Twenty-first century projected summer mean climate in the Mediterranean interpreted through the monsoon-desert mechanism

Abstract

The term “monsoon-desert mechanism” indicates the relationship between the diabatic heating associated with the South Asian summer monsoon rainfall and the remote response in the western sub-tropics where long Rossby waves anchor strong descent with high subsidence. In CMIP5 twenty-first century climate scenarios, the precipitation over South Asia is projected to increase. This study investigates how this change could affect the summer climate projections in the Mediterranean region. In a linear framework the monsoon-desert mechanism in the context of climate change would imply that the change in subsidence over the Mediterranean should be strongly linked with the changes in South Asian monsoon precipitation. The steady-state solution from a linear model forced with CMIP5 model projected precipitation change over South Asia shows a broad region of descent in the Mediterranean, while the results from CMIP5 projections differ having increased descent mostly in the western sector but also decreased descent in parts of the eastern sector. Local changes in circulation, particularly the meridional wind, promote cold air advection that anchors the descent but the barotropic Rossby wave nature of the wind anomalies consisting of alternating northerlies/southerlies favors alternating descent/ascent locations. In fact, the local mid-tropospheric meridional wind changes have the strongest correlation with the regions where the difference in subsidence is largest. There decreased rainfall is mostly balanced by changes in moisture, omega and in the horizontal advection of moisture.

Appendix: Decomposition of projected precipitation change

The precipitation difference between 21C and 20C climatologies can be decomposed as:

$$\begin{aligned} P^{'} = - \left\langle \omega ^{c} \frac{\partial q^{'}}{\partial p} \right\rangle - \left\langle \omega ^{'} \frac{\partial q^{c}}{\partial p} \right\rangle - \langle \mathbf{v} \cdot \nabla q \rangle ^{'} + E^{'} + q_{res} \end{aligned}$$

(2)

where P is precipitation, $\omega$ is vertical pressure velocity, q is specific humidity, $\mathbf{v}$ is the horizontal wind vector and E is evaporation. q is measured in J/kg by absorbing the latent heat of vaporization L. As in Eq. 1, the angle brackets stand for mass-weighted vertical integrals (i.e. $\frac{1}{g} \int dp$) in the troposphere, superscript $^{'}$ represents the difference between 21C and 20C summer climatology, while superscript $^{c}$ stands for the 20C climatology.

Equation 2 is built following the decomposition of precipitation anomalies shown by Chou et al. (2009) and applied to climatologies differences as in Cherchi et al. (2011) and Alessandri et al. (2014). The first term on the right hand side of Eq. 2 represents the change in precipitation associated with the changes in the atmospheric moisture content (q-term). The second term is the precipitation change due to differences in the vertical pressure velocity ($\omega$-term) and the third term is the change due to differences in the horizontal moisture advection. The balance between those terms and the changes in evaporation has a residual, identified as $q_{res}$ in Eq. 2, that contains also the contribution from non-linear terms.