Delhi’s Urban Canyons Turn Heatwaves Deadly: Simulations Reveal Where the City Burns Hottest

When a heatwave settles over a megacity, it does not strike every neighborhood equally. New research on Delhi, one of the most densely populated tropical cities on Earth, shows that the very shape of the urban fabric—the height of buildings, the width of streets, the density of concrete, and the scarcity of vegetation—can add several degrees Celsius to the heat that residents actually experience. The study, published in Theoretical and Applied Climatology by Deepak Kumar and Ujjwal Kumar of Doon University, used high-resolution atmospheric simulations to demonstrate that urban morphology is a decisive control on how strongly heatwaves are amplified within the city, and that the hottest pockets at night can run more than eight degrees Celsius above their rural surroundings.

The researchers tackled a long-standing weakness in city-scale weather modeling. Most mesoscale atmospheric models treat cities as broad, uniform slabs of urban land, smoothing away the dramatic contrasts between a compact high-rise business district, a sprawling low-rise neighborhood of tightly packed homes, and a leafy institutional campus. For a heterogeneous megacity like Delhi, such simplification can badly misrepresent the surface energy balance—the split between sensible heat that warms the air and latent heat that evaporates moisture—and the turbulent exchanges that govern how the atmospheric boundary layer grows and mixes through the day. To capture these processes, the team coupled the Weather Research and Forecasting (WRF) model with an Urban Canopy Model that incorporates the Building Effect Parameterization (BEP) and the Building Energy Model (BEM), allowing buildings to interact explicitly with radiation, turbulence, and air conditioning waste heat.

The crucial innovation was the use of Local Climate Zones, or LCZs. Developed by urban climatologists Stewart and Oke, the LCZ scheme classifies landscapes into standardized types based on building height and spacing, surface cover, imperviousness, and anthropogenic heat release. Rather than a single generic urban category, the modelers could distinguish compact high-rise zones from open low-rise areas, heavy industry, dense trees, and scrubland. The team prepared two independent LCZ maps for Delhi—one derived from MODIS satellite land cover data and one from the WUDAPT community mapping protocol—and ran four parallel simulations for the extreme pre-monsoon heat season of April through June 2024: a baseline WRF run without urban physics, a WRF-UCM run, and two WRF-UCM runs integrated with the respective LCZ maps.

The 2024 pre-monsoon season provided a brutal natural experiment. Northern India endured one of its most severe heat episodes in recent memory, with dozens of cities recording temperatures above 45 degrees Celsius and Delhi’s weather stations flirting with national records. Against this regional backdrop, the LCZ-based simulations revealed how the city itself stacked additional heat on top of the large-scale heatwave signal. During the most intense episodes, the nocturnal urban heat island intensity exceeded 4 to 6 degrees Celsius across densely urbanized districts, and the most extreme hotspots surpassed 8 degrees Celsius—meaning that after sunset, when rural areas finally began to cool, residents of Delhi’s compact cores remained trapped in air many degrees hotter.

The mechanism behind this nocturnal amplification is a story of stored energy and blocked escape routes. Compact high-density zones, with their deep street canyons and massive concrete surfaces, absorb enormous quantities of solar radiation during the day. After dark, that thermal reservoir is slowly released, while the canyon geometry restricts ventilation and traps outgoing longwave radiation, weakening the radiative cooling that would normally lower air temperatures. By contrast, the simulations showed that open low-rise zones and vegetated areas warmed far less, because their surfaces store less heat, allow more sky exposure for radiative cooling, and permit stronger airflow. The LCZ framework thus translated familiar neighborhood differences into quantifiable thermal outcomes.

To separate the urban contribution from the regional heatwave itself, the authors computed the Added Heat Load, a measure of how much warmer the city is than what the surrounding regional climate would produce. The analysis indicated that urbanization contributed an additional 1.5 to 2.5 degrees Celsius above the regional heatwave signal—a substantial increment when one considers that heat-related mortality risk rises steeply with each additional degree of sustained heat. In other words, even without any change in the large-scale weather pattern, the built environment of Delhi was manufacturing its own extra heat burden on top of an already dangerous episode.

The study also traced how urban morphology reshapes the atmosphere above the city. In compact urban sectors, the enhanced sensible heat fluxes drove stronger turbulent mixing and deeper daytime boundary layers, fundamentally altering the vertical structure of the lower atmosphere. This has consequences beyond temperature: a deeper, more turbulent boundary layer changes how pollutants disperse, how clouds form, and how the night-time stable layer develops. The finding that thermal burden concentrates in compact zones—where building density, waste heat from cooling and traffic, and restricted airflow converge—provides a physical explanation for why heatwave impacts within a single city can vary so dramatically from one district to the next.

Among the four configurations tested, WRF-UCM integrated with the WUDAPT-derived LCZ map most realistically captured Delhi’s urban thermal heterogeneity and the persistence of its hotspots. This result carries practical weight, because WUDAPT is a global, community-driven initiative designed to make LCZ mapping feasible for cities worldwide, including those with limited observational infrastructure. If a freely available LCZ map noticeably improves heat simulation in a tropical megacity, the same workflow could be applied to other rapidly urbanizing regions of South Asia, Africa, and Latin America where heat risk is escalating but dense weather-station networks are scarce. The authors emphasize that LCZ-informed urban canopy parameterization is essential for credible heat-risk assessment in such settings.

The implications for planning are direct. The simulations identify which morphological types—compact, high-density, low-vegetation zones—act as heat amplifiers, and which configurations buffer heat. That knowledge can guide climate-resilient urban design: widening ventilation corridors through dense districts, expanding vegetated and open zones, managing anthropogenic heat emissions, and prioritizing heat-action resources for the neighborhoods the models flag as persistent hotspots. Delhi’s peak power demand hit record levels during the 2024 heatwave, illustrating how thermal burden feeds back into energy stress, and national heat-health programs in India are actively seeking finer-grained risk maps. Model-based LCZ analysis offers exactly that granularity.

More broadly, the study adds to a growing body of evidence that heatwaves and urban heat islands are not merely additive but synergistic: during heatwaves, the contrast between urban and rural energy budgets intensifies, so the city’s own heat island grows stronger precisely when people are most vulnerable. As global heatwave frequency and intensity continue to rise, the Delhi results make clear that the atmosphere over a megacity is shaped as much by architecture as by weather. The buildings a city builds today will help determine how hot its heatwaves become tomorrow—and with LCZ-based modeling, planners now have a quantitative tool to see that future before it arrives.

Subject of Research: LCZ-based WRF urban canopy simulation of heatwave amplification and thermal burden in megacity Delhi

Article Title: Urban morphology controls heatwave amplification and thermal burden in megacity Delhi: insights from LCZ-based WRF-BEP/BEM simulations

Article References: Kumar, D., & Kumar, U. (2026). Urban morphology controls heatwave amplification and thermal burden in megacity Delhi: insights from LCZ-based WRF-BEP/BEM simulations. Theoretical and Applied Climatology, 157(11), Article 694. https://doi.org/10.1007/s00704-026-06631-y

Image Credits: AI Generated

DOI: 10.1007/s00704-026-06631-y

Keywords: urban heat island, heatwave, Delhi, Local Climate Zones, WRF model, urban canopy model, BEP, BEM, urban morphology, boundary layer, thermal burden, WUDAPT