The Learning the Earth with Artificial Intelligence and Physics (LEAP) Center will launch in 2021. What was the central mission of this center?
LEAP’s mission is to extend the reliability, utility, and scope of climate predictions by integrating climate and data science. Our main strategy is to improve short-term climate forecasting through the continuous integration of physical modeling and machine learning, from our expertise in climate science and climate modeling to our state-of-the-art machine learning algorithms. is. This is really helpful for both climate and data science communities. Climate scientists and modelers struggle to fully integrate large existing datasets into their models. Machine learning algorithms like ChatGPT are good at emulating but not so good at extrapolating or making extreme predictions. Combining both approaches, we expect that his LEAP will greatly advance data science algorithms applied to physical problems. The center incorporates physics and causal mechanisms into machine learning algorithms to improve generalization and extrapolation while optimally using the wealth of data available in climate science to better predict the future. I’m here.
Can you mention some of the major projects that LEAP is focusing on? Any breakthroughs?
LEAP addresses various aspects of climate and data science, covering not only research but also knowledge transfer and education. Some of the recent research breakthroughs at our center show that AI can be used to discover new, previously unknown physics, such as clouds and ocean turbulence. We hope to use this “new” physics in climate models to improve accuracy, especially in extreme cases. Essentially, we use vast amounts of data, such as data from satellites, to refine climate models and their assessments to improve forecasts. It is also important to share this information with the public and private sectors in a user-friendly way.
We also want to break down some of the historical silos in climate research. Climate research is so technical and difficult to use that it cannot be easily translated to the public or private sector. We’ve created a cloud platform that can provide climate data more broadly, and we’re working with colleagues on and off the ground to see what actually works for them. For example, a company may want to know how the frequency of floods and heatwaves will change in the future so they can adapt their business. With LEAP, we hope to be able to refine our models to provide more accurate predictions.
AI is getting a lot of attention, and AI is currently being applied to things like climate research. How has AI revolutionized climate modeling, and what is the current state of it?
Over the past five years, there has been an explosion in the use of AI to better understand climate models and better represent physical processes such as clouds, oceans, terrestrial carbon cycles, and ocean turbulence. The next big challenge is how to integrate these AI algorithms into climate models that have historically used empirical equations.
The information we predict into the future is truly uncertain. There are many reasons for this. For example, all the processes that we are trained to do in building climate models are very complex. Therefore, we aim to reduce and narrow these uncertainties in order to provide accurate climate forecasts to everyone, from policy makers to business leaders to educators, to inform decision-making. Emphasis on For example, in the agricultural sector, being able to provide accurate information about future climates can have a significant impact on crop productivity and yields.
The goal is to improve climate modeling so that we can predict how many days of drought will last, or the likelihood of flooding in New York City and certain other low-lying areas. These are very important questions. Currently, the range of estimates is so dramatic that it is difficult to implement plans in practice. The battle must begin.
What excites you most about where the field is headed and how LEAP’s work will impact our future?
We are witnessing a real transformation. It is data-centric, using observational and simulated data to try to answer new hypotheses and questions that were previously unaddressed. Of course, we still need to be cautious when it comes to new areas and make sure our results are sound and reproducible, but we are witnessing an incredible pace of progress where this area is headed. I think we are witnessing a revolution in climate science.
What can we tell our children this Earth Day 2023 about how we invest in our planet, how important climate change research is, and what we want to leave behind for them and future generations? do you want to
What I tell my three children is to think of others. We need to use less resources, limit our footprint, recycle and reduce our emissions. In other words, we need to mitigate climate change. But we also need to adapt. Climate change is now part of our daily lives, as evidenced by the recent explosion in the number and intensity of extreme events (droughts, floods, etc.). It is clear that we as a society need to change. This change must be accepted by all and requires broad support (government, policy) to enable whole nations, not just small segments of society, to adapt to these changes. Otherwise, we will fail as a society and as a nation. Social justice and climate change are inextricably linked, and we need both to reach our goals.
