Stefano Nerozzi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Stefano Nerozzi has a total of 110 co-authors. Here are the top co-authors by number of shared publications:
- John W. Holt
Lunar and Planetary Laboratory University of Arizona Tucson AZ USA
5 shared publications - Roberto Jose Aguilar
Lunar and Planetary Laboratory University of Arizona Tucson AZ USA
5 shared publications - Tyler M. Meng
Department of Earth, Environmental and Planetary Sciences Washington University in St. Louis St. Louis MO USA
5 shared publications - Isaac B. Smith
Planetary Science Institute
4 shared publications Jack Holt
4 shared publicationsMichael Christoffersen
Georgia Institute of Technology
4 shared publications
Latest Publications
- Open Access
- 16 October 2026
- Published by Wiley
The Search for Accessible Martian Ice with Drone Radar: From Terrestrial Analog Surveys to Electromagnetic Simulations for Mars
- 16 October 2026
- Published by Wiley
Open AccessThe Search for Accessible Martian Ice with Drone Radar: From Terrestrial Analog Surveys to Electromagnetic Simulations for Mars
- Open Access
- 9 April 2026
- Published by Wiley
Advancing Subsurface Characterization of Mars-Analog Glaciers with Multi-System Drone-Based GPR
Martian debris-covered glaciers (DCGs) contain large quantities of water ice beneath a protective layer of rock and dust. This is supported by the observations from the Shallow Radar (SHARAD) sounding radar orbiting Mars. Properties of the overlying regolith such as density and depth to ice are critical parameters for guiding in-situ resource utilization (ISRU) of water and coring missions targeting potential habitats. Englacial debris layers that progressively outcrop the surface could also provide access to enable shallow sampling of ice ages sequences. Ground penetrating radar (GPR) is capable of imaging through the debris layer on the surface, allowing us to quantify debris thickness and composition, glacier thickness, ice purity, and the presence of englacial debris. Prior studies have demonstrated the potential of drone-based GPR for resolving these shallow features on terrestrial DGCs using an 80 MHz system (Aguilar et al., 2026), but higher-frequency surveys to better resolve the supraglacial debris and near-surface stratigraphy have not been tested. To address this gap, we conducted drone-based GPR tests at multiple frequencies at Galena Creek Rock Glacier, the best studied terrestrial analog for Martian DCGs. Our platform consists of a DJI Matrice 350 RTK drone and AeroZond LF radar operating at a center frequency of 100, 150, and 400 MHz. Given the complex topography, obstacles, and surface irregularities, we followed terrain at 3 m with a laser rangefinder altimeter. This altitude allows a high signal-to-noise ratio and reduces clutter, without compromising the safety of the drone. Preliminary results over previously imaged sites demonstrate that subsurface features were consistently resolved with greater detail than previous aerial systems. Our study demonstrates that drone-based GPR systems operating at multiple frequencies offer a robust method for surveying near-surface features in debris-covered glaciers. This approach can enhance our understanding of ice-rich landforms on Earth and inform the design of future aerial GPR platforms for Mars, particularly those targeting ISRU assessments and search of life missions. * This work was presented at the Summit on Drone Geophysics 2025. References: Aguilar, R. J., Holt, J. W., Christoffersen, M. S., Meng, T. M., & Nerozzi, S. (2026). Revealing the internal structure of Mars-analog glaciers from drone-based radar sounding. Journal of Geophysical Research: Planets, 131, e2025JE009208. https://doi.org/10.1029/2025JE009208
- 9 April 2026
- Published by Wiley
Open AccessAdvancing Subsurface Characterization of Mars-Analog Glaciers with Multi-System Drone-Based GPR
Martian debris-covered glaciers (DCGs) contain large quantities of water ice beneath a protective layer of rock and dust. This is supported by the observations from the Shallow Radar (SHARAD) sounding radar orbiting Mars. Properties of the overlying regolith such as density and depth to ice are critical parameters for guiding in-situ resource utilization (ISRU) of water and coring missions targeting potential habitats. Englacial debris layers that progressively outcrop the surface could also provide access to enable shallow sampling of ice ages sequences. Ground penetrating radar (GPR) is capable of imaging through the debris layer on the surface, allowing us to quantify debris thickness and composition, glacier thickness, ice purity, and the presence of englacial debris. Prior studies have demonstrated the potential of drone-based GPR for resolving these shallow features on terrestrial DGCs using an 80 MHz system (Aguilar et al., 2026), but higher-frequency surveys to better resolve the supraglacial debris and near-surface stratigraphy have not been tested. To address this gap, we conducted drone-based GPR tests at multiple frequencies at Galena Creek Rock Glacier, the best studied terrestrial analog for Martian DCGs. Our platform consists of a DJI Matrice 350 RTK drone and AeroZond LF radar operating at a center frequency of 100, 150, and 400 MHz. Given the complex topography, obstacles, and surface irregularities, we followed terrain at 3 m with a laser rangefinder altimeter. This altitude allows a high signal-to-noise ratio and reduces clutter, without compromising the safety of the drone. Preliminary results over previously imaged sites demonstrate that subsurface features were consistently resolved with greater detail than previous aerial systems. Our study demonstrates that drone-based GPR systems operating at multiple frequencies offer a robust method for surveying near-surface features in debris-covered glaciers. This approach can enhance our understanding of ice-rich landforms on Earth and inform the design of future aerial GPR platforms for Mars, particularly those targeting ISRU assessments and search of life missions. * This work was presented at the Summit on Drone Geophysics 2025. References: Aguilar, R. J., Holt, J. W., Christoffersen, M. S., Meng, T. M., & Nerozzi, S. (2026). Revealing the internal structure of Mars-analog glaciers from drone-based radar sounding. Journal of Geophysical Research: Planets, 131, e2025JE009208. https://doi.org/10.1029/2025JE009208
- Open Access
- 24 March 2026
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Planets
- Vol. 131 (3)
Revealing the Internal Structure of Mars‐Analog Glaciers From Drone‐Based Radar Sounding
Martian debris‐covered glaciers (DCGs) contain large quantities of water ice beneath a protective layer of rock and dust. Properties of the overlying regolith such as density and depth to ice are critical parameters for guiding in situ resource utilization (ISRU) of water and coring missions targeting potential habitats. Englacial debris layers that progressively outcrop at the surface could also provide access to enable shallow sampling of ice age sequences. To assess the capabilities of future drone‐based ground‐penetrating radar for detecting supraglacial debris thickness and near‐surface stratigraphy—properties not resolvable from orbit with the Shallow Radar sounder—we conducted tests over two terrestrial analogs for Martian DCGs. Our platform consists of a DJI Matrice 600 Pro drone and a MALA Geodrone radar operating at a center frequency of 80 MHz. We detected a bulk glacier thickness of up to 28.5 m in Sourdough rock glacier (RG), Alaska, and 48.6 m in Galena Creek RG, Wyoming. We also resolved the supraglacial debris thickness, with a mean thickness of 1.5 m in Sourdough, and in Galena Creek, with a mean thickness of 0.8 m in extensional sections and 1.3 m in compressional sections. Furthermore, we detected layers outcropping at the surface in age sequences within the cirque of Galena Creek RG. We validated the internal reflectors with clutter simulations to discriminate genuine subsurface reflectors from off‐nadir surface reflection “clutter.” Finally, we provide recommendations for optimal survey design, including solutions to increase the signal‐to‐noise ratio and reduce clutter.
