Editing IPCC:AR6/SROCC/Chapter-6 (section)

=== 6.6.1 Key Processes and Feedbacks, Observations, Detection and Attribution, Projections ===

<div id="section-6-6-1key-processes-and-feedbacks-observations-detection-and-attribution-projections-block-1"></div>

In the last two decades, total water transport from the Pacific to the Indian Ocean and the Indian Ocean to the Atlantic Ocean has increased ''(high confidence'' ). Increased ITF has been attributed to Pacific cooling and basin-wide warming in the Indian Ocean. The ITF annual average is 15 x 10 6 m 3 s –1 (Susanto et al., 2012). ITF varies from intraseasonal to decadal time scales. On seasonal time scale, South China Sea Throughflow controls freshwater flux and modulates the main ITF (Fang et al., 2010 <sup>[[#fn:r606|606]]</sup> ; Susanto et al., 2013 <sup>[[#fn:r607|607]]</sup> ; Lee et al., 2019 <sup>[[#fn:r608|608]]</sup> ; Wang et al., 2019 <sup>[[#fn:r609|609]]</sup> ; Wei et al., 2019 <sup>[[#fn:r610|610]]</sup> ). During the extreme El Niño of 1997–1998, the ITF transport was reduced to 9.2 x 10 6 m 3 s –1 . Based on observations and proxy records from satellite altimetry and gravimetry, in the last two decades, 1992–2012, ITF has been stronger (Sprintall and Revelard, 2014 <sup>[[#fn:r611|611]]</sup> ; Liu et al., 2015a <sup>[[#fn:r612|612]]</sup> ; Susanto and Song, 2015 <sup>[[#fn:r613|613]]</sup> ), which translates to an increase in ocean heat-flux into the Indian Ocean (Lee et al., 2015b <sup>[[#fn:r614|614]]</sup> ). Exchanges of heat and fresh water between ocean basins are important at the global scale (Flato et al., 2013 <sup>[[#fn:r615|615]]</sup> ). ITF may have played a key role in the slowdown of the Pacific SST warming during 1998–2013, and the rapid warming in the surface and subsurface Indian Ocean during this period (Section 6.5.1.2; Makarim et al., 2019 <sup>[[#fn:r616|616]]</sup> ), by transferring warm water from the western Pacific into the Indian Ocean (Lee et al., 2015b <sup>[[#fn:r617|617]]</sup> ; Dong and McPhaden, 2018 <sup>[[#fn:r618|618]]</sup> ).

Under 1.5°C warming both El Niño and La Niña frequencies may increase (see Section 6.5) and hence ITF variability may also increase. ITF is also influenced by the IOD events, with an increase in transport during a positive IOD and vice-versa during a negative IOD event (Potemra and Schneider, 2007 <sup>[[#fn:r619|619]]</sup> ; Pujiana et al., 2019 <sup>[[#fn:r620|620]]</sup> ). Positive IODs are projected to increase threefold in the 21st century as a response to changes in the mean state rather than changes in the El Niño frequency (Section 6.5.1.2; Cai et al., 2014b <sup>[[#fn:r621|621]]</sup> ) and this may have an impact on the ITF, additional to the changes due to increasing extreme ENSO events. In response to greenhouse warming, climate models predict that on interannual time scale, it is ''likely'' that the mean ITF may decrease due to wind variability (Sen Gupta et al., 2016), but recent observation trend tends to strengthen which has led to speculations about the fidelity of the current climate models (Chung et al., 2019 <sup>[[#fn:r622|622]]</sup> ). On multidecadal and centennial timescales, it is ''likely'' that mean ITF decreases which is not associated with wind variability but due to reduction of net deep ocean upwelling in the tropical South Pacific (Sen Gupta et al., 2016; Feng et al., 2017 <sup>[[#fn:r623|623]]</sup> ; Feng et al., 2018 <sup>[[#fn:r624|624]]</sup> ). Due to a lack of long-term sustained ITF observations, their impacts on Indo-Pacific climate varibility, biogeochemisty, ecosystem as well as society are not fully understood.

