The Dielectric Properties of Martian Regolith at the Tianwen-1 Landing Site

Abstract

Mars' surface is characterized by a weathered layer of regolith and exposed rock exposures that are the results of long-term geological processes. The Mars Rover Penetrating Radar (RoPeR) on board the Zhurong rover of China's first Mars mission (Tianwen-1) has been investigating the fine structure and dielectric properties of the martian regolith in the southern Utopia Planitia. The permittivity of the regolith within 5 m of the landing zone is $$, and the average loss tangent is 0.0060 ± 0.0002, 0.0087 ± 0.0002, and 0.0114 ± 0.0002 using the geometric compensation of R², R³, R⁴, respectively. The permittivity distribution map has been derived to show permittivity varying with depth and location. The high dispersion of both permittivity and loss tangent values along the traverse path indicates relatively heterogeneous material distribution on the landing site compared to an airless body such as the Moon.

Key Points

• The dielectric properties of martian regolith at the Tianwen-1 landing site suggest that water ice is not the dominant component

• The radar image and permittivity and loss tangent distribution maps show a high heterogeneity of subsurface materials

• Compared to the lunar regolith, the martian regolith shows more diverse surface processes and weaker space weathering effects

Plain Language Summary

A layer of regolith covers the surface of Mars, which is the result of geologic processes that occurred over millions to billions of years. Compared to the observations from satellites, the Zhurong rover of China's first Mars mission (Tianwen-1) had a closer look at the properties of the regolith layer in the explored region within southern Utopia Planitia. There is evidence that the exposed materials might be related to aqueous activities. Local landforms on the surface suggest the possible presence of buried volatiles, like water ice. The radar instrument (RoPeR) on board the rover can expose subsurface structures and the dielectric properties of the regolith layer at high-resolution, to assess their composition. The loss tangent results suggest that water ice is not the main component of the local martian regolith at some depth. The scattering distribution of radar profile along the traveling path and heterogeneous subsurface features show more diverse surface processes and weaker space weathering effects on Mars than those on the airless Moon.

1 Introduction

On Mars, compared to an airless body such as the Moon, the weathering layer (“regolith”, a general term for the layer of fragmental and unconsolidated rock material, whether residual or transported and of highly varied character, that nearly everywhere forms the surface of the land and overlies or covers bedrock) underwent complicated geological processes in addition to weaker impact and space weathering modifications. A fuller understanding of the stratigraphy and properties of the martian regolith would unravel the local evolution history and help address key geological questions, including the potentiality of liquid water on the surface or near-surface (Christensen et al., 2008). The characterization of the dielectric properties of the weathering layer represents a key target of this quest.

Mars Advanced Radar for Subsurface and Ionospheric Sounding (MARSIS) and the Mars Express orbiter and the Shallow Radar (SHARAD) have revealed significant features beneath the surface and also obtained the subsurface dielectric characteristics to constrain the composition of the materials. A 3–5 MHz global permittivity map has been derived from MARSIS data, providing insights into the physical properties within the first ∼60–80 m below the surface (Mouginot et al., 2012). The permittivity of the first few meters of the martian regolith calculated by the SHARAD surface echoes shows a significant correspondence with the geological dichotomy: high permittivity (7–10) on the highland side but lower (3–7) on the lowland side (Castaldo et al., 2017). The loss tangent value inferred from radar data has also been used to infer the possible presence of water ice (e.g., Campbell et al., 2021; Campbell & Morgan, 2018). Though the MARSIS and the SHARAD can detect the reflections on the surface and subsurface down to a depth of hundreds of meters, they have a limited ability to discriminate the presence of internal structures in the shallow subsurface regolith.

