Sustained Decadal Warming Phase in the Southwestern Indian Ocean since the mid-1990s

+ Author Affiliations + Find other works by these authors
  • Corresponding author: Jingzhi SU, sujz@cma.gov.cn
  • Funds:

    Supported by the National Key Research and Development Program of China (2016YFA0600602) and National Natural Science Foundation of China (41776039)

  • doi: 10.1007/s13351-020-0112-9
  • Note: This paper has been peer-reviewed and is just accepted by J. Meteor. Res. Professional editing and proof reading are underway. Please use with caution.

PDF

  • Regardless of the slowdown in global warming during the hiatus period, sea surface temperatures (SSTs) in the southwestern Indian Ocean (SWIO) have experienced sustained decadal warming for more than two decades since the mid-1990s. The SWIO SSTs warmed steadily during 1996-2016, causing a warming hot spot of 0.4 K decade−1 in a large region east of Madagascar. An upper-layer heat budget analysis indicated that heat advection by ocean currents was the greatest contributor to the warming of the SWIO SSTs. The existence of an anticyclonic geostrophic current along the western boundary of the SWIO tended to maintain warming by transporting warmer water from the west into the SWIO region. In addition, net positive heat transport by ocean currents also occurred at the southern boundary of the SWIO as the climatological northward transport of cold water from the Southern Ocean weakened. This reduction in northward ocean currents at the surface was caused by local wind stress changes, leading to a southward Ekman current. Below the surface, an anticyclonic geostrophic current pattern existed around the warming center near the southeastern SWIO, which reduced the transport of cold waters from the Southern Ocean and warmed the SWIO. These processes near the two boundaries formed a self-sustaining positive feedback mechanism and favored the maintenance of sustained warming in the SWIO. More attention is needed to analyze sustained long-lasting warming in the SWIO, as it is a unique phenomenon occurring under the background of ongoing global warming.
  • 加载中
  • Fig. 1.  Spatial distribution of the linear trend in four SST datasets in the Indian Ocean region during 1996-2016 (shading; K decade−1). (a) HadISST. (b) OISST. (c) Kaplan SST. (d) ERSST.v5. The robustness of the trend is indicated by a green contour line of “8” for the ratio of the mean value to the standard deviation of 36 trends. The 36 trends are calculated based on the six starting years (1995, 1996, 1997, 1998, 1999, and 2000) and the six ending years (2014, 2015, 2016, 2017, 2018, and 2019). The red contour is the area that passes the 95% significance test for the warming trend. The SWIO basin is indicated by the blue box (54°-92°E, 34°-13°S).

    Fig. 2.  Time series of annual SST anomalies (K) in HadISST (blue), OISST (red), Kaplan SST (yellow), and ERSST.v5 (black), averaged over the SWIO basin. A three-point moving average was applied.

    Fig. 3.  (a) Diagnostic items for the three-dimensional spatially averaged ocean advection terms for the upper 100 m in the SWIO, where the average values are the climatological annual cycle variables, and the anomaly variables are the difference between the mean values in the last 11 years (2006-2016) and those in the previous 11 years (1996-2006). (b) Diagnostic items in the heat budget Eq. (1) (K year−1). The vertical current $ w $ is corrected by the OSCAR dataset, and the air-sea heat flux is the sum of the shortwave radiation flux and longwave radiation flux of the JRA dataset and the latent heat flux and sensible heat flux of the OAFlux dataset.

    Fig. 4.  (a) Diagnostic items for the three-dimensional spatially averaged ocean advection terms for the upper 50 m in the SWIO, where the average values are the climatological annual cycle variables, and the anomaly variables are the difference between the mean values in the last 11 years (2006-2016) and those in the previous 11 years (1996-2006). (b) Diagnostic items in the heat budget Eq. (1) (K year−1). The vertical current $ w $ is corrected by the OSCAR dataset, and the air-sea heat flux is the sum of the shortwave and longwave radiation flux of the JRA dataset and the latent heat flux and sensible heat flux of the OAFlux dataset.

    Fig. 5.  (a) Spatial distribution of sea temperature trend (shading; K decade−1) and climatic current (vectors; m s−1) averaged over 0-50 m in the Indian Ocean. (b) Spatial distribution of sea temperature trend (shading; K decade−1) and climatic current (vectors; m s−1) averaged over 0-100 m in the Indian Ocean. The purple dots indicate the region with positive $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ (K year−1). The red contour is the area that passes the 95% significance test. The SWIO is enclosed by the blue lines between 54° and 92°E and between 34° and 13°S.

    Fig. 6.  (a) Sea temperature trend (shading, with intervals of 0.05 K decade−1; K decade−1) during 1996-2016 and climatic zonal and vertical currents (vectors; m s−1). All the values are averaged within 34°-13°S. The area between the blue dotted lines indicates the location of the SWIO (54°-92°E). (b) Sea temperature trend (shading, with intervals of 0.05 K decade−1; K decade−1) and climatic meridional and vertical currents (vectors; m s−1). All the values are averaged within 54°-92°E. The purple dots indicate the region with positive $ -\overline{{\rm{w}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{z}}} $ (K year−1). The red contour is the area that passes the 95% significance test. The area between the blue dotted lines indicates the location of the SWIO (34°-13°S).

    Fig. 7.  (a) Sea temperature trend (shading, with intervals of 0.05 K decade−1; K decade−1), ocean zonal and vertical current trends (vectors; m s−1 decade−1) during 1996-2016, and climatic sea temperature (contour, with intervals of 2 K; K). All the values are averaged within 34°-13°S. The area between the blue dotted lines indicates the location of the SWIO (54°-92°E). (b) Sea temperature trend (shading, with intervals of 0.05 K decade−1; K decade−1), ocean meridional and vertical current trends (vectors; m s−1 decade−1) during 1996-2016, and climatic sea temperature (contour, with intervals of 2 K; K). All the values are averaged within 54°-92°E. The red contour is the area that passes the 95% significance test, and only current vectors significant at the 95% confidence level are plotted in black; the rest are gray vectors. The area between the blue dotted lines indicates the location of the SWIO (34°-13°S).

    Fig. 8.  (a) Spatial distribution of climatic sea temperature (shading; K) and current trend (vectors; m s−1 decade−1) averaged over 0-100 m in the Indian Ocean. Only current vectors significant at the 95% confidence level are plotted in black, and the rest are gray vectors. (b) Spatial distribution of climatic sea temperature (shading; K) and climatic current (vectors; m s−1) averaged over 0-100 m in the Indian Ocean. The SWIO is enclosed by the blue lines between 54° and 92°E and between 34° and 13°S.

    Fig. 9.  (a) The climatological heat flux (PW; 1 PW = 1015 W) along all four vertical sections, the sea surface section, and the bottom section at 100 m for the SWIO during 1996-2016. (b) Same as (a) but for the anomalous heat flux (PW). The red arrow indicates that the SWIO gains heat, and the blue arrow indicates that the SWIO loses heat.

