Physics-informed AI for Earth science

Wildfire figure

Real-world responses to critical hazards or climate challenges demand models that are not only accurate but also physically interpretable and transferable across scales. Traditional machine learning often struggles with limited observations or poor generalization, functioning as black boxes, whereas purely process-based models may overlook hidden patterns in complex data and suffer from limited accuracy. Our lab explores physically-interpretable machine learning and geospatial reasoning schemes that integrate domain knowledge for high accurate and transferable geospatial process modeling even with limited data observations. We are particulally interested in:

  • Physics-informed / knowledge-guided machine learning – incorporating process understanding into ML frameworks
  • Causal inference – identifying cause–effect relationships in complex Earth system interactions
  • Spatiotemporal interpretable and explainable AI – ensuring transparency and trustworthiness in Earth predictions
  • Earth foundation models – developing scalable AI models for cross-domain applications in Earth and environmental sciences

Selected Papers

  1. Peng, D., Gui, Z., Wei, W., Li, F., Gui, J., Wu. H., Gong, J. (2025). Sampling-enabled scalable manifold learning unveils the discriminative cluster structure of high-dimensional data. Nature Machine Intelligence

  2. Liu, H., Li, F., Dashti, H., & Chen, M. (2025). Hyperspectral surface reflectance improves GPP estimation in terrestrial biosphere modeling using model–data fusion. Remote Sensing of Environment, 330, 114989.

  3. Ji, F., Li, F., Dashti, H., Hao, D., Townsend, P. A., Zheng, T., You, H., & Chen, M. (2024). Leveraging transfer learning and leaf spectroscopy for leaf trait prediction with broad spatial, species, and temporal applicability. Remote Sensing of Environment.

  4. Li, F., Zhu, Q., Yuan, K., Ji, F., Paul, A., Lee, P., Radeloff, V. C., & Chen, M. (2024). Projecting large fires in the western US with an interpretable and accurate hybrid machine learning method. Earth’s Future, 12, e2024EF004588.

  5. Li, F., Zhu, Q., Riley, W. J., Yuan, K., Wu, H., & Gui, Z. (2022). Wetter California projected by CMIP6 models with observational constraints under a high GHG emission scenario. Earth’s Future, 10(4), e2022EF002694.

  6. Yuan, K., Zhu, Q., Li, F., Riley, W. J., Torn, M., Chu, H., v. G., Chen, M., Knox, S., Delwiche, K., Wu, H., Baldocchi, D., Ma, H., Desai, A. R., Chen, J., Sachs, T., Ueyama, M., Sonnentag, O., Helbig, M., Tuittila, E., Jurasinski, G., Koebsch, F., Campbell, D., Schmid, H. P., Lohila, A., Goeckede, M., Nilsson, M. B., Friborg, T., Jansen, J., Zona, D., Euskirchen, E., Ward, E., Bohrer, G., Jin, Z., Liu, L., Iwata, H., Goodrich, J., & Jackson, R. (2022). Causality guided machine learning model on wetland CH₄ emissions across global wetlands. Agricultural and Forest Meteorology, 324, 109115.

  7. Zhu, Q., Li, F., Riley, W. J., Xu, L., Zhao, L., Yuan, K., Wu, H., Gong, J., & Randerson, J. (2022). Building a machine learning surrogate model for wildfire activities within a global Earth system model. Geoscientific Model Development, 15(5), 1899–1911.

  8. Li, F., Zhu, Q., Riley, W., Zhao, L., Xu, L., Yuan, K., Chen, M., Wu, H., Gui, Z., Gong, J., & Randerson, J. (2023). AttentionFire v1.0: interpretable machine learning fire model for burned area predictions over the tropics. Geoscientific Model Development, 16(3), 869–884.

  9. Gui, Z., Sun, Y., Yang, L., Peng, D., Li, F., Wu, H., Guo, C., Guo, W., & Gong, J. (2021). LSI-LSTM: An attention-aware LSTM for real-time driving destination prediction by considering location semantics and location importance of trajectory points. Neurocomputing, 440, 72–88.

  10. Li, F., Gui, Z., Zhang, Z., Peng, D., Tian, S., Yuan, K., Sun, Z., Wu, H., Gong, J., & Lei, Y. (2020). A hierarchical temporal attention-based LSTM encoder-decoder model for individual mobility prediction. Neurocomputing, 403, 153–166.

  11. Li, F., Gui, Z., Wu, H., Gong, J., Wang, Y., Tian, S., & Zhang, J. (2018). Big enterprise registration data imputation: Supporting spatiotemporal analysis of industries in China. Computers, Environment & Urban Systems, 70, 9–23.