On 15 May 2021, the lander from China's first martian mission, Tianwen-1 (TW-1) touched down in the southern part of the Utopia Planitia (109.925°E, 25.066°N) and west of the Elysium volcanic province. The martian rover (Zhurong) was deployed and drove southwards of the landing area (Figure 1a). The Mars Rover Penetrating Radar (RoPeR) on board the Zhurong rover enables in-situ measurement of the fine-scale subsurface structure and dielectric characteristics of subsurface materials using two different frequency channels. Channel one (CH1) operates in the frequency range between 15 to 95 MHz, and channel two (CH2) is 0.45 to 2.15 GHz. CH1 revealed two sets of weak-to-strong reflection variation patterns in the radar image, suggesting the occurrence of episodic hydraulic flooding sedimentation (Li et al., 2022). CH1 is designed to penetrate Martian materials to depths of 10–100 m with a resolution of a few meters. CH2 can probe the depth range between 3 and 10 m with a resolution in centimeters (Zhou et al., 2020).

Figure 1

Traverse path of the Zhurong rover and images taken by the NaTeCam. (a) Traverse path of the Zhurong rover of the first 191 martian days. (b)–(i) Images were taken by the NaTeCam. The image indexes of b-i are listed in Table S1 in Supporting Information S1. Panel (b) shows the enlarged view of the small pit created during the landing process. Green circles and arrows point to the impact craters in (c, e, g). Blue arrows indicate the dunes in (d) and (f). White arrows in (h) (i) mark the platy rocks.

In this study, we investigate the dielectric properties of the martian regolith of the landing area with the first 191 sols of RoPeR CH2 data. The permittivity and loss tangent distribution of the Martian regolith within 5 m depth has been estimated and analyzed. We compared the results with those obtained by the in-situ radar experiments performed on the Moon to discuss the potential effect of surface processes and atmosphere that might contribute to the differences in material dielectric properties.

2 Materials and Methods

2.1 Landing Site

TW-1 landed in a region near the dichotomy boundary separating the southern Noachian highlands and the younger northern lowlands within the Late Hesperian lowland unit (lHl), dated to ∼3.1–3.36 billion years ago (Ivanov et al., 2017; Stopar et al., 2014). The local surface age estimated with craters distributed in the neighboring area of the landing site is between ∼757 Ma and ∼1.12 Ga (Wu et al., 2022; Zhao et al., 2021). Hence, regional depositional/resurfacing processes might have occurred following the deposition of the widespread Vastitas Borealis Formation (VBF) materials in the lHl unit during the Amazonian period. The landforms in the landing area such as kilometer-scaled polygonal troughs, pitted cones, and rampart/pancake craters possibly indicative of the presence of subsurface water-ice (Ye et al., 2021; Zhao et al., 2021). Both the TES albedo (0.234) and the DCI (0.941 ± 0.011) show that the surface of the study area is covered by dust (Christensen et al., 2001; Ruff & Christensen, 2002). The permittivity of surface materials in the Utopia Planitia is ∼3–4, as derived from SHARAD (Castaldo et al., 2017), and ∼4–5 from MARSIS data (Mouginot et al., 2012). This disparity is due to different radar wavelengths being employed and corresponding detection depths.

2.2 RoPeR Data and Processing Procedures

We report the initial ∼1,200 m of radar results of RoPeR CH2 channel. The IDs for all radar data are listed in Table S2. The processing steps are similar to those for the Chang’E-3 and Chang’E-4 LPR data (Zhang et al., 2020, 2021) and are detailed below:

1. Convert data format. RoPeR data comply with the standard storage format in aviation and spaceflight fields (.PDS) (Tan et al., 2021). We convert the raw data into a binary format for subsequent processing, analysis, and imaging (Figure S1 in Supporting Information S1).

2. Delete self-check traces. During the data collection process, the RoPeR on board the Zhurong rover conducts a self-check on the performance of the instrument. These self-check traces affect the radar image. This step deletes the self-check traces (Figure S2 in Supporting Information S1).

3. Remove the background noise for the whole data. The strong reflection of the surface and the lateral noise of the instrument can have a detrimental effect on the data. This step removes these unwanted horizontal background signals by subtracting the average trace (Figure S3 in Supporting Information S1).

4. Band-pass filtering. The bandwidth of the transmitted signals of high-frequency data is limited to the range of 0.45–2.15 GHz (Figure S4 in Supporting Information S1).