    Fig. 10.  (a) The climatological heat flux (PW; 1 PW = 1015 W) along all four vertical sections, the sea surface section, and the bottom section at 50 m for the SWIO during 1996-2016. (b) Same as (a) but for the anomalous heat flux (PW). The red arrow indicates that the SWIO gains heat, and the blue arrow indicates that the SWIO loses heat.

    Fig. 11.  Spatial pattern of the trend in sea surface heat flux (W m−2 decade−1, downward positive) in the Indian Ocean region: (a) Net shortwave radiation flux of the JRA dataset. (b) Net longwave radiation flux of the JRA dataset. (c) Net latent heat flux based on the mean value of the JRA and OAFlux datasets. (d) Net sensible heat flux based on the mean value of the JRA and OAFlux datasets. The red contour is the area that passes the 95% significance test. The red (blue) shading indicates that there is heat absorption (loss) in the ocean, and the blue lines indicate the SWIO: 54°-92°E, 34°-13°S.

    Fig. 12.  (a) Spatial distribution of the SSH trend (shading; m decade−1) and the sea surface geostrophic current trend (vectors; m s−1 decade−1) in the Indian Ocean region during 1996-2016. (b) Spatial distribution of the wind stress trend (blue vectors; N m−2 decade−1) and Ekman current trend (black vectors; m s−1 decade−1) in the Ekman layer (set as 50 m) over the Indian Ocean region during 1996-2016, where the light blue arrow is the regression of the wind stress distribution. The red contour is the area that passes the 95% significance test. The current vectors significant at the 95% confidence level are plotted in black, and the rest are gray vectors. Only wind stress vectors significant at the 95% confidence level are plotted in blue. The SWIO region is indicated by the blue lines (54°-92°E, 34°-13°S). The SSH data are obtained from GODAS.

    Fig. 13.  Spatial distribution of sea temperature trends (shading; K decade−1) and current trends (vectors; m s−1 decade−1) at 120 m in the Indian Ocean. The red contour is the area that passes the 95% significance test, and only current vectors significant at the 95% confidence level are plotted in black; the rest are gray vectors. The SWIO is enclosed by the blue lines (54°-92°E, 34°-13°S).

    Fig. 14.  (a) Time series of sea temperature anomalies (K year−1) in the GODAS dataset, averaged over the upper 100 m in the SWIO basin. (b) Time series of sea surface zonal wind stress (N m−2) in the GODAS dataset, averaged over the region of the southern boundary of the SWIO (70°-92°E, 34°-28°S). (c) is the same as (b) but for the meridional ocean current (m s−1) in the upper 100 m. A three-point moving average has been applied.

  • [1]

    Allan, R. P., C. L. Liu, N. G. Loeb, et al., 2014: Changes in global net radiative imbalance 1985-2012. Geophys. Res. Lett., 41, 5588–5597. doi: 10.1002/2014gl060962.
    [2]

    Behera, S. K., and T. Yamagata, 2001: Subtropical SST dipole events in the southern Indian Ocean. Geophys. Res. Lett., 28, 327–330. doi: 10.1029/2000GL011451.
    [3]

    Cao, J., P. Yao, L. Wang, et al, 2014: Summer rainfall variability in low-latitude highlands of China and subtropical Indian Ocean dipole. J. Climate, 27, 880–892. doi: 10.1175/JCLI-D-13-00121.1.
    [4]

    Chambers, D. P., B. D. Tapley, and R. H. Stewart, 1999: Anomalous warming in the Indian Ocean coincident with El Niño. J. Geophys. Res. Oceans, 104, 3035–3047. doi: 10.1029/1998JC900085.
    [5]

    Chen, Z. S., Y. Du, Z. P. Wen, et al., 2019: Evolution of south tropical Indian Ocean warming and the climatic impacts following strong El Niño events. J. Climate, 32, 7329–7347. doi: 10.1175/JCLI-D-18-0704.1.
    [6]

    Chowdary, J. S., C. Gnanaseelan, and S. P. Xie, 2009: Westward propagation of barrier layer formation in the 2006-07 Rossby wave event over the tropical southwest Indian Ocean. Geophys. Res. Lett., 36, L04607. doi: 10.1029/2008GL036642.
    [7]

    Du, Y., and S. P. Xie, 2008: Role of atmospheric adjustments in the tropical Indian Ocean warming during the 20th century in climate models. Geophys. Res. Lett., 35, L08712. doi: 10.1029/2008GL033631.
    [8]

    Han, W. Q., J. Vialard, M. J. McPhaden, et al., 2014: Indian Ocean decadal variability: A review. Bull. Amer. Meteor. Soc., 95, 1679–1703. doi: 10.1175/BAMS-D-13-00028.1.
    [9]

    Hoerling, M. P., J. W. Hurrell, and T. Y. Xu, 2001: Tropical origins for recent North Atlantic climate change. Science, 292, 90–92. doi: 10.1126/science.1058582.
    [10]

    Hu, S. N., and A. V. Fedorov, 2019: Indian Ocean warming can strengthen the Atlantic meridional overturning circulation. Nat. Climate Change, 9, 747–751. doi: 10.1038/s41558-019-0566-x.
    [11]

    Huang, B. Y., P. W. Thorne, V. F. Banzon, et al., 2017: Extended reconstructed sea surface temperature, version 5 (ERSSTv5): Upgrades, validations, and intercomparisons. J. Climate, 30, 8179–8205. doi: 10.1175/JCLI-D-16-0836.1.
    [12]

    Johnson, E. S., F. Bonjean, G. S. E. Lagerloef, et al., 2007: Validation and error analysis of OSCAR sea surface currents. J. Atmos. Oceanic Technol., 24, 688–701. doi: 10.1175/JTECH1971.1.
    [13]

    Kaplan, A., M. A. Cane, Y. Kushnir, et al., 1998: Analyses of global sea surface temperature 1856-1991. J. Geophys. Res. Oceans, 103, 18567–18589. doi: 10.1029/97jc01736.
    [14]

    Kaufmann, R. K., H. Kauppi, M. L. Mann, et al., 2011: Reconciling anthropogenic climate change with observed temperature 1998-2008. Proc. Natl. Acad. Sci. USA, 108, 11790–11793. doi: 10.1073/pnas.1102467108.
    [15]

    Lee, S. K., W. Park, M. O. Baringer, et al., 2015: Pacific origin of the abrupt increase in Indian Ocean heat content during the warming hiatus. Nat. Geosci., 8, 445–449. doi: 10.1038/ngeo2438.
    [16]

    Li, Y. L., W. Q. Han, and L. Zhang, 2017: Enhanced decadal warming of the southeast Indian Ocean during the recent global surface warming slowdown. Geophys. Res. Lett., 44, 9876–9884. doi: 10.1002/2017GL075050.
    [17]

    Liu, W., S. P. Xie, and J. Lu, 2016: Tracking ocean heat uptake during the surface warming hiatus. Nat. Commun., 7, 10926. doi: 10.1038/ncomms10926.
    [18]

    Luo, J. J., W. Sasaki, and Y. Masumoto, 2012: Indian Ocean warming modulates Pacific climate change. Proc. Natl. Acad. Sci. USA, 109, 18701–18706. doi: 10.1073/pnas.1210239109.
    [19]