- 24 March 2026
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Planets
- Vol. 131 (3)
Open AccessRevealing the Internal Structure of Mars‐Analog Glaciers From Drone‐Based Radar Sounding
Martian debris‐covered glaciers (DCGs) contain large quantities of water ice beneath a protective layer of rock and dust. Properties of the overlying regolith such as density and depth to ice are critical parameters for guiding in situ resource utilization (ISRU) of water and coring missions targeting potential habitats. Englacial debris layers that progressively outcrop at the surface could also provide access to enable shallow sampling of ice age sequences. To assess the capabilities of future drone‐based ground‐penetrating radar for detecting supraglacial debris thickness and near‐surface stratigraphy—properties not resolvable from orbit with the Shallow Radar sounder—we conducted tests over two terrestrial analogs for Martian DCGs. Our platform consists of a DJI Matrice 600 Pro drone and a MALA Geodrone radar operating at a center frequency of 80 MHz. We detected a bulk glacier thickness of up to 28.5 m in Sourdough rock glacier (RG), Alaska, and 48.6 m in Galena Creek RG, Wyoming. We also resolved the supraglacial debris thickness, with a mean thickness of 1.5 m in Sourdough, and in Galena Creek, with a mean thickness of 0.8 m in extensional sections and 1.3 m in compressional sections. Furthermore, we detected layers outcropping at the surface in age sequences within the cirque of Galena Creek RG. We validated the internal reflectors with clutter simulations to discriminate genuine subsurface reflectors from off‐nadir surface reflection “clutter.” Finally, we provide recommendations for optimal survey design, including solutions to increase the signal‐to‐noise ratio and reduce clutter.
- Open Access
- 14 March 2026
- Published by Copernicus GmbH
Unveiling the Origin and Ice vs Lithic Composition of the Mars North Polar Basal Unit with Multiband Radar Analyses
The basal unit (BU) of Planum Boreum (PB) on Mars is an ice-rich sedimentary deposit between the Late Amazonian North Polar Layered Deposits (NPLD) and the Late Hesperian Vastitas Borealis interior unit. Its two subunits, rupēs and cavi, represent records of polar geologic and climatic processes across most of the Amazonian (~3.3 Ga). The cavi unit likely consists of alternating sand and ice sheet remnants of past polar caps, reflecting volatile–sedimentary interplay, while little is known about rupēs. Thanks to recent advances in radar data processing and dense coverage by the Mars Advanced Radar for Subsurface and Ionosphere Sounding (MARSIS), it is now possible to reconstruct the stratigraphy and composition of the BU, and reveal the enigmatic nature of the rupēs unit.We analyzed over 600 MARSIS profiles at 3, 4, and 5 MHz, leveraging optimized ionospheric corrections and deep penetration to map the full thickness of the BU and retrieve its frequency-dependent complex dielectric permittivity. We find that the rupēs unit spans the western half of PB and part of Olympia Planum as a continuous body beneath the cavi unit with a pole-facing upper unconformity, occupying ~191,000 km³ (~53% of BU volume). Dielectric inversions yield a real permittivity ε’ = 4.0±0.8 (consistent at all frequencies) and a frequency-dependent loss tangent tanδ = 0.017±0.006 (3 MHz) to 0.012±0.006 (5 MHz). Both components of the dielectric permittivity exhibit strong spatial heterogeneity, with values increasing toward Hyperborea Lingula (ε’ > 6, tanδ > 0.02).These results indicate that the rupēs composition differs substantially from that of the cavi unit, with large loss tangent values indicating the presence of significant amounts of lithic materials despite the low real permittivity. Basalt alteration products with tanδ > 0.02 are required to explain the high loss tangent measurements, while their strong frequency-dependence matches the water ice imaginary permittivity behavior. We find a best match of real dielectric permittivity and loss tangent results using a mixture of 85-90% water ice and 10-15% basalt alteration products like hydrated sulfates (e.g., gypsum), clays, and ferric oxides, which are supported by spectroscopic detections at visible exposures. Rupēs lithic materials may have been transported from lower latitude sources, where aqueous alteration is more viable than at polar latitudes. However, the strong spatial heterogeneities suggest that significant localized alteration occurred in situ during the Amazonian period, perhaps facilitated by warmer high-obliquity periods predicted to occur during the last 3 Gyr. Regardless of their source, the volume of these materials corresponds to a 24 cm–thick global layer, indicating that the rupes unit constitutes a substantial sediment reservoir, not merely one of water ice. Finally, the high loss tangent measured in Hyperborea Lingula explains the lack of rupēs basal detections by SHARAD despite the relatively low thickness (i.e., 150-200 m) of the rupēs unit at that location.
- 14 March 2026
- Published by Copernicus GmbH
Open AccessUnveiling the Origin and Ice vs Lithic Composition of the Mars North Polar Basal Unit with Multiband Radar Analyses
The basal unit (BU) of Planum Boreum (PB) on Mars is an ice-rich sedimentary deposit between the Late Amazonian North Polar Layered Deposits (NPLD) and the Late Hesperian Vastitas Borealis interior unit. Its two subunits, rupēs and cavi, represent records of polar geologic and climatic processes across most of the Amazonian (~3.3 Ga). The cavi unit likely consists of alternating sand and ice sheet remnants of past polar caps, reflecting volatile–sedimentary interplay, while little is known about rupēs. Thanks to recent advances in radar data processing and dense coverage by the Mars Advanced Radar for Subsurface and Ionosphere Sounding (MARSIS), it is now possible to reconstruct the stratigraphy and composition of the BU, and reveal the enigmatic nature of the rupēs unit.We analyzed over 600 MARSIS profiles at 3, 4, and 5 MHz, leveraging optimized ionospheric corrections and deep penetration to map the full thickness of the BU and retrieve its frequency-dependent complex dielectric permittivity. We find that the rupēs unit spans the western half of PB and part of Olympia Planum as a continuous body beneath the cavi unit with a pole-facing upper unconformity, occupying ~191,000 km³ (~53% of BU volume). Dielectric inversions yield a real permittivity ε’ = 4.0±0.8 (consistent at all frequencies) and a frequency-dependent loss tangent tanδ = 0.017±0.006 (3 MHz) to 0.012±0.006 (5 MHz). Both components of the dielectric permittivity exhibit strong spatial heterogeneity, with values increasing toward Hyperborea Lingula (ε’ > 6, tanδ > 0.02).These results indicate that the rupēs composition differs substantially from that of the cavi unit, with large loss tangent values indicating the presence of significant amounts of lithic materials despite the low real permittivity. Basalt alteration products with tanδ > 0.02 are required to explain the high loss tangent measurements, while their strong frequency-dependence matches the water ice imaginary permittivity behavior. We find a best match of real dielectric permittivity and loss tangent results using a mixture of 85-90% water ice and 10-15% basalt alteration products like hydrated sulfates (e.g., gypsum), clays, and ferric oxides, which are supported by spectroscopic detections at visible exposures. Rupēs lithic materials may have been transported from lower latitude sources, where aqueous alteration is more viable than at polar latitudes. However, the strong spatial heterogeneities suggest that significant localized alteration occurred in situ during the Amazonian period, perhaps facilitated by warmer high-obliquity periods predicted to occur during the last 3 Gyr. Regardless of their source, the volume of these materials corresponds to a 24 cm–thick global layer, indicating that the rupes unit constitutes a substantial sediment reservoir, not merely one of water ice. Finally, the high loss tangent measured in Hyperborea Lingula explains the lack of rupēs basal detections by SHARAD despite the relatively low thickness (i.e., 150-200 m) of the rupēs unit at that location.
