Weather-Regime-Based Heatwave Risk Typing and Urban Climate Resilience Assessment in New Delhi (1997–2016)

Extreme heat across the North Indian Plain has intensified in recent decades, with the temperature in Delhi repeatedly exceeding 48 °C. We present a physically interpretable and computationally efficient typology of heatwave risk using aggregated station observations of daily mean temperature, relative humidity, wind speed, and pressure from 1997 to 2016. Quality-controlled, standardized daily features (PCA-verified) were clustered with k-means; internal validity indices (Silhouette, Calinski–Harabasz, and Davies–Bouldin) identified an optimal partition with k = 3, defining three distinct weather regimes. Coupling these regimes with an absolute heatwave criterion (daily mean ≥30 °C for ≥3 days) revealed a pronounced gradient: a dry–hot, high-pressure regime (41% of days) accounted for 63% of heatwave days (mean 33.4 °C; median duration ≈17 days); a mild–humid background (59%) yielded ~8% incidence; and a rare blocking-driven dry intrusion (<1%) produced heatwaves each time, with mean temperatures of >35 °C and episodes persisting for ≥30 days. Regime–heatwave relationships were statistically significant and robust across sensitivity tests, including variations in k, alternative clustering algorithms, and bootstrap resampling. This four-stage workflow consists of data preparation, feature extraction, regime classification, and heatwave risk attribution and provides a transparent basis for regime-aware early warning, demand-side energy management, and public health protection in Delhi and is transferable to other rapidly urbanizing regions.