    Nigam, S., and H. S. Shen, 1993: Structure of oceanic and atmospheric low-frequency variability over the tropical pacific and Indian Oceans. Part I: COADS observations. J. Climate, 6, 657–676. doi: 10.1175/1520-0442(1993)006<0657:sooaal>2.0.co;2.
    [20]

    Onogi, K., J. Tsutsui, H. Koide, et al., 2007: The JRA-25 reanalysis. J. Meteor. Soc. Japan, 85, 369–432. doi: 10.2151/jmsj.85.369.
    [21]

    Rao, S. A., and S. K. Behera., 2005: Subsurface influence on SST in the tropical Indian Ocean: Structure and interannual variability. Dyn. Atmos. Oceans, 39, 103–135. doi: 10.1016/j.dynatmoce.2004.10.014.
    [22]

    Rayner, N. A., D. E. Parker, E. B. Horton, et al., 2002: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res. Atmos., 108, 4407. doi: 10.1029/2002JD002670.
    [23]

    Reynolds, R. W., and T. M. Smith, 1994: Improved global sea surface temperature analyses using optimum interpolation. J. Climate, 7, 929–948. doi: 10.1175/1520-0442(1994)007<0929:IGSSTA>2.0.CO;2.
    [24]

    Roxy, M. K., K. Ritika, P. Terray, et al., 2014: The curious case of Indian Ocean warming. J. Climate, 27, 8501–8509. doi: 10.1175/JCLI-D-14-00471.1.
    [25]

    Saha, S., S. Nadiga, C. Thiaw, et al., 2006: The NCEP climate forecast system. J. Climate, 19, 3483–3517. doi: 10.1175/jcli3812.1.
    [26]

    Saji, N. H., B. N. Goswami, P. N. Vinayachandran, et al., 1999: A dipole mode in the tropical Indian Ocean. Nature, 401, 360–363. doi: 10.1038/43854.
    [27]

    Song, Q., A. L. Gordon, and M. Visbeck, 2004: Spreading of the Indonesian throughflow in the Indian Ocean. J. Phys. Oceanogr., 34, 772–792. doi: 10.1175/1520-0485(2004)034<0772:SOTITI>2.0.CO;2.
    [28]

    Su, J. Z., R. H. Zhang, and H. J. Wang, 2017: Consecutive record-breaking high temperatures marked the handover from hiatus to accelerated warming. Sci. Rep., 7, 43735. doi: 10.1038/srep43735.
    [29]

    Su, J. Z., R. H. Zhang, T. Li, et al., 2010: Causes of the El Niño and La Niña amplitude asymmetry in the equatorial eastern Pacific. J. Climate, 23, 605–617. doi: 10.1175/2009JCLI2894.1.
    [30]

    Tourre, Y. M., and W. B. White, 1995: ENSO signals in global upper-ocean temperature. J. Phys. Oceanogr., 25, 1317–1332. doi: 10.1175/1520-0485(1995)025<1317:ESIGUO>2.0.CO;2.
    [31]

    Webster, P. J., A. M. Moore, J. P. Loschnigg, et al., 1999: Coupled ocean-atmosphere dynamics in the Indian Ocean during 1997-98. Nature, 401, 356–360. doi: 10.1038/43848.
    [32]

    Xie, S. P., H. Annamalai, F. A. Schott, et al., 2002: Structure and mechanisms of South Indian Ocean climate variability. J. Climate, 15, 864–878. doi: 10.1175/1520-0442(2002)015<0864:SAMOSI>2.0.CO;2.
    [33]

    Xie, S. P., K. M. Hu, J. Hafner, et al., 2009: Indian Ocean capacitor effect on Indo-Western Pacific climate during the summer following El Niño. J. Climate, 22, 730–747. doi: 10.1175/2008jcli2544.1.
    [34]

    Yamagata, T., S. K. Behera, J. J. Luo, et al., 2004: Coupled ocean-atmosphere variability in the tropical Indian Ocean. Earth’s Climate: The Ocean-Atmosphere Interaction, Volume 147, C. Wang, S. P. Xie, and J. A. Carton, Eds., American Geophysical Union, Washington, D.C., 189–211, doi: 10.1029/147gm12.
    [35]

    Yu, L. S., X. Z. Jin, and R. A. Weller, 2008: Multidecade Global Flux Datasets from the Objectively Analyzed Air-Sea Fluxes (OAFlux) Project: Latent and Sensible Heat Fluxes, Ocean Evaporation, and Related Surface Meteorological Variables. OAFlux Project Technical Report (OA-2008-01), Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, 64 pp.
    [36]