- Open Access
- 21 January 2026
- Published by Wiley
Revealing the internal structure of Mars-analog glaciers from drone-based radar sounding
Martian debris-covered glaciers (DCGs) contain large quantities of water ice beneath a protective layer of rock and dust. Properties of the overlying regolith such as density and depth to ice are critical parameters for guiding in-situ resource utilization (ISRU) of water and coring missions targeting potential habitats. Englacial debris layers that progressively outcrop at the surface could also provide access to enable shallow sampling of ice ages sequences. To assess the capabilities of future drone-based ground-penetrating radar for detecting supraglacial debris thickness and near-surface stratigraphy—properties not resolvable from orbit with the Shallow Radar (SHARAD) sounder—we conducted tests over two terrestrial analogs for Martian DCGs. Our platform consists of a DJI Matrice 600 Pro drone and a MALA Geodrone radar operating at a center frequency of 80 MHz. We detected a bulk glacier thickness of up to 28.5 m in Sourdough Rock Glacier, Alaska, and 48.6 m in Galena Creek Rock Glacier, Wyoming. We also resolved the supraglacial debris thickness, with a mean of 1.5 m in Sourdough, and in Galena Creek, a mean thickness of 0.8 m in extensional sections and 1.3 m in compressional sections. Furthermore, we detected layers outcropping at the surface in age sequences within the cirque of Galena Creek Rock Glacier. We validated the internal reflectors with clutter simulations to discriminate genuine subsurface reflectors from off-nadir surface reflection “clutter.” Finally, we provide recommendations for optimal survey design, including solutions to increase the signal-to-noise ratio and reduce clutter.
- 21 January 2026
- Published by Wiley
Open AccessRevealing the internal structure of Mars-analog glaciers from drone-based radar sounding
Martian debris-covered glaciers (DCGs) contain large quantities of water ice beneath a protective layer of rock and dust. Properties of the overlying regolith such as density and depth to ice are critical parameters for guiding in-situ resource utilization (ISRU) of water and coring missions targeting potential habitats. Englacial debris layers that progressively outcrop at the surface could also provide access to enable shallow sampling of ice ages sequences. To assess the capabilities of future drone-based ground-penetrating radar for detecting supraglacial debris thickness and near-surface stratigraphy—properties not resolvable from orbit with the Shallow Radar (SHARAD) sounder—we conducted tests over two terrestrial analogs for Martian DCGs. Our platform consists of a DJI Matrice 600 Pro drone and a MALA Geodrone radar operating at a center frequency of 80 MHz. We detected a bulk glacier thickness of up to 28.5 m in Sourdough Rock Glacier, Alaska, and 48.6 m in Galena Creek Rock Glacier, Wyoming. We also resolved the supraglacial debris thickness, with a mean of 1.5 m in Sourdough, and in Galena Creek, a mean thickness of 0.8 m in extensional sections and 1.3 m in compressional sections. Furthermore, we detected layers outcropping at the surface in age sequences within the cirque of Galena Creek Rock Glacier. We validated the internal reflectors with clutter simulations to discriminate genuine subsurface reflectors from off-nadir surface reflection “clutter.” Finally, we provide recommendations for optimal survey design, including solutions to increase the signal-to-noise ratio and reduce clutter.
- Open Access
- 1 January 2026
- Published by Wiley
Revealing the Internal Structure of the Rock Glacier - Debris Covered Glacier Continuum from Drone-Based Ground-Penetrating Radar
Rock glaciers (RGs) and Debris-Covered Glaciers (DCGs) are end-members of a continuum of ice-rich landforms found in alpine and polar environments. Although often considered more resilient to climate warming than debris-free glaciers, recent studies have shown that RG-DCGs also exhibit accelerated movement and degradation under warming conditions. Understanding these responses requires surface monitoring and subsurface characterization, the latter commonly achieved with ground-penetrating radar (GPR), a geophysical method capable of revealing the internal stratigraphy of these glaciers. In particular, GPR allows us to quantify the supraglacial debris thickness, total landform thickness, ice purity, and the presence of englacial stratigraphy. However, traditional surface-based GPR can be time-consuming and often hazardous, requiring manual traversal across rough, steep, and icy terrain—some of which may be inaccessible. To overcome these limitations, we integrated a drone-based GPR (DGPR) that consists of a MALA Geodrone 80 GPR mounted on a DJI Matrice 600 Pro, with a UgCS SkyHub system for automated terrain following. We tested our system at two ice-cored RGs: Sourdough, Alaska, and Galena Creek, Wyoming. At both sites, we compare the 80 MHz DGPR with a surface-based GPR pulseEKKO system at different center frequencies (50 MHz, 100 MHz, and 200 MHz), providing a range of vertical resolutions. We validated the internal reflectors with clutter simulations to discriminate genuine subsurface reflectors from off-nadir surface reflections (see figure below). DGPR resolved a maximum bulk glacier thickness of 28.5 m at Sourdough and 48.6 m at Galena Creek. Mean supraglacial debris thickness in Sourdough was 1.5 m, At Galena Creek, we found a mean debris thickness of 0.8 m in extensional sections and 1.4 m in compressional sections, which result from glacial flow stress-strain spatial patterns. We also detected internal debris layers outcropping at the cirque of Galena Creek. We estimated ages of ice and debris deposition by combining the outcrop locations with surface velocity data. We demonstrated that DGPR provides a safe and efficient method that enables more extensive and denser surveys than traditional GPR, improving the characterization of RG-DCGs as dynamic components of the cryosphere.
