1. Introduction
As one of the most significant aspects of climate change, sea level rise is an urgent concern of both scientific communities and the general public, with growing threats to the vast population residing in coastal and island regions (e.g., Douglas 2001; Church et al. 2001; Woodworth et al. 2004; Nicholls and Cazenave 2010; Han et al. 2018). Over the past three decades, the observed global-mean sea level rise was at an average rate of ∼3 mm yr−1 (Willis et al. 2015; Chen et al. 2017; Frederikse et al. 2020) and projected to persist or further accelerate in the upcoming future as the greenhouse gas–forced global warming continues (Collins et al. 2013; Kirtman et al. 2013; Slangen et al. 2017; Oppenheimer et al. 2019; Chen et al. 2021). The global-mean sea level rise has been attributed primarily to ocean thermal expansion and mass loss of glaciers (e.g., Church et al. 2001, 2013; Oppenheimer et al. 2019; Frederikse et al. 2020). Yet, the regional sea level trend shows complex geographical distributions (e.g., Stammer and Hüttemann 2008; Cazenave and Llovel 2010; Han et al. 2010, 2014; Merrifield et al. 2012; Hamlington et al. 2020) and tends to agree with the steric sea level (SSL) trend associated with ocean temperature and salinity changes (e.g., Qiu and Chen 2012; Han et al. 2014; Llovel and Lee 2015; Frederikse et al. 2020).
The SSL estimate–based on in situ observations of ocean subsurface temperature and salinity–serves as the primary approach for the quantification of broad-scale regional sea level change prior to the satellite altimetry era. The SSL change is linearly decomposed into the thermal and haline components, dubbed the thermosteric sea level and halosteric sea level (TSL and HSL), respectively (e.g., Levitus et al. 2005; Antonov et al. 2002; Köhl 2014). Previous studies have revealed prominent SSL variabilities from seasonal to decadal time scales associated with the thermal component, whereas the effect of ocean salinity is less appreciated (e.g., Antonov et al. 2002; Vinogradov et al. 2008; Qiu and Chen 2012; Stammer et al. 2013; Köhl 2014; Han et al. 2014). Existing assessments suggested that the global-mean HSL trend since the mid-twentieth century is at least one order smaller than TSL (e.g., Gregory and Lowe 2000; Antonov et al. 2002; Frederikse et al. 2020).
Reliable quantification and in-depth understanding of the ocean salinity effect have been hindered by the shortage of observational data. The launch of the Argo Program (Roemmich and Gilson 2009) since the early 2000s has brought about salinity observation records over 0–2000 m with unprecedented sampling coverage and provided insights into salinity changes on various time scales. Recent studies suggested that the halosteric effect has a considerable contribution to regional sea level changes in many regions (e.g., Köhl 2014; Levitus et al. 2005; Nidheesh et al. 2013; Llovel and Lee 2015; Jyoti et al. 2019). Among them, the southeast Indian Ocean (SEIO) is located to the west of Australia and south of Indonesia. The rapid sea level rise here exerts potential stress to the western Australian and Indonesian coasts and may worsen the damages of marine hazards such as marine heatwaves that frequently occur in this region during the past two decades (e.g., Pearce and Feng 2013; Feng et al. 2015b; Frölicher et al. 2018). On the other hand, the SEIO is of a unique ocean circulation system that is substantially modulated by the Pacific-origin inflow of the Indonesian Throughflow (ITF) (e.g., Feng et al. 2003; Wijffels and Meyers 2004; Schott et al. 2009). The SEIO is also affected by the Subantarctic Mode Water (SAMW) and Antarctic Intermediate Water (AAIW) (e.g., McCartney and Talley 1982; Hanawa and Talley 2001; Sallée et al. 2006; Koch-Larrouy et al. 2010; Schmidtko and Johnson 2012; Hong et al. 2020; Zhang et al. 2021) originating from the Southern Ocean. Under the combined influences of the Pacific Ocean, the Southern Ocean, and local air–sea interaction, the SEIO is characterized by multiple time scale variabilities in ocean temperature (e.g., Wijffels and Meyers 2004; Feng et al. 2008; Li et al. 2017, 2019; Guo et al. 2020a) and salinity (e.g., Du et al. 2015, 2019; Feng et al. 2015a; Hu and Sprintall 2016; Hu et al. 2019). These changes link the regional sea level change in the SEIO to global climate phenomena such as greenhouse gas warming and water cycle alterations (Durack et al. 2012; Durack 2015).