    Zhang, Y., M. Feng, Y. Du, et al., 2018: Strengthened Indonesian throughflow drives decadal warming in the Southern Indian Ocean. Geophys. Res. Lett., 45, 6167–6175. doi: 10.1029/2018GL078265.
  • [1] WU Guoxiong, MAO Jiangyu, DUAN Anmin, ZHANG Qiong. Current Progresses in Study of Impacts of the Tibetan Plateau on Asian Summer ClimateJournal of Meteorological Research,
    [2] Dao-Yi GONG, Dong GUO, Sang LI, Seong-Joong KIM. Winter AO/NAO Modifies Summer Ocean Heat Content and Monsoonal Circulation over the Western Indian OceanJournal of Meteorological Research,  doi: 10.1007/s13351-017-6175-6.
    [3] Yiwei ZHANG, Guangzhou FAN, Wei HUA, Yongli ZHANG, Bingyun WANG, Xin LAI. Differences in Atmospheric Heat Source between the Tibetan Plateau–South Asia Region and the Southern Indian Ocean and Their Impacts on the Indian Summer Monsoon OutbreakJournal of Meteorological Research,  doi: 10.1007/s13351-017-6042-5.
    [4] . INDIAN OCEAN DIPOLE AND ITS RELATIONSHIP WITH ENSO MODEJournal of Meteorological Research,
    [5] MU Mingquan, LI Chongyin. INDIAN OCEAN DIPOLE AND ITS RELATIONSHIP WITH ENSO MODE*Journal of Meteorological Research,
    [6] TAN Zhemin, FANG Juan, WU Rongsheng. Nonlinear Ekman Layer Theories and Their ApplicationsJournal of Meteorological Research,
    [7] CHAO Jiping, CHAO Qingchen, LIU Lin. The ENSO Events in the Tropical Pacific and Dipole Events in the Indian OceanJournal of Meteorological Research,
    [8] YANG Mingzhu, DING Yihui, LI Weijing, MAO Hengqing, HUANG Changxing. The Leading Mode of Indian Ocean SST and Its Impacts on Asian Summer MonsoonJournal of Meteorological Research,
    [9] ZHANG Tao, WU Guoxiong, GUO Yufu. ENERGY BUDGET BIAS IN GLOBAL COUPLED OCEAN-ATMOSPHERE-LAND MODEL*Journal of Meteorological Research,
    [10] Wu Rongsheng. EFFECTS OF OROGRAPHY AND AIR FLOW IN EKMAN BOUNDARY LAYER*Journal of Meteorological Research,
    [11] She Fengning, Chao Jiping, Wang Lizhi. NUMERICAL EXPERIMENT OF THE INFLUENCES OF HEAT AND WIND STRESS ON THE EQUATORIAL OCEAN CIRCULATION*Journal of Meteorological Research,
    [12] MA Zhanhong, FEI Jianfang, HUANG Xiaogang, CHENG Xiaoping, LIU Lei. Evaluation of the Ocean Feedback on Height Characteristics of the Tropical Cyclone Boundary LayerJournal of Meteorological Research,  doi: 10.1007/s13351-013-0611-z.
    [13] Ni Yunqi, Qian Yongfu, Lin Yuanbi. THE EFFECTS OF SEA SURFACE TEMPERATURE OVER THE EQUATORIAL WESTERN PACIFIC AND THE INDIAN OCEAN ON THE ASIAN SUMMER MONSOONJournal of Meteorological Research,
    [14] ZHAO Shanshan, ZHOU Tianjun, YANG Xiuqun, ZHU Yimin, TAN Yanke, SUN Xuguang. Interdecadal Change of the Relationship Between the Tropical Indian Ocean Dipole Mode and the Summer Climate Anomaly in ChinaJournal of Meteorological Research,  doi: 10.1007/s13351-011-0021-z.
    [15] Junhu ZHAO, Jie ZHOU, Kaiguo XIONG, Guolin FENG. Relationship between Tropical Indian Ocean SSTA in Spring and Precipitation of Northeast China in Late SummerJournal of Meteorological Research,  doi: 10.1007/s13351-019-9026-9.
    [16] Qingchi ZHANG, Tim LI, Jia LIU. Contrast of Evolution Characteristics of Boreal Summer and Winter Intraseasonal Oscillations over Tropical Indian OceanJournal of Meteorological Research,  doi: 10.1007/s13351-019-9015-z.
    [17] Yixuan SHEN, Yuan SUN, Zhong ZHONG, Kefeng LIU, Jian SHI. Sensitivity Experiments on the Poleward Shift of Tropical Cyclones over the Western North Pacific under Warming Ocean ConditionsJournal of Meteorological Research,  doi: 10.1007/s13351-018-8047-0.
    [18] Zanmin XU, Yuan SUN, Tim LI, Zhong ZHONG, Jia LIU, Chen MA. Tropical Cyclone Size Change under Ocean Warming and Associated Responses of Tropical Cyclone Destructiveness: Idealized ExperimentsJournal of Meteorological Research,  doi: 10.1007/s13351-020-8164-4.
    [19] Xu Yinzi, Wu Rongskeng. THE ADJUSTMENT OF WIND TO EKMAN FLOW WITHIN THE PLANETARY BOUNDARY LAYERJournal of Meteorological Research,
    [20] ZHAO Yongping, WU Aiming. Numerical experiments for the influences of sst anomalies over the south china sea-eastern tropical indian ocean on the south china sea summer monsoon onsetJournal of Meteorological Research,
  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Sustained Decadal Warming Phase in the Southwestern Indian Ocean since the mid-1990s

    Corresponding author: Jingzhi SU, sujz@cma.gov.cn
  • Institute of Climate System, Chinese Academy of Meteorological Sciences, Beijing 100081, China
Funds: Supported by the National Key Research and Development Program of China (2016YFA0600602) and National Natural Science Foundation of China (41776039)

Abstract: Regardless of the slowdown in global warming during the hiatus period, sea surface temperatures (SSTs) in the southwestern Indian Ocean (SWIO) have experienced sustained decadal warming for more than two decades since the mid-1990s. The SWIO SSTs warmed steadily during 1996-2016, causing a warming hot spot of 0.4 K decade−1 in a large region east of Madagascar. An upper-layer heat budget analysis indicated that heat advection by ocean currents was the greatest contributor to the warming of the SWIO SSTs. The existence of an anticyclonic geostrophic current along the western boundary of the SWIO tended to maintain warming by transporting warmer water from the west into the SWIO region. In addition, net positive heat transport by ocean currents also occurred at the southern boundary of the SWIO as the climatological northward transport of cold water from the Southern Ocean weakened. This reduction in northward ocean currents at the surface was caused by local wind stress changes, leading to a southward Ekman current. Below the surface, an anticyclonic geostrophic current pattern existed around the warming center near the southeastern SWIO, which reduced the transport of cold waters from the Southern Ocean and warmed the SWIO. These processes near the two boundaries formed a self-sustaining positive feedback mechanism and favored the maintenance of sustained warming in the SWIO. More attention is needed to analyze sustained long-lasting warming in the SWIO, as it is a unique phenomenon occurring under the background of ongoing global warming.

    • Since the end of the 20th century, the warming of global mean sea surface temperatures (SSTs) has slowed, and this change is referred to as the hiatus period (Kaufmann et al., 2011; Allan et al., 2014). As a special case, the entire Indian Ocean has been continuously warming in the upper 700 m since the 1950s (Yamagata et al., 2004; Han et al., 2014). The basin-wide warming in the tropical Indian Ocean has occurred at a faster rate than that in other tropical oceans since the 1950s (Du and Xie, 2008). In particular, the tropical western Indian Ocean has experienced significant warming for more than a century (Roxy et al., 2014), which has contributed to an increase in the global average SST.

      Indian Ocean warming has received much attention because of its global atmospheric impacts. One famous example is its impact on the positive phase shift of the North Atlantic Oscillation (Hoerling et al., 2001). Additionally, Indian Ocean warming is found to influence the Walker Circulation in the tropical Pacific (Luo et al., 2012). More recently, Indian Ocean warming has been proposed to play a role in sustaining the Atlantic meridional overturning circulation (Hu and Fedorov, 2019).

      At the same time, the causes of the changes in the upper Indian Ocean over the past two decades have attracted considerable attention. Generally, it is hypothesized that the accumulated heat in the Pacific Ocean can be transported into the Indian Ocean by the Indonesian through flow (ITF; Song et al., 2004), resulting in a sudden increase in heat content in the Indian Ocean (Lee et al., 2015; Liu et al., 2016). Several regions of the Indian Ocean have undergone significant warming during the last two decades. Li et al. (2017) illustrated that the upper 400 m of heat in the Indian Ocean was concentrated in the southeastern Indian Ocean, forming a warming hot spot of 0.8-1.2 K decade-1 during the hiatus period of 2000-2012. Zhang et al. (2018) emphasized the significant warming in the upper 700 m in the southern Indian Ocean (Africa-115°E, 32°-5°S) during 1998-2015.