- 1 January 2026
- Published by Wiley
Open AccessRevealing the Internal Structure of the Rock Glacier - Debris Covered Glacier Continuum from Drone-Based Ground-Penetrating Radar
Rock glaciers (RGs) and Debris-Covered Glaciers (DCGs) are end-members of a continuum of ice-rich landforms found in alpine and polar environments. Although often considered more resilient to climate warming than debris-free glaciers, recent studies have shown that RG-DCGs also exhibit accelerated movement and degradation under warming conditions. Understanding these responses requires surface monitoring and subsurface characterization, the latter commonly achieved with ground-penetrating radar (GPR), a geophysical method capable of revealing the internal stratigraphy of these glaciers. In particular, GPR allows us to quantify the supraglacial debris thickness, total landform thickness, ice purity, and the presence of englacial stratigraphy. However, traditional surface-based GPR can be time-consuming and often hazardous, requiring manual traversal across rough, steep, and icy terrain—some of which may be inaccessible. To overcome these limitations, we integrated a drone-based GPR (DGPR) that consists of a MALA Geodrone 80 GPR mounted on a DJI Matrice 600 Pro, with a UgCS SkyHub system for automated terrain following. We tested our system at two ice-cored RGs: Sourdough, Alaska, and Galena Creek, Wyoming. At both sites, we compare the 80 MHz DGPR with a surface-based GPR pulseEKKO system at different center frequencies (50 MHz, 100 MHz, and 200 MHz), providing a range of vertical resolutions. We validated the internal reflectors with clutter simulations to discriminate genuine subsurface reflectors from off-nadir surface reflections (see figure below). DGPR resolved a maximum bulk glacier thickness of 28.5 m at Sourdough and 48.6 m at Galena Creek. Mean supraglacial debris thickness in Sourdough was 1.5 m, At Galena Creek, we found a mean debris thickness of 0.8 m in extensional sections and 1.4 m in compressional sections, which result from glacial flow stress-strain spatial patterns. We also detected internal debris layers outcropping at the cirque of Galena Creek. We estimated ages of ice and debris deposition by combining the outcrop locations with surface velocity data. We demonstrated that DGPR provides a safe and efficient method that enables more extensive and denser surveys than traditional GPR, improving the characterization of RG-DCGs as dynamic components of the cryosphere.
- Open Access
- 1 December 2025
- Published by American Astronomical Society in The Planetary Science Journal
- Vol. 6 (12) , 307
Crater-based Dating of Hebrus Valles and Hephaestus Fossae, Mars
Hebrus Valles (HV) and Hephaestus Fossae (HF) are connected outflow channel systems in Utopia Planitia, Mars. Previous investigations concluded that these features were carved by liquid water into a tectonically fractured terrain in the Late Hesperian or Early Amazonian (>3 Ga). However, their formation ages were poorly constrained. We use crater-based surface dating to determine model ages of the Utopia Planitia geologic units incised by HV and HF. For these units, we obtain model formation ages of μ 3 . 6 − 0.07 + 0.06 – μ 3 . 1 − 0.3 + 0.2 Ga and model resurfacing ages of μ 2 . 1 − 0.2 + 0.2 – μ 1.8 − 0.1 + 0.2 Ga, where μ represents uncertainties in absolute model age calibration. Using the buffered crater counting method, we also obtain model ages of μ 2 . 2 − 0.4 + 0.4 Ga and μ 1 . 9 − 0.3 + 0.4 Ga for the channels, coinciding with the peak of volcanic activity in the neighboring Elysium province. This confirms that Mars had geologically recent (Amazonian period), volcanically driven outflow channel activity, with implications for climate and habitability.
- 1 December 2025
- Published by American Astronomical Society in The Planetary Science Journal
- Vol. 6 (12) , 307
Open AccessCrater-based Dating of Hebrus Valles and Hephaestus Fossae, Mars
Hebrus Valles (HV) and Hephaestus Fossae (HF) are connected outflow channel systems in Utopia Planitia, Mars. Previous investigations concluded that these features were carved by liquid water into a tectonically fractured terrain in the Late Hesperian or Early Amazonian (>3 Ga). However, their formation ages were poorly constrained. We use crater-based surface dating to determine model ages of the Utopia Planitia geologic units incised by HV and HF. For these units, we obtain model formation ages of μ 3 . 6 − 0.07 + 0.06 – μ 3 . 1 − 0.3 + 0.2 Ga and model resurfacing ages of μ 2 . 1 − 0.2 + 0.2 – μ 1.8 − 0.1 + 0.2 Ga, where μ represents uncertainties in absolute model age calibration. Using the buffered crater counting method, we also obtain model ages of μ 2 . 2 − 0.4 + 0.4 Ga and μ 1 . 9 − 0.3 + 0.4 Ga for the channels, coinciding with the peak of volcanic activity in the neighboring Elysium province. This confirms that Mars had geologically recent (Amazonian period), volcanically driven outflow channel activity, with implications for climate and habitability.
- Open Access
- 9 July 2025
- Published by Copernicus GmbH
Advanced Radar Sounding and Imaging on SmallSat and CubeSat Missions Across the Solar System with Vector Sensor Antennas
Introduction: Orbital radar sounding and imaging are unique tools in planetary studies, with an extensive array of possible targets. Flight-proven instruments are currently limited to a few examples based on conventional technologies with inherent limitations in terms of mass/power/volume and performance. For example, current and planned spaceborne radar sounders use single dipole antennas for both transmit and receive, which are physically large and linearly polarized in just one direction, precluding studies based on polarization, such as ice detection [1]. They also have broad radiation patterns and are susceptible to off-nadir reflections from the surface (clutter) that can be confused with a nadir subsurface echo since there is no ability to determine the source direction of the echo [2, 3].Given the wide range of scientific objectives across our solar system that can addressed with radar sounding, plus the increased emphasis on low-cost, small, low mass, and low power consumption missions with flexible launch options, there is a need for radar sounder technology that can deployed in a CubeSat/SmallSat platform, or as part of an instrument suite onboard a larger spacecraft but without dominating the payload. Such a radar must provide fully polarimetric information for the characterization of materials including their scattering properties and have directivity to determine the source of echoes.Vector Sensor Antenna: The Vector Sensor Antenna (VSA, Fig. 1) is a transformative solution for radar sounding with the potential to address nearly all limitations of current technology. The VSA measures both the electric and magnetic field vectors of an electromagnetic wave at a common phase center [4], enabling the determination of polarization and angle-of-arrival for multiple discrete sources in a small payload of 2U while stowed.Leveraging heritage from NASA’s AERO-VISTA mission to develop and launch a CubeSat VSA for passive (receive-only) study of the aurora and other galactic radio sources [5], this technology is now sufficiently mature to develop an active VSA-based radar capable of addressing a wide range of planetary science objectives.Figure 1: VSA-based sounding radars have full polarization for ice detection and can determine echo direction of arrival for clutter discrimination, in a compact 6-8U, ~6 kg package. The crossed dipoles serve to transmit either linear or circular polarization while all elements operate in receive mode.ACORN: The Advanced Compact Orbiting Radar for luNar sounding (ACORN) will be based on a 2x2x1 m deployable VSA (Fig. 1) and will operate at 40-70 MHz in low lunar orbit. The ACORN VSA and all radar subsystems are compact and deployable from a 6U volume, enabling the possibility of a complete radar in