/fetotratio in terrestrial and extraterrestrial alkali-and iron-rich silicate glasses using raman spectroscopy.[4] Vetere, F., Iezzi, G., Behrens, H., Holtz, F., Ventura, G., Misiti, V., ... & Dietrich, M. (2015). Glass forming ability and crystallisation behaviour of sub-alkaline silicate melts. Earth-science reviews, 150, 25-44.[5] Rossi, S., Petrelli, M., Morgavi, D., Vetere, F. P., Almeev, R. R., Astbury, R. L., & Perugini, D. (2019). Role of magma mixing in the pre-eruptive dynamics of the Aeolian Islands volcanoes (Southern Tyrrhenian Sea, Italy). Lithos, 324, 165-179.[6] Treiman, A.H. (2005) The nakhlite meteorites: Augite-rich igneous rocks from Mars. Chemie der Erde 65, 203-270 [7] Maturilli, A., Helbert, J., Ferrari, S., Davidsson, B., and D’Amore, M. (2016). Characterization of asteroidanalogues by means of emission and reflectance spectroscopy in the 1-to 100-μm spectral range.Earth,Planets and Space, 68(1):113

Abstract Understanding magma differentiation and formation of eruptible magmas is one of the key issues in Earth sciences. Many studies have either focused on mixing or crystallization, but none have studied these two processes simultaneously. Here, we conduct an innovative experimental study investigating the simultaneous occurrence of crystallization and dynamic mixing, using basaltic and dacitic end members at sub-liquidus conditions. We reproduce the injection of mafic magma into felsic magma and their mixing while crystallization occurs. Our results indicate that crystallization of basaltic magmas occurs faster than mixing between basalt and dacite leading to the formation of crystal-rich mafic enclaves within a felsic magma and the development of basaltic andesitic to andesitic melts. Then, convection promotes stretching and folding that favor chemical and physical magma mixing, disaggregation of enclaves and formation of clusters of crystals in disequilibrium with the surrounding melt. Magma mixing is the predominant process after the initial crystallization event. Our results provide insights into pre-eruptive dynamics, which is crucial for improving volcanic hazard assessment.

In this study, we proposed a general workflow that aims to enhance the ML-based geothermobarometer modelling. Our workflow focuses on three key areas. Firstly, we developed a robust pre-processing pipeline that addresses data imbalance, feature engineering, and data augmentation. Secondly, we assessed modelling errors using a Monte Carlo approach to quantify the impact of analytical uncertainties on the final pressure and temperature estimates. Thirdly, we implemented a robust strategy to validate and test the ML models to avoid over- and under-fitting issues while correcting biases associated with the application of specific ML models (i.e., tree-based ensembles). To facilitate the use of our workflow, we have developed a web app (https://bit.ly/ml-pt-web) and a Python module (https://bit.ly/ml-pt-py). The robustness of this strategy has been tested on two calibrations: clinopyroxene (cpx) and clinopyroxene-liquid (cpx-liq). Our results show a significant reduction in errors compared to the baseline model, as well as good generalization ability on an independent external dataset. The Root Mean Squared Errors are 57 degrees C and 2.5 kbar for the cpx calibration, and 36 degrees C and 2.1 kbar for the cpx-liq calibration. Finally, our models show improved outcomes on the external dataset compared to existing ML and classical cpx and cpx-liq thermobarometers.