      In addition to the abovementioned warming regions in the Indian Ocean, the SST in the southwestern Indian Ocean (SWIO; 54°-92°E, 34°-13°S) has experienced decadal warming for more than two decades since the mid-1990s (Fig. 1), which has rarely been mentioned in previous studies. Although there have been some studies on the phenomena and causes of warming in the Indian Ocean, most of these efforts have focused on the warming of the Indian Ocean as a whole (Han et al., 2014; Li et al., 2017; Zhang et al., 2018) or on the seasonal scale of sea surface warming affected by El Niño (Rao and Behera, 2005; Chowdary et al., 2009; Xie et al., 2002, 2009; Chen et al., 2019). The longer-term warming trend since the 1950s does not show the strongest warming in this SWIO region. Hence, this paper aims to investigate the processes and formation mechanisms causing sustained decadal warming in the SWIO. Such an investigation may help us better understand the characteristics and mechanisms of the SST changes in the Indian Ocean, broaden our understanding of the internal mechanisms of the climate system, and improve climate prediction skills.

      Figure 1.  Spatial distribution of the linear trend in four SST datasets in the Indian Ocean region during 1996-2016 (shading; K decade−1). (a) HadISST. (b) OISST. (c) Kaplan SST. (d) ERSST.v5. The robustness of the trend is indicated by a green contour line of “8” for the ratio of the mean value to the standard deviation of 36 trends. The 36 trends are calculated based on the six starting years (1995, 1996, 1997, 1998, 1999, and 2000) and the six ending years (2014, 2015, 2016, 2017, 2018, and 2019). The red contour is the area that passes the 95% significance test for the warming trend. The SWIO basin is indicated by the blue box (54°-92°E, 34°-13°S).

    2.   Data and methods
    • The National Oceanic and Atmospheric Administration Extended Reconstructed SST version 5 (NOAA ERSST.v5) dataset provides monthly data on SST, with a resolution of 2.0° × 2.0° (Huang et al., 2017). Three other monthly SST datasets are obtained: the Hadley Centre Global Sea Ice and SST (HadISST) with a resolution of 1.0° × 1.0° (Rayner et al., 2002), NOAA Optimum Interpolation SST version 2 (NOAA OISST.v2) with a resolution of 1.0° × 1.0° (Reynolds and Smith, 1994), and Kaplan Extended SST version 2 with a resolution of 5.0° × 5.0° (Kaplan et al., 1998). The Global Ocean Data Assimilation System (GODAS) provides the monthly mean reanalysis dataset of sea temperature (40 levels in the upper 4500 m), currents (40 levels in the upper 4500 m), sea surface height (SSH) and wind stress with a resolution of 1.0° × 0.33° (Saha et al., 2006). Another monthly mean sea surface current data set is the Ocean Surface Current Analysis (OSCAR) dataset with a resolution of 1.0° × 1.0° (Johnson et al., 2007). The Japanese Meteorological Agency-Tokyo (RSMC) provides monthly surface heat flux data with a resolution of 1.25° × 1.25° (Onogi et al., 2007). The Objectively Analyzed air-sea Fluxes (OAFlux) of the Woods Hole Oceanographic Institution (WHOI) provides monthly surface heat flux data with a resolution of 1.0° × 1.0° (Yu et al., 2008). Unless noted otherwise, all anomalies are defined as the departure from the 1981 to 2010 climatological values for each month. The SWIO is defined between 54° and 92°E and between 34° and 13°S.

      To better understand the mechanism that causes the change in SSTAs, an upper-layer heat budget analysis method (Su et al., 2010) was used in this study. The heat budget is formulated as:

      where $ T $ is the upper-layer temperature; $ V=(u,v,w) $ represents the three-dimensional ocean current; $ \triangledown =\left(\dfrac{\partial }{\partial {\rm{x}}}, \dfrac{\partial }{\partial {\rm{y}}}, \dfrac{\partial }{\partial {\rm{z}}}\right) $ represents the three-dimensional gradient operator; $ \left( \right){'} $ represents the anomaly of a variable, which is the difference between the mean values in the last 11 years (2006-2016) and the previous 11 years (1996-2006); $ \overline{\left( \right)} $ represents the climatological annual cycle variables; $ -\left({V}^{{{'}}}\cdot\triangledown \overline{T}+\overline{V}\cdot\triangledown {T}^{{{'}}}\right) $ represents a three-dimensional linearized advection term; $ -({{\rm{V}}}^{{{'}}}\cdot\triangledown {T}^{{{'}}}) $ represents a three-dimensional eddy advection term; $ Q{’}_{net} $ is the net air-sea heat flux anomaly, including solar radiation flux, longwave radiation flux, sensible heat flux, and latent heat flux; $ \rho $ (assumed constant at 1026 kg m-3) and $ {C}_{p} $ (assumed constant at 3850 J (kg K)-1) represent the density and heat capacity of seawater, respectively; $ H $ is the depth of the mixed layer (50 m) or of the upper layer (100 m); and $ R $ is the residual.

      The heat flux formula is defined as:

      where $ \rho $ is the density of seawater (1026 kg m-3), $ u $ is the current velocity, $ T $ is the sea temperature, and $ S $ is the area of a single grid in different sections of the SWIO. Then, the heat flux $ F $ of the total area of the boundary cross sections in the six sections of the SWIO is obtained. The mean value of temperature averaged within the SWIO three-dimensional box is first removed from the sea temperature along each section before these heat flux calculations.

      To analyze the causes of the changes in ocean currents, geostrophic currents and Ekman currents are calculated. The formula for geostrophic currents is as follows:

      The formula for Ekman currents is as follows:

      where gravitational acceleration $ g $ is 9.8 m s−2, the Coriolis parameter $ f=2\varOmega {\rm{sin}}\varphi $ rad s−1 ($ {\rm{\varphi }} $ is latitude), $ h $ is the SSH, $ \tau $ is wind stress, $ \rho $ is the density of seawater (1026 kg m−3), and $ {{\rm{d}}}_{\rm{E}} $ is the depth of the Ekman layer of the SWIO (50 m).

    3.   Results
    • The four observed SST datasets (HadISST, OISST, Kaplan SST, and ERSST.v5) showed general warming across the entire Indian Ocean in recent decades, but the most significant warming trend was centered over the SWIO, with a warming hot spot of approximately 0.4 K decade−1 (Fig. 1a). The SWIO warming pattern obtained here is different from the generally accepted SST modes in the Indian Ocean proposed by previous studies, e.g., the Indian Ocean basin mode (IOBM; Nigam and Shen, 1993; Tourre and White, 1995; Chambers et al., 1999), Indian Ocean dipole (IOD; Saji et al., 1999; Webster et al., 1999), subtropical IOD mode (SIOD mode; Behera and Yamagata, 2001) and subtropical IOD-like mode (SIOD-like mode; Cao et al., 2014). The spatial pattern of sustained warming is not sensitive to the choice of the beginning/ending year when calculating the trend. The time series of SST anomalies (SSTAs) in SWIO (Fig. 2) revealed that the four datasets are consistent. The SSTAs from 1996 to 2016, with no strong interannual variability signal, are in a sustained and decadal warming phase, even during the hiatus period. As the SSTA in 1996 is the lowest value of these neighboring years, the year 1996 is chosen as the beginning of the decadal turning point of the SWIO’s warming. Therefore, a period of 21 years from 1996 to 2016 was selected to study the sustained and decadal phases in the SWIO.

