OU Junior Leads Study Revealing Urban Climate Model Biases Across U.S. Cities

By

Date

NORMAN, OKLA. – Liam Thompson, a junior in the University of Oklahoma College of Atmospheric and Geographic Sciences, is the lead author of a new study published in Urban Climate. The study was an assessment of a hydroclimate model’s ability to simulate temperature, dew point and wind speed evaluated against ground-level observations.

“Part of the reason we do studies like model evaluations is to figure out the obvious errors and biases in these models,” said Thompson.

The research investigates the performance of CONUS404, a hydroclimate modeling dataset of the contiguous United States. While scientists are interested in understanding how cities modify hydroclimate and ecosystems, CONUS404’s ability to capture conditions in urban areas is uncertain. CONUS404 uses a simplified urban surface representation and does not capture details such as building geometry or diverse types of land cover, factors that impact temperature and humidity. Urban areas are pivotal in shaping regional climate, driven by their unique surface characteristics and concentrated infrastructure.

Thompson’s research was conducted under the mentorship of Chenghao Wang, Ph.D., a professor in OU’s Department of Geography and Environmental Sustainability and School of Meteorology, and supported by undergraduate research funding from their department and the university’s Honors Research Opportunities Program.

Working in collaboration with scientists at the National Science Foundation’s National Center for Atmospheric Research, Thompson analyzed 10 years of hourly observations across dozens of U.S. cities, comparing them with CONUS404 outputs. Importantly, the study also assessed how model performance differs between urban and surrounding rural areas.

The results show biases in the model: a slight warm bias in temperature, a pronounced dry bias in dewpoint, and a systematic underestimation of wind speed in urban settings. These discrepancies varied seasonally and diurnally, revealing patterns that suggest areas where the model could be refined.

“Liam’s work fills a critical gap in how we evaluate high-resolution climate models at the continental scale. His findings help us understand where these models capture urban climate conditions well, and where there is still room for improvement,” said Wang. “That’s very important if we want to further improve how cities are represented in future simulations. It’s opportunities like this that make meaningful undergraduate research possible.”

Thompson was recently named a Goldwater Scholar for his research into CONUS-404. He has also won two conference poster awards for the work.