Kay Wohlfarth

An accurate thermal correction applied to Moon Mineralogy Mapper (M ³ ) visible/near-infrared (VNIR) spectroscopy data is necessary for proper interpretation of lunar surface hydration (i.e., OH ⁻ /H ₂ O) signatures. We compare thermal correction techniques applied to both global and targeted M ³ image data cubes to assess correction effects on M ³ 3 μ m region observations, with implications for interpreting the presence of hydrated materials on the lunar surface. We present M ³ VNIR spectra collected at five different lunar highlands study sites, acquired from separately prepared images with different thermal corrections applied. The recently applied thermal correction by K. Wohlfarth et al. reveals VNIR spectral signatures consistent with the presence of hydration at all study region sites. The observed hydration signatures are different and generally weaker or absent in M ³ Level 2 global and targeted reflectance data currently available on the Planetary Data System.

Wildfire detection and characterization from space critically depend on accurate thermal infrared measurements across a wide dynamic range. OroraTech’s SAFIRE payloads feature mid-wave infrared (MWIR) and long-wave infrared (LWIR) bands optimized for this purpose, yet the volume and power constraints of many CubeSat platforms preclude the use of onboard calibration sources. This reflects a broader limitation of current Earth observation systems: the absence of robust calibration and validation methodologies at high brightness temperatures relevant to active fires. We evaluate the potential of using thermophysical models (TPM) of the Moon as a vicarious calibration reference for calibration transfer between instruments. The Moon reaches surface temperatures of up to 400 K, offering a stable target in the thermal domain that is visible from many different orbits. Previous research has established the use of TPMs for Moon observations in the infrared; however, uncertainties in surface properties remain, particularly in the mid-wave infrared. We investigate these effects using SAFIRE observations of the lunar surface acquired over a wide range of lunar phase angles (from waxing -81.5° to waning +122.2°). We constrain the TPMs using observations from the Sentinel-3 Sea and Land Surface Temperature Radiometer (SLSTR) fire channels and uncover a 5.2 % inter-sensor disagreement in the MWIR between Sentinel-3A and Sentinel-3B at high brightness temperatures, corresponding to ~2.2 K at 400 K. Applying the same methodology to observations of our SAFIRE payloads bounds the calibration transfer error at 5.8 % in the MWIR and 4.2 % in the LWIR for lunar phase angles within ±45°. These results establish lunar vicarious calibration as a viable approach in the thermal domain for high-temperature applications, providing a pathway toward improved fire radiative power retrievals and enhanced global wildfire monitoring.

IntroductionLunar swirls are a unique class of surface features defined by sinuous, high-albedo markings that lack distinct topographic or stratigraphic expression. These features are influenced by crustal magnetism, regolith maturation, and space weathering, yet decades of orbital observations have not resolved the competing hypotheses for their formation. Proposed models include solar wind shielding by localized mini-magnetospheres [1, 2], surface scouring from cometary impacts [3], and electrostatic transport of fine-grained regolith [4, 5]. Previous efforts to distinguish among these hypotheses have primarily relied on qualitative interpretation, mineralogical correlations [6], and linear regression methods [7]. However, these traditional approaches are limited in their capacity to capture the complex, high-dimensional, and nonlinear relations present in multimodal planetary datasets. To address the need for advanced data integration in planetary science, this study introduces a multimodal machine learning model that synthesizes distinct categories of orbital datasets, such as spectral and radar measurements, to identify the unified physical signatures of lunar swirls.MethodologyDrawing on recent breakthroughs in foundational vision-language models (VLA), a state-of-the-art multimodal Masked Autoencoder (MAE) based on a Transformer architecture was developed and trained on 12 distinct remote sensing modalities. The dataset covers 56 sites across 8 recognized lunar swirls (see Figure 1). As shown in