Recent studies based on Argo data revealed that for the decadal sea level rise since 2005, the contribution of the HSL in the SEIO is of at least equal importance to that of TSL (Llovel and Lee 2015; Huang et al. 2020). By analyzing Argo data in conjunction with an ocean reanalysis product, Jyoti et al. (2019) suggested that the halosteric component could explain ∼40% of the total sea level rise of the south Indian Ocean from the 1990s to the early 2010s. The recent HSL rise was associated with a decadal freshening in the upper 300 m of the SEIO (Llovel and Lee 2015). Strengthening of the ITF since the 1990s (Sprintall and Révelard 2014; Lee et al. 2015) may have enhanced the freshwater transport from the west Pacific and led to this freshening trend in the SEIO (e.g., Du et al. 2015, 2019; Hu and Sprintall 2016; Jyoti et al. 2019). Huang et al. (2020) pointed out that the meridional exchange between the tropical low-salinity water and the subtropical high-salinity water is also an important process driving changes of ocean salinity and HSL in the SEIO. In addition to sea level change, ocean salinity also plays an essential role in regulating the geostrophic circulation of the SEIO. The intensity of the Eastern Gyral Current (EGC), through which a portion of ITF water merges into the Leeuwin Current, is largely determined by the meridional salinity gradient (Menezes et al. 2013). In these regards, the SEIO is an ideal region for investigating the salinity effect on sea level change.
Despite existing studies of the salinity effect on sea level change, the contribution of ocean salinity to the long-term sea level rise in the SEIO since the mid-twentieth century has not been evaluated yet. This contribution is quite likely to exist, given that a robust multidecadal subsurface freshening has been detected in the southern Indian Ocean (e.g., Wong et al. 1999; Bindoff and McDougall 2000; Durack and Wijffels 2010; Cheng et al. 2020). Existing studies have also revealed the complexity in the relationship between the TSL and HSL changes. In some regions (mostly at low and midlatitudes), subsurface temperature and salinity changes are dominated by the advection of density-compensated water-mass property changes, dubbed the spicing mode (e.g., Stommel 1962; Bindoff and McDougall 1994; Huang 2020). As such, the TSL and HSL usually act in a density-compensated manner and tend to be negatively correlated (Munk 2003; Köhl 2014). In other regions such as high-latitude oceans, the TSL and HSL may work in concert to drive sea level changes (e.g., Köhl 2014; Levitus et al. 2005; Nidheesh et al. 2013).
This study utilizes four updated ocean observational datasets and ocean model-based datasets (hindcasts and reanalyses) to systematically investigate the multidecadal regional sea level rise and explore the role played by ocean salinity. The SEIO region is chosen as an example to understand the impact of ocean salinity on regional sea level rise in a rapidly changing climate. This paper is organized as follows. Section 2 describes the datasets and methodology. Section 3 presents the characteristics of SSL, TSL, and HSL changes during 1960–2018 in the SEIO and quantifies the contribution of ocean salinity. Section 4 investigates temperature and salinity changes at different depths and quantifies their contributions to the total SSL change. Section 5 explores the causes of temperature and salinity changes. Finally, section 6 provides conclusions and discussion.
2. Data and methods
a. Datasets
Temperature and salinity fields from four observational datasets are used to characterize the multidecadal SSL, TSL, and HSL changes in the SEIO. They are the NCEI (the National Centers for Environmental Information; Levitus et al. 2005, 2012), the IAP (the Institute of Atmospheric Physics ocean analysis data; Cheng and Zhu 2016; Cheng et al. 2017, 2020), the Ishii (Ishii et al. 2017), and the EN4 (Good et al. 2013) datasets. All the four datasets provide 1° × 1°, standard-level temperature and salinity fields for 0–2000 m. Among them, NCEI provides pentad-mean data in annual intervals. Monthly fields of IAP, Ishii, and EN4 are resampled to annual-mean data for our analysis. These datasets are based on all-available historical data collected by various in situ measurement platforms instruments. IAP was reconstructed by using a mapping technology of ensemble optimal interpolation (EnOI) (Cheng et al. 2017), which actually implemented the ideas of how to create smoothed three-dimensional global salinity field from incomplete in situ observations (Cheng and Zhu 2016). IAP overcomes key shortcomings of earlier reconstruction products with uncertainties being better constrained (Cheng et al. 2020). The analysis of the present study is focused on the 1960–2018 period, which is covered by all the NCEI, IAP, Ishii, and EN4 datasets.
