Internship 4
Host name: Kara Lamb
School and Department: Earth and Environmental Engineering
Internship title: Causal Machine Learning to Evaluate the Climatological Effectiveness of Historic Weather Modification Activities
Number of interns to be hosted: up to 2
Types of support offered:
- Stipend: $3500/month
- Access to computational resources and opportunities to interact with researchers and attend seminars on climate data science at the Learning the Earth with Artificial Intelligence and Physics (LEAP) Center at Columbia.
Internship description
As the western United States faces increasingly severe droughts and wildfire activity, cloud seeding is receiving growing attention as a potential strategy to augment water supplies and reduce wildfire risk. Ten states and provinces across western North America currently operate cloud-seeding programs, yet their overall effectiveness remains uncertain because it is difficult to distinguish seeding-induced precipitation from natural meteorological variability. We have recently developed a dataset of weather modification activities across the western United States, compiled from federal records spanning 2000–2025 (Donohue and Lamb, Nature Scientific Data, 2025).During this internship, you will combine these records with hydrological and climatological observations and apply machine-learning methods (including transformer-based architectures and causal machine learning approaches) to investigate whether watersheds associated with substantial cloud-seeding programs exhibit measurable changes in streamflow relative to comparable control watersheds. The project will provide opportunities to develop new technical skills, shape the direction of the analysis, and pursue related research questions aligned with your interests.
Skills required
Python programming, familiarity with data science and/or machine learning methods in python, previous experience working with geospatial data sets, experience with atmospheric physics or hydrology .