AI for Climate Action

Artificial Intelligence for Climate Action

Leveraging the power of artificial intelligence and mathematics to spur innovation and novel solutions to challenges at the intersection of climate change and national security

Our Contribution

APL engineers and scientists are building novel AI capabilities to address the emerging national security challenges arising from climate change.

Visit APL’s Climate Security page to learn more about how the Lab is bringing its core competencies to bear on this critical challenge area.


Remote Sensing and Deep Learning for Detection and Forecasting

Researchers in APL’s Intelligent Systems Center (ISC) are building on the Lab’s existing strengths in remote sensing and deep learning to pinpoint the sources of climate change and increase situational awareness in a rapidly changing environment. For instance, APL is characterizing global greenhouse gas emissions from the road transportation sector as part of the Climate TRACE coalition. City-level results are now available on the coalition website. APL is also detecting and forecasting Arctic sea ice to enable navigation in a rapidly changing Arctic.

Related Publications

  • Mukherjee, R., D. M. Rollend, G. A. Christie, A. Hadžić, S. Matson, A. Saksena, M. Hughes, “Towards Indirect Top-Down Road Transport Emissions Estimation,” 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 1092-1101 (2021).
  • Rollend, D., K. Foster, T. Kott, R. Mocharla, R. Muñoz, N. Fendley, C. Ashcraft, F. Willard, M. Hughes. “Machine Learning for Activity-Based Road Transportation Emissions Estimation,” Tackling Climate Change with Machine Learning Workshop at NeurIPS 2022.
  • Keller, Mary Ruth, Christine Piatko, Mary Versa Clemens-Sewall, Rebecca Eager, Kevin Foster, Christopher Gifford, Derek Rollend, Jennifer Sleeman, “Short-Term Beaufort Sea-Ice Extent, Forecasting with Deep Learning,” submitted to Artificial Intelligence for the Earth Systems (Fall 2022).

Accelerated Earth Systems Models

Jennifer Sleeman (right) and Jay Brett
Artificial intelligence researcher Jennifer Sleeman (right) works with oceanographer Jay Brett (left) to explore the parameter space of models for the Atlantic Meridional Overturning Circulation (AMOC) tipping point.

Credit: Johns Hopkins APL

ISC researchers are developing operational forecasts and tools for scientific exploration by accelerating Earth systems models. Existing physics models require extensive computational resources and can be time consuming to run, limiting the exploration of possible futures and characterization of uncertainty. We are collaborating with Earth systems researchers to develop:

  • Deep-learning models to accelerate air quality forecasting
  • Tools to characterize climate tipping points using surrogate models, generative adversarial networks, and neuro-symbolic reasoning
  • Enhanced sea-ice models built with physics-inspired neural networks

Related Publications

  • Sleeman, Jennifer A., David Chung, Chace Ashcraft, Anshu Saksena, G. Jay Brett, Marisa Hughes, Anand Gnanadesikan, Yannis Kevrekidis, Marie-Aude Sabine Pradal, Thomas W. N. Haine, Renske Gelderloos, “Using Deep Learning for Climate Tipping Point Discovery to Understand Atlantic Meridional Overturning Circulation Collapse,” AGU (December 2022).
  • Halem, Milton, Jan Mandel, Adam Kochanski, Sen Chiao, Zhifeng Yang, Eugenia Kalnay, Jennifer A. Sleeman, Adam Bargteil, James P. Mackinnon, John Edward Dorband, Yaacov Yesha, Safa Motesharrei, Cheng Da, John Sorkin, Andrea Iorga, Samit Shivadekar, “Towards a Dynamic Multiscale Wildfire Digital Twin,” AGU (December 2022).
  • Hamer, Sophia, Jennifer A. Sleeman, Ivanka Stajner, Milton Halem, Christoph Keller, Raffaele Montuoro, Kai Wang, Jian-Ping Huang, Ho-Chun Huang, David Allured, James M. Wilczak, Irina Djalalova, Jeffrey McQueen, Barry Baker, Vladimir Krasnopolsky, “Forecast Aware Model-Driven Deep Learning Bias Correction for Improved Operational Air Quality Forecasting,” AGU (December 2022).
  • Brett, Genevieve, Larry H. White, Anand Gnanadesikan, Marie-Aude Sabine Pradal, Renske Gelderloos, Thomas W. N. Haine, Yannis Kevrekidis, Jennifer A. Sleeman, “AMOC Dynamics: Can a Box Model Explain a GlobalModel?” AGU (December 2022).
  • Sleeman, J, D. Chung, A. Gnanadesikan, J. Brett, Y. Kevrekidis, M. Hughes, T. Haine, M. A. Pradal, R. Gelderloos, C. A. Ashcraft, “Generative Adversarial Network for Climate Tipping Point Discovery (TIP-GAN),” arXiv preprint arXiv:2302.10274 (2023).
  • Sleeman, J; D. Chung, C. Ashcraft, J. Brett, A. Gnanadesikan, Y. Kevrekidis, M. Hughes, T. Haine, M. A. Pradal, R. Gelderloos, C. Tang, A. Saxsena, L. White, “Using Artificial Intelligence to aid Scientific Discovery of Climate Tipping Points.” arXiv preprint arXiv:2302.06852 (2023).
  • Ashcraft, C; J. Sleeman, J. Brett, A. Gnanadesikan, “A Bidirectional Neuro-symbolic Methodology for Translating Between Generative Latent Representations and Natural Language Questions,” AAAI Spring Symposium (2023).

Recent Media Coverage

Reinforcement Learning for Efficiency

A reinforcement learning agent's view of improving crop growth.
A reinforcement learning agent’s view of improving crop growth.

ISC researchers are developing AI methods to reduce the carbon footprint of various activities while maintaining or improving effectiveness. This includes leveraging reinforcement learning to train more efficient systems for adaptive HVAC control and optimizing crop yield.

Related Publications

  • Markowitz, J., N. Drenkow, “Efficient HVAC Control with Deep Reinforcement Learning and EnergyPlus,” ICLR 2023 Workshop on Tackling Climate Change with Machine Learning (2023).
  • Ashcraft, C., K. Karra, “Machine Learning aid Crop Yield Optimization,” Where AI Meets Food Security Workshop at AAAI Fall Symposium Series 2021. 

Advanced Systems Modeling and Integration

Diagram of an integrated modeling framework connecting predictions of population, rainfall, water demand, and emissions for 2050.
Diagram of an integrated modeling framework connecting predictions of population, rainfall, water demand, and emissions for 2050.

To combat climate and resource challenges, it is essential to understand and capture the complex interdependencies between different scales and areas of interest. The ISC’s approach to this problem includes using modular modeling frameworks to capture system interdependencies, applying network analysis to green power and food systems, and creating geographic overlays of Arctic infrastructure and anticipated climate changes.

Related Publications

  • Hughes, M., et al., “System Integration with Multiscale Networks (Simon): A Modular Framework for Resource Management Models,” 2020 Winter Simulation Conference (WSC), pp. 656–667, doi:10.1109/WSC48552.2020.9383983 (2020).
  • Reilly, E. P., S. Agarwala, M. T. Kelbaugh, A. Ciesielski, H.-J. M. Ebeid, M. Hughes, “Modeling the Relationship Between Food and Civil Conflict,” 2020 Winter Simulation Conference (WSC), pp. 715–726, doi:10.1109/WSC48552.2020.9384007 (2020).