Abstract Number: 513 Working Group: Aerosols, Clouds, and Climate
Abstract The radiative balance of the earth system is strongly dependent on the reflectivity of clouds, which arises from the cloud droplet number concentration (CDNC). Aerosol acting as cloud condensation nuclei (CCN) can modulate CDNC by providing additional sites for droplet formation; thus, the aerosol concentration from precursor gas or primary emissions can indirectly perturb climate. Many global climate models include a representation of this microphysical process, but the assessment of the relative influence of specific emissions sources remains a significant computational challenge. In recent years, this information has proven valuable enough to warrant expensive investigations including traditional sensitivity analysis (e.g., emissions toggling) and more sophisticated approaches, such as building model emulators. We present an operationally efficient alternative.
The adjoint of a model allows expedient and accurate determination of the sensitivity of the CDNC of a specific region with respect to each aerosol and precursor emissions simultaneously. For the first time, a global chemical transport model adjoint (Henze et al., 2007) fully coupled with the adjoint of a droplet activation parameterization (Karydis et al., 2012) determines the relative influence of aerosol and precursor gas emissions on CDNC. The now widely used GEOS-Chem adjoint is equipped with ANISORROPIA (Capps et al., 2012) to ensure that the sensitivities of inorganic precursor gases and aerosol are modeled with the same accuracy as the forward GEOS-Chem model. The adjoint of a droplet activation parameterization that models absorption of water on soluble aerosol and adsorption of it on insoluble dust (Kumar et al., 2009) is fully integrated into the GEOS-Chem adjoint framework with a set of prescribed size distributions. This novel system is used to assess the relative influence of global emissions on regions where stratocumulus cloud decks persist.