The Río Tinto is a river located in the Iberian Pyritic Belt (IPB) region, a mineralized geological body covering an area of approximately 250 km to a maximum of 60 km in southwestern Spain (Martin-Izard et a., 2016), formed by hydrothermal activity (Leistel et al., 1997) (Figure 1). The IPB corresponds to an area of ​​Devonian-Carboniferous volcanic and sedimentary rocks containing about 90 deposits of massive polymetallic sulfides that have been mined since the Chalcolithic, i.e., Copper Age, 4500-3000 BC (Leblanc et al., 2000). The Río Tinto originates in Peña de Hierro and flows through the mineral deposits of the IPB (mainly iron and copper sulfide) and reaches the Atlantic Ocean in Huelva, creating the largest acid province in the world. The waters of the Tinto River have a high concentration of ferric iron together with other heavy metals (Cu, Zn, As, Mn, Cr, etc.) kept in solution at an acid pH (average 2.3, Amils et al., 2007) and a high level of microbial diversity, mainly eukaryotic (Aguilera et al., 2006a). Although Río Tinto’s acidic waters were once thought to result from environmental damage from mining, research in the area has shown that such an environment predates human activity (Fernández-Remolar, 2003; Amils et al., 2007).

The Tinto River maintains acidic conditions and high iron and other metal ion concentrations along its entire length, despite seasonal climate changes and dilution by tributaries and rainfall. The extreme acidity and high concentration of heavy metals present in the Río Tinto ecosystem are a direct consequence of the biological processes of ferrification and sulfur oxidation of the microbes growing in the rich complex of metal sulfide deposits of the IPB (González-Toril et al., 2003). The microbial and physical characteristics of the river have been suggested as evidence for a subterranean ecosystem sustained by chemolithotrophic anaerobic metabolism of iron and sulfur minerals (Fernández-Remolar et al., 2005; Amils et al., 2007). The important role played by the iron-oxidizing microorganisms is to maintain a high concentration of ferric iron through a mechanism of aerobic respiration: Fe²⁺ + ½O₂ + 2H⁺ → Fe³⁺ + H₂O, which is responsible for maintaining the constant pH of the water thanks to its buffer potential Fe³⁺ + 3H₂O → Fe(OH)₃ + 3H⁺. Microbiological studies identified the prokaryotic phylogenetic belonging of the acidophilic microorganisms of the river. The 80% of the prokaryotic diversity present in the water column corresponds to three bacterial genera: Leptospirillum spp., Acidithiobacillus, and Acidiphilium spp., which are important members of the iron cycle (González-Toril et al., 2003). In addition to the extreme physical and chemical characteristics of the water, Río Tinto exhibits a unique and unexpected degree of eukaryotic diversity in its waters (Amaral Zettler et al., 2002; Aguilera et al., 2006a), and the fact that eukaryotic organisms account for more than 65% of the biomass in the habitat (i.e., the ecological paradox, Aguilera et al., 2013). Members of the phylum Chlorophyta, such as Chlamydomonas, Chlorella, and Euglena, are the most frequent species together with two filamentous algae belonging to the genera Klebsormidium and Zygnemopsis (Aguilera et al., 2006b; Aguilera et al., 2007).

The more acidic part of the river is inhabited by a eukaryotic community dominated by two species belonging to the genera Dunaliella and Cyanidium, known for their high tolerance to the presence of metals and acids (Visviki and Santikul, 2000). Also, pinnate diatoms (genus Pinnularia) are present in the river, which form large brown biofilms. Ifhomycetes are also very abundant in this environment, managing to grow in the extreme conditions of the river. Fungal species such as Hortaea werneckii and Acidomyces acidophilum have been identified in the Río Tinto with molecular techniques (Amaral Zettler et al., 2002). The mixotrophic community is dominated by microorganisms belonging to the genera Bodo, Ochromonas, Labyrinthula, and Cercomonas. The community of consumer protists is characterized by two different species of ciliates (Oxytrichia and Euplotes). Amoebae belonging to the genera Valhkampfia and Naegleria are also frequently found in the more acidic parts of the river and a heliozoan species belonging to the genus Actinophyris appears to be the main characteristic predator of the benthic food web of the river. However, not only single-celled eukaryotic systems grow in the extreme conditions of the Tinto River. Several plants grow in acidic riverbank soils demonstrating resistance to heavy metal concentrations in the soils in which they grow, even using the strategy of concentrating metals in their tissues (Rodríguez et al., 2005). This suggests that the management of heavy metals in general and of iron, in particular, is very complex (Schmidt, 2003). Furthermore, these results demonstrate that complex multicellular systems can develop even in extreme conditions, such as those existing in Río Tinto.

