Atmospheric Sensing in Tropical Forests Using Unmanned Aerial Vehicles

Abstract

Tropical regions play indispensable roles in the energy balance, biogeochemical cycles, and air quality at regional and global scales. The Amazon basin attracts great research interest because of its immense size and complex biosphere-atmosphere interactions. Atmospheric studies benefit from advance in technologies in the past few decades, such as mass spectrometry and satellite. Observational platforms for investigating intermediate-scale phenomena like river winds, however, have been lacking. Local atmospheric recirculation flows (i.e., river winds) induced by thermal contrast between wide Amazon rivers and adjacent forests can affect pollutant dispersion. The transport of urban, fire, and forest pollutant emissions by river winds has implications for the health of the millions of river-side inhabitants. The combination of miniature sensor and unmanned aerial vehicles (UAVs) opens a new horizon for atmospheric sensing. Herein, different types of low-cost sensors were evaluated and deployed on UAVs to collect data in tropical forests. Commercial metal oxide semiconductor (MOS) sensors were optimized and evaluated for detection of trace-level biogenic volatile organic compounds (BVOC). Selectivity among major BVOC species was evaluated and improved with gas separation membrane upstream of a MOS sensor. Mixed matrix membranes composed of metal organic framework and polymer were able to separate a chemically similar mixture of isoprene and α-pinene. Electrochemical sensors and other meteorological sensors were also pre-calibrated and deployed on the UAVs in the tropical forest. Daytime vertical profiles of meteorological variables and chemical concentrations up to 500 m were collected by a copter-type unmanned aerial vehicle (UAV) operated from a boat during the 2019 dry season. Cluster analysis showed that a river-forest recirculation flow occurred from surface to 300 m for 23% (13 of 56) of the profiles whereas it was completely absent for 21% of them. In fair weather, the thermally driven river winds fully developed for synoptic wind speeds below 4 m s 1, and during these periods the vertical profiles of carbon monoxide (CO) and total oxidants (Ox, defined as O3 + NO2) were significantly altered. Under some conditions, the river winds recirculated pollution back toward the riverbank and its point of origin. Numerical experiments through large-eddy simulation were validated for river winds simulation with observational data. Simulation results supported that under some conditions river winds can enhance pollutant mixing perpendicular to the river orientation even when the synoptic-scale winds are parallel. The impacts on interpretation of riparian surface measurement depend on locations and should be considered carefully in the future. The results of this dissertation advance an urgent need to quantify the occurrence and properties of river winds in respect to mesoscale chemical dispersion, air quality, and human health. There are significant implications for the many human settlements along the rivers throughout northern Brazil.