Snapress
Live
Back to all stories
4d ago•⏱️ 2 min read
science

University of Tokyo Researchers Develop Ultra-High Resolution Global Atmospheric Model That Erases Popcornlike Rain

Phys.org Science•Published: Oct 8, 2026 06:30 PM IST
Read full source article
Share:
University of Tokyo Researchers Develop Ultra-High Resolution Global Atmospheric Model That Erases Popcornlike Rain
A research team at the University of Tokyo has developed a groundbreaking global atmospheric model that narrows grid spacing to just 220 meters, successfully eliminating artificial simulation errors known as "popcornlike" rain. By utilizing advanced supercomputing, the model explicitly simulates individual cloud updrafts and downdrafts on a global scale rather than relying on coarser parameterizations. This breakthrough enables far more accurate predictions of extreme local weather events and storm dynamics.
  • Ultra-Fine Grid: The new model divides the atmosphere into sections measuring just 220 meters, or about the length of two soccer fields.
  • Eliminating Simulation Bias: The higher resolution successfully eradicated "popcornlike" rain, a long-standing model error involving unrealistic, intense localized precipitation bursts.
  • Massive Computational Scale: Running the global large-eddy simulation required nearly one trillion three-dimensional grid points across the globe.

The Context & Background

Traditional climate models use coarse grids spanning tens to hundreds of kilometers, forcing scientists to estimate smaller atmospheric phenomena using generalized formulas. As climate change drives increasingly unpredictable and severe weather events worldwide, advanced high-resolution modeling tools have become vital for improving disaster preparedness and early warning systems. This new global large-eddy simulation marks a major milestone in capturing complex atmospheric interactions.

Why it matters:More accurate weather and climate models are crucial for anticipating extreme storms and flooding driven by climate change. Eliminating persistent simulation errors like popcornlike rain allows meteorologists to deliver sharper local forecasts and improve early warning systems, potentially saving lives and mitigating future climate-related disasters.

Enjoyed this 2-minute breakdown?
Share it with your colleagues and friends!
Share:

Original Sources & Coverage (1)

This event has been analyzed and clustered from the following reporting outlets. Click to read their direct coverage:

More in science →