The selected landing site of the Rosalind Franklin rover mission, Oxia Planum, situated at the Martian dichotomy boundary, is part of a wide basin characterized by the extensive presence of Noachian Fe/Mg-rich phyllosilicates (Mandon et al., 2021; Quantin-Nataf et al., 2021). Hydrated silica and Al-rich phyllosilicates may be present to a lesser extent in the eastern sector of the site (Carter et al., 2016) where Hesperian alluvial and deltaic sediments overlap the Fe-/Mg-rich phyllosilicates. Al-rich clays were also found in the Oxia’s catchment area (16.7°N, 337°E), indicating the presence of various clay mineralogies (Turner et al., 2021; Brossier et al., 2022). The deposition of this unit is thought to have occurred either 1) in a subaerial environment with sediments deposited in lacustrine or marine environments); or 2) in a subaerial/surface environment through the deposition of volcanic products or aeolian deposits (Carter et al., 2016; Quantin-Nataf et al., 2021). In the latter case, groundwater alteration or pedogenesis is the most likely scenario for the formation of the Fe/Mg-rich clay minerals observed at the site (Carter et al., 2016). This mineralogical evidence, combined with other morphological features, suggests that Oxia Planum has been affected by long-lasting surface water activity, a factor consistent with conditions favorable for the development of life that make this a Martian site of certain astrobiological interest.

On Mars, obviously, not only the presence of clays testifies the interaction between the pre-existing rocks and the water. In Meridiani Planum or Mawrth Vallis for example, the mineralogy is quite different from that of the Rosalind Franklin rover landing site. The past presence of an aqueous environment is evidenced by the presence of iron-bearing sulfates (Squyres et al., 2004; Farrand et al., 2009). These minerals could result from the breakdown of sulfides (Zolotov and Shock, 2005; Bigham et al., 1996) and testify the presence of a low pH aqueous environment. Similarly, also the detection of hematite at a regional scale (Christensen et al., 2001) could indicate that pre-existing rocks have undergone chemical erosion in aqueous conditions (Rieder et al., 2004; Squyres et al., 2004). Iron oxides associated with sulfates are the characteristic minerals that form in modern sediments and young terraces of the Río Tinto: they are hydrojarosite, schwertmannite, copiapite, coquimbite, natrojarosite, gypsum, and other sulfate minerals. Goethite and hematite are the predominant minerals in the ancient terraces of the Tinto basin (Fernández-Remolar et al., 2005).

From these studies, it emerges that some mineralogies detected in the Río Tinto are consistent with those detected in various parts of the Mars surface but could seem far from the mineralogies detected in Oxia Planum. In this case, the non-detection of sulfates and iron-bearing oxides could be linked to the low spatial resolution of the instruments that operated in remote sensing. In situ investigations that will be carried out at the Rover Rosalind Franklin scale and in particular the acquisitions made by Ma_MISS (with a spatial resolution of 120 µm) could reveal the presence of sulfate veins such as those revealed in Gale Crater by the ChemCam instrument (Nachon et al., 2014). A further indication supporting the identification of minerals that form in acid conditions could be supported by the lack of detection of carbonates in Oxia Planum. Analogies between Oxia Planum and Río Tinto can be represented by the aqueous alteration processes that formed clays. On the other hand, differences in the resulting clay assemblage could be due to the different environmental conditions (e.g., pH, temperature, etc.) as shown by laboratory experiments (Krzesinska et al., 2020).