Pacific SST cooling trends and strengthened the equatorial Pacific trade winds have been linked to anomalously warm tropical Indian and Atlantic oceans. The period following the mid-1990s saw a marked strengthening of both the easterly trade winds in the central equatorial Pacific (Figure 6.7) and the Walker circulation (L’Heureux et al., 2013; England et al., 2014 <sup>[[#fn:r625|625]]</sup> ). Both the magnitude and duration of this trend are large when compared with past variability reconstructed using atmosphere reanalyses. (The 1886–1905 extreme weakening trend is poorly constrained by observations and we note the disparity between reanalysis products going back in time.) Moreover, it is very unusual when model simulations are used as an estimate of internal climate variability (Figure 6.7; England et al., 2014 <sup>[[#fn:r629|629]]</sup> ; Kociuba and Power, 2015 <sup>[[#fn:r630|630]]</sup> ). The slowdown in global surface warming is dominated by the cooling in the Pacific SSTs, which is associated with a strengthening of the Pacific trade winds (Kosaka and Xie, 2013 <sup>[[#fn:r631|631]]</sup> ). This pattern leads to cooling over land and possibly to additional heat uptake by the ocean, although recent studies suggest that ocean heat uptake may even slow down during surface warming slowdown periods (Xie et al., 2016 <sup>[[#fn:r632|632]]</sup> ; von Känel et al., 2017). The intensification of the Pacific trade winds has been related to inter-ocean basin SST trends, with rapid warming in the Indian (see section 6.5.1.2) and Atlantic Oceans both hypothesised as drivers (Kucharski et al., 2011 <sup>[[#fn:r633|633]]</sup> ; Luo et al., 2012 <sup>[[#fn:r634|634]]</sup> ; McGregor et al., 2014 <sup>[[#fn:r635|635]]</sup> ; Zhang and Karnauskas, 2017 <sup>[[#fn:r636|636]]</sup> ). While the extreme event of strengthening trade winds are potentially a result of natural internal variability, a role of anthropogenic contribution has not been ruled out. Nevertheless, the CMIP5 models indicate no general change in trends into the future (Figure 6.7), giving more weight to natural internal variability as an explanation.

<div id="section-6-6-1key-processes-and-feedbacks-observations-detection-and-attribution-projections-block-2"></div>

<span id="figure-6.7"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 6.7'''

<span id="figure-6.7-running-twenty-year-trends-of-zonal-wind-stress-over-the-central-pacific-area-averaged-over-8os8on-and-160oe150ow-in-coupled-model-intercomparison-project-phase-5-cmip5-models-and-three-reanalyses-european-centre-for-medium-range-weather-forecasts-ecmwf-interim-re-analysis-era-interim-dee-et-al.-2011-ecmwf-20th-century-reanalysis-era-20c-poli-et-al.-2016-and"></span>
<!-- IMG CAPTION -->
'''Figure 6.7 | Running twenty-year trends of zonal wind stress over the central Pacific (area-averaged over 8oS–8oN and 160oE–150oW) in Coupled Model Intercomparison Project Phase 5 (CMIP5) models and three reanalyses: European Centre for Medium-Range Weather Forecasts (ECMWF) Interim re-analysis, ERA-Interim (Dee et al. 2011), ECMWF 20th century reanalysis, ERA-20C (Poli et al. 2016), and […]'''
<!-- IMG FILE -->
[[File:9c147c1c3b9157292755a5fbc7d5533a IPCC-SROCC-CH_6_7.jpg]]

