While a network of low-cost sensors (LCS) boosts the capability in monitoring of air quality as covering a wide area, its fixed positioning hinders pinpointing of an unknown emitting source, its accuracy is variable over time, and its operation requires more periodic calibration than the USEPA's reference monitors. Alternatively, for short-term concentration mapping and regulatory compliance assessment, a drone equipped with a single reference sensor to detect the target air pollutant may provide a highly accurate concentration mapping tool as it offers flexibility to maneuver over hard-to-access terrains while addressing the abovementioned limitations. In this modeling study, we focused on a hypothetical point source emitting a certain air pollutant steadily inside a 1 km by 1 km study area. With respect to true concentrations generated from such source using USEPA's SCREEN3 Gaussian dispersion model and an inverse distance weighing (IDW) approach, we developed R and Python scripts to plot concentration and reliability maps for when a drone-based reference sensor/a stationary network of LCS is deployed to reconstruct the concentration maps. First, we compared the impact of the drone flight path on 1-hr-averaged concentration mapping of an air pollutant for locating a hidden point source or concentration accuracy determination at certain regions. Then, we conducted sensitivity analyses on the drone-based mapping reliability under different number of drone passes, wind speeds, and wind directions. Our results showed that compared to S-, N-, Z-, and diagonal-shape drone travel paths for the same travel total time, the time-averaged mapping accuracy of a zigzag drone flight is the best for pinpointing an unknown point source. For concentration mapping of known air polluting sources, which is intended to provide the utmost accuracy, an increase in the number of drone passes from the source to the downstream receptor points improves reliability of the projected dispersing plume. When flown under a windy condition (wind speeds above 3 m/s), concentration maps are underestimated for the most part, and the bias reduces when the drone travels in parallel to the wind direction. The outcomes of this study shed light on optimized development of the drone-based air quality sensors as proof-of-concept of a field-ready device to facilitate accurate short-term air quality monitoring or regulatory compliance of the air pollutants with hourly thresholds.
Read full abstract