5. Amplitude recovery. Compensate and correct the attenuation of radar electromagnetic wave energy (Figure S5 in Supporting Information S1).

6. Time-depth conversion (Figure S6 in Supporting Information S1). ε = 3.6 is used for time-to-depth conversion. Based on the derived permittivity range, the uncertainty of the converted depth can reach 30%.

After the above processing steps, the data of the first 191 solar days were processed, and the radar profile of more than 1,200 m was obtained (Figure 2a). The radar files with the appropriate X-Y ratio are shown in Figures S7–S9 in Supporting Information S1. The dielectric properties of the surveyed materials including the relative permittivity ε and the loss tangent tanδ can affect the propagation of radar waves. Hence, we can derive these two parameters from RoPeR data as described below.

Figure 2

(a) Radar image of the high-frequency channel of RoPER data. a-1 is the whole profile of the first 191 sols data. Panel (a-2, a-3 and a-4) show the details of some specific areas which show randomly distributed strong echos caused by clusters of scatters. (b) Permittivity distribution map beneath the exploration region of Zhurong rover. (c) Loss tangent distribution map along the traversing path. Each loss tangent data represents the average value of data traces collected every 10 m. ε = 3.6 is used for time-to-depth conversion.

2.3 Permittivity Calculation

The hyperbola fitting method is used to extract the average permittivity. The geometric model employed here considers the antenna height (Fa, 2020) (Figure S10 in Supporting Information S1). The calculation equations are:

$$ (1)

$$ (2)

where R is the one-way apparent range from the antenna to the target, h is the height of the antenna from the surface, d is the depth of the embedment, θ_i is the angle of incidence, θ is the emergence angle and x is the horizontal distance of the target. Around 100 hyperbola-shaped signals were extracted manually to estimate the average permittivity values at different depths based on Equations 1 and 2 (Table S3). The permittivity of the martian regolith within 5 m of the landing zone is $$. The permittivity ε distribution map is shown in Figure 2b.

2.4 Loss Tangent Calculation

The loss tangent tanδ, the ratio of the real part and the imaginary part of the dielectric permittivity, represents a key constraint on the composition of the targets (Campbell & Morgan, 2018). To analyze the subsurface material properties, we calculated the loss tangent of each unit (Robert et al., 2006). The loss tangent was calculated from the radar range equation and the relationship between loss tangent and attenuation (Robert et al., 2006; Zhang et al., 2022), a methodology that has been successfully applied in lunar loss tangent calculations relating to LPR data (Lai et al., 2019). The calculation procedures of the loss tangent are listed as follows:

1. Obtain the mean square of all data traces.

2. Apply model-dependent gain function, and normalized traces. Three models of the reflection targets should be considered: for geometric spreading and the backscatter cross-section jointly, P_r/P_t ∝ 1/R²; for reflections to be due to small (subwavelength) spheres, P_r/P_t ∝ 1/R⁴; for the intermediate case, P_r/P_t ∝ 1/R³ (Robert et al., 2006).

3. A suitable range would be selected, and then the attenuation follows simply from the least-squares fit of the two-way distance versus power. We follow the same criteria to select local data segments. The first is to avoid the part heavily affected by surface clutters and interface reflection to mitigate noises. The second is to select the segment with echo strength continuously decreasing with the depth.

4. According to the relationship between loss tangent and attenuation, the loss tangent is obtained from the attenuation.

$$ (3)

The estimated loss tangent of the martian regolith at the Tianwen-1 site is 0.0060 ± 0.0002, 0.0087 ± 0.0002, 0.0114 ± 0.0002 using average radar power (black line) with geometric compensation of R², R³, R⁴ (Figure 3a), respectively. R³ correction applied to the rough interface is assumed in this work as the material heterogeneity observed in Figure 1b. The horizontal variations of loss tangent value along the traversing path are shown in Figure 2c.