Figure 2, the model processes heterogeneous input modalities, including image-like geophysical observations (such as LRO WAC reflectance, LRO Mini-RF radar backscatter, M3-derived elemental abundance maps, and brightness temperature) and textual spatial position information, which are jointly routed through a shared Transformer encoder-decoder. To approach the unique characteristics of each input, image-based modalities are compressed into discrete latent tokens using a dedicated, pre-trained VQ-GAN-style tokenizer, while textual coordinate modalities employ a structured, regex-based tokenization method. This brings the training dataset up to 39 million individual tokens. The Transformer encoder processes a visible subset of these tokens, and the decoder reconstructs the masked subset for each modality, compelling the network to acquire robust cross-modal geophysical relationships. The model underwent rigorous validation by repeated stratified cross-validation, followed by deployment as an exploratory tool using Leave-One-Out (LOO) ablations, Leave-One-In (LOI) evaluations, and Representational Similarity Analysis (RSA).ResultsDeploying the validated Transformer as an exploratory probe revealed quantitative constraints on the physical nature of lunar swirls. Leave-One-Out and Leave-One-In ablations identified plagioclase abundance as the most critical modality for boundary delineation, while Wide Angle Camera (WAC) reflectance emerged as the most independently sufficient dataset. Representational Similarity Analysis further demonstrated that the network encodes geophysical content along directions geometrically distinct from spatial coordinates. This confirms the model successfully captures a shared physical context across swirl sites, rather than relying on location-specific biases. By extending this ablation logic to the spatial domain, we found that the dependence of optical reflectance on mineralogical inputs is significantly stronger within swirl interiors than in the surrounding terrain. Crucially, no comparable enhancement in cross-modal coupling was observed for thermal emission or Mini-RF radar backscatter. This distinct asymmetry provides the empirical signature predicted by the maturation-retardation hypothesis, clearly separating it from competing scenarios, including cometary impacts or dust transport.Discussion and ConclusionThe enhanced optical-compositional coupling observed inside lunar swirls is the empirical signature uniquely predicted by the maturation-retardation hypothesis [8, 9]. Reduced solar wind exposure preserves the structural relationship between regolith composition and observed brightness, a relationship that is otherwise obscured by accumulated nanophase iron in normally weathered background terrain [8, 11]. Alternative scenarios, such as cometary impacts [3], electrostatic dust transport [4], or thermophysical modification, do not predict this specific asymmetry. Instead, these alternative formation models would suggest that thermal or radar properties should exhibit similarly enhanced coupling to other physical domains within the swirls. This work characterizes the analytical limits of current orbital archives, highlighting that orbital data alone cannot definitively distinguish solar-wind shielding from other mechanisms that equivalently slow regolith maturation. Resolving the deeper structural and magnetic origins of swirls requires in-situ measurements, such as those planned for the upcoming Lunar Vertex mission [10]. Our results show that machine learning can extract complex, non-linear relationships from existing planetary datasets. This not only maximizes the data from legacy missions but also helps prioritize targets for future surface exploration.References[1] Bamford, R., et al., "Minimagnetospheres above the lunar surface and the formation of lunar swirls", Physical Review Letters, 2012.[2] Deca, J., et al., "Reiner gamma albedo features reproduced by modeling solar wind standoff", Communications Physics, 2018.[3] Schultz, P.H., Srnka, L.J., "Cometary collisions on the moon and mercury", Nature, 1980.[4] Garrick-Bethell, I., et al., "Spectral properties, magnetic fields, and dust transport at lunar swirls", Icarus, 2011.[5] Domingue, D., et al., "Topographic correlations within lunar swirls in mare ingenii", Geophysical Research Letters, 2022.[6] Domingue, D., et al., "Spectrophotometric and topographic correlations within the mare ingenii...

