LES-based evaluation of UAV flight patterns to quantify local scale carbon emissions

Understanding carbon flux processes and controls is crucial to constrain greenhouse gas exchanges of different ecosystems under a changing climate. However, over heterogeneous landscapes (e.g., Arctic permafrost regions), carbon exchange fluxes (CO₂, CH₄) show significant variations even on very small spatial scales. As a result, upscaled carbon fluxes from eddy covariance towers and flux chambers hold the potential to be biased due to their limited spatial representativeness. Therefore, quantifying carbon fluxes at different scales is needed to improve understanding of spatial variability, and the interaction of processes over heterogeneous landscapes. Constraining carbon fluxes with unmanned aerial vehicles (UAVs) carrying greenhouse gas analyzers has the potential to bridge this scaling gap since UAVs allow to monitor large areas with high spatial resolution. Nevertheless, only few guidelines are available on UAV flight strategies and methods to accurately quantify the surface-atmosphere carbon exchange fluxes.

In this study, we conducted synthetic UAV flights using a Large Eddy Simulation (LES) model to evaluate various carbon flux quantification methods based on UAV observations. These methods, including e.g. mass balance and flux gradient approaches, were tested with different flight strategies to find the optimum setup that maximizes information gain for a given flight time. In addition, we conducted experiments to improve the accuracy of the UAV-based carbon flux estimation from a campaign conducted in a subarctic heterogeneous ecosystem in which the UAV platform was able to collect in-situ atmospheric CO₂ and CH₄ concentrations, and environmental parameters such as 2D wind speed, air temperature, humidity, and pressure. Replicating these UAV flight strategies within the LES model enabled us to quantify the uncertainties and provide guidelines for future UAV flight campaigns in heterogeneous landscapes.