Response of the Intertropical Convergence Zone to Climate Change: Location, Width, and Strength

Abstract

Purpose of review: The intertropical convergence zone (ITCZ) is a planetary-scale band of heavy precipitation close to the equator. Here, we consider the response of the ITCZ structure to climate change using observations, simulations, and theory. We focus on the substantial yet underappreciated projected changes in ITCZ width and strength, and highlight an emerging conceptual framework for understanding these changes.

Recent findings: Satellite observations and reanalysis data show a narrowing and strengthening of precipitation in the ITCZ over recent decades in both the Atlantic and Pacific basins, but little change in ITCZ location. Consistent with observations, coupled climate models predict no robust change in the zonal-mean ITCZ location over the twenty-first century. However, the majority of models project a narrowing of the ITCZ and weakening mean ascent. Interestingly, changes in ITCZ width and strength are strongly anti-correlated across models.

Summary: The ITCZ has narrowed over recent decades yet its location has remained approximately constant. Climate models project further narrowing and a weakening of the average ascent within the ITCZ as the climate continues to warm. Following intense work over the last ten years, the physical mechanisms controlling the ITCZ location are now well understood. The development of complementary theories for ITCZ width and strength is a current research priority. Outstanding challenges include understanding the ITCZ response to past climate changes and over land versus ocean regions, and better constraining all aspects of the ITCZ structure in model projections.

Electronic supplementary material: The online version of this article (10.1007/s40641-018-0110-5) contains supplementary material, which is available to authorized users.

Keywords: Atmospheric dynamics; Climate change; Intertropical convergence zone; Models; Observations; Theory; Tropical precipitation.

Fig. 1

a Global Precipitation Climatology Project (GPCP; version 2.3) 2.5° × 2.5° annual-mean precipitation climatology from 1979–2017. b Trends in de-seasonalized GPCP monthly-mean precipitation over 1979–2017

Fig. 2

Vertically averaged annual- and zonal-mean meridional streamfunction (700 to 300 hPa with mass weighting) in the historical simulation (1985–2004) for the CNRM-CM5 model. Features of the ITCZ structure are indicated: location (ϕ_ITCZ), width (W_ITCZ), total mass transport (Ψ_ITCZ), and the northern and southern edges (ϕ_N and ϕ_S). The ITCZ strength is defined as ω_ITCZ = −gΨ_ITCZ/A_ITCZ, where the ITCZ area is given by Eq. 3

Fig. 3

Projected changes in ITCZ a location, b width, and c strength for 32 CMIP5 models between the historical (1985–2004) and RCP8.5 (2079–2098) simulations. The red lines indicate the median model changes, the boxes show the interquartile ranges, and the whiskers show the full model ranges. Note that by the small-angle approximation, sinϕ ≈ ϕ close to the equator implying that the fractional changes in ITCZ width shown here are very similar to fractional changes in ITCZ area.

Fig. 4

a Multimodel-median change at each latitude in (minus) annual- and zonal-mean mid-tropospheric vertical velocity between the historical (1985–2004) and RCP8.5 (2079–2098) simulations (black line). The red shading indicates the interquartile range in vertical velocity changes across models at each latitude. The vertical velocities have been vertically averaged (with mass weighting) from 700hPa to 300hPa. The red vertical lines show the multimodel-median northern and southern edges of the ITCZ in the historical simulations, as defined using the mass streamfunction method. b Scatterplot of fractional changes in ITCZ strength versus fractional changes in ITCZ width between the historical and RCP8.5 simulations. The fractional changes in strength and width have been normalized by each model’s global-mean surface-air temperature change. The black dots indicate individual CMIP5 models and the red dot shows the median model changes. The correlation coefficient across models is r = − 0.85