# Katy Elsam, UCL Geography Chris Brierley, UCL Geography

Abstract

Understanding future tropical circulation is fundamental for adapting to its associated changes in precipitation patterns. Observations indicate that the Hadley circulation is widening, causing enhanced subtropical drying. Until now, research has mostly focused on recent and near-future Hadley cell changes, with little exploration into the additional knowledge from past climate states. Here we explore the Hadley circulation in an ensemble of past, present and future climate simulations. Our results find a systematic widening and weakening of the Hadley cell under future scenarios, with a robust narrowing and strengthening during the Last Glacial Maximum. We find a strong association between the Northern Hemisphere temperature gradient and the strength of its tropical Hadley cell. Despite this connection between the tropical circulation and the Arctic, we find little correlation between two of the most well-known model biases: their underestimation of both summer sea ice loss and the expansion of the Tropics.

## Introduction

\label{sec_intro}

A growing body of literature has shown extensive observational evidence that the outer boundaries of the Hadley circulation have shifted poleward by 1-3° latitude in either hemisphere since at least 1979 (Lucas 2013). The magnitude of such expansion, however, appears consistently weaker under model simulations compared to observations (Bindoff 2013, Tao 2015). Though the degree of tropical expansion and consequential subtropical drying is uncertain, it is expected to continue throughout the 21st century under global warming. Therefore, it is essential to understanding the mechanisms governing the nature of the Hadley circulation (Lu 2007, Frierson 2007, Karnauskas 2014). Expanded Hadley cell(World Map of the Köpp...) latitudinal width translates to broader tropics, with poleward shifts in subtropical dry zones (Nguyen 2015). Ultimately, this may cause desertification to occur further poleward than existing boundaries, adding pressure to existing human settlements and ecosystems under the descending branches of the HC (Johanson 2009).

Causes of recent poleward expansion and weakening of Hadley cell strength are still debated. However suggestions include black carbon (BC) aerosols (Allen 2012, Wang 2007), greenhouse gases (Lu 2007, Frierson 2007, Tao 2015), and stratospheric ozone depletion (Polvani 2011). Here we use Coupled Model Intercomparison Project Phase 5 (CMIP5) and Palaeoclimate model intercomparison project phase 3 (PMIP3) (Taylor 2012) simulations to assess the nature of the Hadley circulation under various models and climatic scenarios.

## Methods

We use output from six participating climate models in the CMIP5 ensemble to assess the nature of the Hadley circulation under different scenarios; namely GISS-E2-R, MIROC-ESM, CCSM4, CNRM-CM5, IPSL-CM5A-LR, CSIRO-Mk3-6-0. The Hadley circulation is characterised by the zonal mean meridional streamfunction, $$\Psi$$, which is calculated from the long-term zonal average of the meridional velocity. Internal variability is suspected to be important for observed trends in the Hadley circulation (Garfinkel 2015, Nguyen 2013). It is assessed by sampling 20-year segments of each models preindustrial control simulations. The strength of the circulation is indicated by the maximum $$\Psi$$ in the N.H. cell, and minimum $$\Psi$$ in the SH cell (Oort 1996). The poleward edges of the Hadley cells are defined as the first latitude in either hemisphere where $$\Psi$$ at 500hPa ($$\Psi_{500}$$) is equal to 0kg/s, poleward of the cell’s maximum strength (Lu 2007, Frierson 2007, Hu 2007). The centre of the circulation identifies as the position of the Intertropical Convergence Zone (ITCZ)(ITCZ). The total width of the Hadley circulation is calculated as the difference between the two poleward edge latitudes  (Tao 2015).

\label{fig_HCabsolute}Zonal mean meridional mass streamfunction, $$\Psi$$. (A) The average circulation between 1981-2000 in the NCEP2 Reanalysis. (B) The multi-model mean circulation in the historical simulations of the PMIP3 model subset over the same time-frame.