SSL computed with the above-mentioned datasets are compared against the total sea level records of the 1/4° × 1/4° multisatellite altimeter (Ablain et al. 2009) product. Geophysical and environmental corrections have been applied to this sea level anomaly product. In this study, we refer to it as the AVISO (Archiving Validation, and Interpretation of Satellite Oceanography) data. Figure S1a in the online supplemental material shows the patterns of sea level trend during 2005–18 from AVISO satellite data, in which the global mean trend of ∼32.5 mm decade−1 during this period has been removed. This global-mean trend is close to the 33.2 mm decade−1 trend of 1993–2018 estimated by Frederikse et al. (2020). During this period, there are many regions showing rates significantly exceeding the global mean (positive regional sea level trend in Fig. S1a), and the SEIO is among them. The gridded Argo (Roemmich and Gilson 2009) temperature and salinity products for 2005–18 are used to verify the other four datasets. The SSL trends estimated by Argo, NCEI, IAP, Ishii, and EN4 are shown in Figs. S1b–f for comparison. Their spatial patterns and typical magnitudes are in good agreement with those of AVISO. There are regional differences that may arise from eustatic sea level rise associated with the ice-mass loss, contribution of the deep ocean below 2000 m, and insufficient sampling of observations (e.g., Miller and Douglas 2004; Lorbacher et al. 2012; Church et al. 2013; Llovel et al. 2014; Llovel and Lee 2015). Nevertheless, the SSL based on 0–2000-m ocean temperature and salinity is a good proxy for regional sea level and useful in understanding regional sea level change (Willis et al. 2008; Llovel et al. 2010; Cazenave and Llovel 2010; Stammer et al. 2013; Llovel and Lee 2015). Moreover, the SSL trends from NCEI, IAP, Ishii, and EN4 (Figs. S1c–f) are consistent with those of Argo data (Fig. S1b), suggesting that the four datasets are suitable for the investigation of SSL change.
We also analyzed the 1960–2010 hindcast of the LICOM3 model [Laboratory of Atmospheric Sciences and Geophysical Fluid Dynamics (LASG)/IAP Climate Ocean Model version 3; Lin et al. 2020]. LICOM3 adopts a global ocean model domain with eddy-resolving resolutions (∼0.1°) and is forced by the ERA-20C [the European Centre for Medium-Range Weather Forecasts (ECMWF) Twentieth-Century Reanalysis; Poli et al. 2016] fields. More details of LICOM3 configurations are available in Li et al. (2020). In addition to LICOM3, there are three other oceanic reanalysis products and three atmospheric reanalysis products used to explore causes of ocean temperature and salinity changes. The three oceanic reanalysis products are the 0.25° × 0.4° SODA2.2.4 (Simple Ocean Data Assimilation, version 2.2.4; Giese and Ray 2011), the 1° × 1/3° GECCO2 (German contribution to Estimating the Circulation and Climate of the Ocean project, version 2; Köhl 2014), and the 1° × 1° ORA-S4 (Ocean Reanalysis System, version 4; Balmaseda et al. 2013) for 1960–2010. The three atmospheric reanalysis products are the NOAA-20CR (National Oceanic and Atmospheric Administration Twentieth Century Reanalysis version 2; Compo et al. 2011), the combination of ERA-40 (Uppala et al. 2005) for 1960–88 and ERA-Interim (Dee et al. 2011) for 1989–2010 (dubbed the ECMWF data here), and the NCEP–NCAR (National Centers for Environmental Prediction–National Center for Atmospheric Research product; Kalnay et al. 1996). Note that the ERA-20C, NOAA-20CR, NCEP–NCAR, and ECMWF are the surface forcing fields for LICOM3, SODA2.2.4, GECCO2, and ORA-S4, respectively. The monthly temperature, salinity, and current fields from LICOM3, SODA2.2.4, GECCO2, and ORA-S4 along with monthly surface fields of ERA-20C, NOAA-20CR, NCEP–NCAR, and ECMWF are all resampled to 1° × 1° grids for our analysis.
b. Definitions
In the Southern Hemisphere, a positive WSC corresponds to a downward Ekman pumping velocity that drives upper-ocean convergence and depression of isopycnal surfaces.