Since the presence of iron-bearing minerals detected at the Río Tinto is not only due to the acid drainage of exposed minerals but, on the contrary, can be attributed to the role of bacteria that play a key role in making this place an extreme environment, the Río Tinto represents a peculiar site from an astrobiological point of view where to collect and measure samples useful for the scientific data interpretation.

The Río Tinto analog site is currently used by international research centers and industry (e.g., Wood Hole Oceanographic Institute, NASA Ames Research Center, University of Washington at St. Louis, Honey Bee Robotics) and is among analog sites that Europlanet Transnational Access (TA) programme supports, promoting the visit the field sites for European and international researchers (e.g., Europlanet TA1 facility 2).

In this terrestrial analog site, we performed a VIS-NIR measurement campaign intending to detect Martian analog compositions and testing acquisition procedures in support of the Ma_MISS instrument scientific activity.

The Rosalind Franklin rover mission will be launched in 2028 to determine whether life ever existed or is still active on Mars and characterize the water/geochemical environment as a function of depth in the shallow subsurface. The rover payload (i.e., Pasteur payload) consists of panoramic instruments, contact instruments, the Analytical Laboratory Drawer (ALD), and two subsystems (Vago et al., 2017). Among the panoramic instruments, there are: 1) The Panoramic Camera (PanCam) composed of two wide-angle stereo cameras (WAC) and one high-resolution camera (HRC) to investigate the landing site environment and geology, it will be used to image collected samples (Coates et al., 2017); 2) a context Infrared Spectrometer that will substitute the Russian one ISEM (Korablev et al., 2017) capable of detecting the infrared spectrum of a given target and determining its composition; 3) the radar-GPR WISDOM with surface penetration for shallow depths, capable of detecting the presence of water, ice, and stratification down to depths of 3 m, with a vertical resolution of a few centimeters (Ciarletti et al., 2017); and 4) ADRON a gamma rays and neutrons detector that will investigate the number of thermal and epithermal neutrons scattered (with a range from 0.01 eV to ∼100 keV) in the subsurface, to determine the possible presence of ice, hydrated minerals, and hygroscopic and/or capillary water (Mitrofanov et al., 2017).

Among the contact instruments, there is 1) the CLose-UP Imager (CLUPI) which is devoted to investigating rock targets, and cuttings produced during drilling operations, and imaging the collected samples with sub-millimeter resolution (Josset et al., 2017), and Ma_MISS (Mars Multispectral Imager for Subsurface Studies; De Sanctis et al., 2017; De Sanctis et al., 2022), the miniaturized visible and near-infrared (VIS-NIR) spectrometer (0.5–2.3 μm; spatial resolution 120 μm; spectral resolution ∼20 nm) integrated into the drilling system devoted to Martian subsurface exploration. The spectral data collected by Ma_MISS on unexposed rocks will be crucial for 1) the determination of the composition of the subsurface rocks, 2) the optical and physical properties of materials (i.e., grain size), and 3) the reconstruction of a stratigraphic column, getting clues about subsurface geological processes. In addition, the Ma_MISS data could give some hints about the presence of biomarkers in the subsurface because the presence of organic matter can affect some spectral parameters (Ferrari et al., 2023). Ma_MISS will run periodically during pauses of the drilling activity and depending on the adopted scientific strategies, it will perform “columnar” or “ring” acquisitions using the vertical translation or the rotation of the drill reconstructing the hyperspectral image of the drill’s borehole (De Sanctis et al., 2017).

The ALD contains the instruments devoted to the analysis of the collected samples, these are: 1) MircOmega a NIR hyperspectral camera for characterizing mineralogy and organics at the micrometric scale (Bibring et al., 2017); 2) the Raman Laser Spectrometer (RLS) that provides information both on the mineralogy and to establish the presence of organic and inorganic carbon (Rull et al., 2017) and 3) the Mars Organic Molecule Analyzer (MOMA) that can identify a wide range of organic molecules, even if they have a low concentration (Goesmann et al., 2017).

