Spectral Performance Assessment of Enfys for Mineral Discrimination at Oxia Planum

Introduction

Accurate mineral identification at Oxia Planum is central to the geological objectives of the ESA Rosalind Franklin rover [1], particularly in the characterisation of the phyllosilicate-bearing units associated with past aqueous environments [2]. This study evaluates the spectral performance of the onboard short-wave infrared (SWIR) spectrometer, Enfys, for reliable mineral identification on the martian surface [3].

 

Methods

We have assessed the performance of Enfys using laboratory-derived spectral libraries, including, RELAB [4], MICA [5], VISOR [6], and USGS [7], alongside Oxia Planum CRISM observations [8]. To evaluate Enfys’ detection capabilities, the high-resolution library spectra have been resampled to Enfys’ spectral parameters (Figure 1) using SPTK [9].

Figure 1: Comparison of original library spectra (solid lines) and resampled Enfys spectra (dashed lines) from the mica files [5], the Enfys-like spectra were resampled at a nyquist sampling rate using sptk [9].

 

Continuum removal and smoothing techniques can be incorporated to isolate absorption features, making it easier to compare spectra with different mineral assemblages. These preprocessing steps are useful for normalising spectral shapes across varying viewing and environmental conditions, improving the comparability of laboratory, field, and martian data. 

A variety of error metrics have been calculated (Figure 2), that can quantify the spectral performance of Enfys by measuring overall numerical deviation and spectral feature preservation, and to aid comparison with SWIR spectrometers on previous and ongoing Mars rovers.

Figure 2: Error Metrics for the Enfys resampled mineral spectra in the mica files with mineral ID labelled [2]. The six error metrics indicate preservation of shape and overall numerical deviation. Left to right: root mean square error (RMSE), mean absolute error (MAE), mean square error (MSE), spectral angle mapper (SAM), spectral information divergence (SID) and cross-correlation coefficient (CCC).

To support automated mineral classification, spectral parameters such as band centres, depths, and widths will be extracted after resampling. Characterising these parameters for expected minerals at Oxia Planum will aid post-landing mineral detection and target selection.

Results

Preliminary analysis shows that Enfys can identify key mineral groups, such as phyllosilicates and preserve key diagnostic absorption features. These results confirm the value of Enfys for in situ mineralogical analysis on Mars, improving our ability to select promising drill targets for biosignature detection. More broadly, this work provides a framework for optimising spectral data acquisition in future planetary exploration missions.

 

 

[1] Vago, J.L, et al. (2017) Astrobiology, 17(6-7), pp.471–510. [2] Quantin-Nataf, C. et al. (2021) Astrobiol. 21, 345–366. [3] Grindrod, R., et al. (2025). LPSC 56, #1840.  [4] Milliken, R.E., Hiroi, T., Patterson, W. (2016) LPSC 47, #2058. [5] Viviano, C.E. et al. (2014) JGR Planets 119, 1403–1431. [6] Million, C.C. et al. (2022) LPSC 53, #2678. [7] Kokaly, R.F. et al. (2017) USGS Data Series 1035. [8] Murchie, S., et al. (2007) Journal of Geophysical Research, 112(E5).  [9] Stabbins, R., Grindrod, P. (2024) Zenodo, doi:10.5281/ZENODO.10694286.