Estimating Carbon Dioxide Emissions from Cities using In-situ and Remote Measurements on Low Cost Drones

Carbon dioxide (CO₂) emissions, especially from large cities have resulted in the build-up of greenhouse gas concentrations. Existing sensor technology is too expensive for large-scale use and cannot distinguish between anthropogenic and biogenic emissions. Current platforms for measurement are limited in height and stability. In this paper, I develop a low-cost in-situ sensor for direct CO₂ measurements and a spectral imaging system to remotely monitor vegetation phenology. The sensor and camera were installed on a drone for greater mobility and mathematical models used collected data to estimate emission inventories. The low-cost sensor compared favorably with state-of-the-art instruments (correlation factor = 0.99) and exhibited expected annual, seasonal and diurnal trends. The normalized difference vegetation index was computed from spectral data, and correlated with plant transpiration and carbon exchange. The predicted CO₂ emission inventory for Washington DC showed a large contribution from the transportation sector. In summary, I demonstrate a methodology to measure and monitor city-wide CO₂ emissions by combining low-cost in-situ measurements, multispectral imaging, small drone technology and puff modeling.