. Introduction:Mercury is a fascinating planet for thermal modelers. With extreme temperatures ranging from 100 K at night to over 700 K during the day [1, 2], and its 3:2 spin-orbit resonance with the sun [3], modelling Mercury’s thermal environment presents a unique challenge. Because of its proximity to the Sun and the orbit-spin resonance, Mercury also experiences differences in regional temperature identified as “hot” and “cold” poles. Various factors influence the temperature on Mercury’s surface like the albedo variation of impact craters, topography, surface morphology, distance to the sun, density and conduction of the material etc.Over the years, many studies have attempted to model the thermal properties of Mercury, often by using the Moon as a reference due to both being an airless body [4, 5]. From Mariner 10 to BepiColombo, the accuracy of these thermal models for Mercury has significantly improved, thanks to mission data. However, due to limited knowledge about the surface composition and other constraints, thermal models still require further refinement.In this study, we present a thermal model aimed at improving our understanding of Mercury’s local temperature behavior, ahead of MERTIS’s arrival at Mercury.2. MERTIS and the observation area:MERTIS (Mercury Radiometer and Thermal infrared Imaging Spectrometer), onboard the ESA-JAXA mission BepiColombo, is a thermal infrared spectrometer (TIS) and radiometer (TIR) that is part of the Mercury Planetary Orbiter payload [6]. TIS has a spectral range of 7-14 μm with a resolution of 90 nm, and TIR, a radiometric range of 7-40 μm, divided into two bands - 8-14 μm and 7-40 μm [6].One of the primary objectives of MERTIS is to study surface temperature distribution on both day and night side of Mercury. It first observed Mercury during BepiColombo’s 5th flyby on 1st December, 2024. The spacecraft achieved its closest approach at 37,000 km and MERTIS was able to observe the planet, through its space baffle, for 37 minutes, with a pixel resolution of 27 km for TIS and 100 km for TIR.The observed area extended from 140°E to 230°E longitude on Mercury and covered major craters such as Bashō crater in the south and parts of Caloris Basin in the north including the hot pole at 180°. MERTIS was able to observe the planet with an incidence angle of as little as 0.2° to as high as 95.00°, where the Sun is below the horizon, covering therefore a large range of local time.3. Methodology:The thermal model in our study takes the energy balance thermal equation for airless bodies as a reference to calculate the temperature distribution within the observed area, where absorbed solar energy equals emitted thermal radiation.Equation 1: Energy balance equation for airless bodies.Where: T = Temperature A = Bond Albedo - 0.058 S = Solar Flux 𝜎 = Stefan-Boltzmann Constant 𝜀 = Surface emissivityThe incoming radiation, or solar flux is calculated using a constant bond albedo [2], the distance from the sun and the incidence angle. The incidence angles of more than 90° are ignored as this indicates that the sun is below the horizon.Equation 2: Incident solar energy or solar flux.Where:F0 = Solar constant for Mercury at distance from the sun.𝜃 = Incidence angle (

IntroductionOn December 1st, 2024, the ESA/JAXA spacecraft BepiColombo performed its fifth Mercury swing-by (MSB#5), during which the onboard Mercury Radiometer and Thermal Infrared Spectrometer (MERTIS) [1,2] observed Mercury’s surface for the first time [3]. MERTIS consists of a radiometer (TIR, two channels, 7-14 µm and 7-40 µm) and an imaging spectrometer (TIS, 80 channels, 7-14 µm). A thermal model is required to retrieve emissivity spectra from the TIS data and to address the anisotropic thermal emission caused by surface roughness. Possible approaches are simple semi-analytical roughness models fit to an emission phase function (EPF) [4,5] or fully numerical models. For this work, we use our numerical thermophysical roughness model [6], which has been validated for the lunar surface with Gaofen-4 [7], Diviner [8], and MERTIS [6] data. We derive essential parameters to tailor the model to Mercury and retrieve spectral emissivities. Performing Principal Component Analysis (PCA) and clustering allows us to gain first insights into Mercury’s emissivity variations.The planet baffle of MERTIS was blocked during MSB#5 due to BepiColombo’s stacked cruise configuration. The observations were instead taken through the space baffle. While the planet baffle was extensively calibrated before launch [9], the space baffle was initially not intended for target observations. The effects of the baffle complicate data calibration. Hence the results presented here should be considered preliminary. The team is currently investigating and improving the calibration [3].Thermal ModelMany airless planetary bodies (e.g., Mercury, asteroids, and the Moon) are covered by a rough and highly insulating regolith. Different surface orientations on small spatial scales (mm-cm) lead to significant temperature variations on those scales. Consequently, the observed thermal emission is a superposition of multiple Planck functions. The thermal emission produced by a rough surface can significantly