      Figure 2.  Time series of annual SST anomalies (K) in HadISST (blue), OISST (red), Kaplan SST (yellow), and ERSST.v5 (black), averaged over the SWIO basin. A three-point moving average was applied.

    • The warming trend in the SWIO is mainly confined within the upper 100 meters, with a mean value of 0.040 K year−1, and the maximum warming area of the SWIO was found to be approximately 50-70 m. The three-dimensional oceanic advection terms in the SWIO region are calculated [Eq. (1)], where the average values are the climatological annual cycle variables, and the anomaly variables are the difference between the mean values in the last 11 years (2006-2016) and the previous 11 years (1996-2006). According to the heat budget analysis of three-dimensional ocean advection terms in the SWIO (Figs. 3, 4), including linear and nonlinear advection terms of ocean currents, the results show that the major terms that affect sea temperature warming in the upper 100 m are $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}},-\overline{{\rm{w}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{z}}}, -{\rm{u}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{x}}} $ and $ -{\rm{v}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{y}}} $, and those in the upper 50 m are $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}}, -{\rm{u}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{x}}} $ and $ -{\rm{v}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{y}}} $, where $ -{\rm{u}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{x}}} $ and $ -{\rm{v}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{y}}} $ are two key active contributions for local warming in the SWIO, indicating that the mechanism of SWIO warming in these two levels is consistent (Figs. 3b, 4b).

      Figure 3.  (a) Diagnostic items for the three-dimensional spatially averaged ocean advection terms for the upper 100 m in the SWIO, where the average values are the climatological annual cycle variables, and the anomaly variables are the difference between the mean values in the last 11 years (2006-2016) and those in the previous 11 years (1996-2006). (b) Diagnostic items in the heat budget Eq. (1) (K year−1). The vertical current $ w $ is corrected by the OSCAR dataset, and the air-sea heat flux is the sum of the shortwave radiation flux and longwave radiation flux of the JRA dataset and the latent heat flux and sensible heat flux of the OAFlux dataset.

      Figure 4.  (a) Diagnostic items for the three-dimensional spatially averaged ocean advection terms for the upper 50 m in the SWIO, where the average values are the climatological annual cycle variables, and the anomaly variables are the difference between the mean values in the last 11 years (2006-2016) and those in the previous 11 years (1996-2006). (b) Diagnostic items in the heat budget Eq. (1) (K year−1). The vertical current $ w $ is corrected by the OSCAR dataset, and the air-sea heat flux is the sum of the shortwave and longwave radiation flux of the JRA dataset and the latent heat flux and sensible heat flux of the OAFlux dataset.

      It should be noted that the selection of the ocean current dataset and the heat flux dataset can have a great influence on the heat budget results. The values of the OGDAS current are compared with those of OSCAR, which is an independent frequently used dataset. The zonal and meridional current trends of OSCAR $ (u, v) $ are the same as those of GODAS $ (u, v) $ (figures omitted), indicating that the horizontal currents of OSCAR data are consistent with those of GODAS. Furthermore, the vertical current $ w $ at 15 m was determined by calculating the horizontal currents of OSCAR $ (u, v) $, and the $ w $ trend of OSCAR was found to be 1.8 times smaller than that of GODAS at 15 m, indicating that the $ w $ trend of GODAS is larger. Therefore, the $ w $ trend of GODAS at all depths was revised by multiplying by 0.56. In addition, the shortwave and longwave radiation fluxes of the OAFlux dataset are only available until 2009. Therefore, the heat flux of the sum of the JRA(无全称) dataset (shortwave and longwave radiation fluxes) and OAFlux dataset (latent heat flux and sensible heat flux) was used to calculate the air-sea heat budget. In this way, the heat budget analysis is almost in balance. The calculated result for the upper 50 m is affected by the depth level, since the negative value of $ -\overline{{\rm{w}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{z}}} $ in the upper 50 m is due to the maximum warming center being below 50 m and the climatic vertical current being downward ($ \overline{w} $ < 0), which cannot directly heat the upper surface seawater (Fig. 4a).

    • According to the diagnosis results, the major positive terms for the SWIO warming are $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}}, -{\rm{u}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{x}}} $ and $ -{\rm{v}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{y}}} $ for 0-50 m and $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}},-\overline{{\rm{w}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{z}}}, -{\rm{u}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{x}}} $ and $ -{\rm{v}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{y}}} $ for 0-100 m. The physical meaning for those terms is explained as follows.

      The distribution of advection term $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ at 0-50 m showed that the positive value center of $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ is mainly distributed in two regions (Fig. 5a): 1. In the northwestern sector of the SWIO (north of 20°S), where the climatic zonal current flows westward ($ \overline{u} $ < 0), and the anomaly of sea temperature increases with longitude ($ \dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ > 0); therefore, $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ > 0, indicating that the heat anomalies are transmitted from the northeast of the SWIO region via the westward current; 2. in the east-central area (20°-30°S) of the SWIO, where the climatic current flows eastward ($ \overline{u} $ > 0), the SWIO center is the most significant area of SST warming, and its SST anomaly decreases with increasing longitude ($ \dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ < 0); hence, $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ > 0, indicating that the heat anomalies are transmitted from the central sector of the SWIO to the east, which leads to warming in the eastern region. The distribution of $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ in the upper 100 m is similar to the distribution in the upper 50 m (Fig. 5b).

      Figure 5.  (a) Spatial distribution of sea temperature trend (shading; K decade−1) and climatic current (vectors; m s−1) averaged over 0-50 m in the Indian Ocean. (b) Spatial distribution of sea temperature trend (shading; K decade−1) and climatic current (vectors; m s−1) averaged over 0-100 m in the Indian Ocean. The purple dots indicate the region with positive $ -\overline{{\rm{u}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{x}}} $ (K year−1). The red contour is the area that passes the 95% significance test. The SWIO is enclosed by the blue lines between 54° and 92°E and between 34° and 13°S.

      The profile of the sea temperature and current trend (Fig. 6) indicates that the advection term $ -\overline{{\rm{w}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{z}}} $ is negative from 0-50 m and positive from 50-100 m. Since the maximum warming center is at approximately 60 m, the climatic downward vertical current ($ \overline{w} $ < 0) can carry warmer water from near 60 m downward, thus heating the seawater at 60-100 m. The negative value of $ -\overline{{\rm{w}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{z}}} $ in the upper 50 m is due to the maximum warming center being below the level of 50 m; thus, it cannot directly heat the upper surface seawater. Since the warming trend and the climatic downward vertical current ($ \overline{w} $ < 0) in the upper 50 m, it will have a positive contribution to the 60-100 m warming. In addition, the warming of the subsurface layer comes from the upper layer. The warming of 0-100 m will further deepen the accumulation of heat in the subsurface layer below 100 m, but as the depth increases, the warming trend gradually weakens.