a very small package. It is a novel approach to providing spatial resolution and penetration (Fig. 2) that provides a new alternative to distributed arrays of elements or large directive apertures.It has applications to addressing science and exploration priorities that have been set forth by the Planetary Science Decadal Survey, the Lunar Exploration Analysis Group, and the Artemis Science Definition Team. These objectives include (1) the distribution of ice to several meters depth, (2) the depth and structure of regolith, (3) the detection and characterization of lava tubes and stratigraphy of volcanic deposits, and (4) three-dimensional geophysical context for landed and human missions with in-situ instrumentation. Figure 2: (a) On the Moon, ACORN can penetrate through all regolith types and ice up to several km in thickness [e.g., 6, 7], and tens to hundreds of meters of high loss maria basalts [e.g., 8]. The penetration depth was estimated as a function of loss tangent and real dielectric permittivity based on the methods of [9, 10] and VSA radar performance at 100 km altitude (panel b). WHISPR: The Wideband Hf Ice and Subsurface Penetrating Radar (WHISPR) builds on the 4x4x2 m deployable VSA of AERO-VISTA to perform radar sounding at a 10-15 MHz frequency range from a compact 8U volume compatible with SmallSat platforms. This very small form factor, mass (6 kg), and power requirements (
- 9 July 2025
- Published by Copernicus GmbH
Open AccessAdvanced Radar Sounding and Imaging on SmallSat and CubeSat Missions Across the Solar System with Vector Sensor Antennas
Introduction: Orbital radar sounding and imaging are unique tools in planetary studies, with an extensive array of possible targets. Flight-proven instruments are currently limited to a few examples based on conventional technologies with inherent limitations in terms of mass/power/volume and performance. For example, current and planned spaceborne radar sounders use single dipole antennas for both transmit and receive, which are physically large and linearly polarized in just one direction, precluding studies based on polarization, such as ice detection [1]. They also have broad radiation patterns and are susceptible to off-nadir reflections from the surface (clutter) that can be confused with a nadir subsurface echo since there is no ability to determine the source direction of the echo [2, 3].Given the wide range of scientific objectives across our solar system that can addressed with radar sounding, plus the increased emphasis on low-cost, small, low mass, and low power consumption missions with flexible launch options, there is a need for radar sounder technology that can deployed in a CubeSat/SmallSat platform, or as part of an instrument suite onboard a larger spacecraft but without dominating the payload. Such a radar must provide fully polarimetric information for the characterization of materials including their scattering properties and have directivity to determine the source of echoes.Vector Sensor Antenna: The Vector Sensor Antenna (VSA, Fig. 1) is a transformative solution for radar sounding with the potential to address nearly all limitations of current technology. The VSA measures both the electric and magnetic field vectors of an electromagnetic wave at a common phase center [4], enabling the determination of polarization and angle-of-arrival for multiple discrete sources in a small payload of 2U while stowed.Leveraging heritage from NASA’s AERO-VISTA mission to develop and launch a CubeSat VSA for passive (receive-only) study of the aurora and other galactic radio sources [5], this technology is now sufficiently mature to develop an active VSA-based radar capable of addressing a wide range of planetary science objectives.Figure 1: VSA-based sounding radars have full polarization for ice detection and can determine echo direction of arrival for clutter discrimination, in a compact 6-8U, ~6 kg package. The crossed dipoles serve to transmit either linear or circular polarization while all elements operate in receive mode.ACORN: The Advanced Compact Orbiting Radar for luNar sounding (ACORN) will be based on a 2x2x1 m deployable VSA (Fig. 1) and will operate at 40-70 MHz in low lunar orbit. The ACORN VSA and all radar subsystems are compact and deployable from a 6U volume, enabling the possibility of a complete radar in a very small package. It is a novel approach to providing spatial resolution and penetration (Fig. 2) that provides a new alternative to distributed arrays of elements or large directive apertures.It has applications to addressing science and exploration priorities that have been set forth by the Planetary Science Decadal Survey, the Lunar Exploration Analysis Group, and the Artemis Science Definition Team. These objectives include (1) the distribution of ice to several meters depth, (2) the depth and structure of regolith, (3) the detection and characterization of lava tubes and stratigraphy of volcanic deposits, and (4) three-dimensional geophysical context for landed and human missions with in-situ instrumentation. Figure 2: (a) On the Moon, ACORN can penetrate through all regolith types and ice up to several km in thickness [e.g., 6, 7], and tens to hundreds of meters of high loss maria basalts [e.g., 8]. The penetration depth was estimated as a function of loss tangent and real dielectric permittivity based on the methods of [9, 10] and VSA radar performance at 100 km altitude (panel b). WHISPR: The Wideband Hf Ice and Subsurface Penetrating Radar (WHISPR) builds on the 4x4x2 m deployable VSA of AERO-VISTA to perform radar sounding at a 10-15 MHz frequency range from a compact 8U volume compatible with SmallSat platforms. This very small form factor, mass (6 kg), and power requirements (
- Open Access
- 9 July 2025
- Published by Copernicus GmbH
Measuring the Mars Global Surface Reflectivity at 1.3-5.5 MHz with MARSIS
Introduction: One of the main goals of Mars science and exploration is to explore the shallow subsurface (i.e., 10s-100s meters) and reveal its geologic nature and composition [1]. Although many remote sensing techniques can probe the subsurface, there is currently a sensing gap at depths between ~1 and 100s of meters. Radar surface reflectivity analyses can fill this gap thanks to sensitivity to changes in composition and geologic structures at ~5-15 m depth with the Shallow Radar (SHARAD, [2, 3]) and ~50-150 m depth with the Mars Advanced Radar for Subsurface and Ionosphere Sounding (MARSIS, [4, 5]). Historically, MARSIS surface reflectivity analyses were limited by insufficient coverage and ionospheric effects, which caused attenuation and distortion of the radar echoes. Thanks to new processing techniques and nearly global coverage after continuing MARSIS observations for nearly 20 years [6], it is now possible to probe the Martian near-subsurface at four distinct frequencies ranging from 1.3 MHz to 5.5 MHz.Methods: We employ a recently released PDS MARSIS dataset that corrects ionospheric phase distortions and provides three empirical parameters associated with ionospheric total electron content (TEC), in turn related to radar echo delay and attenuation [7, 8]. MARSIS operates at four distinct frequency bands, centered at 1.8, 3, 4, and 5 MHz, each with 1 MHz bandwidth, resulting in a free space vertical resolution of 150 m. This allows us to construct four reflectivity maps, each corresponding to a MARSIS band and probing to larger depths as frequency decreases. We process the surface echo power for each frequency through the following steps: Automatic picking of surface echoes based on global topography and “clutter” simulations [9]. Path loss correction to account for surface echo power changes due to spacecraft altitude. Ionospheric loss correction via normalization of surface power with respect to solar zenith angle (SZA) and the three TEC parameters (example in Fig. 1). Removal of surface geometry effects via normalization of surface power