deviate from that of a smooth surface (corresponding to a single Planck function).We model this subpixel roughness with fractal surfaces [6], including self-scattering and self-heating between the facets. Computing the model for every TIS pixel is computationally expensive. Instead, we evaluate the model for a uniform grid in the parameter space and interpolate the results according to the MSB#5 geometry. We estimate three main model parameters:The bolometric directional-hemispherical albedo Adh, which controls the reflected solar irradiance. The bolometric hemispherical emissivity εh, which controls the total emitted thermal radiance. The surface roughness θ, which alters the directional thermal emission characteristic. Model Parameter EstimationA theoretical analysis of laboratory data suggests that the bolometric emissivity of Mercury’s surface is lower than that of the Moon, which is often assumed to be εh = 0.95 [11]. Figure 1 shows modeled radiances for different global values of the directional-hemispherical albedo compared to the measured radiances. Even when Adh is set to values outside the plausible range, the model does not match the data. By reducing the bolometric emissivity to εh = 0.75 , we can reproduce the measured radiances (see Figure 2).Figure 1. Radiances at 8.6 µm with εh = 0.95Figure 2. Radiances at 8.6 µm with εh = 0.75We derived individual Adh values for each MERTIS measurement from MDIS data [10] to incorporate spatial albedo variations. Averaging the values yields 0.07.Figure 3 shows the same profile as Figure 1 and 2 for different surface roughness parameter values. Near the subsolar point, a rough model behaves like a smooth model, and the parameter has no effect. However, a smooth model cannot fit the data for increasing latitudes due to thermal beaming [12]. We find that a roughness between 30° and 35° is required to explain the data.Figure 3. Radiances at 8.6 µm for different roughness values.Figure 4 shows the radiances gridded to a perspective projection. While the model agrees with the data in the latitudinal direction, there is an offset in the longitudinal direction. However, the highest temperature of a slowly rotating body like Mercury should be centered on the subsolar point and decrease in all directions with increasing incidence angle. This offset is likely caused by the unusual observation conditions. We used a wavelength-independent and pixel-position-dependent fourth-order polynomial fit to correct the radiance scaling without affecting the shape of the spectra.Figure 4. Radiances at 8.6 µm. The black star marks the subsolar point. The dashed area is investigated in Figure 5.High-Level Emissivity AnalysisSelecting only frames where all 100 TIS pixels saw Mercury allows us to use Fourier filtering to suppress the striping artefacts. To reduce the dimensionality of the data and visualize spatial variations, we applied a PCA to the retrieved spectral emissivities. We only use the first two components PC1 and PC2. Figure 5 shows a false color image with PC1 in red, PC2 in green and a constant value of 0.5 in blue. Further, we used a self-organizing feature map (SOM) to cluster the features PC1 and PC2 into 9 clusters.Figure 5. Comparison between maps derived from MDIS [10,13] and MERTIS data. The Caloris basin is visible in the PCA image and cluster index map. Several show up in the emissivity-derived maps. The low reflectance material around Tolstoj vanishes, and the Tolstoj interior is assigned to the same cluster as part of the Caloris interior.References[1] Hiesinger, H., et al. 2010, Planet Space Sci. 58 [2] Hiesinger, H., et al. 2020, Space Sci. Rev., 216, 147 [3] Adeli, S., et al., EPSC 2025. [4] Jhoti, E., et al. (2023), TherMoPS IV, 48 [5] Powell, T., et al. 2025, LPSC 2025 [6] Wohlfarth, K., et al., 2023, Astronomy & Astrophysics, 674, A69. [7] Wu, Y., et al. 2021,...

Introduction: The MERTIS (MErcury Radiometer and Thermal Infrared Spectrometer) is a mid-infrared imaging instrument onboard the BepiColombo ESA/JAXA mission to Mercury expected to arrive in 2026. Part of the instrument suite is MERTIS, the Thermal Infrared spectrometer (TIS), covering the wavelength range from 7 to 14 µm, which will map the mineralogy of the surface of Mercury [1,2].MERTIS provided the first thermal infrared data of the hermean surface from a spacecraft after Mariner 10 [3], since it was among the few instruments used during the 5th flyby at Mercury. Owing to the distance of nearly 40000 km during the flyby, the pixel footprint is relatively large (~30 km), but already allows distinguishing surface details.The main challenge even with the first batch of spectra is to obtain quantitative mineralogical information from the vast amount of data. The standard methods of quantitative mineralogical data analysis are unmixing routines. However, for the accurate modal mineralogy, these routines but are time-intensive and require additional inputs [e.g., 4]. Band ratios are a powerful alternative to obtain a fast first impression of the surface mineralogy and perform mapping.Techniques: The characteristic bands of minerals are difficult to identify in mixtures as expected for surface regolith owing to overlapping features and additional