Earth’s atmosphere undergoes a dramatic transformation each day at sunrise. As sunlight hits the planet, near-surface air is heated and rises upward through the planetary boundary layer, the layer of air that directly interacts with the ground, buildings, vegetation, and people. The rise of warm air transitions the boundary layer from a cool, stable state to a more turbulent one that lasts throughout the day.

This morning transition is a critical window for weather forecasting. It influences how air pollution disperses, when morning fog lifts, the conditions that drive wildfire behavior, and where afternoon thunderstorms might eventually form. But current weather models often struggle to get the timing of this wake-up call right, leading to errors in daily forecasts that can affect everything from aviation safety to air quality alerts. 

A recent study looked into why this is: CIRES researchers and colleagues in NOAA’s Global Systems Laboratory used field observations to understand how the atmosphere truly behaves during the morning transition. They collected data from the Department of Energy’s Atmospheric Radiation Measurement site in Oklahoma, the world’s largest and most extensive atmospheric research facility. They then compared the observations to output from a one-dimensional atmospheric model and found the planetary boundary layer starts mixing about an hour later in the model than it does in the real world. 

The researchers traced part of this delay back to the model’s code that simulates how ground temperatures and water content change. In the model, on calm mornings, the soil warms up too slowly, delaying the moment the ground could begin warming the air. Even small systematic errors like these could have important impacts in weather predictions for hours afterward, according to the study. 

“These model biases lead to a delay in the development of the vertical mixing in the lower atmosphere as the sun rises, which can lead to erroneous predictions of how smoke and other pollutants may disperse during the day,” said CIRES researcher Julia Simonson, the study’s lead author. 

These findings are driving model developers to investigate the cause of these biases and fix them. Their goal is to improve short-term forecasts used for high-impact events like sudden severe storms, wildfires, or air quality alerts. Weather model improvements can better prepare the public for potential health hazards and help officials more effectively route commercial transportation that can be affected by low visibility or turbulence. 

Importantly, understanding how the boundary layer behaves during the morning transition is also crucial for fighting wildland fires: responders and emergency officials need to know exactly how and when conditions will change. Firefighters also need to anticipate boundary layer changes in order to effectively fly uncrewed aerial systems for fire monitoring based on more precise forecasts of the turbulent air mixing that can ground these devices. 

More broadly, by identifying that the ground-to-air connection is a primary source of error, scientists can now refine the math behind how soil and vegetation are represented. Improving these processes means better heatwave predictions, better timing for agricultural planning, and more accurate warnings for emergency responders who rely on knowing exactly when and how the morning air will start to stir.