Under the Uzès sun: the climate we remember no longer exists

A data scientist analyses Uzès weather.

For generations, the people living around Uzès in southern France have spoken about the weather with the certainty of memory. Summers were always hot, yes — but bearable. Nights cooled down. Rain came often enough. The rhythm of life, agriculture, and architecture had quietly adapted to a climate that felt stable, almost permanent.

Then someone decided to ask a deceptively simple question: what if memory is wrong?

In his article “Under the Uzès Sun: When Historical Data Reveals Climate Change,” data scientist Marc Polizzi turns to an unusual witness — historical weather records. Rather than relying on broad global climate models or abstract international averages, he digs into decades of temperature data collected locally around Uzès. The objective is not political argument or theoretical modeling. It is to let the numbers tell their own story.

At first glance, annual averages do not seem dramatic. Climate change often hides behind statistics because averages smooth away the extremes. A year that is slightly hotter than the previous one does not feel revolutionary. But Polizzi approaches the problem like a detective rather than a casual observer.

He begins comparing temperature records across long time horizons, examining not just averages but patterns: how many very hot days occur each summer, whether heatwaves last longer, whether nighttime temperatures remain elevated, and how frequently temperature thresholds once considered exceptional are now crossed routinely.

The historical baseline becomes crucial. Weather data from previous decades effectively acts as a time machine, showing what “normal” once meant for the region.

And gradually the evidence becomes unmistakable.

The number of extreme heat days has risen steadily. Summers are not simply warmer by a degree or two — they behave differently. Heat events cluster more often. Warm periods extend longer into the season. Critically, nights no longer cool the landscape the way they once did, reducing natural recovery time for ecosystems, crops, buildings, and human bodies.

This matters profoundly in a place like Uzès. The region’s stone houses, vineyards, olive groves, and Mediterranean vegetation all evolved around historical climatic assumptions. Agriculture that functioned reliably for generations begins facing new stress patterns. Water demand rises. Drought cycles intensify.

The deeper lesson of the article is methodological.

Climate change is often debated through ideology, distant scientific reports, or global projections. But local historical data strips away abstraction. It transforms climate change from something happening somewhere else into a measurable shift occurring in one specific place people know intimately.

The sun above Uzès has always shone brightly. What changed is not the sun itself, but the system surrounding it.

And what decades of forgotten numbers quietly reveal is unsettling: the climate people remember no longer exists.

The future is not arriving.

It has already rewritten the past.

Marc Polizzi is a French engineer and co-founder of icCube, a Swiss analytics company. He specializes in data science, software architecture, and using historical data to explore real-world phenomena such as climate change.

Summary created with ChatGPT from the original article on Towards data science.