Inequalities in urban air pollution in sub-Saharan Africa: an empirical modeling of ambient NO and NO2 concentrations in Accra, Ghana.

Road traffic has become the leading source of air pollution in fast-growing sub-Saharan African cities. Yet, there is a dearth of robust city-wide data for understanding space-time variations and inequalities in combustion related emissions and exposures. We combined nitrogen dioxide (NO2) and nitric oxide (NO) measurement data from 134 locations in the Greater Accra Metropolitan Area (GAMA), with geographical, meteorological, and population factors in spatio-temporal mixed effects models to predict NO2 and NO concentrations at fine spatial (50 m) and temporal (weekly) resolution over the entire GAMA. Model performance was evaluated with 10-fold cross-validation (CV), and predictions were summarized as annual and seasonal (dusty [Harmattan] and rainy [non-Harmattan]) mean concentrations. The predictions were used to examine population distributions of, and socioeconomic inequalities in, exposure at the census enumeration area (EA) level. The models explained 88% and 79% of the spatiotemporal variability in NO2 and NO concentrations, respectively. The mean predicted annual, non-Harmattan and Harmattan NO2 levels were 37 (range: 1-189), 28 (range: 1-170) and 50 (range: 1-195) µg m-3, respectively. Unlike NO2, NO concentrations were highest in the non-Harmattan season (41 [range: 31-521] µg m-3). Road traffic was the dominant factor for both pollutants, but NO2 had higher spatial heterogeneity than NO. For both pollutants, the levels were substantially higher in the city core, where the entire population (100%) was exposed to annual NO2 levels exceeding the World Health Organization (WHO) guideline of 10 µg m-3. Significant disparities in NO2 concentrations existed across socioeconomic gradients, with residents in the poorest communities exposed to levels about 15 µg m-3 higher compared with the wealthiest (p < 0.001). The results showed the important role of road traffic emissions in air pollution concentrations in the GAMA, which has major implications for the health of the city's poorest residents. These data could support climate and health impact assessments as well as policy evaluations in the city.

Beyond here and now: Evaluating pollution estimation across space and time from street view images with deep learning.

Advances in computer vision, driven by deep learning, allows for the inference of environmental pollution and its potential sources from images. The spatial and temporal generalisability of image-based pollution models is crucial in their real-world application, but is currently understudied, particularly in low-income countries where infrastructure for measuring the complex patterns of pollution is limited and modelling may therefore provide the most utility. We employed convolutional neural networks (CNNs) for two complementary classification models, in both an end-to-end approach and as an interpretable feature extractor (object detection), to estimate spatially and temporally resolved fine particulate matter (PM2.5) and noise levels in Accra, Ghana. Data used for training the models were from a unique dataset of over 1.6 million images collected over 15 months at 145 representative locations across the city, paired with air and noise measurements. Both end-to-end CNN and object-based approaches surpassed null model benchmarks for predicting PM2.5 and noise at single locations, but performance deteriorated when applied to other locations. Model accuracy diminished when tested on images from locations unseen during training, but improved by sampling a greater number of locations during model training, even if the total quantity of data was reduced. The end-to-end models used characteristics of images associated with atmospheric visibility for predicting PM2.5, and specific objects such as vehicles and people for noise. The results demonstrate the potential and challenges of image-based, spatiotemporal air pollution and noise estimation, and that robust, environmental modelling with images requires integration with traditional sensor networks.

Characterisation of urban environment and activity across space and time using street images and deep learning in Accra.

The urban environment influences human health, safety and wellbeing. Cities in Africa are growing faster than other regions but have limited data to guide urban planning and policies. Our aim was to use smart sensing and analytics to characterise the spatial patterns and temporal dynamics of features of the urban environment relevant for health, liveability, safety and sustainability. We collected a novel dataset of 2.1 million time-lapsed day and night images at 145 representative locations throughout the Metropolis of Accra, Ghana. We manually labelled a subset of 1,250 images for 20 contextually relevant objects and used transfer learning with data augmentation to retrain a convolutional neural network to detect them in the remaining images. We identified 23.5 million instances of these objects including 9.66 million instances of persons (41% of all objects), followed by cars (4.19 million, 18%), umbrellas (3.00 million, 13%), and informally operated minibuses known as tro tros (2.94 million, 13%). People, large vehicles and market-related objects were most common in the commercial core and densely populated informal neighbourhoods, while refuse and animals were most observed in the peripheries. The daily variability of objects was smallest in densely populated settlements and largest in the commercial centre. Our novel data and methodology shows that smart sensing and analytics can inform planning and policy decisions for making cities more liveable, equitable, sustainable and healthy.

Spatial modelling and inequalities of environmental noise in Accra, Ghana.

Noise pollution is a growing environmental health concern in rapidly urbanizing sub-Saharan African (SSA) cities. However, limited city-wide data constitutes a major barrier to investigating health impacts as well as implementing environmental policy in this growing population. As such, in this first of its kind study in West Africa, we measured, modelled and predicted environmental noise across the Greater Accra Metropolitan Area (GAMA) in Ghana, and evaluated inequalities in exposures by socioeconomic factors. Specifically, we measured environmental noise at 146 locations with weekly (n = 136 locations) and yearlong monitoring (n = 10 locations). We combined these data with geospatial and meteorological predictor variables to develop high-resolution land use regression (LUR) models to predict annual average noise levels (LAeq24hr, Lden, Lday, Lnight). The final LUR models were selected with a forward stepwise procedure and performance was evaluated with cross-validation. We spatially joined model predictions with national census data to estimate population levels of, and potential socioeconomic inequalities in, noise levels at the census enumeration-area level. Variables representing road-traffic and vegetation explained the most variation in noise levels at each site. Predicted day-evening-night (Lden) noise levels were highest in the city-center (Accra Metropolis) (median: 64.0 dBA) and near major roads (median: 68.5 dBA). In the Accra Metropolis, almost the entire population lived in areas where predicted Lden and night-time noise (Lnight) surpassed World Health Organization guidelines for road-traffic noise (Lden <53; and Lnight <45). The poorest areas in Accra also had significantly higher median Lden and Lnight compared with the wealthiest ones, with a difference of ∼5 dBA. The models can support environmental epidemiological studies, burden of disease assessments, and policies and interventions that address underlying causes of noise exposure inequalities within Accra.