Artificial Intelligence and Earth Sciences

Artificial intelligence is coming alive at a strange moment in human history. While computer scientists have been unpacking the mysteries of logic and reasoning toward a brighter future for all mankind, scientists in climatology have been making more disturbing discoveries.

Global climate change is just one aspect of the mysteries of earth sciences, but it is the one that may trump all the others. Affecting almost every kind of geophysical planetary system, it has impacts on everything from hydrology to meteorology.

And, of course, it’s going to have just as much impact on the human race as artificial intelligence will.

So it may be fitting that AI is poised to help earth scientists decipher and develop solutions to some of the most pressing problems of our time.

What Are the Earth Sciences?

Earth sciences is a broad term that takes in all the various fields of study of nature and the systems that make up the planet. It includes:

Moreover, earth sciences takes in all the various ways in which those individual fields connect and affect one another.

While it’s a big picture kind of field, it’s also one that embraces details. No piece of information is too small to ignore. The smallest leaf can paint the picture of the health of the whole forest.

How Artificial Intelligence Can Unravel the Mysteries of the Planet

The complexity of all the systems involved in the study of earth sciences has already made computational analysis and big data breakthroughs indispensable in these fields. With billions of data points swirling around natural events as basic as a rain shower, only computers can track and analyze all the details.

Machine learning, a vital subset of artificial intelligence, is already off to the races in earth sciences research.

AI in Meteorology: Weather forecasters have relied on computers for decades, building elaborate forecast models that are essentially huge, computationally-heavy algorithms. AI forecasts developed by companies like Google and Nvidia, however, have gone toe-to-toe with those models, and done so much faster. In the future, it’s likely that AI will be combined with traditional forecasts to offer better predictions for extreme weather events and more accurate local forecasts.
AI in Ecology: Predictive analysis of complex systems is red meat for AI algorithms, and ecology offers some of the most complex systems on the planet. Nonlinear results from interactions and feedback in ecosystems are both tough and critical to predict. AI’s ability to churn through millions or billions of simulated events quickly and with fine statistical accuracy makes it an ideal tool to evaluate the health of ecosystems and model various impacts.
AI in Geology: Geologists have jumped on AI as a tool for mineral analysis, mapping, and exploration. Feeding seismic and core sampling data to machine learning algorithms can identify similarities and hidden features worth additional exploration, or flag sites that likely contain important mineral resources. Computer vision techniques are being used to easily visualize geological structures and soil properties.
AI in Climatology: Climatology, of course, is the vital science of the decade. As the planet climbs past 1.5°C in warming, climatologists are scrambling to nail down answers to impacts on every other aspect of earth sciences. It’s the mother of all complex problems. Artificial intelligence may offer the insights to help unravel it. With better predictions, more confidence will enter the conversation to head off the worst catastrophes climate change has on the menu for humanity.

AI is also already putting in good work in tying together some of the toughest connections in those various areas, such as long-range forecasts for ENSO: the El Niño Southern Oscillation, and its little sister, La Niña. These mysterious shifts in ocean temperature are caused by meteorological patterns and influenced by deep oceanography. In turn, they pass along effects such as drought, strengthened Atlantic hurricanes, and flooding in different regions.

Traditional statistical models tend to fall apart when asked about El Niño odds more than a year out. But probabilistic forecasts analyzed by an AI model is accurate to as much as 18 months ahead, offering valuable warning to planners.

The kind of predictive ability offered by AI is not only vital for dealing with the worst of the coming effects of climate change, but also in helping to mitigate the phenomena itself.

Other ways that AI boosts earth sciences will come indirectly, through other areas of AI development. For example, conducting vital research in key trigger points like the Arctic, Antarctic, and deep ocean are beyond the physical capabilities of human researchers. But advances in autonomous systems and robotics can deliver research assistants in the form of machines that can work longer, more accurately, and in more extreme conditions than human scientists.

Similarly, improvements in computer vision have boosted satellite capabilities in tracking vital real-time information about the state of the ice pack, deforestation, and other important inputs in climate models.

The data they gather may become critical in future calculations.

A Look at the Jobs in AI That Will Emerge in Earth Sciences

The kind of work for AI professionals in earth sciences will vary quite a bit from field to field.

