Inter-hemispheric differences in energy budgets and cross-equatorial transport anomalies during the 20th century

Abstract

We analyze the evolution of inter-hemispheric asymmetries in the energy budgets (EBs) and near-surface temperature anomalies during the 20th century, as given in Coupled Model Inter-comparison Project, phase 5 (CMIP5) simulations. We also consider the cross-equatorial energy transports (CET) in the atmosphere and in the oceans, in order to evidence how EB asymmetries affect the redistribution of energy between the two hemispheres. Two different experimental settings have been considered, one including only the spatially homogeneous evolving greenhouse gas forcing (GHG), and another one a realistic superposition of all known evolving forcings (ALL), such as aerosols and volcanic eruptions. This study shows that, according to the CMIP5 models, the response of the climate system to the ongoing forcing during the 20th century has differed substantially from what would have resulted from an increase in GHG concentration alone. In the GHG ensemble the Northern Hemisphere (NH) warms more than the Southern Hemisphere (SH), while both hemispheres exhibit similar and positive EB anomalies at the TOA, mainly due to increasing shortwave absorption and with no significant variations of cross-equatorial energy transports. On the contrary, in the ALL ensemble the two hemispheres warm similarly, while the SH exhibits a positive EB anomaly twice as large as in the NH, due to a reduced LW emission (Outgoing Longwave Radiation, OLR) in the SH, with oceanic CET anomalies directed towards the NH. The EB asymmetry in ALL is ascribed to the asymmetry in OLR changes, which is explained by the different role of clouds in the two hemispheres. The ocean heat content (OHC) tendency per unit surface area is similar in the two hemispheres, so that the asymmetries in ALL EB determine CET changes. We evidence that CET changes in the ALL ensemble are associated with the inter-hemispheric asymmetry in the aerosol forcing, which is stronger in the NH than in the SH. We find no significant relation between CETs and inter-hemispheric near-surface temperature asymmetries in GHG, partly due to the large model spread. Generally, deficits in modeled CET for present-day conditions are not ascribed to forcings and feedbacks, rather they are intrinsic to the models.

Appendix: The computation of ocean heat content

The computation of the vertically integrated Ocean Heat Content (OHC) in CMIP5 models follows from (Palmer and McNeall 2014). It is accomplished by using equation:

$$\begin{aligned} OHC = \int _V \rho C_p \theta _o dV, \end{aligned}$$

(5)

where $\rho$ is a reference density, amounting to 1025 kg/m$^3$ and $C_p$ a specific heat capacity, set to 3985 J/kg/K, $\theta _o$ is the ocean potential temperature. The volume dV is either provided as an output of the model, or it is computed from the surface area multiplied by the thickness of the respective vertical level (cfr. Palmer and McNeall 2014). Equation (5) is used for the computation of OHC in each hemisphere.

The estimation of OHC tendencies has not been performed on all models, given that not all model outputs provided both $\theta _o$ and the volume/area of the ocean grid cell. We thus retrieved OHC tendencies for 21 out of 31 models in ALL, 13 out of 14 models in GHG.

Before computing OHC tendencies in ALL and GHG, one has to account for spurious model drifts, that are a major source of uncertainty for many models. In order to do so, trends in OHC are computed from piControl runs for each model, then are subtracted from the ALL and GHG respective runs.

The rate of change in OHC $\displaystyle \frac{1}{\varSigma } \frac{\partial {OHC}}{\partial t}$, i.e. the flux of energy entering the ocean from its surface $\varSigma$, is then computed by the detrended OHC tendency. In this way fluxes are obtained in W/m$^2$, i.e. per unit area. The amount of energy asymmetry which is not available for $OHT_{eq}$ ($\varDelta OHC$, cfr. Eq. 4) is given by:

$$\begin{aligned} {}\langle \varDelta OHC\rangle =\frac{1}{2\varSigma } \left[ \varSigma _{SH,o} \frac{\partial OHC_{SH}}{\partial t} - \varSigma _{NH,o} \frac{\partial OHC_{NH}}{\partial t} \right] \end{aligned}$$

(6)

where $\varSigma _{SH,o}$ and $\varSigma _{NH,o}$ are the surface areas of the SH and NH oceans, respectively.