      Figure 6.  (a) Sea temperature trend (shading, with intervals of 0.05 K decade−1; K decade−1) during 1996-2016 and climatic zonal and vertical currents (vectors; m s−1). All the values are averaged within 34°-13°S. The area between the blue dotted lines indicates the location of the SWIO (54°-92°E). (b) Sea temperature trend (shading, with intervals of 0.05 K decade−1; K decade−1) and climatic meridional and vertical currents (vectors; m s−1). All the values are averaged within 54°-92°E. The purple dots indicate the region with positive $ -\overline{{\rm{w}}}\dfrac{\partial {{\rm{T}}}^{{{'}}}}{\partial {\rm{z}}} $ (K year−1). The red contour is the area that passes the 95% significance test. The area between the blue dotted lines indicates the location of the SWIO (34°-13°S).

      Climatologically, the South Equatorial Current (SEC) is driven by the southeast trade winds. Relatively warm waters are transported by the SEC from the equatorial region to the SWIO, and the zonal temperature gradient decreases from west to east in the SWIO region. In the linear trend pattern, the anomalous meridional currents in the SWIO are directed southward (Fig. 7b), which reduces the climatic northward transport of cold water from the Southern Ocean into the SWIO region and leads to a net warming effect on the temperature trend in the SWIO (Fig. 7b). The anomalous meridional advection term associated with the mean meridional temperature gradient became positive ($ -{\rm{v}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{y}}}>0 $), which was a major contributor to the temperature warming in the SWIO. In addition, there is a counterclockwise subtropical gyre (STG) in the SWIO region. During 1996-2016, the trend of ocean currents changed, forming a counterclockwise circulation in the western SWIO, and the STG strength was enhanced in the western subtropical Indian Ocean. For the linear trend pattern, there were anomalous eastward currents near the western boundary of the SWIO (Fig. 8a), and the average zonal current anomaly along the western boundary was positive (approximately 4 × 10-4 m s-1), which indicated that the transport of the eastward current was stronger than that of the westward current. These anomalous currents transported the climatological warm water from the west into the SWIO region, warming the upper-layer temperature in this region (Fig. 8a). Hence, the anomalous zonal advection term associated with the climatological zonal temperature gradient became positive ($ -{\rm{u}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{x}}}>0 $). A similar case also existed at the southern boundary of the SWIO. Climatologically, colder water is located in the Southern Ocean, and this colder water can be transported northward by the climatological ocean current (Fig. 8b). In summary, the major contributors to the upper 50/100 m ocean warming in the SWIO were the two anomalous positive heat advection terms: meridional heat transport at the southern boundary ($ -{\rm{v}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{y}}}>0 )$ and zonal heat transport at the western boundary ($ -{\rm{u}}{{'}}\dfrac{\partial \overline{{\rm{T}}}}{\partial {\rm{x}}}>0 $).

      Figure 7.  (a) Sea temperature trend (shading, with intervals of 0.05 K decade−1; K decade−1), ocean zonal and vertical current trends (vectors; m s−1 decade−1) during 1996-2016, and climatic sea temperature (contour, with intervals of 2 K; K). All the values are averaged within 34°-13°S. The area between the blue dotted lines indicates the location of the SWIO (54°-92°E). (b) Sea temperature trend (shading, with intervals of 0.05 K decade−1; K decade−1), ocean meridional and vertical current trends (vectors; m s−1 decade−1) during 1996-2016, and climatic sea temperature (contour, with intervals of 2 K; K). All the values are averaged within 54°-92°E. The red contour is the area that passes the 95% significance test, and only current vectors significant at the 95% confidence level are plotted in black; the rest are gray vectors. The area between the blue dotted lines indicates the location of the SWIO (34°-13°S).

      Figure 8.  (a) Spatial distribution of climatic sea temperature (shading; K) and current trend (vectors; m s−1 decade−1) averaged over 0-100 m in the Indian Ocean. Only current vectors significant at the 95% confidence level are plotted in black, and the rest are gray vectors. (b) Spatial distribution of climatic sea temperature (shading; K) and climatic current (vectors; m s−1) averaged over 0-100 m in the Indian Ocean. The SWIO is enclosed by the blue lines between 54° and 92°E and between 34° and 13°S.

      For each section along the four vertical sections, the sea surface section, and the bottom section at 100 m for the SWIO region, the time-averaged heat flux is calculated [Eq. (2)] to quantify the heat budgets. The results indicate that the SWIO heat flux in the climatic state (Fig. 9a) is mainly caused by warmer water from the equator at the northern boundary and that it loses heat at all other boundaries. For the anomalous heat flux terms (Fig. 9b), a strong eastward current anomaly at the western boundary can enhance the positive heat transfer of warmer water from the west throughout the entire SWIO. The positive anomalous heat flux along the southern boundary also contributes to a net heat gain for the SWIO by reducing the input of cold water from the Southern Ocean into the SWIO. The heat flux values at each boundary of 0-50 m (Fig. 10) show that the magnitude distribution is similar to that of 0-100 m.

      Figure 9.  (a) The climatological heat flux (PW; 1 PW = 1015 W) along all four vertical sections, the sea surface section, and the bottom section at 100 m for the SWIO during 1996-2016. (b) Same as (a) but for the anomalous heat flux (PW). The red arrow indicates that the SWIO gains heat, and the blue arrow indicates that the SWIO loses heat.

      Figure 10.  (a) The climatological heat flux (PW; 1 PW = 1015 W) along all four vertical sections, the sea surface section, and the bottom section at 50 m for the SWIO during 1996-2016. (b) Same as (a) but for the anomalous heat flux (PW). The red arrow indicates that the SWIO gains heat, and the blue arrow indicates that the SWIO loses heat.

      The trend distribution of the four heat fluxes (net shortwave radiation flux of the JRA dataset, net longwave radiation flux of the JRA dataset, latent heat flux is the mean value of the JRA and OAFlux datasets, and sensible heat flux is the mean value of the JRA and OAFlux datasets) in the SWIO (Fig. 11) revealed that the trend of the shortwave radiation flux is negative, which indicates that the ocean loses heat. However, the net longwave radiation, latent heat flux and sensible heat flux showed a positive trend, indicating that the ocean absorbs heat. It is calculated that the net surface radiation heat flux would cause a warming of 0.0071 K year-1 in the SWIO.

      Figure 11.  Spatial pattern of the trend in sea surface heat flux (W m−2 decade−1, downward positive) in the Indian Ocean region: (a) Net shortwave radiation flux of the JRA dataset. (b) Net longwave radiation flux of the JRA dataset. (c) Net latent heat flux based on the mean value of the JRA and OAFlux datasets. (d) Net sensible heat flux based on the mean value of the JRA and OAFlux datasets. The red contour is the area that passes the 95% significance test. The red (blue) shading indicates that there is heat absorption (loss) in the ocean, and the blue lines indicate the SWIO: 54°-92°E, 34°-13°S.

