Akhil Paulraj ’25
The Interface of Hydrology and Machine Learning: Generating Better Information for Decision-makers and Educating the Decision-makers of the Future
Certificate(s): Applied and Computational Mathematics, Statistics and Machine Learning
Groundwater is an increasingly important water resource, especially as drought and climate change make other sources of water more scarce. Mapping water table depth (WTD) and understanding the sensitivities of input parameters to WTD are of great use for decision-making, as well as hydrological modeling. As part of my project, I develop a deep learning emulator to predict steady-state WTD across the contiguous United States. I focused on exploring the relationships between WTD, hydraulic conductivity, and precipitation minus evapotranspiration. I created uncertainty distributions for the aforementioned variables by injecting Gaussian noise into the emulator, which enables an assessment of input parameter sensitivities. As I continue working on this project, my long-term goal is to calibrate the simulated steady-state WTD map to observation data using simulation-based inference to create an improved, continuous map of WTD that better matches field observations. Through this opportunity, I gained confidence in my practical understanding of machine learning, and I am now aware of the diverse ways in which machine learning can be applied to different scientific challenges. I also enjoyed teaching at The Watershed Institute’s Water and Climate Academy; it was a fulfilling and enriching experience that reinforced my passion for service.
Water and the Environment
Integrated GroundWater Modeling Center, Department of Civil and Environmental Engineering, Princeton University - Princeton, New Jersey
Reed Maxwell, Professor of Civil and Environmental Engineering and the High Meadows Environmental Institute; Lisa Gallagher, Education and Outreach Manager, High Meadows Environmental Institute