Estimation of urban methane concentration from remote sensor data

Abstract

Methane (CH4) is the second more important greenhouse gas (GHG), respecting its potential global warming. Although cities represent only 2% of the global surface, they are responsible for 70% of the GHGs emissions. Thus, it is necessary to study their atmospheric concentration variations to identify the main sources and mitigate their emissions. The main objective of this study is to estimate the CH4 urban concentration using satellite products. To do this, first the atmospheric CH4 concentration was analyzed in 16 sites in the city of Tandil (Argentina) for one year; thus, the observed data could be registered. It was found that in winter and autumn, the concentrations were higher than in summer and spring. Then, the data from Landsat 8 satellite were used to obtain the Normalized Difference Vegetation Index (NDVI) and Land Surface Temperature (LST). Linear regression was applied, taking into account the seasonal CH4 concentration as the dependent variable, and the NDVI and LST as the independent variables. The adjusted R2 was 0.53, and the principal variable that affected the CH4 concentration was NDVI, which is related to urbanization. Finally, the mathematical expression from the regression was applied to obtain CH4 urban concentration, which allows us to analyze the temporal and spatial variations.