3. Multidecadal steric sea level change in the SEIO
The SSL estimates from IAP, Ishii, NCEI, and EN4 reach consensus in exhibiting significant multidecadal rising trends during 1960–2018 over most areas of the globe (left panels of Fig. 1). Trends of TSL and HSL (middle and right panels of Fig. 1) represent the contributions of ocean temperature and salinity changes to SSL change, respectively. The trend of TSL is generally larger than that of HSL and largely determines the SSL trend pattern. However, in the SEIO, the trends of TSL and HSL are of comparable magnitudes. These datasets show significant positive HSL trends in the subpolar North Atlantic, the subtropical Pacific, and the SEIO, which enhance the total SSL rise. By contrast, the north Indian Ocean and the tropical and subtropical Atlantic Oceans show significant negative trends that act to attenuate the total SSL rise. Ocean salinity change is not negligible in sea level changes of these regions.
Linear trend maps of (a) steric sea level (SSL), (b) thermosteric sea level (TSL), and (c) halosteric sea level (HSL) for the 1960–2018 period derived from IAP. (d)–(f),(g)–(i),(j)–(l) As in (a)–(c), but derived from Ishii, NCEI, and EN4, respectively. Stippling indicates significant at 90% confidence level based on a Mann–Kendall test. Black lines denote the southeast Indian Ocean (SEIO) region (90°–120°E, 8°–32°S).
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Linear trend maps of (a) steric sea level (SSL), (b) thermosteric sea level (TSL), and (c) halosteric sea level (HSL) for the 1960–2018 period derived from IAP. (d)–(f),(g)–(i),(j)–(l) As in (a)–(c), but derived from Ishii, NCEI, and EN4, respectively. Stippling indicates significant at 90% confidence level based on a Mann–Kendall test. Black lines denote the southeast Indian Ocean (SEIO) region (90°–120°E, 8°–32°S).
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Linear trend maps of (a) steric sea level (SSL), (b) thermosteric sea level (TSL), and (c) halosteric sea level (HSL) for the 1960–2018 period derived from IAP. (d)–(f),(g)–(i),(j)–(l) As in (a)–(c), but derived from Ishii, NCEI, and EN4, respectively. Stippling indicates significant at 90% confidence level based on a Mann–Kendall test. Black lines denote the southeast Indian Ocean (SEIO) region (90°–120°E, 8°–32°S).
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
In the Indian Ocean sector, we see enhanced rising trends of SSL in the 30°–50°S band (left panels of Fig. 1), which are associated with enhanced ocean heat uptake (e.g., Frölicher et al. 2015; Swart et al. 2018) and TSL rising (middle panels of Fig. 1). Using ORA-S4 data, Jyoti et al. (2019) proposed another maximal rising band of SSL within 20°–30°S in the Indian Ocean. However, this unique feature is likely exclusive for ORA-S4 and not seen in any of the four observational datasets (Fig. S2). There seems to be an overestimation of SSL rise over 20°–30°S; particularly, the increase of HSL is by far stronger in ORA-S4 (Fig. S2c) than in observational datasets. Durack et al. (2014) identified negative SSL trends for 1950–2008 in the low-latitude sector of the SEIO (north of 20°S) using the gridded dataset of Durack and Wijffels (2010); they also found that these trends are weaker and insignificant in the old-version Ishii data (see their Fig. 1). We revisit this feature here with multiple datasets (Fig. S3). For the SSL trend to 2008, the four datasets show an overall positive trend over the SEIO region, with only weak and insignificant negative trends northwest of Australia, which arises mainly from the TSL component. This feature is not seen in the trend maps of 1960–2018 and 1960–2017 (Fig. 1; see also Fig. S2). Therefore, the negative trends near the northwestern Australian coast reported by Durack et al. (2014) are likely associated with the ending year choice of 2008; our analysis presented below will show that there is a short-term drop in SSL and TSL during 2001–08 (Fig. 2), which leaves influence in the trend estimate. Based on the comparisons described above, our following analysis will focus on the robust SSL rising trend of 1960–2018 and the role of ocean salinity, over which the four datasets reach consensus.
