Reconciling Compensating Errors Between Precipitation Constraints and the Energy Budget in a Climate Model

Abstract

Precipitation microphysics and the effective radiative forcing due to aerosol-cloud interactions (ERF aci) contribute to some of the largest uncertainties in general circulation models (GCMs) and are closely interrelated. This study shows that a sophisticated, two-moment prognostic precipitation scheme can simultaneously represent both warm rain characteristics consistent with satellite observations and a realistic ERF aci magnitude, thus reconciling compensating errors between precipitation microphysics and ERF aci that are common to many GCMs. The enhancement of accretion from prognostic precipitation and accretion-driven buffering mechanisms in scavenging processes are found to be responsible for mitigating the compensating errors. However, single-moment prognostic precipitation without the explicit prediction of raindrop size cannot capture observed warm rain characteristics. Results underscore the importance of using a two-moment representation of both clouds and precipitation to realistically simulate precipitation-driven buffering of the cloud response to aerosol perturbations. Plain Language Summary Aerosol-cloud-precipitation interactions are among the most uncertain processes in general circulation models (GCMs). Many GCMs tend to represent cloud-to-rain conversion too frequently because of the use of simplified schemes for aerosolcloud interactions (ACI). It has been shown that model tuning cannot improve the simulated precipitation characteristics without unrealistically perturbing the energy budget through ACI. This indicates the presence of compensating errors in GCMs. These errors must be understood at a fundamental level and be mitigated in GCMs for reliable predictions of future climate change. This study shows that a sophisticated representation of precipitation, referred to as prognostic precipitation, can effectively mitigate these compensating errors in the model by incorporating precipitation-driven buffering mechanisms in the cloud response to aerosol perturbations. Results indicate that incorporating prognostic precipitation in GCMs leads to more reliable climate simulations. This issue can be attributed, at least in part, to the simplified treatment of precipitation in GCMs (Gettelman et al., 2013; Wang et al., 2012). Most GCMs treat precipitation diagnostically (hereafter referred to as DIAG),

Opportunistic experiments to constrain aerosol effective radiative forcing

“…Further uncertainty then follows by different choices made during the quality checks applied to N d retrievals. This is exemplified by inconsistent estimates of in the subtropical stratocumulus regions ( Michibata and Suzuki, 2020 ; Gryspeerdt et al, 2019a ; Possner et al, 2020 ). These estimates stem from the same retrievals.…”

New insights on the prevalence of drizzle in marine stratocumulus clouds based on a machine learning algorithm applied to radar Doppler spectra

“…However, this study shows the presence of significant drizzle drops during low reflectivity conditions (lower than -20 dBZ) and a lack of considering this may lead to a certain degree of the retrieval uncertainty; 2) Drizzle production mechanisms are widely regarded as a critical missing piece of the puzzle in warm cloud research (Takahashi et al, 2017). Particularly, the parameterization schemes of the autoconversion/accretion processes in numerical models have large variations among each other, leading to significant uncertainty in future climate predictions (Michibata and Suzuki, 2020;Wood, 2005b). The results presented in this study can be used to verify the proposed parameterization schemes by comparing the drizzle climatology.…”

“…(Paluch and Lenschow, 1991;Yamaguchi et al, 2017). A misrepresentation of the autoconversion process in GCM's can affect not only the hydrological cycle but also generate compensating errors in the aerosol-cloud interactions (Michibata and Suzuki, 2020).…”

Snow-induced buffering in aerosol–cloud interactions

“…If needed, these experiments were retuned so that the imbalance of the radiative flux at the TOA remained within 1.0 W m −2 . Model tuning was conducted by modifying the scale factor for accretion rate but not autoconversion for the warm rain process because the latter can influence the magnitude of ACI due to the direct relation to droplet number (Michibata and Takemura, 2015;Jing et al, 2019) and thus the precipitation initiation (Mülmenstädt et al, 2020). This is effective for modifying SW radiation, but if needed, cloud ice and snow processes were also tuned for modifying LW radiation by changing scale factors for the fall speed of hydrometeors, which may be uninfluential on ERF aci because they are not involved directly in the hydrometeor number densities.…”