with respect to slope measured from the Mars Orbiter Laser Altimeter (MOLA, [10]) 463 m/pixel global DEM and surface roughness calculated at 0.6 km, 2.4 km, and 19.2 km baselines from [11] (example in Fig. 2). Finally, we generate surface reflectivity maps for each MARSIS band by gridding corrected power values at 5 km/pixel resolution and accounting for the variable footprint of each echo, measured as the radius of the 1st Fresnel zone.Figure 1: Example of ionospheric effect corrections based on SZA (top) and TEC1 parameter (bottom) applied to the 3 MHz surface echo power data. Figure 2: Example of surface geometry corrections based on slope (463 m/pixel) and roughness (at 0.6 km baseline, [11]) applied to the 3 MHz surface echo power data. Results: We find a strong surface echo power dependence on the local SZA (Fig. 1), with lower frequencies experiencing loss at higher SZA values, an expected behavior for MARSIS [5, 8]. After applying the SZA correction, we still find a non-linear dependence on all TEC parameters (Fig. 1), showing that up to 3-4 dB of loss occurs even at low TEC values. There is also a significant dependence on slope and 0.6 km baseline roughness (Fig. 2), suggesting that MARSIS echo power is most susceptible to short wavelength roughness.Figure 3: Preliminary MARSIS surface reflectivity maps at the 3 and 4 MHz bands, plotted on MOLA shaded relief. After empirical correction, the preliminary surface reflectivity maps (examples in Fig. 3) reveal several prominent features, such as high surface reflectivity associated with lava flows in Tharsis and the Elysium/Amazonis Planitiae, and low reflectivity zones corresponding to the friable Medusae Fossae Formation. The Martian northern plains reveal a wide range of reflectivity likely associated with variable terrain types including ice-rich deposits [12, 13] and the latitude-dependent mantle [14].Discussion: The consistency of surface power measured across adjacent profiles in our preliminary maps indicates that the correction of path and ionospheric losses is effective at all SZA values, thus eliminating the need for filtering. Despite the low number of observations and significant ionospheric losses, we also obtained the first Mars surface reflectivity map at 1.3-2.3 MHz.There is a weak residual surface power dependence on roughness, likely due to simplifications in our preliminary approach. To address this, we are currently running a detailed analysis of several roughness and slope parameters within the 1st Fresnel zone of each MARSIS echo, leveraging the entire MOLA shot point dataset [10]. This will allow us to capture the effects of short wavelength roughness highlighted by the preliminary results at 0.6 km baseline (Fig. 2).All reflectivity maps show enigmatic “sinuous stripes” across the southern highlands with low power values compared to surrounding areas. These features are consistent across adjacent profiles and their location and shape resemble ionospheric anomalies associated with remnant crustal magnetization observed by SHARAD [15]. A simple empirical correction based on the total crustal magnetic field [17] appears to remove most of these features, suggesting that MARSIS is particularly sensitive to ionospheric phenomena linked to crustal magnetization. We will further test this hypothesis via analyses of TEC parameters and magnetic anomaly data.Overall, these results offer new insights into volcanic, sedimentary, and ice-related processes. They also lay the groundwork for future missions targeting subsurface ice or landing sites.Acknowledgements: This work was supported by NASA MDAP grant 80NSSC22K1079.References: [1] MEPAG Science Goals document. [2] Campbell et al. (2013)JGR: Planets. [3] Seu et al. (2007)JGR: Planets. [4] Mouginot et al. (2010)Icarus. [5] Jordan et al. (2009)PSS. [6] Orosei et al. (2015)PSS. [7] McMichael et...
- 9 July 2025
- Published by Copernicus GmbH
Open AccessMeasuring the Mars Global Surface Reflectivity at 1.3-5.5 MHz with MARSIS
Introduction: One of the main goals of Mars science and exploration is to explore the shallow subsurface (i.e., 10s-100s meters) and reveal its geologic nature and composition [1]. Although many remote sensing techniques can probe the subsurface, there is currently a sensing gap at depths between ~1 and 100s of meters. Radar surface reflectivity analyses can fill this gap thanks to sensitivity to changes in composition and geologic structures at ~5-15 m depth with the Shallow Radar (SHARAD, [2, 3]) and ~50-150 m depth with the Mars Advanced Radar for Subsurface and Ionosphere Sounding (MARSIS, [4, 5]). Historically, MARSIS surface reflectivity analyses were limited by insufficient coverage and ionospheric effects, which caused attenuation and distortion of the radar echoes. Thanks to new processing techniques and nearly global coverage after continuing MARSIS observations for nearly 20 years [6], it is now possible to probe the Martian near-subsurface at four distinct frequencies ranging from 1.3 MHz to 5.5 MHz.Methods: We employ a recently released PDS MARSIS dataset that corrects ionospheric phase distortions and provides three empirical parameters associated with ionospheric total electron content (TEC), in turn related to radar echo delay and attenuation [7, 8]. MARSIS operates at four distinct frequency bands, centered at 1.8, 3, 4, and 5 MHz, each with 1 MHz bandwidth, resulting in a free space vertical resolution of 150 m. This allows us to construct four reflectivity maps, each corresponding to a MARSIS band and probing to larger depths as frequency decreases. We process the surface echo power for each frequency through the following steps: Automatic picking of surface echoes based on global topography and “clutter” simulations [9]. Path loss correction to account for surface echo power changes due to spacecraft altitude. Ionospheric loss correction via normalization of surface power with respect to solar zenith angle (SZA) and the three TEC parameters (example in Fig. 1). Removal of surface geometry effects via normalization of surface power with respect to slope measured from the Mars Orbiter Laser Altimeter (MOLA, [10]) 463 m/pixel global DEM and surface roughness calculated at 0.6 km, 2.4 km, and 19.2 km baselines from [11] (example in Fig. 2). Finally, we generate surface reflectivity maps for each MARSIS band by gridding corrected power values at 5 km/pixel resolution and accounting for the variable footprint of each echo, measured as the radius of the 1st Fresnel zone.Figure 1: Example of ionospheric effect corrections based on SZA (top) and TEC1 parameter (bottom) applied to the 3 MHz surface echo power data. Figure 2: Example of surface geometry corrections based on slope (463 m/pixel) and roughness (at 0.6 km baseline, [11]) applied to the 3 MHz surface echo power data. Results: We find a strong surface echo power dependence on the local SZA (Fig. 1), with lower frequencies experiencing loss at higher SZA values, an expected behavior for MARSIS [5, 8]. After applying the SZA correction, we still find a non-linear dependence on all TEC parameters (Fig. 1), showing that up to 3-4 dB of loss occurs even at low TEC values. There is also a significant dependence on slope and 0.6 km baseline roughness (Fig. 2), suggesting that MARSIS echo power is most susceptible to short wavelength roughness.Figure 3: Preliminary MARSIS surface reflectivity maps at the 3 and 4 MHz bands, plotted on MOLA shaded relief. After empirical correction, the preliminary surface reflectivity maps (examples in Fig. 3) reveal several prominent features, such as high surface reflectivity associated with lava flows in Tharsis and the Elysium/Amazonis Planitiae, and low reflectivity zones corresponding to the friable Medusae Fossae Formation. The Martian northern plains reveal a wide range of reflectivity likely associated with variable terrain types including ice-rich deposits [12, 13] and the latitude-dependent mantle [14].Discussion: The