Quantifying the impact of El Niño-driven variations in temperature and precipitation on regional atmospheric CO2 growth rate variations

Quantifying the climatic drivers of variations in atmospheric CO2 observations over a range of timescales is necessary to develop a mechanistic understanding of the global carbon cycle that will enable prediction of future changes. Here, we combine NOAA cooperative global air sampling network CO2 observations, remote sensing data, and a flux perturbation model to quantify the feedbacks between interannual variability in physical climate and the atmospheric CO2 growth rate. In particular, we focus on the differences between the 1997/1998 El Niño and the 2015/2016 El Niño during which atmospheric CO2 increased at an unprecedented rate. The flux perturbation model was trained on data from 1997 to 2012, and then used to predict regional atmospheric CO2 growth rate anomalies for the period from 2013 through 2016. Given gridded temperature anomalies from the Hadley Center's Climate Research Unit (CRU), precipitation anomalies from the Global Precipitation Climatology Project (GPCP), and fire emissions from the Global Fire Emissions Database (GFEDv4s), the model was able to the reproduce regional growth rate variations observed at marine boundary layer stations in the NOAA network, including the rapid CO2 growth rate in 2015/2016. The flux perturbation model output suggests that the carbon cycle responses differed for1997 and 2015 El Niño periods, with tropical precipitation anomalies causing a much larger net flux of CO2 to the atmosphere during the latter period, while direct fire emissions dominated the former. The flux perturbation model also suggests that high temperature stress in the Northern Hemisphere extratropics contributed almost one-third of the CO2 growth rate enhancement during the 2015 El Niño. We use satellite-based metrics for atmospheric column CO2, vegetation, and moisture to corroborate the regional El Niño impacts from the flux perturbation model. Finally, we discuss how these observational results and independent data on ocean air-sea flux anomalies, couched in an empirical model, may be useful for evaluating the fidelity of mechanistic land models.