Colombia will launch a USD 25 million artificial intelligence-based climate modeling center
Last summer, intense wildfires, heat waves, floods and droughts occurred all over the world. How much extreme weather can we expect in the next few decades? Hard to say. Climate change is complex, and the farther we look, the more fuzzy our predictions will be. Globally, this uncertainty is often used as an excuse to postpone expensive or unpopular measures, such as building seawalls, retreating inland, upgrading or relocating power grids, transportation systems, and other critical infrastructure.
In order to improve the accuracy of climate modeling and encourage society to prepare for the inevitable future destruction, The National Science Foundation (NSF) chose Colombia Lead a climate modeling center called “Learning the Earth with Artificial Intelligence and Physics” (LEAP). The center will work with the National Center for Atmospheric Research (NCAR) and NASA’s Goddard Institute for Space Research (GISS) to develop next-generation data-driven, physically-based climate models. It will also train a new wave of students who are proficient in climate science and using large data sets and modern machine learning algorithms. The larger goal of the center is to provide society with actionable information to adapt to climate change and protect the most vulnerable groups.
“Until climate models can provide more accurate predictions, it will be difficult to make the multi-billion-dollar investments required for adaptation at the regional level in making planning decisions,” said Columbia University President Lee C. Bollinger. “I can’t think of a better university than Columbia University. Its interdisciplinary focus is on leading the way in solving climate prediction problems.”
The center will be led by the Columbia School of Engineering, its School of Sciences, and the Lamont-Dougherty Earth Observatory, in collaboration with Teachers College, Columbia Business School, Social Work School, and the new interdisciplinary Columbia Climate Institute.
Global climate models agree that the earth will continue to warm in the next 40 years. But they disagree on the magnitude and severity of the impact from rising sea levels to increased floods and droughts. Most of the problems boil down to trying to incorporate the details of complex physical and biological processes—such as clouds that reflect sunlight into space or trees that absorb carbon from the air—into the model. Processes interact, and many processes are poorly understood.
With the help of big data and machine learning, researchers will study these processes in depth and update models with new knowledge to improve climate predictions. Researchers will use existing algorithms to analyze satellite images and other large-scale observation data that are missing from today’s models. They will also develop new algorithms to make detailed observations and generalize them to a wider context, discover causal relationships in data, and find better equations to describe the processes represented in the model. As new knowledge is incorporated into the model, researchers will use machine learning tools to test their predictions.
In addition to NCAR and GISS, Columbia University researchers will also collaborate with colleagues from New York University, University of California Irvine, Minnesota, and Montreal to update NSF funding and NCAR-based Community Earth System Model.
More accurate modeling will enable researchers to look into the future in a greater geographic range. “Suppose I want to know the number of heatwave days in New York City in 2050,” the center director said Pierre Genting, Professor of Earth and Environmental Engineering. “Some models may say 10 days. Some say 20 days. How will you adjust your grid to avoid power outages?”
“We still have these huge uncertainties,” the deputy director added Galen McKinley, Professor of Earth and Environmental Sciences, at the Lamont-Dougherty Campus of the Columbia Institute of Climatology. “Our goal is to use data from observations and simulations to better represent the basic physics, chemistry, and biology of the Earth’s climate system. More accurate models will help us have a clearer understanding of the future.”
Processing massive amounts of data requires modern infrastructure. Cooperate with Google Cloud and Microsoft, Ryan Abernathy, Lamont-Doherty’s associate professor of earth and environmental sciences will create a platform for researchers to share and analyze data.
To achieve its mission, the center will create a new discipline that combines climate science with data science and artificial intelligence. Carl VonderekAssistant Professor of Computer Science, will lead the team to develop new algorithms to advance climate change science and improve existing models. Tian ZhengThe statistics professor and dean will lead a team to develop undergraduate and graduate education programs that combine climate science and data science, classroom learning and research experience.
Climate change will affect all mankind, but so far, most climate science research has failed to make full use of the existing diversity of people and ideas. In order to solve this imbalance, the center will put diversity, fairness and inclusiveness as the core of its research and education mission. Courtney D. Cogburn, Associate Professor of Social Work and a member of the Data Science Institute, will oversee diversity, equity and inclusiveness initiatives and the transfer of knowledge to communities most vulnerable to climate change.
“Tackling the threat of climate change requires our brightest minds and the best technology,” said the Columbia Provost Mary Boyce“The LEAP Center’s research and educational missions are intertwined and will play a key role in helping us achieve our goals.”
Colombia will play a role in three of the six science and technology centers selected by the NSF this year for funding. In addition to LEAP, Columbia University will also join the Woods Hole Oceanographic Institute to study how marine microorganisms affect the carbon cycle and climate, as well as the University of Washington to develop photovoltaic materials.
This story originally Appeared in Columbia News.