Cities know exactly what is coming. They are still not built for it.

The most damning detail from Hamburg's Extreme Weather Congress was not the death toll from Europe's summer heat waves. It was that the forecasts were right. Germany's warning systems worked; the science performed precisely as designed. People died anyway - because the gap that kills is no longer between knowledge and prediction, but between prediction and pavement.

This should reframe how the urban profession talks about adaptation. For two decades the field has invested in better climate models, finer-grained risk maps, earlier warnings - the comfortable, fundable end of the problem. The uncomfortable end is retrofit: shading streets, cooling buildings and homes, rewriting building codes, and finding money for maintenance rather than ribbon-cuttings.

This week's evidence elsewhere in this edition points the same way. Reporting from the United States shows hundreds of small, cheap, distributed flood fixes outperforming expectations (Page 4); the sponge-city study in Nature Cities finds drainage investment quietly cutting disease risk, with the largest gains precisely where infrastructure was weakest. The tools that work are unglamorous, incremental and local.

COP31's Turkish presidency has made "resilient cities" a headline theme for Antalya. The temptation will be another declaration. The better outcome would be duller: adaptation finance that flows to municipal balance sheets, measured in retrofitted buildings and shaded streets per year - not in pledges. Cities have run out of excuses built on uncertainty. The thermometer is no longer surprising anyone.