physical effects (e.g., temperature). Our goal is to find easy to identify bands in complex mineral mixtures.In the first step to identify such band ratios, 28 synthetic mixtures with exactly defined modal mineralogy were used [5,6]. We used only the spectra of the finest size fraction (0-25µm), which is probably the dominant grain size on the hermean surface [2]. We computed emissivity spectra from our reflectance spectra using Kirchhoffs‘ law [7]. While this is a simplified approach, future studies will take the directional hemispherical laboratory setup into account.In order to avoid absolute spectral intensities, we modelled band ratios between features, where the integrated area of two bands is divided. MERTIS spectra consist of 80 channels, binned by a factor of 2.Using a Python code, ,we calculated all possible ratio combinations for the spectra of all mixtures – starting with bands consisting of one channel, to bands with a width of 40 channels. The results are 28 arrays with all possible band ratio intensities for each spectrum.To relate these ratios with mineralogical composition, all the band ratios for a given synthetic mixture were correlated with the known modal mineralogy for 6 phases of this mixture (Glass, Forsterite, Diopside, Plagioclase, Enstatite and Quartz). Thus, we identified the band ratios with the strongest correlations to a given mineral phase (Table 1). Data Processing: We used Python to develop routines for data processing. Open-source package we used were Pandas, NumPy, SciPy, and Matplotlib [8-11].Results: For all six phases, remarkably high correlations r = 0.89 – 0.99 were found (r=correlation coefficient). Table 1 presents the wavelength ranges for the two bands of the highest correlations of each mineral phase.Summary and Outlook: We have identified characteristic band ratios based on synthetic laboratory spectra for the use on hermean surface spectra. In the following step, we will apply these ratios to create mineral maps of the hermean surface using the processed MERTIS emissivity spectra of the 5th flyby.Acknowledgments: MPR, JHP, MPR, IW, AM, KEB, and JHP were funded by DLR grant number 50 QW 2201 A. KW and MT were partially funded by DLR grant number 50 QW 2201 B. References: [1] Benkhoff J. et al. (2010) Planetary and Space Science 58, 2-20 [2] Hiesinger H. et al. (2020) Space Science Reviews, 216, 1-37, 115498 [3] Chase, S. C. (1976) Icarus 28, 565-578 [4] Bauch K.E. et al. (2023) LPSC 54, 2247 [5] Morlok A. et al. (2024) Icarus 425, 116078 [6] Morlok A. et al. (2023) Icarus 396 [7] King, P. et al. (2004) In: Mineral, Assoc. of can. Short Course Ser. 33. Min. Ass. of Canada, Ottawa, 93-133 [8] McKinney, W. (2010) Proceedings 9th Python Sci. Conf. 56-61 [9] Harris et al. (2020) Nature 585,357-362 [10] Virtanen, P. (2020) Nature Methods 17, 261-272 [11] Hunter, J.D. (2007) Computing in Sci. & Eng. 9, 90-95 Phase r BAND 1 BAND 2 ID 158 Glass 0.89 8.49-10.41 8.84-10.59 ID 249 Forsterite 0.97 7.96-8.84 7.53-9.19 ID 22 Diopside 0.96 9.19-12.08 7.18-9.63 ID 28 Plagioclase 0.94 12.16-12.86 12.78-13.83 ID 53 Enstatite 0.98 7.18-12.43 7.18-13.83 ID 13 Quartz 0.99 7.18-8.31 7.35-8.40 Table 1: Band pairs BAND 1 and BAND 2 (range of each band in µm) of the highest correlations. r = corre-lation coefficient. ID = database identification number.Figure 1: Comparison of laboratory spectra. For presentation purposes we show the data in reflectance. The light gray and pink shaded areas: range for the two bands having the highest correlation with the given mineral phase (Table 1). ID = database identification number

Determining which rocky exoplanets have atmospheres, and why, is a key goal for the James Webb Space Telescope. So far, emission observations of individual rocky exoplanets orbiting M stars (M-Earths) have not provided definitive evidence for atmospheres. Here, we synthesize emission data for M-Earths and find a trend in measured brightness temperatures (ratioed to its theoretical maximum value) as a function of instellation. However, the statistical evidence of this trend is dependent on the choice of stellar model, and we consider its identification tentative. We show that this trend can be explained by either the onset of thin/tenuous (2-rich atmospheres on colder worlds, or a population of bare rocks with stronger space weathering and/or coarser regolith on closer-in worlds. Such grain coarsening may be caused by sintering near the melting point of rock or frequent volcanic resurfacing. Furthermore, we highlight considerations when testing rocky planet hypotheses at the population level, including the choice of instrument, stellar modeling, and how brightness temperatures are derived. We also find that fresh (unweathered) fine-grained surfaces can serve as a false positive to the detection of moderate atmospheric heat redistribution through eclipse observations. However, we argue that such surfaces are unlikely given the ubiquity of space weathering in the solar system, the low albedo of solar system airless bodies, and the high stellar wind environments of M-Earths. Emission data from a larger sample of M-Earths will be able to confirm or reject this tentative trend and diagnose its cause through spectral characterization.