DOE Grant Will Recalibrate Current Climate Change Data

Professors Aritra Dasgupta and Chase Wu from the Department of Data Science in NJIT’s Ying Wu College of Computing are developing software to take more advantage of simulation data used to study climate science through a $145,972 grant from the U.S. Department of Energy (DOE). The project aims to improve existing methods traditionally employed by climate scientists to more reliably predict the effects of cloud simulations related to global warming.

The Research and Development Pilot Program awards over $4 million in funding to colleges and universities under-represented in DOE’s foundational climate, Earth, and environmental science research investments. The grants will help build capacity and achieve the goal of broadening institutional participation in DOE’s science investments.

Dasgupta is the principal investigator for the project, A Scientist-in-the-Loop Data Analytics Framework for Intelligent Simulation Model Tuning and Validation. The expected outcome will be interactive and user-friendly software that carefully combines domain knowledge with data science methods, empowering scientists to focus on model development without worrying about the scale and complexity of simulation data. The findings will have a direct impact on addressing urgent issues of climate change, particularly natural disasters such as hurricanes, tornados and flooding, among other extreme events.

Existing models are “good,” the researchers said, but climate scientists lack effective ways to measure just how good they are and fine-tune model performance. 

“In some sense, selecting the best model, identifying critical parameters, and setting appropriate values in the enormous model and parameter space are just as challenging as looking for a needle in a haystack,” says Dasgupta. 

This is what he and Wu will attempt to drastically improve by developing new technology that combines the predictive power of machine learning with expressive visualizations for enabling scientists to proactively supervise, modify and intervene in fixing their models. 

Both professors have significant experience working with climate scientists at DOE-operated national labs, which was a contributing factor to being awarded the competitive grant. Wu has collaborated on developing machine learning methods with the Brookhaven National Laboratory (BNL). Before joining NJIT, Dasgupta was a senior research scientist at Pacific Northwest Laboratory (PNNL). BNL and PNNL have been part of the recent focus of the Biden-Harris government’s investment in basic sciences.

Dasgupta’s prior work at PNNL has previously been highlighted in EurekAlert!, an online platform operated by the American Association for the Advancement of Science (AAAS). He has produced open-source tools and published his results in leading venues of visualization research such as IEEEVIS and ACM CHI.

“I am grateful for the opportunity to work on technological solutions that can have a deep impact on climate change research. This is the most important problem of our lifetime, and complementing data-driven computation with science communication is imperative,” Dasgupta stated. Working with natural and physical scientists, social science experts, and communications professionals, among others, he views data science as having the power to transform existing modalities of practicing science to affect future change. “The time has come to be responsible.” 

The project will be developed and tested over the next 22 months.