(left) Time series of annual SSL (blue), TSL (red), and HSL (green) in the SEIO (region specified in Fig. 1) based on (a) IAP, (b) NCEI, (c) EN4, and (d) Ishii. (right) Linear trends of SSL (blue bar), TSL (red bar), and HSL (green bar) in the SEIO for the periods of 1960–2018 and 1990–2018 derived from (e) IAP, (f) NCEI, (g) EN4, and (h) Ishii. Error bars denote the 90% confidence interval based on an F test.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
(left) Time series of annual SSL (blue), TSL (red), and HSL (green) in the SEIO (region specified in Fig. 1) based on (a) IAP, (b) NCEI, (c) EN4, and (d) Ishii. (right) Linear trends of SSL (blue bar), TSL (red bar), and HSL (green bar) in the SEIO for the periods of 1960–2018 and 1990–2018 derived from (e) IAP, (f) NCEI, (g) EN4, and (h) Ishii. Error bars denote the 90% confidence interval based on an F test.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
(left) Time series of annual SSL (blue), TSL (red), and HSL (green) in the SEIO (region specified in Fig. 1) based on (a) IAP, (b) NCEI, (c) EN4, and (d) Ishii. (right) Linear trends of SSL (blue bar), TSL (red bar), and HSL (green bar) in the SEIO for the periods of 1960–2018 and 1990–2018 derived from (e) IAP, (f) NCEI, (g) EN4, and (h) Ishii. Error bars denote the 90% confidence interval based on an F test.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
For better quantification, we computed the annual time series of SSL, TSL, and HSL averaged over the SEIO region (Fig. 2). SSL shows prominent decadal variability superimposed on a long-term rising trend over the past half-century (Figs. 2a–d). This rising becomes more evident since 1990; in fact, the trend before 1990 is feeble. We calculated the linear trends of SSL for both the 1960–2018 and 1990–2018 periods (Figs. 2e–h). As estimated with IAP, the trends are 9.1 ± 2.6 mm decade−1 (±90% confidence interval based on an F test) for 1960–2018 and 25.4 ± 6.6 mm decade−1 for 1990–2018, respectively. The results of Ishii, NCEI, and EN4 are broadly consistent with IAP, reaching consensus in the acceleration of SSL rise since 1990. By comparing TSL and SSL, we note that decadal variability of SSL arises predominantly from TSL, but the long-term trend of SSL is evidently stronger than that of TSL. In IAP, the TSL trend accounts for ∼58% of the SSL trend (Fig. 2e). In all the four datasets, the HSL shows a persistent rise since the 1960s with limited interannual and decadal fluctuations. The trends of HSL are 1.8–3.9 mm decade−1 among these datasets, accounting for ∼42% of the SSL trends.
During the 1990–2018 period, TSL shows a much larger trend and dominates the total trend of SSL. It reaches a trend as large as 12.3–20.2 mm decade−1 among the four datasets, accounting for 71.9%–84.6% of the total SSL change. The trend of HSL in 1990–2018 is much weaker than TSL and plays a minor role in SSL rise. The HSL trend of 1990–2018 is close to that of 1960–2018, suggesting the persistent contribution of salinity change to sea level rise. On the other hand, the HSL trend is insignificant for the 1990–2018 period. This discrepancy partly arises from decadal variability. Overall, there are stark contrasts between the 1960–2018 and 1990–2018 periods, as shown in Figs. 2e–h. Contributions of temperature and salinity on the SSL change vary with time, with their relative importance altered before and after ∼1990.
4. Changes in temperature and salinity
In this section, we explore temperature and salinity changes associated with TSL and HSL trends. All the datasets show rapid warming in the upper 1000 m since 1990 (Figs. 3a–d). The temperature of the 400–1000-m layer is subjected to multidecadal cycles, showing warming anomalies during 1960–80 and 2000–10 and cooling anomalies during 1980–2000 and 2014–18. The phase transition from cooling to warming conditions during the 1990s contributed to the rapid TSL rise. Ocean salinity exhibits a discernible decreasing trend in the SEIO (Figs. 3e–h). Salinity changes above and below ∼400 m differ drastically. The upper layer (0–400 m) exhibits strong interannual and decadal fluctuations, but its multidecadal trend is hardly discernible. By contrast, there is a persistent long-term freshening trend at 400–1000 m, which is likely the primary origin of the long-term rising trend of HSL over 0–2000 m.