Figure 3

The loss tangent of the materials from the RoPeR. (a) The signal power profiles of the average trace with R², R³, and R⁴ backscatter/spreading corrections, respectively. The best-fit lines are used to calculate the attenuation η (dB/m). The selection of the fitting section was in the range of 2.5–4.5 m, depending on the requirements of the method and the validity of the data (Figure S12 in Supporting Information S1). (b) The estimated loss tangent value with uncertainties.

3 Results

The processed radargram (Figure 2a) shows no evident layered structure except for some strong reflections looking like vertical stripes, probably caused by the surface or subsurface scatters. Most of the echoes diminish around the depth of 5 m. The radargram with proper X-Y scale (Figures S7–S9 in Supporting Information S1) shows randomly distributed strong echoes caused by clusters of scatters. Some resemble a bowl or slope shape that could represent buried impact remnants. To aid the structural analysis, the permittivity ε distribution map (Figure 2b) was obtained by interpolating (the locally weighted scatter plot smooth method) the local ε values estimated with the 100 hyperbola signals (Table S3). The permittivity of the martian regolith within 5 m of the landing zone is $$ and varies significantly with the location. The ε map here only shows the distribution trend of the permittivity rather than the details of local changes because of the low density of sampling data and selected interpolation method. No layered structure such as a continuous interface of permittivity contrast is observed, as in Figure 2b, as per the radargram. The permittivity gradually increases with the depth.

The estimated loss tangent of the martian regolith at the Tianwen-1 site is 0.0060 ± 0.0002, 0.0087 ± 0.0002, 0.0114 ± 0.0002 using the average radar power (black line) with geometric compensation of R², R³, R⁴ (Figure 3a), respectively. Figures 1b–1i show that the martian regolith contains certain amount of small scattered spheres. So, the R³ correction is applied in this work as explained in Section 2.3. In the whole path, the loss tangent fluctuates greatly between 0.0017 and 0.0140 (R³ correction), especially compared to the CE-4 results (Lai et al., 2019). The reasons for such large fluctuation include the selection of local data segments and the differences in local cohesive materials content and their composition. Overall, the loss tangent value of 0.0087 ± 0.0027 is the average value of the Martian regolith.

According to the distribution of relative permittivity ε along with the depth z (Figure 4a), the average ε ranges between 1.6 and 6.7 within the effective detection depth of 4.5 m. It varies with location. Based on the empirical formula built from lunar regolith samples, two fitting Equations 4 and 5 are employed (Carrier, 1991), where z is the depth.

$$ (4)

$$ (5)

Figure 4

Permittivity, density, and loss tangent of the materials from the RoPeR. (a) The permittivity value is calculated from the shape of the hyperbola in the RoPeR image. According to the relationship between the permittivity value and the depth, fitting curves of different forms are fitted. (b) The relationship between density and depth is fitted from the permittivity value in Figure (a). (c) The relationship between the content of TiO₂ + FeO, density and loss tangent. The density of the martian regolith in the Tianwen-1 landing area fluctuates between 1.2 and 2.2 g/cm³ which can be found from Density-Form2 in panel (b). The loss tangent ranges from 0.0060 to 0.0114, according to different correction factors, of which 0.0087 is the loss tangent value using the R³ correction. Based on the empirical formula of lunar regolith, we obtained the TiO₂ + FeO of martian regolith, ranging between 13 wt.% and 22 wt.%. (d) Comparison of density versus depth results. The results are from the ground-penetrating radar data of TW-1, CE-3 (Fa, 2020), the CE-4 (Lai et al., 2019) landing sites, and the lunar regolith Apollo missions samples (Carrier, 1991).

Figure 4a shows the two forms of permittivity-depth profiles, the fitting lines of which have similar χ² per degree of freedom. The average permittivity in Figure 4a relates to the permittivity value averaging within the depth range of z, whereas local permittivity represents the permittivity value at the depth z. “Fitted Average Permittivity-Form 1” in Figure 4 is obtained by the fitted line using the model depicted in Equation 4, whereas “Local Permittivity-Form 1” is generated with Equation 4. The ε is generally closely related to the bulk density ρ (Carrier, 1991). Based on the distribution of ε (Figure 4a) and the relationship between permittivity and density (Equation 6):

$$ (6)

we obtained the density profile of the martian regolith at the Tianwen-1 landing site (Figure 4b).