    • The reasons for the ocean current changes can be theoretically determined by calculating the geostrophic current and Ekman current. At the western boundary of the SWIO, the spatial distribution of ocean currents can be well explained by the distribution of the geostrophic current (Fig. 12a). The SSH trend in the western region of the SWIO (55°-76°E, 20°-26°S) is positive. Around the center of the positive SSH trend, there is an anomalous westward geostrophic flow on the north side and an anomalous eastward geostrophic current on the south side, forming an anticyclonic (Southern Hemisphere) flow field in this area. As the eastward geostrophic current can transport warmer waters from the west into the SWIO, more net heat is retained in the SWIO, leading to an enhanced SSH pattern there. This is a positive feedback process of self-sustainment and self-reinforcement. The southeast trade winds prevail in the southern Indian Ocean throughout the whole year. During 1996-2016, the trade winds weakened in the subtropical southern Indian Ocean, and the trend of wind stress showed a clockwise circulation in the SWIO (Fig. 12b). The surface ocean current changes at the southern boundary can be attributed to the surface Ekman current induced by atmospheric circulation variabilities, which formed an anticyclonic pattern in the southeastern section of the SWIO. Such anomalous wind patterns can lead to southward Ekman current anomalies. This anomalous southward transport of the Ekman current at the southern boundary of the SWIO reduced the transport of cold water from the Southern Ocean into the SWIO. Below the surface, a positive feedback can also be seen near the southeastern SWIO, with a warming center and geostrophic currents around it. This self-sustainment pattern favored southward transport near the southern SWIO boundary, which can reduce the climatic northward transport of cold water from the Southern Ocean and warm the SWIO region.

      Figure 12.  (a) Spatial distribution of the SSH trend (shading; m decade−1) and the sea surface geostrophic current trend (vectors; m s−1 decade−1) in the Indian Ocean region during 1996-2016. (b) Spatial distribution of the wind stress trend (blue vectors; N m−2 decade−1) and Ekman current trend (black vectors; m s−1 decade−1) in the Ekman layer (set as 50 m) over the Indian Ocean region during 1996-2016, where the light blue arrow is the regression of the wind stress distribution. The red contour is the area that passes the 95% significance test. The current vectors significant at the 95% confidence level are plotted in black, and the rest are gray vectors. Only wind stress vectors significant at the 95% confidence level are plotted in blue. The SWIO region is indicated by the blue lines (54°-92°E, 34°-13°S). The SSH data are obtained from GODAS.

      The processes of the two boundaries formed a self-sustaining positive feedback mechanism that occurred on each level above at least 120 m (Fig. 13) and favored the maintenance of a sustained warming trend in the SWIO region. The distribution of sea temperature at each depth layer (figures omitted) shows that apparent warming signals existed from the surface to a depth of 120 m, with a maximum value concentrated at 50-70 m. At each level, the most significant areas of warming were located on the western and southern boundaries of the SWIO. In addition, around each warming center, there were geostrophic-like patterns of ocean currents, indicating positive feedback between warming and geostrophic currents. As the depth increases, the warming phenomenon in the SWIO area gradually disappears below 120 m, and the trend of current circulation also gradually weakens (not shown).

      Figure 13.  Spatial distribution of sea temperature trends (shading; K decade−1) and current trends (vectors; m s−1 decade−1) at 120 m in the Indian Ocean. The red contour is the area that passes the 95% significance test, and only current vectors significant at the 95% confidence level are plotted in black; the rest are gray vectors. The SWIO is enclosed by the blue lines (54°-92°E, 34°-13°S).

      The trend of SST, as well as the 0-100 m temperature, over the SWIO is almost constant during 1996-2006. There are obvious interannual changes in wind field changes, indicating that the surface dynamical forcing is not stable, nor is it the key reason for the stable warming of the 100 m sea temperature. The interannual changes in ocean currents are not large, and the ocean current terms mainly show interdecadal trends, indicating that the heat transported by ocean currents should be the key reason for the stable warming phenomenon. In particular, the meridional/zonal ocean currents near the southern/western boundary of the upper 100 m are stable, which is the fundamental reason for maintaining the sustained warming of the SWIO (Fig. 14).

      Figure 14.  (a) Time series of sea temperature anomalies (K year−1) in the GODAS dataset, averaged over the upper 100 m in the SWIO basin. (b) Time series of sea surface zonal wind stress (N m−2) in the GODAS dataset, averaged over the region of the southern boundary of the SWIO (70°-92°E, 34°-28°S). (c) is the same as (b) but for the meridional ocean current (m s−1) in the upper 100 m. A three-point moving average has been applied.

    4.   Summary and discussion
    • Since the mid-1990s, the SSTs in the SWIO have experienced sustained decadal warming for more than two decades, even during the hiatus period, during which a slowdown in the global warming speed occurred. The linear advection of heat by ocean surface currents, including anomalous eastward heat transport at the western boundary and anomalous northward heat transport at the southern boundary, has made a notable contribution to warming in this region. Corresponding to the local warming in the SWIO, a self-maintaining anticyclonic geostrophic current pattern was formed in the western part of the SWIO region, which could further enhance the eastward heat transport at the western boundary of the SWIO. However, at the southern boundary of the SWIO, the anomalous wind stress drives a southward Ekman flow, which reduces the climatic input of cold water from the Southern Ocean into the SWIO. Below the surface, a self-maintaining anticyclonic geostrophic current pattern also favored local warming near the southeastern SWIO by reducing cold water transport from the Southern Ocean. These processes near the two boundaries formed a self-sustaining positive feedback mechanism and led to a sustained warming trend in the SWIO for more than two decades.

      Although there have been some studies on the mechanism and effects of SST warming in the Indian Ocean, most of these efforts have focused on the overall warming of the Indian Ocean (Han et al., 2014; Li et al., 2017; Zhang et al., 2018) or seasonally scaled sea surface warming affected by El Niño (Rao and Behera, 2005; Xie et al., 2002, 2009; Chowdary et al., 2009; Chen et al., 2019). This study reveals that sustained decadal warming in the SWIO is mainly affected by heat advection from ocean currents. With significant increases in greenhouse gas emissions, the rate of global warming is likely to continue to increase after a brief slowdown (Su et al., 2017). Over the past century, the warming trend in the Indian Ocean has become the most prominent among the world’s oceans, and the SWIO’s warming trend over the past century is also significantly positive. However, this paper highlights that the SWIO has maintained sustained and steady warming over the two decades since the mid-1990s, which is a unique phenomenon of global warming. It should be said that the cause of the SWIO’s sustained and steady warming is a special aspect of global warming. In other words, the warming of the SWIO was bound to occur sooner or later. The fundamental reason lies in the driving force of the global warming background, which is irreversible in the near future. Even if the warming is temporarily slowed on a decadal scale, the temperature will rise definitely again later. On the other hand, this unique phenomenon of sustained and decadal warming in the SWIO should be subject to local oceanic/atmospheric dynamics. In this paper, the dynamic effect of the atmosphere on the ocean is emphasized unilaterally. The unique phenomenon should be analyzed from an integrated perspective of ocean-atmospheric coupling in the future to obtain a more comprehensive understanding.

      Acknowledgments. J.S. conceived the idea and designed the study. J.L. conducted most of the analysis and discovered the main results. All authors contributed to interpreting the results, discussing the associated dynamics and improving this paper. The authors wish to thank two anonymous reviewers for their constructive and helpful comments.

Reference (36)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return