Using the BFAST Algorithm and Multitemporal AIRS Data to Investigate Variation of Atmospheric Methane Concentration over Zoige Wetland of China

Abstract

The monitoring of wetland methane (CH4) emission is essential in the context of global CH4 emission and climate change. The remotely sensed multitemporal Atmospheric Infrared Sounder (AIRS) CH4 data and the Breaks for Additive Season and Trend (BFAST) algorithm were used to detect atmospheric CH4 dynamics in the Zoige wetland, China between 2002 and 2018. The overall atmospheric CH4 concentration increased steadily with a rate of 5.7 ± 0.3 ppb/year. After decomposing the time-series of CH4 data using the BFAST algorithm, we found no anomalies in the seasonal and error components. The trend component increased with time, and a total of seven breaks were detected within four cells. Six were well-explained by the air temperature anomalies primarily, but one break was not. The effect of parameter h on decomposition outcomes was studied because it could influence the number of breaks in the trend component. As h increased, the number of breaks decreased. The interplays of the observations of interest, break numbers, and statistical significance should determine the h value.