consistency of surface power measured across adjacent profiles in our preliminary maps indicates that the correction of path and ionospheric losses is effective at all SZA values, thus eliminating the need for filtering. Despite the low number of observations and significant ionospheric losses, we also obtained the first Mars surface reflectivity map at 1.3-2.3 MHz.There is a weak residual surface power dependence on roughness, likely due to simplifications in our preliminary approach. To address this, we are currently running a detailed analysis of several roughness and slope parameters within the 1st Fresnel zone of each MARSIS echo, leveraging the entire MOLA shot point dataset [10]. This will allow us to capture the effects of short wavelength roughness highlighted by the preliminary results at 0.6 km baseline (Fig. 2).All reflectivity maps show enigmatic “sinuous stripes” across the southern highlands with low power values compared to surrounding areas. These features are consistent across adjacent profiles and their location and shape resemble ionospheric anomalies associated with remnant crustal magnetization observed by SHARAD [15]. A simple empirical correction based on the total crustal magnetic field [17] appears to remove most of these features, suggesting that MARSIS is particularly sensitive to ionospheric phenomena linked to crustal magnetization. We will further test this hypothesis via analyses of TEC parameters and magnetic anomaly data.Overall, these results offer new insights into volcanic, sedimentary, and ice-related processes. They also lay the groundwork for future missions targeting subsurface ice or landing sites.Acknowledgements: This work was supported by NASA MDAP grant 80NSSC22K1079.References: [1] MEPAG Science Goals document. [2] Campbell et al. (2013)JGR: Planets. [3] Seu et al. (2007)JGR: Planets. [4] Mouginot et al. (2010)Icarus. [5] Jordan et al. (2009)PSS. [6] Orosei et al. (2015)PSS. [7] McMichael et...
- Open Access
- 9 July 2025
- Published by Copernicus GmbH
Validating internal reflectors from drone-based GPR in debris-covered glaciers with clutter simulations
Drone-based ground penetrating radar (DGPR) is a novel tool to retrieve internal properties of debris-covered glaciers, including supraglacial debris thickness, bulk thickness, and detection of englacial debris layers, demonstrating a unique potential of drones to study buried ice reservoirs on Mars [1, 2]. We present the results of surface clutter analysis as part of a project involving a GPR Geodrone 80, centered at 80 MHz, mounted on a DJI Matrice 600 drone.Since the DGPR platform is above the surface, identification of subsurface interfaces requires the simulation of off-nadir surface reflections (“clutter”) in order to rule out false positive detections from other sources. There is an inherent uncertainty in confirming internal reflectors in our study sites due to sloped surfaces, the presence of boulders, and the proximity to valley walls and trees (Figure 1). Radar surface clutter simulations are employed to model radar returns from off-nadir surface topography and have been successfully used to validate internal reflectors in glaciers in Antarctica [3] using airborne sounding radar, as well as over lobate debris aprons on Mars from orbital sounding radar [4, 5, 6].Figure 1. Diagram of the DGPR and the detection of subsurface and off-nadir targets. Clutter occurs when the reflection from the off-nadir object returns at a similar travel time to the englacial reflections. (a) A tree A has a return with similar delays to the bottom of the ice B. (b) An off-nadir surface boulder D has a similar return as a buried boulder C. (c) A headwall F has a return with a similar delay as an internal debris layer F.We used digital elevation models (DEMs) derived from drone photogrammetry (~5 cm/px) available for Galena Creek and Sourdough [7] to generate clutter simulations (“cluttergrams”) using state-of-the-art radar surface return simulation software [8. 9]. For each trace in the GPR profile, the clutter simulation software generates a faceted representation of the topography surrounding the drone when the trace was acquired (Figure 2a). Then, the expected reflection power is estimated for each facet using a modified version of the Friis transmission formula (Choudhary et al., 2016) and the two way travel time is calculated from the distance between the drone and the facet center. The reflected power estimates and travel times are assembled into a simulated radar trace for comparison to the GPR data. Given that the antennas of the MALA Geodrone 80 are dipoles (Figure 2b), we use a normalized dipole radiation pattern by multiplying to a factor of the azimuth angle (φ) as shown in equation 1 [10].Figure 2. (a) Setup of the faceted representation of the topography to generate the clutter simulations. The step size of the facets in both the along and across track is 0.5 m. The total length of the along and across track is based on the time window for each acquisition, in case of a window of 800 ns, the maximum distance to a facet is 100 m. (a) Normalized field patterns of a half-wave dipole antenna. (Visser, H.J., 2012). The red center arrow represents the angle . We validated three scenarios with the clutter simulator Although there are off-nadir reflectors coming from the forest, the glacier base is still visible in some sections of the profile (Figure 3). Dipping reflectors interpreted as englacial debris bands are not associated with clutter coming from the headwall (Figure 4). The reflector corresponding to a boulder is not obscuring the englacial reflectors (Figure 5). This reflector could have also been identified as clutter with hyperbola fitting, given that its value is close to the speed of light. Figure 3. DGPR radar profile (a) and its clutter simulation (b) in the lower section of Sourdough Rock Glacier. The base is a feature observed in the radar profile but not in the clutter simulation. (c) Drone-based orthomosaic with the nadir of the clutter simulation in red (d) Echomap with the first return of the clutter simulation in black. (e) Oblique photo of the DGPR takeoff with the forest in the background, location marked as E in panel c.Figure 4. Radar profile and clutter simulation at the cirque of Galena Creek. (a) Radar profile with multiple dipping reflectors interpreted as internal debris layers and basal reflectors. (b) Clutter simulation, no clutter was associated with the dipping reflectors or basal reflectors shown in panel a. (c) Hillshade with the nadir return. (d) Power map of the clutter simulation with the first return. (e) Orthomosaic. (f) Ground view with a section of the profile. Figure 5. Radar profile (a) and clutter simulation (b) of lunch rock in Galena Creek. (c) Hillshade, white arrow indicates the glacier flow direction. (d) Power map of the clutter simulation (e) Orthomosaic indicating the location of the weather station (magenta star) from where the photo in panel f was taken. (f) Ground view. The profile F-F’ in panels c, d, and f shows the first return of the clutter simulation. The yellow arrow indicates the boulder location in all panels. References[1] Aguilar et al. (2024) EPSC 2024, https://doi.org/10.5194/epsc2024‐1271. [2] Aguilar et al. (2025) LPSC 2025, https://www.hou.usra.edu/meetings/lpsc2025/pdf/1693.pdf. [3] Holt et al. (2006) JGR: Planets, https://doi.org/10.1029/2005JE002525. [4] Holt et al. (2008) Science, https://doi.org/10.1126/science.1164246. [5] Plaut et al. (2009) GRL, Https://doi.org/10.1029/2008GL036379. [6] Baker et al. (2019) Icarus, Https://doi.org/10.1016/j.icarus.2018.09.001. [7] Meng et al. (2023b). Journal of Glaciology, https://doi.org/10.1017/jog.2022.90. [8] Choudhary et al. (2016) IEEE GRSL, https://doi.org/10.1109/LGRS.2016.2581799. [9] Christoffersen et al. (2024) Zenodo, https://doi.org/10.5281/zenodo.10595007. [10] Visser (2012) Antenna Theory and Applications, https://doi.org/10.1002/9781119944751.ch5.