The lunar regolith samples show that the loss tangent value is closely related to the iron and titanium content (Carrier, 1991), as shown in Figure 4c. Using empirical formulas derived from the measurements of lunar samples, the calculated density, and loss tangent values, we estimate that the content of iron and titanium in the regolith at 17.5 ± 4.5 wt% at the Tianwen-1 landing site. The remote sensing measurements of FeO abundances at the Tianwen-1 landing site is 18%, and that of the nearby Elysium Planitia is ∼21% (Boynton et al., 2007).

4 Discussion

4.1 Implications on Material Composition at the TW-1 Landing Site

The surface layer across a large area within the northern Utopia Planitia contains a high percentage (estimated at 50%–85% by volume) of water ice (Stuurman et al., 2016). The loss tangent value of the lobate debris apron (LDA), as measured by the SHARAD is 0.002 ± 0.0008, which is dominated by water ice, whereas the Elysium Planitia area in the east of the Tianwen-1 landing site, which has been resurfaced by lava flows, has a loss tangent value of 0.022 ± 0.011 (Campbell & Morgan, 2018). Data from the RoPeR gives the loss tangent value at the Tianwen-1 landing site of 0.0087 ± 0.0027 that sits, between the values of the ice-rich and the basalt units. Considering the local surface temperature, atmosphere pressure and the loss tangent value, the regolith at the Tianwen-1 landing site is unlikely to be dominated by water ice (Zhao et al., 2021). The derived local surface age is younger than the geologic unit age (e.g., Wu et al., 2022; Zhao et al., 2021), suggesting that local resurfacing events might have occurred at the landing site. However, current evidence of a resurfacing event and whether the surface material is either igneous or sedimentary is circumscribed. Local elemental compositions derived from the Laser Induced Breakdown Spectrometer (LIBS) data show that the spectrum of the exposed martian regolith and rocks are closer to the onboard igneous samples used as calibration targets (Y. Liu et al., 2022). Given that the loss tangent is affected by minerology, but also by density, the loss tangent of regolith should be lower than lava flows such as those in the Elysium Planitia area.

The short-wave infrared spectrometer onboard the Zhurong rover, operating in the spectrum range between 0.8 and 2.4 μm across 321 channels, discovered that the landing site may contain a considerable amount of hydrated minerals (Y. Liu et al., 2022). The water-altered materials increase heterogeneity and lead to a stronger attenuation of radar signals (Stillman & Grimm, 2011). However, the iron and titanium derived from loss tangent value using the empirical formulas of lunar basalt samples is comparable with the remote sensing measurements of the FeO/ilmenite abundances at the landing site, suggesting the fraction of hydrated minerals to be minor. Consequently, the loss tangent value estimated from radar data fits the estimations based on the Mars Surface Composition Detector (MarSCoDe) data, which also support the hypothesis that the local martian regolith might be derived from igneous rocks and some might have experienced a low degree of chemical alteration by aqueous activities. However, sedimentary materials also have relatively similar loss tangent (Williams & Greeley, 2004) and permittivity values ranging in ∼2–6 (Meng et al., 2020), hence, the RoPeR could not exclude this possibility; additional infrared spectral data or other sources would help clarify this issue.

4.2 Comparison With Regolith at Other Martian Landing Sites

The Radar Imager for Mars' Subsurface Experiment (RIMFAX) (150–1,200 MHz) aboard the Perseverance rover (Hamran et al., 2020) surveyed the Jezero crater (Hamran et al., 2022). The average permittivity detected by the RIMFAX on Perseverance in the Jezero Crater is 9.0 ± 2.8, with a mean depth of 1.9 ± 1.1 m (Casademont et al., 2022). The observed slopes, thicknesses, and internal morphology of the inclined stratigraphic sections by RIMFAX can be interpreted either as a magmatic layering formed in a differentiated igneous body or as a sedimentary layering commonly formed in aqueous environments on Earth (Hamran et al., 2022). Rocks of various sizes can be seen on the surface, which accounts for a very high proportion of the weathering layer and generates a high permittivity value. However, the average value of the permittivity at the Tianwen-1 landing zone is much lower (Figure 4) indicating either different material composition or different degrees of weathering. It is consistent with the inversed permittivity value derived from the SHARAD data (Castaldo et al., 2017).

