Resolved coastal dynamics reduce the Peru coastal warm SST bias in an eddy-resolving ICON simulation

The ocean off the coast of Peru is a place of paradox. While the sun beats down on the surface, the water remains surprisingly cold, a chill that sustains one of the most productive fishing grounds on Earth. This coldness is not a accident of the weather but a result of a constant, invisible engine: winds blowing along the shore push the surface water away, allowing deep, frigid water to rise and replace it. This process, known as upwelling, creates a sharp boundary between the cold coastal water and the warmer ocean further out. For decades, scientists have tried to simulate this delicate system on computers to understand how the climate works and how it might change. However, their models have consistently failed in this specific region, predicting water that is far too warm. This error is not just a minor glitch; it throws off the entire model's prediction of rainfall, clouds, and ocean currents across the Pacific, making it difficult to trust forecasts for the future.

A new study by researchers at the Max Planck Institute for Meteorology has finally identified why these computer models have been so wrong and how to fix them. By running a simulation with a much finer level of detail than ever before, the team discovered that the error stems from how the computer sees the shape of the land. In older, coarser models, the steep Andes mountains and the jagged coastline were smoothed out, like a low-resolution photograph where sharp edges become blurry. This blurriness weakened the winds that drive the cold water up. The new, high-resolution simulation, which acts like a high-definition camera for the atmosphere and ocean, captured the true sharpness of the terrain. This realistic shape allowed the winds to blow stronger and more directly along the coast, which in turn pulled up more cold water from the depths, correcting the temperature error. The result is a model that finally gets the cold coastal waters right, revealing that the key to solving this long-standing climate puzzle was simply seeing the world with greater clarity.

The researchers compared two versions of their climate model, the ICON system. The first version was a standard configuration with a grid spacing of about 160 kilometers in the atmosphere and 40 kilometers in the ocean. In this version, the computer could not "see" the fine details of the South American coast. The second version was an eddy-resolving simulation, a term that describes a model fine enough to capture swirling ocean currents and small-scale weather patterns, with grid spacings of just 10 kilometers in the air and 5 kilometers in the sea. When they ran both models over the same historical period, the difference was stark. The coarse model showed the coastal water as being nearly 2.4 degrees Celsius warmer than reality, a massive error that mirrored the failures of almost all other global climate models. The high-resolution model, however, flipped this result. Instead of being too warm, it showed the water to be about 0.7 degrees Celsius too cold. While this is still not perfect, it represents a dramatic improvement, shifting the model from a state of significant error to one that is remarkably close to the observed reality.

To understand exactly what changed, the team broke down the physics of the coastal system. They found that the coarse model smoothed the Andes mountains and the coastline, which weakened the winds blowing parallel to the shore. In the real world, these winds are strong and focused, pushing surface water out to sea and forcing deep water to rise. In the coarse model, the winds were weaker and spread out over a wider area, failing to generate enough upwelling. The high-resolution model, by contrast, resolved the steep mountains and the narrow coast with precision. This realistic geography funneled the winds, making them stronger and more concentrated right along the shore. The stronger winds pushed the surface water away more effectively, which triggered a much stronger upwelling of cold, deep water. This process cooled the surface layer, directly counteracting the warming bias that had plagued previous simulations.

The study also examined what happens to the heat once it enters the ocean. In both the old and new models, the sun and the atmosphere actually try to warm the surface water. The surface heat exchange acts as a warming force, pushing the temperature up. The only thing that keeps the water cold is the ocean itself. The high-resolution model showed that the ocean's internal movements—specifically the horizontal flow of water and the vertical rising of deep water—create a powerful cooling effect that overpowers the warming from the sun. In the coarse model, this cooling mechanism was too weak to fight the warming, leading to the hot bias. In the new model, the cooling was strong enough to win, though it was so effective that it slightly overcorrected, leaving the water a bit colder than it should be. This suggests that the high-resolution model has successfully restored the cooling engine, but perhaps turned it up just a little too high.

The researchers also looked at the role of clouds and short-term weather fluctuations. They found that while clouds do reflect sunlight and cool the ocean, and while swirling eddies do move heat around, neither of these factors was the primary reason for the improvement. The coarse model had weak clouds, and the high-resolution model had better ones, but the difference in cloud cover was not the main driver of the temperature fix. Similarly, the swirling eddies in the high-resolution model were active and complex, but they mostly just shuffled heat around locally rather than providing the steady, large-scale cooling needed to fix the bias. The true hero of the story was the mean ocean circulation—the steady, large-scale flow driven by the wind. The high-resolution model simply allowed this flow to work as nature intended, creating a realistic chain reaction from the shape of the mountains to the strength of the wind, to the rise of the cold water, and finally to the temperature of the sea.

This discovery has broader implications for how we understand climate models. The study shows that the long-standing error in the Peru coast was not due to a single missing piece of physics, but to the inability of coarse grids to represent the coupled relationship between the land, the wind, and the ocean. When the land is smoothed out, the wind fails, the ocean stops rising, and the water gets too warm. By sharpening the view of the world, the model fixes the entire chain. The researchers noted that this improvement is not unique to their specific model; other high-resolution simulations show similar patterns, suggesting that this is a general rule for climate modeling. As models continue to get finer, the warm bias in the eastern Pacific is likely to disappear, replaced by a more accurate, and perhaps slightly cold, representation of the coast. This shift marks a turning point where the challenge is no longer just to generate enough cooling, but to balance it perfectly against the warming forces of the sun and the atmosphere.

The study concludes that resolving the detailed topography of the coast is central to fixing the climate of the eastern Pacific. The Peru coastal upwelling system is a narrow band, but its influence stretches far beyond, affecting cloud cover, rainfall, and the temperature of the entire tropical Pacific. By getting this small region right, the high-resolution model improves the global picture. The researchers acknowledge that their simulation is still a work in progress, as it slightly overcools the water, indicating that the restored cooling engine is perhaps a bit too strong. However, the path forward is clear: the key to a better climate model lies in capturing the sharp, real-world details of the Earth's surface and the complex dance of air and water that follows. This study provides a concrete demonstration that when we look closer at the world, the answers to our biggest climate puzzles often become visible.