The two subsystems onboard the Rosalind Franklin rover are the Drill capable of obtaining samples up to 2 m depth and the Sample Preparation and Distribution System (SPDS) which manage the collected samples and delivers them to the analytical laboratory instruments (Vago et al., 2017).

In this framework, we performed a field campaign in the Río Tinto area where we performed a set of VIS-NIR measurements using the portable spectrometer ASD FieldSpec4 both on acidic alteration minerals and on biosignature-bearing geological samples.

In addition, for each analyzed mineral/rock on-field, we collected representative samples that we measured by micro-Raman spectroscopy at the INAF-IAPS laboratories (Rome, Italy). This work will provide crucial preparation for the exploitation and interpretation of the scientific data that the Ma_MISS instrument will acquire during the active phase of the next Mars mission.

The “Ravine of the fools” (37°42′30.6″N 6°33′29.4″W) is located along a freshwater stream (Figure 3A) that carves a valley immediately north of a deposit of mining tailing deposit, which we will here refer to as the Pequeño Grand Canyon. Along this valley, the stream activity has generated fluvial terraces where the stratigraphy is exposed. These river terraces represent ad hoc sites where it is possible to collect a series of spectra across the stratification (Figure 3B). This type of on-field acquisition results useful to simulate a Ma_MISS columnar acquisition in the Martian subsurface. Figure 3B shows an exposed wall of a fluvial terrace situated on the orographic left of the stream. The blue dots represent the 37 acquisition points of the VIS-NIR spectrometer. The distance (Δ)_z between each acquisition point is about 1 cm. The total length of this survey is about 38 cm. It crosses six different lithotypes, almost all of which appear in horizontal stratification (A, B, C, D, E) except for type F, which represents a black spot embedded in the stratification. This acquisition was made going from the top (layer A) to the bottom (layer E) including the F black spot. All the collected spectra are reported in Figure 3C, the large noise near 1,400 and 1,900 nm is due to the atmospheric water that is difficult to compensate with the reference target measurement using the sunlight as a light source during open-air acquisitions.

In Figure 4A the full-range average spectra of each lithotype are reported. Layers A, C, and E show spectra with similar characteristics especially in the visible range and near 900 nm related to Charge Transfer Absorption (CTA) and Crystal Field Absorption (CFA) effects as highlighted by the gray and orange bands respectively in Figure 4B. In contrast, layers B and D do not show these strong bands at the same locations and D shows a high level of reflectance in the visible range. Layer F shows a lower reflectance level with a strong red slope in the NIR range; Figure 4C shows that also in the NIR field the layers A, C, and E are like each other showing a band close to 2,200 nm which in the case of E is wider than the others. In addition to this band, the D layer shows absorption at 2,265 nm which is also present in the B layer spectra although it appears to be a weaker absorption. The D layer in addition to this band shows a doublet with band centers at 2,165 and 2,210 nm and a band near 2,285. The black spot (F) shows no absorption at the same wavelength as the other lithotypes but appears to have a very broad absorption centered near 2,230 nm.

The Pequeño Grand Canyon (37°42′09.5″N 6°33′10.1″W) is a large mining tailings basin where surface waters flowing into the Río Tinto have carved a small canyon. This deposit appears fine-grained and intensely layered with a random succession of two different lithotypes. The grey layers have a very fine grain size and plastic behavior, whereas the white ones have a larger grain size and sandy texture. The thickness of grey and white layers is comparable and ranges from a few millimeters to a few centimeters with sub-horizontal bedding throughout the entire area. Figure 5A shows the south sector of the Pequeño Grand Canyon where we performed the spectroscopic survey across the stratigraphy; Figure 5B shows the exposed wall of the mining waste deposit where, in this case, we collected 11 acquisition points with variable spacing (Δ)_z collecting one spectrum per layer for a total length of the survey of about 15 cm crossing the stratigraphy. This acquisition was made going from the top (white layer) to the bottom (grey layer). All the collected spectra are reported in Figure 5C, the large noise near 1,400 and 1,900 nm is due to a bad compensation of the atmospheric water.

