New causal analysis reveals surprising non-factors in Tehran's urban heat islands

A new peer-reviewed study on arXiv (arXiv:2605.24459) proposes a statistical method based on Hotelling's T-square test to causally evaluate contributing factors to urban heat islands (UHI). The research, led by Sadaf Alavi Soltani and colleagues, analyzed 22 municipal districts of Tehran across 19 years (2003-2021). They examined time series of weather conditions (precipitation, NDSI, NDWI), vegetation cover (NDVI, EVI), and urban density (NDBI) against nighttime land surface temperature (LST) as the UHI metric. Unlike typical correlation-based studies, this approach tests whether differences between districts with increasing versus non-increasing UHI trends are significant, providing evidence of causal association.

The results revealed that all districts in Tehran exhibited stable or increasing LST trends. Surprisingly, temporal changes in NDWI (water indices) and NDBI (built-up index) did not show a causal relationship with UHI intensification. Meanwhile, variations in vegetation indices (NDVI, EVI) and precipitation were identified as key causal drivers. This suggests that in Tehran's context, adding water bodies or altering urban density alone may not mitigate UHI as effectively as increasing vegetation cover. The method offers a replicable framework for other cities to identify location-specific causal factors, improving targeted urban climate interventions.