Spatial Analytics Pipeline for Thermal Trend Diagnostics Across American Megacities: A Multi-Decadal ERA5 Assessment

Rapid urbanization and climate change increase heat exposure in large cities, and reproducible baselines are needed to compare their long-term thermal behavior. We derive 1994–2024 near-surface air temperature (T2m) climatologies and trends for eight American megacities from hourly ERA5 reanalysis. The multiannual mean, hourly 90th percentile, interannual standard deviation, Mann–Kendall test, and Sen’s slope were computed on the native 0.25° grid and weighted by area, with 2km bilinear regridding used only for boundary-conformal cartographic visualization. Compared with one NOAA ISD airport station per city, ERA5 achieved CC=0.92–0.99 and RMSE=1.24–2.04 °C in seven cities, with values of 0.84 and 2.66 °C in Bogotá. Multiannual means ranged from 12.2 (Bogotá) to 23.6 °C (Rio de Janeiro). Warming was robust to serial dependence and the analysis period in New York, Buenos Aires, Mexico City, and Bogotá (+0.032 to +0.045 °C/year), conditional in Los Angeles (+0.048 °C/year; not significant after the Hamed–Rao correction), and not detected in São Paulo, Rio de Janeiro (+0.017 and +0.014 °C/year, p≈0.10), and Lima (+0.001 °C/year, p=0.92), where the 31-year record cannot distinguish these slopes from zero. Robust within-city slope contrasts reached 0.4–0.8 °C per 30 years in Mexico City and Bogotá. Because ERA5 resolves at about 28 km, the results describe city-scale baselines, not intra-urban variability or urban effects.