Evolutions of temperature anomalies in the upper 1200 m averaged over the SEIO region derived from (a) IAP, (b) NCEI, (c) Ishii, and (d) EN4. (e)–(h) As in (a)–(d), but for salinity anomalies.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Evolutions of temperature anomalies in the upper 1200 m averaged over the SEIO region derived from (a) IAP, (b) NCEI, (c) Ishii, and (d) EN4. (e)–(h) As in (a)–(d), but for salinity anomalies.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Evolutions of temperature anomalies in the upper 1200 m averaged over the SEIO region derived from (a) IAP, (b) NCEI, (c) Ishii, and (d) EN4. (e)–(h) As in (a)–(d), but for salinity anomalies.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
To quantify the contributions of temperature and salinity change at different depths to the SSL change, TSL and HSL are separately computed for 0–400 and 400–1000 m and compared with those of 0–2000 m (Fig. 4). During 1960–2018 (Figs. 4a–d), the SSL trends of the 0–400- and 400–1000-m layers are 6.7 ± 2.3 and 1.8 ± 0.7 mm decade−1, respectively, which account for ∼74% and ∼20% of the 0–2000-m SSL trend, as derived from IAP data (Fig. 4a). Similar results are achieved from other datasets (Figs. 4b–d), with the 0–400-m layer playing a more important role than the 400–1000-m layer. The trend of 0–400 m TSL is 4.9 ± 1.7 mm decade−1 from IAP that dominates the TSL change, with the large uncertainty indicating strong interannual and decadal variability. The trend of 400–1000-m TSL is negative in IAP, NCEI, and EN4 (against the total sea level rise) and shows a feeble rising trend in Ishii.
Linear trends of SSL, TSL, and HSL computed for the 0–2000-m (blue bars), 0–400-m (red bars), and 400–1000-m (green bars) layers for the period of 1960–2018, derived from (a) IAP, (b) NCEI, (c) Ishii, and (d) EN4. (e)–(h) As in (a)–(d), but for the 1990–2018 period. Error bars denote the 90% confidence interval based on an F test.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Linear trends of SSL, TSL, and HSL computed for the 0–2000-m (blue bars), 0–400-m (red bars), and 400–1000-m (green bars) layers for the period of 1960–2018, derived from (a) IAP, (b) NCEI, (c) Ishii, and (d) EN4. (e)–(h) As in (a)–(d), but for the 1990–2018 period. Error bars denote the 90% confidence interval based on an F test.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Linear trends of SSL, TSL, and HSL computed for the 0–2000-m (blue bars), 0–400-m (red bars), and 400–1000-m (green bars) layers for the period of 1960–2018, derived from (a) IAP, (b) NCEI, (c) Ishii, and (d) EN4. (e)–(h) As in (a)–(d), but for the 1990–2018 period. Error bars denote the 90% confidence interval based on an F test.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
The trend of 0–400-m HSL from IAP is 1.7 ± 1.0 mm decade−1, accounting for ∼44% of the 0–2000-m HSL trend. This portion of the contribution is mainly from the interannual and decadal variability of the upper-ocean salinity, and the freshening occurring after 2010 greatly contributes to the total HSL rising trend (Fig. 3e). The 400–1000-m HSL trend from IAP is 2.3 ± 0.2 mm decade−1, with a robust long-term trend accounting for ∼59% of the 0–2000-m HSL trend and limited noises from short-term variability (as denoted by the error bars). In three of the four datasets (IAP, NCEI, and EN4), the contribution of the 400–1000-m layer is larger than it is in the HSL rise of 1960–2018 (Figs. 4a,b,d), whereas in Ishii the two layers make equal contributions (Fig. 4c). In all datasets, the 1000–2000-m layer makes a negative contribution to the total HSL.