- 9 July 2025
- Published by Copernicus GmbH
Open AccessValidating internal reflectors from drone-based GPR in debris-covered glaciers with clutter simulations
Drone-based ground penetrating radar (DGPR) is a novel tool to retrieve internal properties of debris-covered glaciers, including supraglacial debris thickness, bulk thickness, and detection of englacial debris layers, demonstrating a unique potential of drones to study buried ice reservoirs on Mars [1, 2]. We present the results of surface clutter analysis as part of a project involving a GPR Geodrone 80, centered at 80 MHz, mounted on a DJI Matrice 600 drone.Since the DGPR platform is above the surface, identification of subsurface interfaces requires the simulation of off-nadir surface reflections (“clutter”) in order to rule out false positive detections from other sources. There is an inherent uncertainty in confirming internal reflectors in our study sites due to sloped surfaces, the presence of boulders, and the proximity to valley walls and trees (Figure 1). Radar surface clutter simulations are employed to model radar returns from off-nadir surface topography and have been successfully used to validate internal reflectors in glaciers in Antarctica [3] using airborne sounding radar, as well as over lobate debris aprons on Mars from orbital sounding radar [4, 5, 6].Figure 1. Diagram of the DGPR and the detection of subsurface and off-nadir targets. Clutter occurs when the reflection from the off-nadir object returns at a similar travel time to the englacial reflections. (a) A tree A has a return with similar delays to the bottom of the ice B. (b) An off-nadir surface boulder D has a similar return as a buried boulder C. (c) A headwall F has a return with a similar delay as an internal debris layer F.We used digital elevation models (DEMs) derived from drone photogrammetry (~5 cm/px) available for Galena Creek and Sourdough [7] to generate clutter simulations (“cluttergrams”) using state-of-the-art radar surface return simulation software [8. 9]. For each trace in the GPR profile, the clutter simulation software generates a faceted representation of the topography surrounding the drone when the trace was acquired (Figure 2a). Then, the expected reflection power is estimated for each facet using a modified version of the Friis transmission formula (Choudhary et al., 2016) and the two way travel time is calculated from the distance between the drone and the facet center. The reflected power estimates and travel times are assembled into a simulated radar trace for comparison to the GPR data. Given that the antennas of the MALA Geodrone 80 are dipoles (Figure 2b), we use a normalized dipole radiation pattern by multiplying to a factor of the azimuth angle (φ) as shown in equation 1 [10].Figure 2. (a) Setup of the faceted representation of the topography to generate the clutter simulations. The step size of the facets in both the along and across track is 0.5 m. The total length of the along and across track is based on the time window for each acquisition, in case of a window of 800 ns, the maximum distance to a facet is 100 m. (a) Normalized field patterns of a half-wave dipole antenna. (Visser, H.J., 2012). The red center arrow represents the angle . We validated three scenarios with the clutter simulator Although there are off-nadir reflectors coming from the forest, the glacier base is still visible in some sections of the profile (Figure 3). Dipping reflectors interpreted as englacial debris bands are not associated with clutter coming from the headwall (Figure 4). The reflector corresponding to a boulder is not obscuring the englacial reflectors (Figure 5). This reflector could have also been identified as clutter with hyperbola fitting, given that its value is close to the speed of light. Figure 3. DGPR radar profile (a) and its clutter simulation (b) in the lower section of Sourdough Rock Glacier. The base is a feature observed in the radar profile but not in the clutter simulation. (c) Drone-based orthomosaic with the nadir of the clutter simulation in red (d) Echomap with the first return of the clutter simulation in black. (e) Oblique photo of the DGPR takeoff with the forest in the background, location marked as E in panel c.Figure 4. Radar profile and clutter simulation at the cirque of Galena Creek. (a) Radar profile with multiple dipping reflectors interpreted as internal debris layers and basal reflectors. (b) Clutter simulation, no clutter was associated with the dipping reflectors or basal reflectors shown in panel a. (c) Hillshade with the nadir return. (d) Power map of the clutter simulation with the first return. (e) Orthomosaic. (f) Ground view with a section of the profile. Figure 5. Radar profile (a) and clutter simulation (b) of lunch rock in Galena Creek. (c) Hillshade, white arrow indicates the glacier flow direction. (d) Power map of the clutter simulation (e) Orthomosaic indicating the location of the weather station (magenta star) from where the photo in panel f was taken. (f) Ground view. The profile F-F’ in panels c, d, and f shows the first return of the clutter simulation. The yellow arrow indicates the boulder location in all panels. References[1] Aguilar et al. (2024) EPSC 2024, https://doi.org/10.5194/epsc2024‐1271. [2] Aguilar et al. (2025) LPSC 2025, https://www.hou.usra.edu/meetings/lpsc2025/pdf/1693.pdf. [3] Holt et al. (2006) JGR: Planets, https://doi.org/10.1029/2005JE002525. [4] Holt et al. (2008) Science, https://doi.org/10.1126/science.1164246. [5] Plaut et al. (2009) GRL, Https://doi.org/10.1029/2008GL036379. [6] Baker et al. (2019) Icarus, Https://doi.org/10.1016/j.icarus.2018.09.001. [7] Meng et al. (2023b). Journal of Glaciology, https://doi.org/10.1017/jog.2022.90. [8] Choudhary et al. (2016) IEEE GRSL, https://doi.org/10.1109/LGRS.2016.2581799. [9] Christoffersen et al. (2024) Zenodo, https://doi.org/10.5281/zenodo.10595007. [10] Visser (2012) Antenna Theory and Applications, https://doi.org/10.1002/9781119944751.ch5.