A Distributed Network of 100 Black Carbon Sensors for 100 Days of Air Quality Monitoring in West Oakland, California.

Ambient particulate matter (PM) pollution is a major environmental health risk in urban areas. Dense networks of low-cost air quality sensors are emerging to characterize the spatially heterogeneous concentrations that are typical of urban settings, but are not adequately captured using traditional regulatory monitors at central sites. In this study, we present the 100×100 BC Network, a 100-day deployment of low-cost black carbon (BC) sensors across 100 locations in West Oakland, California. This 15 km2 community is surrounded by freeways and affected by emissions associated with local port and industrial activities. We assess the reliability of the sensor hardware and data collection systems, and identify modes of failure to both quantify and qualify network performance. We illustrate how dynamic, local emission sources build upon background BC concentrations. BC concentrations varied sharply over short distances (∼100 m) and timespans (∼1 hour), depending on surrounding land use, traffic patterns, and downwind distance from pollution sources. Strong BC concentration fluctuations were periodically observed over the diurnal and weekly cycles, reflecting the impact of localized traffic emissions and industrial facilities in the neighborhood. Overall, the results demonstrate how distributed sensor networks can reveal the complex spatiotemporal dynamics of combustion-related air pollution within urban neighborhoods.

In-Use Performance and Durability of Particle Filters on Heavy-Duty Diesel Trucks.

Diesel particle filters (DPFs) are standard equipment on heavy-duty diesel trucks with 2007 and newer engines in the U.S. This study evaluates the performance and durability of these filters. Black carbon (BC) emission rates from several thousand heavy-duty trucks were measured at the Port of Oakland and Caldecott Tunnel over multiple years as California regulations accelerated the adoption of DPFs. As DPF use increased, fleet-average BC emissions decreased, and emission factor distributions became more skewed. Relative to 2004-2006 engines without filters, DPFs reduced BC emission rates by 65-70% for 2007-2009 engines and by >90% for 2010+ engines. Average BC emission rates for 2007-2009 engines increased by 50-67% in 2015 relative to measurements made 1-2 years earlier. Some trucks in this cohort have become high-emitters, indicating that some DPFs are no longer working well. At the Port, where DPFs were universal in 2015, high-emitting 2007-2009 engines (defined here as emitting >1 g BC kg-1) comprised 7% of the fleet but were responsible for 65% of the total BC emitted. These observations raise concerns about DPF durability and the prospects for fully mitigating adverse effects of diesel particulate matter on human health and the environment.

A New Black Carbon Sensor for Dense Air Quality Monitoring Networks.

Low-cost air pollution sensors are emerging and increasingly being deployed in densely distributed wireless networks that provide more spatial resolution than is typical in traditional monitoring of ambient air quality. However, a low-cost option to measure black carbon (BC)-a major component of particulate matter pollution associated with adverse human health risks-is missing. This paper presents a new BC sensor designed to fill this gap, the Aerosol Black Carbon Detector (ABCD), which incorporates a compact weatherproof enclosure, solar-powered rechargeable battery, and cellular communication to enable long-term, remote operation. This paper also demonstrates a data processing methodology that reduces the ABCD's sensitivity to ambient temperature fluctuations, and therefore improves measurement performance in unconditioned operating environments (e.g., outdoors). A fleet of over 100 ABCDs was operated outdoors in collocation with a commercial BC instrument (Magee Scientific, Model AE33) housed inside a regulatory air quality monitoring station. The measurement performance of the 105 ABCDs is comparable to the AE33. The fleet-average precision and accuracy, expressed in terms of mean absolute percentage error, are 9.2 ± 0.8% (relative to the fleet average data) and 24.6 ± 0.9% (relative to the AE33 data), respectively (fleet-average ± 90% confidence interval).