Figure 6.7 | Running twenty-year trends of zonal wind stress over the central Pacific (area-averaged over 8oS–8oN and 160oE–150oW) in Coupled Model Intercomparison Project Phase 5 (CMIP5) models and three reanalyses: European Centre for Medium-Range Weather Forecasts (ECMWF) Interim re-analysis, ERA-Interim (Dee et al. 2011 <sup>[[#fn:r626|626]]</sup> ), ECMWF 20th century reanalysis, ERA-20C (Poli et al. 2016 <sup>[[#fn:r627|627]]</sup> ), and the National Oceanic and Atmospheric Administration’s (NOAA) 20th century reanalysis, NOAA 20CR v2c (Compo et al. 2011 <sup>[[#fn:r628|628]]</sup> ). The 66% and 100% ranges of all available CMIP5 historical simulations with Representative Concentration Pathway (RCP)8.5 extension are shown.

<!-- END IMG -->
<div id="section-6-6-1key-processes-and-feedbacks-observations-detection-and-attribution-projections-block-3"></div>

Among the number of potential causes of this decadal variability in surface global temperature, a prolonged negative phase of the Pacific Decadal Oscillation/Interdecadal Pacific Oscillation (PDO/IPO) was suggested as a contributor. Because of the magnitude and duration of this Pacific-centred variability (Figure 6.7), it is identified as an extreme decadal climate event. One line of research has explored the role of the warm tropical Atlantic decadal variability in forcing the trade wind trends and associated cooling Pacific SST trends (Kucharski et al., 2011 <sup>[[#fn:r637|637]]</sup> ; McGregor et al., 2014 <sup>[[#fn:r638|638]]</sup> ; Li et al., 2016b <sup>[[#fn:r639|639]]</sup> ). It appears that climate models may misrepresent this link due to tropical Atlantic biases (Kajtar et al., 2018 <sup>[[#fn:r640|640]]</sup> ; McGregor et al., 2018 <sup>[[#fn:r641|641]]</sup> ) and thus potentially underestimate global mean temperature decadal variability. Nevertheless, there is no indication that such an underestimation of global temperature variability is evident in the models (Flato et al., 2013 <sup>[[#fn:r642|642]]</sup> ; Marotzke and Forster, 2015 <sup>[[#fn:r643|643]]</sup> ). The impact of modes of natural variability on global mean temperature decadal variability remains an active area of research.

In the Indian Ocean, water exits the Indonesian Seas mostly flowing westward along with the South Equatorial Current, and some supplying the Leeuwin Current. The South Equatorial Current feeds the heat and biogeochemical signatures from the Indian Ocean into the Agulhas Current, which transports it further into the Atlantic Ocean. Observations of Mozambique Channel inflow from 2003 – 2012 measured a mean transport of 16.7 x 10 6 m 3 s –1 with a maximum in austral winter, and IOD related interannual variability of 8.9 x 10 6 m 3 s –1 (Ridderinkhof et al., 2010 <sup>[[#fn:r644|644]]</sup> ). A multidecadal proxy, from three years of mooring data and satellite altimetry, suggests that the Agulhas Current has been broadening since the early 1990s due to an increase in eddy kinetic energy (Beal and Elipot, 2016 <sup>[[#fn:r645|645]]</sup> ). Numerical model experiments suggest an intensification of the Agulhas leakage since the 1960s, which has contributed to the warming in the upper 300 m of the tropical Atlantic Ocean (Lübbecke et al., 2015 <sup>[[#fn:r646|646]]</sup> ). Agulhas leakage is found to covary with the AMOC on decadal and multi-decadal timescales and has ''likely'' contributed to the AMOC slowdown (Biastoch et al., 2015 <sup>[[#fn:r647|647]]</sup> ; Kelly et al., 2016 <sup>[[#fn:r648|648]]</sup> ). Meanwhile, climate projections indicate that Agulhas leakage is ''likely'' to strengthen and may partially compensate the AMOC slowdown projected by coarse-resolution climate models (Loveday et al., 2015 <sup>[[#fn:r649|649]]</sup> ).

<span id="impacts-on-natural-and-human-systems"></span>