Figure 6A shows the average spectra of the gray and white layers as the result of the spectral diversity found in the Pequeño Grand Canyon. The average spectrum of the gray layers shows a 430 nm band and a lower reflectance level than that of the white layers. In contrast, the average spectrum of the white layers has a higher reflectance level and shows no absorption band in the visible range. Figure 6B shows the same spectra after continuum subtraction. This operation on the spectra allows us to highlight the differences in the depth of some absorption bands common to the two different layers. Both layers show bands near 1,400 and 1,900 nm, which in the case of the gray layers are stronger. The continuum removal procedure normalizes reflectance spectra to facilitate the evaluation of individual absorption features from a common baseline. We subtracted a baseline from each spectrum. The baseline was drawn using straight-line segments that connect local spectra maxima individually for each spectrum by selecting anchor points in the intervals where no absorption features were observed. Figure 6B also highlights a difference in terms of the depth of the band at 2,206 nm, which in the case of the representative spectrum of the white layer is considerably deeper.

Berrocal is located about 10 km south of the town of Nerva (37°35′35.1″N 6°33′03.5″W). In this location, on the riverbed, there are pebbles partially rounded by the fluvial activity that are visibly altered by the acidic environment (Figure 7A), with surface colorations tending to reddish and with grey/beige alteration features along fractures and pebble boundaries (Figure 7B). These pebbles, probably consisting of a volcanic rock, appear to be cemented to the ground by salty crusts that make the substrate a kind of conglomerate consisting of these centimetric pebbles with a fine-grained matrix. At this location, we decide to collect data from a fragment of these rock pebbles cemented to the ground to evaluate by VIS-NIR spectroscopy both the composition of the rock, trying to derive information on metamorphic minerals related to hydrothermal processes (Large et al., 2001), and the late alteration related to the effect of the acidic environment resulting from exposure to the acidic river water.

The location of the analyzed spots on the pebble and the resulting spectra are reported in Figures 7C, D respectively. The spectrum of the un-altered rock (green spectrum in Figure 7D) is characterized by a broad absorption centered at 1,065 nm with a left shoulder at 750 nm. A smaller absorption band is located at 660 nm. Another weak but resolvable absorption is centered at about 400 nm. In the NIR range the spectrum is characterized by a V-shaped absorption with a center at ∼1,550 nm. The most prominent feature of the spectrum is a doublet with band centers at 2,257 and 2,350 nm. The other three spectra resulting from acidic aqueous alteration show spectral features completely different from the unaltered surface. The spectrum collected in correspondence with the blue spot (blue spectrum in Figure 7D) is characterized by a deep absorption band at 930 nm with a left shoulder at 775 nm. Smaller absorption features are found at 680 and 500 nm. In the NIR range it is possible to see an inflection at 1,900 nm and very weak features at 2,257 and 2,350 nm which are the same absorption bands observed in the un-altered rock spectrum (green in Figure 7D) but almost completely obliterated by the alteration process. Spectrum collected in the red spot position, (red spectrum in Figure 7C), the main absorption is found at 905 nm. It is characterized by a strong intensity and a marked asymmetry. In the visible region, the only other observable absorption band is located at 435 nm. The NIR region is characterized by weak absorptions superimposed to high noise regions, so it is not easy to assess their shape and position. There is a small absorption feature at 2,265 nm.

