Codifying intuition in the search for life

Discovering life on Mars will require missions with specialized payloads and well-informed search strategies. A new study combines geology, statistical ecology, and machine learning to deliver a methodological framework to codify latent scientific intuition in the search for life.
Published in Astronomy
Codifying intuition in the search for life
Like

Is there life on Mars?

The presence of life on Mars was considered a forgone conclusion in the 19th century. After all, what else could have created the enormous 'canals' that scarred the surface besides an advanced Martian civilization? At least that's how some preeminent astronomers of the time reasoned, such as Giovanni Schiaparelli and Percival Lowell. The "Advanced Martian Civilization" hypothesis lurked in the scientific zeitgeist until the 1970s when relatively high-resolution observations from the Viking missions relegated it the realm of pseudoscience. The popularity of what might seem an absurd hypothesis given our present knowledge of Mars can be explained by a lack of data at the appropriate spatial scales combined with a desire to believe there is life on Mars. Isn't a reality in which Mars teems with life more interesting than one in which Mars is sterile?

We don't know...

The search for life on Mars continues in the era of modern space exploration. We now know Mars was once habitable for the types of microbial lifeforms that took root on early Earth, but we lack the data necessary to determine if microbes once colonized Mars - a challenge similar to that which confronted Schiaparelli and Lowell and prone to similar over-interpretations of data. Convincing evidence for microbial life on Mars won't be easy to find, and despite the many robotic missions to the martian surface we have yet to make a bona fide "life detection" effort. Such an effort would benefit from a robust search strategy, but none have been developed. Our recent publication combines the expertise of geologists, microbial ecologists, and artificial intelligence researchers to start building a coherent methodological framework to aid in the search for life. A repeatable methodology backed by statistics and predictive models is a first step toward standardizing astrobiology search strategies. With a standard methodology, results can be more easily cross-compared, repeated, and validated here on Earth before extra-terrestrial deployment.

But we are searching

The exploration strategy of NASA is 'flyby, orbit, rove, return'. In the 'orbit' stage of martian exploration, knowledge gained from studying Earth is leveraged to understand what makes an environment habitable and what compositional and geological indicators of such environments are preserved in the rock record. Planetary Scientists use this knowledge to comb through the many orbital remote sensing datasets of Mars and identify past habitable environments, such as ancient lakes, rivers, and hydrothermal vents. It is in these locations that we 'rove'; that is, send Mars rovers to conduct detailed characterizations at the surface.

Once landed in a past habitable environment, the challenge of searching for signs of life begins in earnest. At the orbital scale, we have an immense amount of data from both Earth and Mars that enables us to locate past habitable environments. We do not have the same luxury at the scale of rover exploration. We lack vast troves of rover-scale data - from Earth and Mars - that we can use to guide the search for life within a past habitable environment. Put simply, there is no roadmap for a rover team to follow that will lead them to the best place to look for signs of life.

Is it possible to develop such a biosignature roadmap? Our team decided to try this out at Salar de Pajonales - a Mars-analog, salt-encrusted, paleo-lake in the Andean Altiplano Puna Plateau. Our goal was not only to map the probability of finding signs of life from orbit to the ground across our field site (Fig. 1), but also to prototype the process by which biosignature guides can be generated at any analog site. The longterm vision of our effort is for the astrobiology community to contribute biosignature roadmaps - constructed in a similar fashion across as many Mars analog sites as possible - to a biosignature-roadmap library. A meta-analysis of such a library would be invaluable for codifying the latent intuition held by geologists and microbial ecologists for how to maximize the probability of finding biosignatures in any harsh environment, including Mars.

CNN-derived biosignautre probability maps
Fig. 1 Biosignature probability maps from Convolutional Neural Network (CNN) models and statistical ecology data | a) aerial-scale biosignature probability map for macrohabitats at Salar de Pajonales (SdP). b) A visible image of a dome and biosignature probability maps for each microhabitat within the gypsum dome; “true” shows ground-truth data (top row); CNN predictions for presence of biosignatures and microhabitats (middle row); Classification accuracy (y-axis) across 10 randomized runs, summarized by boxplots, each defined by mean, standard deviation, minimum, and maximum (bottom row). 'Test" refers to the performance on the CNN test dataset and "val" indicates the performance on the validation dataset.

A biosignature roadmap for a single environment will not deliver the universal guide to life that we desire, but it is a start. Perhaps there are overarching rules, patterns, and principles that life follows in extreme environments that we can discover and leverage in the search for life in the Solar System. We will only discover these patterns if we have a coherent strategy for mapping microbes across harsh environments to build a library of rover-relevant data. Only then can we hope to apply what we've learned in a way that will deliver results at Mars. We have come a long way since the days of Schiaparelli and Lowell, but much ground remains to be traversed before we can answer the question, "is there life on Mars?" 

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Subscribe to the Topic

Astronomy, Cosmology and Space Sciences
Physical Sciences > Physics and Astronomy > Astronomy, Cosmology and Space Sciences

Related Collections

With collections, you can get published faster and increase your visibility.

Progress towards the Sustainable Development Goals

The year 2023 marks the mid-point of the 15-year period envisaged to achieve the Sustainable Development Goals, targets for global development adopted in September 2015 by all United Nations Member States.

Publishing Model: Hybrid

Deadline: Ongoing

Wind, water and dust on Mars

In this Collection, we bring together recent work, and invite further contributions, on the nature and characteristics of the Martian surface, the processes at play, and the environmental conditions both in the present-day and in the distant past.

Publishing Model: Hybrid

Deadline: Ongoing