Detection of carbon monoxide trends in the presence of interannual variability

Abstract

AbstractTrends in fossil fuel emissions are a major driver of changes in atmospheric CO, but detection of trends in CO from anthropogenic sources is complicated by the presence of large interannual variability (IAV) in biomass burning. We use a multiyear model simulation of CO with year‐specific biomass burning to predict the number of years needed to detect the impact of changes in Asian anthropogenic emissions on downwind regions. Our study includes two cases for changing anthropogenic emissions: a stepwise change of 15% and a linear trend of 3% yr−1. We first examine how well the model reproduces the observed IAV of CO over the North Pacific, since this variability impacts the time needed to detect significant anthropogenic trends. The modeled IAV over the North Pacific correlates well with that seen from the Measurements of Pollution in the Troposphere (MOPITT) instrument but underestimates the magnitude of the variability. The model predicts that a 3% yr−1 trend in Asian anthropogenic emissions would lead to a statistically significant trend in CO surface concentration in the western United States within 12 years, and accounting for Siberian boreal biomass‐burning emissions greatly reduces the number of years needed for trend detection. Combining the modeled trend with the observed MOPITT variability at 500 hPa, we estimate that the 3% yr−1 trend could be detectable in satellite observations over Asia in approximately a decade. Our predicted timescales for trend detection highlight the importance of long‐term measurements of CO from satellites.