CSIR Study Develops Advanced Model for Delhi's Environmental Noise Mapping

A new study conducted under India's Council of Scientific and Industrial Research (CSIR) has developed a model that combines geographical data with real-time traffic information to map environmental noise across Delhi, one of the world's most traffic-congested cities.

Titled "Integrated land use regression (iLUR) model with road traffic characteristics for environmental noise prediction and mapping in urban regions with heterogeneous traffic conditions," the study aims to overcome the limits of conventional noise-monitoring methods, which struggle with Delhi's highly mixed urban landscape — residential colonies sit alongside markets and commercial establishments, schools and hospitals stand close to busy roads, different types of vehicles share the same roads, and honking is a constant feature of traffic.

To build the model, researchers collected noise observations from 201 locations across Delhi over nine months, from October 2023 to June 2024, to also capture seasonal variation. They combined this with data on hourly traffic counts, vehicle speeds, honking events, roads, land use, population density, built-up areas, metro infrastructure, railway stations, bus stops and traffic crossings, and used it to build four separate models predicting daytime noise, nighttime noise, 24-hour equivalent noise and day-night average noise.

Hourly traffic count was the most influential predictor of noise across all four models, and the number of honking events was also among the most influential predictors, particularly for daytime, 24-hour and day-night noise levels. Spatial variation in noise was strongly associated with residential and major roads, while at night, proximity to roads and railway stations became more important. The researchers also found a negative association between the number of bus stops and noise levels, since roads with frequent bus stops tend to have lower vehicle speeds, while high-speed traffic shifts to alternate routes with fewer stops.

Using the four models, the researchers generated high-resolution noise maps for the New Delhi region, which they say could be compared against Indian noise standards and international guidelines to identify hotspots and target interventions — such as traffic-management measures where traffic volume is the dominant factor, or stronger enforcement where honking contributes significantly.