This site is located This location is along the course of the Río Tinto (37°35′00.6″N 6°33′00.5″W) at just under a kilometer downstream from the location described in the previous section (3.3 Berrocal). The spot where we made the VIS-NIR measurements and collected the samples is in correspondence with the inlet of a freshwater tributary into the Río Tinto. The fluvial sediments that characterize the location appear stratified with grain size and color varying from yellow to gray and with a preponderance of purple sediments (Figure 8A). In this case, we chose to characterize the purple sediments that at their base have a slightly less sorted grain size layer consisting of a very fine light gray part along with grains of different nature of higher in size (i.e., the left side of the sample in Figure 8B).

In the case of the sample characterized by a purple color collected in the site “La Casa de Los Sordos,” the spectra are characterized by well-defined electronic absorption in the visible spectral range. The spectrum collected in the bottom grey layer has slightly different spectral features with respect to the purple top layer. In all the spectra we see a strong absorption centered at near 880 nm with a left shoulder near 775 nm. In the bottom layer, this band has a center at higher wavelengths (935 nm). Other well-defined absorptions occur near 550 nm and 440 nm for the top layer while the bottom layer does not show the band near 550 nm (Figure 8C). The NIR region of the spectrum is characterized by several absorption bands. Unfortunately, the absorption occurring around 1,400 nm and 1,800–1,950 nm are superimposed to the high noise region of the on-site spectra due to a bad compensation for the atmospheric water. Despite the presence of a noisy signal in the on-site spectra, it is possible to see a clear inflection roughly centered near 1,400 nm, another complex absorption around 1,900 nm features, and a doublet at 2,205 and 2,265 nm.

The tunnel (Figure 9A) is located near the Río Tinto birth (37°43′17.1″N 6°33′03.9″W) and is a structure above which is the road to Peña de Hierro and below which flows the Río Tinto, which in this site, has the appearance of a stream. Previous studies that characterized the ecosystem of this site (Gómez et al., 2012) highlight its astrobiological relevance. The tunnel is a man-made artifact probably made for logistic purposes useful to mining extraction and was built using rocks that have mineralogical characteristics entirely akin to those of the mining area. Internally, the tunnel has walls covered with saline/mineral inflorescence due to the percolation of meteoric water through the embarkment that contains the tunnel itself. The water percolating through the rocks alters/oxidizes the metal-rich minerals (i.e., polymetallic sulfides) and deposits salts at the interface between the tunnel wall and the open air. The inner walls of the tunnel, where not covered by saline inflorescence seem to be made of rock blocks with characteristics entirely like those of the surrounding area held together by cement mortar. Figure 9B shows a detail of the left wall (referring to the left riverbank) covered with saline efflorescences of different colors. There are orange and white efflorescences forming a surface below which is possible to find a green patina due to the presence of alleged organic material. On these three different material types, we collected VIS-NIR spectra, the measured points are indicated by the circles in Figure 9B with the colors corresponding to the related spectra shown in Figure 7C. In this case, we show the spectra only in the range 400–700 nm since the poor illumination conditions inside the tunnel caused a high noise in the NIR range. The spectrum collected on the orange efflorescences (orange spectrum in Figure 9C) shows strong absorption in the region of 400–500 nm and near 850 nm. The spectrum collected on the white efflorescence (grey spectrum in Figure 9C) shows a higher level of reflectance and an absorption band near 430 nm and a shallower absorption at around 500 nm with respect to the orange one. The spectrum collected on the green patina (green spectrum in Figure 9C) shows a different spectral shape to the previous ones with absorption bands near 430, 620, and 675 nm, with a scattering peak in the green region at 532 nm and one at 650 nm. In the case of the spectra collected on the green material, the absorption band close to 430 nm could be consistent with the presence of spots of white material, which are not discernable among the green patina at the spatial resolution of the ASD FieldSpec4. This issue could be fixed by calculating the band centers in the region of 430 nm on both white and green spectra (Figure 11).

