Population-Centric Optimization of Air Quality Monitoring Networks in Data-Sparse Urban Regions: A Weighted K-Means Approach

Air quality monitoring networks in the Global South often face challenges due to data scarcity and limited resources, which can hinder effective air quality management. This study presents a methodology for optimizing the spatial distribution of air quality monitoring stations in urban areas of the Global South, aiming to improve data representativeness for population exposure and inform evidence-based policies.

Building upon existing monitoring network design literature (Kanaroglou et al., 2005; Gupta et al., 2018), our approach integrates high-resolution population data with a modified K-means clustering algorithm. We combine the center of gravity concept with standard K-means to determine monitor locations, prioritizing areas of high population density. The methodology incorporates a two-tier approach, categorizing areas into low and high population density zones and applying weighted K-means clustering separately to each category. To enhance applicability across diverse urban landscapes, we implement geospatial considerations in distance calculations, addressing limitations of standard Euclidean distance-based methods in geographic coordinate systems.

We applied this methodology to rapidly growing urban centers including Lahore (Pakistan), Lagos (Nigeria), and Dhaka (Bangladesh). Results suggest potential improvements in representing population exposure compared to current monitoring configurations.

Limitations of our approach include its reliance on population data, which may overlook other important air quality determinants. The current method also does not account for land use patterns, emission sources, or meteorology. However, the proposed methodology provides a foundation for further development of air quality monitoring network design, potentially enhancing urban air quality management by optimizing air quality monitor placement in data-sparse regions of the Global South.