PhD in Earth and Planetary Sciences (M/F) – Large Scale Study of Seasonal Dark Spots on Mars: Potential Implication for Regional Climate

Seasonal ice processes are major drivers of present-day surface activities on Mars. Each year, about one third of the atmospheric CO2, the main component of the Martian atmosphere, condenses onto the high latitude surfaces to form seasonal ice cap deposits [1]. In spring, sublimation of CO2 ice deposits coincides with widespread surface changes and enhanced dust transport. Among the most striking features are millions of transient dark spots, covering vast areas of Mars [2, 3], observed in the form of darker albedo blotches and fan-shaped deposits, typically tens of meters in size. 

To date, their formation is best explained by the Kieffer model [2] which propose that solar radiation penetrates the seasonal translucent CO2 slab ice, heating the substrate and triggering basal sublimation. The build-up pressure would eventually cause ruptures in the slab, releasing gas and dragging dust to the surface and potentially into the atmosphere. 

Despite their ubiquity, key questions remain regarding their formation environment and mechanisms and interaction with dust. Due to their large number, the area they cover and their transient nature, no (or partial) comprehensive mapping and characterisation of the dark spots has been performed so far [3, 4, 5, 10], which prevents us from fully understanding their characteristics. Dark spots develop in diverse environments with different substrates, pressure-temperature conditions, dust content and seasonal ice properties, yet they are absent from some regions with seasonal CO2, suggesting that specific conditions are necessary for their formation [3, 6, 7] but not clearly established yet. Finally, dust is known to strongly influences Martian climate [8], ice properties and stability, and surface albedo. The contribution of the dark spot process to the dust budget, influence for the atmosphere and the feedback on the seasonal processes are not constrained. 

Therefore, the objectives of the thesis project are the following:

• Automatic mapping and statistical analysis of the dark spot properties

Dark spots exhibit diverse sizes, morphologies and spectral properties [2, 7, 10]. We have developed a deep learning framework to automatically detect dark spots in mid-resolution CTX images, enabling the first global spatiotemporal inventory of their distribution [3]. We aim to pursue this work by identifying areas of interest for detailed investigation. The development of an image segmentation framework will be used to perform automatic mapping and statistical analysis of spectral characteristics and morphometric parameters using large datasets of spectral and high-resolution images (CaSSIS, HiRISE, CRISM, MOC). This work will complement the methodological efforts already conducted by P. Bessin (LPG-Le Mans) on periodic bedform mapping and will be conduct in collaboration with V. Bickel (CSH, UniBern).

• Identification of favourable environments for dark spot development

The global and high-resolution mapping will be cross-referenced with complementary datasets to investigate links between dark spots properties (distribution, morphology, colour) and their formation environment (geology, geomorphology, topography and roughness, atmospheric properties, seasonal ice properties). Statistical methods (e.g., correlation analysis, PCA) and predictive modelling approaches (e.g., tree-based algorithms) will be applied to identify the key variables controlling dark spot formation and variability on Mars and will be conducted in collaboration with A. Pommerol (UniBerne).

• Implications for Mars' active volatile and dust cycles

The interaction between the atmosphere, dust and seasonal ice deposits on Mars is complex. Dark spot processes involve volatiles and dust transport, with dust potentially lifted up to hundreds of meters [11] and representing a non negligeable amount (potentially about 1013 kg) [12]. A fraction of this material redeposits directly on the seasonal ice surface, reducing its albedo and so its thermal equilibrium. This effect is not yet taken into account in the Martian atmospheric models and will be explored to assess its impact on the volatiles cycle and the local climate (U, V, T) [13], as well as potential feedback since dark spot activity appears sensitive to atmospheric dust content [3]. Another fraction of the ejected dust may be entrained by winds into local atmospheric circulation [11] but this potential source of dust has never been quantified so far. Numerical simulations will be conducted to evaluate the mass of dust that can be injected into the atmosphere via this process (in collaboration with S. Carpy, LPG), improving our knowledge on the dust reservoirs and cycles.

