A possible explanation for the divergent projection of ENSO amplitude change under global warming

Abstract

The El Niño-Southern Oscillation (ENSO) is the greatest climate variability on interannual time scale, yet what controls ENSO amplitude changes under global warming (GW) is uncertain. Here we show that the fundamental factor that controls the divergent projections of ENSO amplitude change within 20 coupled general circulation models that participated in the Coupled Model Intercomparison Project phase-5 is the change of climatologic mean Pacific subtropical cell (STC), whose strength determines the meridional structure of ENSO perturbations and thus the anomalous thermocline response to the wind forcing. The change of the thermocline response is a key factor regulating the strength of Bjerknes thermocline and zonal advective feedbacks, which ultimately lead to the divergent changes in ENSO amplitude. Furthermore, by forcing an ocean general circulation mode with the change of zonal mean zonal wind stress estimated by a simple theoretical model, a weakening of the STC in future is obtained. Such a change implies that ENSO variability might strengthen under GW, which could have a profound socio-economic consequence.

Appendix: A simple theoretical model

Based on a simplified Lindzen–Nigam model (Lindzen and Nigam 1987; Wang and Li 1993), boundary-layer zonal and meridional wind fields may be written as:

$$Eu-fv=A\frac{\partial T}{\partial x}$$

(7)

$$Ev+fu=A\frac{\partial T}{\partial \text{y}}$$

(8)

in which u (v) denotes the mean zonal (meridional) wind field averaged in the boundary layer, T denotes the mean sea surface temperature, f represents the Coriolis parameter, E represents the frictional coefficient (10⁻⁵ s⁻¹), $\text{A}=R\frac{{{p}_{s}}-{{p}_{e}}}{2{{p}_{e}}}$ is the SST-gradient forcing coefficient, R is the gas constant for dry air (287 J K⁻¹ kg⁻¹), p _e denotes the pressure at top of the boundary layer (850 hPa), and p _s denotes the pressure at bottom of the boundary layer (1000 hPa).

Applying a zonal (0°–360°E) average operator to Eqs. (7, 8), we have

$$E[u]-f[v]=0$$

(9)

$$E[v]+f[u]=A\frac{\partial [T]}{\partial y}$$

(10)

where “[]” represents the zonal average. Thus, we can obtain

$$[u]=A\frac{\partial [T]}{\partial y}/(\frac{{{E}^{2}}}{f}+f)$$

(11)

Assuming coefficients A and E are constant in the present-day and future global warming climate states, one may obtain

$$\delta [u]=A\frac{\partial \delta [T]}{\partial y}/\left(\frac{{{E}^{2}}}{f}+f\right)$$

(12)

where δ means the future changes (using GW minus PD).

The change of zonally-averaged mean zonal wind stress (i.e., $\delta [{{\tau }_{x}}]$) may be determined by the following equation

$$\delta [\tau _{x} ] = \rho _{a} C_{D} \left| {\overset{\lower0.5em\hbox{$\smash{\scriptscriptstyle\rightharpoonup}$}}{V} } \right|\delta [u]{\text{ = }}\rho _{a} C_{D} \left| {\overset{\lower0.5em\hbox{$\smash{\scriptscriptstyle\rightharpoonup}$}}{V} } \right|A\frac{{\partial \delta [T]}}{{\partial y}}/\left(\frac{{E^{2} }}{f} + f\right)$$

(13)

where ${{\tau }_{\text{x}}}$ denotes the mean zonal wind stress field, ${{\rho }_{a}}$ denotes the atmosphere density (1.29 kg m⁻³), C _D denotes the drag coefficient (1.5 × 10⁻³), and $\left| {\vec{V}} \right|$ denotes the surface wind speed in PD.