Use Case KJS-CES-01 · Climate, Energy & Sustainability Data partner: India Meteorological Department, Pune
Heatwave Intelligence Platform
Somaiya Vidyavihar University
Phase V · Decision support

Heatwave Watch Dashboard

Region-wise severity grading against the local long-period normal, spatial hotspot clusters, and the population currently exposed to heat-wave-grade conditions or worse.

Highest forecast Tmax
46.4°C
Interior Peninsula
Largest departure
+7.0°C
Above long-period normal
Regions on alert
4/ 7
Heat wave grade or above
Population exposed
9.9cr
Within alerted regions

Hotspot map

Schematic view; marker value is the hotspot cluster count

Normal Caution Heat Wave Severe

Forecast maximum temperature by region

Bars are coloured by the severity class assigned to each region

Region-wise heatwave watch Severity graded per IMD criteria against the region long-period normal
IMD region Forecast Tmax Departure Severity Hotspots Exposed
Severity classification scale Thresholds applied by the classification model
ClassCriterion

How severity is assigned

Each grid cell is compared against its own long-period normal for the same calendar window. A region is escalated when a contiguous cluster of cells crosses the threshold, not when a single station spikes, which suppresses false alarms from instrument error.

Hotspot clusters are detected with a Getis-Ord Gi* statistic over the residual field, then grouped with DBSCAN so that adjacent significant cells are reported as one actionable area rather than as scattered points.

See the models behind these numbers