Considering the close relationship between permittivity and bulk density, the density of the martian regolith at the Tianwen-1 landing area is in the range of 2.0 ± 1.0 g/cm³, which is also much lower than the value of 3.2 g/cm³ obtained by the RIMFAX in the Jezero Crater area (Casademont et al., 2022). Data from the Viking-1 mission produced a regolith density of about 1–1.6 g/cm³ through multiple field measurements (Shorthill et al., 1976). The Microwave Radiometer on Mars-3 and the Radio Telescope on Mars-5 radars operating at the wavelengths of 3.4 cm measured the surface density of the regolith to 1.37 ± 0.33 g/cm², which is consistent with the inferred density of the surface layer (within a depth of 0.2 m) measured by the RoPeR, specifically the fitting curve given in form-2 (Equation 5).

4.3 Comparison With Regolith at CE-3 and CE-4 Lunar Landing Sites

Compared to the permittivity at the shallow surface in the Perseverance landing area, the Tianwen-1 is closer to the permittivity range of the lunar regolith as measured by the Chang'e-3 and Chang'e-4 ground-penetrating radars (LPRs) (Dong et al., 2020; Feng et al., 2017; Lai et al., 2019). The ε ranges from $$ within the first ∼4.5 m in depth, a somehow similar value to the Chang'e-3 results of 2.9 ± 0.4 (Dong et al., 2017) and Chang'e-4's 3.6 ± 0.3 (Lai et al., 2019). The loss tangent obtained at the Tianwen-1 landing site is 0.0087 ± 0.0027, a smaller value than Chang'e-3's 0.011–0.017 (Ding et al., 2020) and a larger value than Chang'e-4's 0.0039 ± 0.0002 (Lai et al., 2019).

Figures 4a and 2b show that the relative permittivity at the same depth measured by the RoPeR has a larger fluctuation range, which differs from the condition of the lunar regolith measured by Chang'e-4's LPR. The large deviation could be due to the heterogeneous distribution of the martian regolith along Zhurong's traveling path while the lunar regolith layer at the Chang'e-4 landing site is relatively uniform. The TW-1 landing site presents a larger heterogeneity in both the underground structure (Chen et al., 2023) and the permittivity distribution. It suggests that the space weathering degree (defined to include the continuous impact of large and small meteoroids and the steady bombardment of the surface by charged atomic particles from the sun and the stars.) of the lunar regolith is higher than the martian counterpart, which may be caused by the protection offered by the martian atmosphere and the consequential fewer surface impacts compared to the Moon.

The fitting curve of density against depth calculated by the Tianwen-1 RoPeR was compared with the density results measured by both the Chang'e-3 and the Chang'e-4 LPRs (Figure 4d). The density range at the three sites is comparable. The key difference is that the lunar regolith density changes marginally when the depth exceeds two meters, indicative of the high space weathering degree of the lunar regolith, while the density at the Tianwen-1 landing area continues to increase with depth. This may be caused by the heterogeneity of the regolith due to the reduced space weathering regime on Mars.

NaTeCam images show that the landing area is covered by various landforms, including scattered rocks (Figures 1b–1i), impact craters (Figures 1c, 1e and 1g), dunes (Figures 1 a, 1d, and 1f), and platy rocks (Figures 1h and 1i), indicating that some local geologic processes, be them aeolian, impact, aqueous activity, etc., may have affected the composition/weathering-degree of the area's surface. The heterogeneity of dielectric properties of the martian regolith in the landing zone reveals the complexity and diversity of its composition and formation processes, especially compared to the lunar regolith, which is also supported by spectral and camera data (Ding et al., 2022; J. J. Liu et al., 2022; Gou et al., 2022).