In a different part of the left wall of the tunnel, the efflorescences are thicker and appear to be tougher with respect to those previously analyzed. By collecting a sample of this type, we detached a thick mineral crust (Figure 10A) incorporating in the mineral matrix a 2-mm-thick, green-colored layer immediately below the yellowish surface. Internally, where the green layer is not present, the mineral crust appears white, while on the side where it was attached to the wall (wall side, left side of the sample, Figure 10A) it has a reddish color. Once collected, the sample was taken out of the tunnel to have the maximum possible solar illumination, useful for collecting the spectrum over the entire Fieldscpec4 range. The spectra collected on the reddish side, in the inner white part of the mineral crust, and on the green layer are shown in Figure 10B. The white crust spectrum shows a higher reflectance than the other two spectra and shows an absorption band near 430 nm and three strong absorption bands near 1,195, 1,450, and 1,750 nm in addition to a shallower absorption near 985 nm. The spectra collected on the green layer, in the visible part of the range (Figure 10C), show the same spectral characteristics as that collected on the green patina on the tunnel wall (Figure 9C) but at higher wavelengths it shows the same spectral shape and absorptions as that of white crust. In particular, the spectra collected on the white mineral crust and that on the green layer show absorption with two minima in the band at 1,195 nm and three minima in the band near 1,750 nm. The spectrum collected on the reddish layer has a very different spectral shape from the other two. It has a lower level of reflectance and strong absorptions near 400, 650, and 950 nm. This spectrum also shows an absorption feature near 1,440 nm but unlike the other two, this band does not show three distinct minima.

To better understand whether there is an actual difference in the position of the bands near 430 nm in the spectra of the green layer and that of the white crust we present data collected on the tunnel wall (Figure 11A) and those of the mineral crust sample (Figure 11B) after the spectra continuum removal. Also in this case we subtracted a straight-line segment baseline from each spectrum. This procedure allows a more accurate assessment of the band positions, however, mixing effects between the two components of the spectrum (white layer and green patina) due to the spatial resolution of the ASD FieldSpec4 could be present. In both cases, data collected on the tunnel wall show the band centers at 435 nm in the case of the white material and at 439 nm in the case of the green ones. This determination will be useful in the characterization of the organic material that grows below the white mineral crust surface.

To better characterize the mineralogy and the nature of organics, the types of material measured with the VIS-NIR spectroscopy in the field were measured in the laboratory with the micro-Raman spectrometer. Figure 12 shows the Raman spectra collected on the two types of efflorescences on the wall and described in Section 3.5. In the case of the orange efflorescence (Figure 12A), Raman peaks were detected at 271, 303, 474, 607, 990, 1,020, and 1,107 cm⁻¹. The case of the white efflorescence sample shows peaks at 208, 282, 501, 598, 1,023, 1,096, and 1,198 cm⁻¹. The Raman data confirms a clear mineralogical difference between the two types of efflorescence that characterize the tunnel wall confirming the spectral diversity found in the VIS-NIR data.

The sample collected inside the tunnel (i.e., the sample in Figure 10A) in the white parts shows a crystal structure consisting exclusively of colorless tabular crystals but also prismatic crystals that have dimensions in the range of 50–300 microns (Figure 13A). In correspondence with the green layer, the colorless crystals coexist with green-colored material consisting of spheroidal elements with dimensions of about five microns. The Raman spectrum collected in this area of the sample and shown in Figure 13B shows peaks at 414, 493, 619, 670, 1,008, and 1,135 cm⁻¹, differing greatly from the spectra collected on the green material which show Raman peaks at 1,106, 1,158 and 1,522 cm⁻¹ (Figure 13C).

The results of laboratory Raman investigations performed on the measured types/samples described in the previous sections (Sections 3.1–3.5) are summarized in Table 2. In the Ravine of Fools site, we found abundant quartz associated with oxides (goethite) and sulfates (natrojarosite). The sample collected in Berrocal is characterized by a complex mineralogical association: augite, titanite, epidote, K-feldspar, and quartz are representative of the unaltered rock while natrojarosite and hematite are only found in the external crust of the sample. In the samples collected in the sites: Pequeño Grand Canyon and La casa de los Sordos we found evidence of clay minerals (illite) in association with coquimbite, quartz, feldspar, barite, and hematite.