For the most part, these roles will lean heavily toward research and pure science. They will often be found in universities or at government labs, assisting professionals in the relevant sciences with their own modeling. In other cases, they’ll be found at startups offering predictive services in various sciences for commercial clients.

For the most part, these positions won’t necessarily splash AI or ML right in the title. Instead, they fall into familiar categories for such researchers:

AI engineers working with earth sciences researchers may be assigned to help create and run the models needed by scientists.

In some cases, however, there are private sector opportunities for AI in earth sciences. In geology, for example, large corporations in resource extraction will be scrambling to find applied uses for AI systems. Similarly, meteorology and forecasting is a field that holds the interest of organizations ranging from shipping companies to airlines. If they can get a competitive edge with AI tools, you can bet they’ll be hiring their own experts to build those systems.

These can come with titles like:

What Degrees Are the Best for AI Training in Earth Sciences?

Clearly, work in artificial intelligence for earth sciences takes place at some of the most advanced levels of research and discovery. It’s unlikely that anyone with only a bachelor’s degree will make much headway in any of these fields.

Instead, this is the realm of master’s and doctoral graduates in AI.

The gold standard for AI degrees with a focus on earth sciences would be something like a Master of Science in Intelligent Systems Engineering Environmental Engineering track. On top of the standard coursework in statistical analysis, machine learning, and programming, such degrees offer additional classes in areas like:

They are also likely to involve research projects that line up directly with various studies in today’s hottest topics in earth sciences.

Environmental engineering is just one aspect of earth sciences. Other tracks in similar majors may be more applicable to different areas of earth sciences.

One specialization that is almost universally needed, however, is data science. The large amounts of information that AI needs and is expected to process in nearly all these fields mean a familiarity with advanced ML data processing will almost always be useful. A Master of Science in Artificial Intelligence concentration in Data Science or a Master of Science in Engineering AI and Data Science focus deliver exactly the skills needed to handle tough data gathering and analysis in any earth sciences field.

For some specialized earth sciences applications, the best bet may be a Master of Science in Robotics or Master of Science in Computer Vision.

At the doctoral level, there are fewer specializations listed in the earth sciences arena. But that doesn’t mean the degrees aren’t focused in these areas. Instead, the relative academic freedom that comes with doctoral studies allows students to specialize their studies to whatever degree their college and advisors can support. Courses are designed by the student and advisors. Research is devoted to the area in which their dissertation is focused.

Of course, this points to a key consideration in enrolling in a PhD in Artificial Intelligence program… the expertise and research facilities available at a particular university will dramatically influence available options in earth sciences studies. A school that has leading scientists working on issues in the relevant area, or well-respected research institutes in the field, will allow students more in-depth specialization options.

The nature of these studies in some areas, like climate, means that other majors often feature substantial AI and ML training. It’s possible to develop similar expertise through a PhD in Computational Physics, or a PhD in Computational Earth, Atmospheric, and Planetary Sciences, for instance. The cutting edge of those fields is artificial intelligence; some of the most applied uses for the technology are likely to emerge from researchers in those programs.

Certificates in Artificial Intelligence and Earth Sciences Can Complement One Another

Anyone coming into the field with a primarily AI background can benefit from something like a Graduate Certificate in Climate Science or a Certificate in Atmospheric and Oceanic Studies. These offer the kind of interdisciplinary studies in earth science systems necessary to be familiar with to work on accurate modeling in the field.

From the other direction, of course, earth scientists with the right coding and mathematics background can benefit from a Graduate Certificate in Applied Machine Learning, or a Graduate Certificate in Data Mining and Machine Learning. Understanding the advantages and drawbacks of deep neural networks and their applications in crunching large and diverse datasets may spark new ideas in how to explore phenomena of planetary size.

Separate from educational certificates are professional certifications: third-party assessments of knowledge and skill that assure employers that candidates have specific skillsets required in certain positions.

Both because the technology and science are so advanced and specific, and because of the relatively narrow range of positions, there are no professional certifications specific to AI and earth sciences. However, you may find that some more general professional certifications in machine learning are particularly applicable in this area.

Some of the most exuberant boosters of artificial intelligence spend a lot of time talking about how it may be the salvation of the human race. While it’s easy to write off some of their wilder predictions, AI in the earth sciences might actually live up to all the expectations.