5 Conclusions

In conclusion, the RoPeR on board the Zhurong martian rover provides the first direct measurement of the detailed subsurface structure and dielectric properties of the martian regolith in southern Utopia Planitia within a depth of ∼5 m. The derived permittivity distribution map shows a trend where permittivity gradually increases with depth and no obvious horizontal and continuous permittivity contrast is present within such stratigraphic column. A layered structure is also not observed in the radar images. The reason could be due to the thickness of the local regolith layer being greater than ∼5 m, or that the dielectric contrast at the interface is not detectable by the RoPeR.

The permittivity of the Martian regolith within 5 m of the landing zone is $$, and the average loss tangent is 0.0087 ± 0.0027. It also fits with the hypothesis of the presence of volcanic rocks that might contain hydrous minerals formed by the rising groundwater as inferred from infrared spectral data; however, this does not preclude the possibility of sedimentary deposits brought by floods. The large deviation of dielectric properties along the traversing path of the Zhurong rover indicates the heterogeneity and low weathering level of the martian regolith compared to those at the CE-4 lunar landing site, showing more diverse surface processes although also a weak impact gardening effect.

Acknowledgments

We thank the Tianwen-1 payload team for mission operations and China National Space Administration for providing the Tianwen-1 data that made this study possible. This work is supported by the Civil Aerospace Pre-research Project (Grant D020101), the Science and Technology Development Fund of Macau (Grant 0049/2020/A1), Macau SAR (Grant SKL-LPS(MUST)-2021-2023), the National Natural Science Foundation of China (Grant 42104141), and the Guangdong Introducing Innovative and Entrepreneurial Teams, Zhujiang Talent Project Foundation of Guangdong Province (Grant 2017ZT07Z066).

Open Research

Data Availability Statement

The Tianwen-1 data including the Mars Rover Penetrating Radar (RoPeR) data and the Navigation and Terrain Camera image (NaTeCam) used in this study are processed and produced by the Ground Research and Application System (GRAS) of China's Lunar and Planetary Exploration Programme and provided by CNSA at https://clpds.bao.ac.cn/web/enmanager/mars1. The data can be obtained after registration and application on the website. The IDs for NaTeCam data are listed in Table S1 in Supporting Information S1. The IDs for RoPeR data are in Table S2. We have uploaded the processed radar data and all hyperbolas to a publicly accessible repository at https://doi.org/10.5281/zenodo.8035493.

