Variability of atmospheric CO2 in Earth System model large-ensemble simulations with an interactive carbon cycle

<p><span lang="en-US">Atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> concentrations have increased from around 280 parts per million (ppm) in 1800 to over 416 ppm in 2020. This is a direct result of increasing anthropogenic emissions of CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> since the industrial era. Nearly half of the emitted anthropogenic CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> is taken up by the ocean and terrestrial ecosystems, while the remaining half remains in the atmosphere, where it is a heat-trapping greenhouse gas. The growth of atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> varies from year to year with inhomogeneous spatial distribution depending on the CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> uptake by the ocean and land. The CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> uptake by the natural sinks and atmospheric growth are affected by the climate variations and the long-term changes; in turn, the variations of the carbon cycle also modulate global climate change. </span><span lang="en-US">The state-of-the-art large ensemble simulations based on Earth System Models (ESMs) prescribe the concentration of atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US">, but the missing interactive response of atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> changes to the CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> fluxes into the ocean and the land hinders the investigation of the variability in atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US">. Furthermore, such simulations will be insufficient to represent the changes in the efficiency of the land and ocean carbon sinks once emissions start to decline. Based on the low-resolution version of the Max Planck Earth System Model v1.2 (MPI-ESM-1.2-LR), we have done a novel set of 30-member ensemble simulations driven by anthropogenic CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> emissions. In such simulations, atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> concentrations are computed prognostically, modulated by the strength of CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> fluxes to the land and the ocean. While general trends in atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> concentrations for different Shared Socioeconomic Pathways (SSP) are well known, trends in its global dispersion and variations within the seasons of each year have not been investigated in ESMs with an interactive carbon cycle. In this project, we use MPI-ESM-1.2-LR large ensemble simulations under four SSP scenarios, i.e., SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5, together with historical runs to analyze changes of atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> concentrations. We focus on seasonal variability and spatial distribution of atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> changes in the presence of internal climate variability. We address two questions: first, what is the temporal evolution of atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> in regard to its seasonal variability by the end of the century following different emission pathways; and second, how does atmospheric CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> evolve spatially (horizontally across the globe and vertically into the stratosphere) in the historical period and future projections until 2100? This study aims to refine our understanding of the spatial and temporal variations of CO</span>_{<span lang="en-US">2</span>}<span lang="en-US"> in support of activities to monitor and verify decarbonization measures.</span></p>