Studying dark spots is essential for understanding the mechanisms driving Martian seasonal processes and present-day surface activities. Notably, it addresses how the coupling between the atmosphere, the seasonal ice deposits and the dust shape the planet's landscape, offering new insights into the volatile-driven geomorphology, volatiles and dust cycles, dust sources, atmospheric dynamics, and the identification of relevant sites for monitoring surface processes. Ultimately, it will help to test and improve the Kieffer model and climate models.

[1] Hess et al. (1980). The annual cycle of pressure on Mars measured by Viking Landers 1 and 2. Geophysical Research Letters, 7(3), 197-200. https://doi.org/10.1029/GL007i003p00197.

[2] Kieffer (2007). Cold jets in the Martian polar caps. Journal of Geophysical Research, 112(E8), E08005.  https://doi.org/10.1029%2F2006JE002816.

[3] Vignon et al. (in prep.). Dark spots as spatiotemporal indicators of sublimation of CO2 seasonal polar caps on Mars.

[4] McDonnell et al. (2023). Planet Four: A Neural Network’s search for polar spring-time fans on Mars. Icarus, 391, 115308. https://doi.org/10.1016/j.icarus.2022.115308.

[5] Diniega et al. (2025). Holistic Mapping of the Present-day Martian Seasonal CO2 Frost. I. Frost Detection within Global Visible, Thermal, and Spectral Data Sets. The Planetary Science Journal, 6, 209. https://doi.org/10.3847/PSJ/adef07.

[6] Hansen et al. (2024). A comparison of CO2 seasonal activity in Mars’ northern and southern hemispheres. Icarus, 419, 115801. https://doi.org/10.1016/j.icarus.2023.115801.

[7] Thomas et al. (2024) Seasonal and Short Timescale Changes on the Martian Surface: Multi-Spacecraft Perspectives. Space Science Reviews, 221:3. https://doi.org/10.1007/s11214-024-01128-4.

[8] Madeleine et al. (2015). Revisiting the radiative impact of dust on Mars using the LMD Global Climate Model. Journal of Geophysical Research. Planets, 2011, 116, ppE11010. https://doi.org/10.1029/2011JE003855.

[10] Cesar et al. (2022). Seasonal southern circum-polar spots and araneiforms observed with the colour and stereo surface imaging system (CaSSIS). Planetary and Space Science, 224, 105593. https://doi.org/10.1016/j.pss.2022.105593.

[11] Thomas et al. (2011). HiRISE observations of gas sublimation-driven activity in Mars’ southern polar regions: IV. Fluid dynamics models of CO2 jets. Icarus, 212, 66-85. https://doi.org/10.1016/j.icarus.2010.12.016.

[12] Piqueux and Christensen (2008). North and south subice gas flow and venting of the seasonal caps of Mars: A major geomorphological agent. JGR, 113, E06005, https://doi.org/10.1029/2007JE003009.

[13] Lange (2024). Studying and modeling the dynamics of water and CO2 ice on Mars. PhD thesis, Sorbonne University. https://theses.hal.science/tel-04824230v1.

• Master’s degree (or equivalent) in Earth/Planetary Science, Geophysics, Physics, Computational geosciences, or a closely related field.
• Skills in programming languages, Geographic Informatic System, computer vision and data science methods. In particular, some knowledge of Python (and major libraries), geospatial image processing (e.g., scikit-image, Open CV, rasterio, GDAL,…), machine learning/deep learning methods (e.g., scikit-learn, PyTorch, TensorFlow,…) and GIS (e.g., QGIS, ArcGIS) are expected.
• Ability to manage multiple and large datasets.
• Knowledge of statistical analysis.
• Strong interest for planetary topics, physical (e.g., atmosphere, phase change, dust drag) modelling and computer vision methods.
• Good communication skills and the ability to work both independently and collaboratively.
• A sensitivity to FAIR principles would be appreciated.