Texas State tapped by DOE to improve climate change forecasting through use of machine learning

Salah A. Faroughi

Salah A. Faroughi, an assistant professor in the Ingram School of Engineering at Texas State University, has been awarded a U.S. Department of Energy (DOE) grant to accelerate climate change forecasting using advanced machine learning.

The $150,000 grant will support Faroughi’s project, “ESMs Latent Space Exploration for Uncertainty Quantification and Spatiotemporal Downscaling.”

Advanced Earth System Models (ESMs) have been extensively used to forecast the rapidly changing climate, through developing quantitative understanding of the multiscale dynamics of the climate in response to various factors. Despite significant advancements, ESMs require large amounts of computing power and cannot accurately forecast future variations in global or regional climate systems due to two significant challenges: the presence of large knowledge and data uncertainties in scenario testing and predictions; and an inability to resolve mesoscale processes or fully account for the environmental physics that can take anywhere from seconds to centuries to fully manifest on scales ranging from the hyperlocal to global.

Faroughi’s research team will address these issues by leveraging advanced unsupervised machine learning techniques to select representative ESMs (RESMs) from a larger set without a priori knowledge while preserving the uncertainty domain. This approach is expected to improve and accelerate uncertainty quantification and risk analysis through the extraction of principal criteria. Faroughi will then analyze the RESMs to extract hybrid dynamical-statistical downscaling models that map important fine-scale variations in climate models without incurring additional computing expenses.

Drought, megafires, rising sea levels, eroding shorelines, loss of wetlands and deterioration of water quality have all led to widespread concerns about climate change in the world. Improved climate change forecasting is important to enable accurate assessments and informed decision making about climate risks.