Erratum

The originally published version of this article contained errors owing to an issue with one of the parameters used in one of the formulas, resulting in the final calculated loss tangent values being twice the correct value. As a result, errors occurred in the abstract, Sections 3, 4.1, 4.3, and 5 of the text, Figures 2, 3, and 4, the caption of Figure 4, and Figure S12 in the supporting information. In the third sentence of the abstract, “$$” should read “0.0060 ± 0.0002, 0.0087 ± 0.0002, and 0.0114 ± 0.0002 using the geometric compensation of R², R³, R⁴, respectively,” and the fourth sentence of the abstract should be deleted. In list item 3 in Section 2.4, “η = 20log₁₀P_local/P_max” should be deleted. In the first sentence of the last paragraph of Section 2.4, “0.121 ± 0.0004, 0.0174 ± 0.0004, 0.0227 ± 0.0004” should read “0.0060 ± 0.0002, 0.0087 ± 0.0002, 0.0114 ± 0.0002.” In the third sentence, of the first paragraph of Section 3, “echos” should read “echoes.” In the first sentence of the second paragraph of Section 3, “0.121 ± 0.0004, 0.0174 ± 0.0004, 0.0227 ± 0.0004” should read “0.0060 ± 0.0002, 0.0087 ± 0.0002, 0.0114 ± 0.0002.” In the fourth sentence of the second paragraph of Section 3, “0.0034 and 0.0279” should read “0.0017 and 0.0140.” In the last sentence of the second paragraph of Section 3, “0.0174 ± 0.0053” should read “0.0087 ± 0.0027.” In the second sentence of the last paragraph of Section 3 “between 22 and 28” should read “at 17.5 ± 4.5.” In the first sentence of the first paragraph of Section 4.1, “(estimated to be” should read “(estimated at.” In the second sentence of the first paragraph of Section 4.1, “at the northern Utopia Planitia” should read “of the lobate debris apron (LDA),” “(located in ~139°E, ~40°N)” should be deleted, and “0.009 ± 0.004,” should read “0.002 ± 0.0008, which is dominated by water ice.” In the first paragraph of Section 4.1, the third and fourth sentences (“However,…frequencies.”) should be deleted. In the fifth sentence of the first paragraph of Section 4.1, “RoPeR obtains” should read “Data from the RoPeR gives,” “0.174 ± 0.0004 (R³ correction)” should read “0.0087 ± 0.0027 that sits,” and “the neighboring ice-rich” should read “the ice-rich.” In the sixth sentence of the first paragraph of Section 4.1, “Based on these differences, the high loss tangent indicates that” should read “Considering the local surface temperature, atmosphere pressure and the loss tangent value,” “is not dominated” should read “is unlikely to be dominated,” and “and it produces considerable signal attenuation, thus lowering the detection depth potential of the radar” should read “(Zhao et al., 2021).” In the seventh sentence of the first paragraph of Section 4.1, “regional” should read “local.” In the last sentence of the first paragraph of Section 4.1, “which could explain the high loss properties of local materials” should read “Given that the loss tangent is affected by minerology, but also by density, the loss tangent of regolith should be lower than lava flows such as those in the Elysium Planitia area.” In the second paragraph of Section 4.1, the following two sentences should be inserted before the original first sentence (“The iron and titanium…landing site”): “The short-wave infrared spectrometer onboard the Zhurong rover, operating in the spectrum range between 0.8 and 2.4 μm across 321 channels, discovered that the landing site may contain a considerable amount of hydrated minerals (Y. Liu et al., 2022). The water-altered materials increase heterogeneity and lead to a stronger attenuation of radar signals (Stillman & Grimm, 2011).” In addition, the original first and second sentences should be combined to read “However, the iron and titanium derived from loss tangent value using the empirical formulas of lunar basalt samples is comparable with the remote sensing measurements of the FeO/ilmenite abundances at the landing site, suggesting the fraction of hydrated minerals to be minor” and the original third, fourth, and fifth sentences should be deleted. In the sixth sentence of the second paragraph of Section 4.1, “Therefore,” should read “Consequently,” “data that the local” should read “data, which also support the hypothesis that the local,” and “a degree” should read “a low degree.” In the last sentence of the second paragraph of Section 4.1, “relatively high loss tangent” should read “relatively similar loss tangent.” In the third sentence of the first paragraph of Section 4.3, “0.0174 ± 0.0004, a larger value than Chang'e-3's 0.011–0.017 (Ding et al., 2020) and Chang'e-4's 0.0039 ± 0.0002 (Lai et al., 2019)” should read “0.0087 ± 0.0027, a smaller value than Chang'e-3's 0.011-0.017 (Ding et al., 2020) and a larger value than Chang'e-4's 0.0039 ± 0.0002 (Lai et al., 2019).” At the end of the first sentence of the second paragraph of Section 5, “$$” should read “0.0087 ± 0.0027.” In Figures 2, 3, and 4, the values on the y-axes should be reduced by half. In the seventh sentence of caption of Figure 4, “0.121 to 0.0227” should read “0.0060 to 0.0114” and “0.0174” should be “0.0087.” At the end of the eighth sentence of the caption of Figure 4, “22 wt.% to 28 wt.%” should be “13 wt.% and 22 wt.%.” In Figure S12 in the supporting information, the values on the y-axis should be reduced by half. The errors have been corrected, and this may be considered the authoritative version of record.