| Summary: | Satellite-based rapid sweeping screening of localized PM<sub>2.5</sub> hotspots at fine-scale local neighborhood levels is highly desirable. This motivated us to develop a random forest–convolutional neural network–local contrast normalization (RF–CNN–LCN) pipeline that detects local PM<sub>2.5</sub> hotspots at a 300 m resolution using satellite imagery and meteorological information. The RF–CNN joint model in the pipeline uses three meteorological variables and daily 3 m/pixel resolution PlanetScope satellite imagery to generate daily 300 m ground-level PM<sub>2.5</sub> estimates. The downstream LCN processes the estimated PM<sub>2.5</sub> maps to reveal local PM<sub>2.5</sub> hotspots. The RF–CNN joint model achieved a low normalized root mean square error for PM<sub>2.5</sub> of within ~31% and normalized mean absolute error of within ~19% on the holdout samples in both Delhi and Beijing. The RF–CNN–LCN pipeline reasonably predicts urban PM<sub>2.5</sub> local hotspots and coolspots by capturing both the main intra-urban spatial trends in PM<sub>2.5</sub> and the local variations in PM<sub>2.5</sub> with urban landscape, with local hotspots relating to compact urban spatial structures and coolspots being open areas and green spaces. Based on 20 sampled representative neighborhoods in Delhi, our pipeline revealed an annual average 9.2 ± 4.0 μg m<sup>−3</sup> difference in PM<sub>2.5</sub> between the local hotspots and coolspots within the same community. In some cases, the differences were much larger; for example, at the Indian Gandhi International Airport, the increase was 20.3 μg m<sup>−3</sup> from the coolest spot (the residential area immediately outside the airport) to the hottest spot (airport runway). This work provides a possible means of automatically identifying local PM<sub>2.5</sub> hotspots at 300 m in heavily polluted megacities and highlights the potential existence of substantial health inequalities in long-term outdoor PM<sub>2.5</sub> exposures even within the same local neighborhoods between local hotspots and coolspots.
|
|---|