During the 1990–2018 period (Figs. 4e–h), the 0–400-m layer shows an SSL trend of 19.9 ± 6.4 mm decade−1 and dominates the trend of 0–2000-m SSL, as computed with the IAP data. This contribution is predominantly due to upper-ocean warming (TSL), and the role of salinity change is secondary (with a contribution of 20% according to IAP data). The trend of 0–400-m TSL is 15.7 ± 4.8 mm decade−1, which accounts for ∼78% of the 0–2000-m TSL rise, while the 400–1000-m TSL shows a much weaker trend of 2.8 ± 1.2 mm decade−1. In fact, the HSL trend during 1990–2018 (4.2 ± 2.7 mm decade−1) is even stronger than that of 1960–2018 (3.1 ± 1.1 mm decade−1), and the difference arises mainly from the 0–400-m HSL that increases from 1.6 ± 0.9 mm decade−1 during 1960–2018 to 4.1 ± 2.7 mm decade−1 during 1990–2018. The 0–400-m trend of 1990–2018 is subjected to strong influence from interannual and decadal variability (Fig. 3); in particular, the strong decadal freshening anomalies after 2010 greatly enlarge the HSL trend. Results from NCEI, Ishii, and EN4 (Figs. 4f–h) are broadly consistent with IAP, except that EN4 shows a negative trend in 400–1000-m TSL. The results presented in this section overall suggest the complexity of mechanisms underlying the observed regional sea level rise in the SEIO, with time-varying contributions from temperatures and salinities at different depths.
5. Causes of temperature and salinity changes
A question naturally arises as to the causes for the temperature and salinity changes. As shown in section 4, the temperature change of 0–400 m and the salinity change of 400–1000 m are found to be essential. Therefore, we separately explore the processes causing the two features.
a. Freshening of the 400–1000-m layer
Temperature anomalies of (a) the heaving mode, (b) the spicing mode, and (c) their sum in the upper 1200 m averaged over the SEIO, derived from IAP. (d)–(f) As in (a)–(c), but for salinity anomalies.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Temperature anomalies of (a) the heaving mode, (b) the spicing mode, and (c) their sum in the upper 1200 m averaged over the SEIO, derived from IAP. (d)–(f) As in (a)–(c), but for salinity anomalies.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Temperature anomalies of (a) the heaving mode, (b) the spicing mode, and (c) their sum in the upper 1200 m averaged over the SEIO, derived from IAP. (d)–(f) As in (a)–(c), but for salinity anomalies.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Unlike temperature change, the spicing mode is more important for salinity change (Figs. 5d–f). The term SS (the salinity anomaly of the spicing mode) contributes ∼95.8% to the total change. In particular, the freshening trend at 400–1000 m (Fig. 5e) arises from the spicing mode. This freshening trend is paired with a cooling trend in θS (Fig. 5b), characterizing the density-compensated temperature and salinity changes along isopycnals. These results confirm that the freshening trend at 400–1000 m, along with the resultant long-term HSL rise, is primarily associated with subsurface water-mass property changes.
The 400–1000-m layer roughly corresponds to the potential density range of 26.5–27.4 kg m−3 in the SEIO, occupied by the SAMW and AAIW (e.g., McCartney and Talley 1982; Hanawa and Talley 2001; Sallée et al. 2006; Koch-Larrouy et al. 2010; Schmidtko and Johnson 2012; Hong et al. 2020; Zhang et al. 2021). Schmidtko and Johnson (2012) showed that the AAIW, as identified by a salinity minimum between 27.0 and 27.5 kg m−3, can be detected in the SEIO; Zhang et al. (2021) shows that the SEIO region is covered by the northward spreading of the SAMW, which is identified by the low potential vorticity feature at 26.8 kg m−3. Isopycnal temperature and salinity between 26.5 and 27.4 kg m−3 show concurrent decreasing trends in all datasets (Figs. 6a,b). The hindcast of LICOM3 can reproduce these long-term trends, although it fails to simulate decadal fluctuations. Therefore, we seek the source of these trends with the help of LICOM3 output. The trend pattern of isopycnal salinity between 26.5 and 27.4 kg m−3 from LICOM3 (Fig. 6c) shows a prevailing freshening in the subtropical sector of the SEIO as an extension of the stronger freshening of the Southern Ocean, which are consistent with those from IAP and NCEI (Figs. S8c,d). The climatological currents of 1960–2010 (Fig. 6c) on the southern boundary are northwestward between 26.5 and 27.4 kg m−3, favoring the advection of freshening signatures. According to LICOM3 hindcast, these northwestward currents were not significantly enhanced during 1960–2010 (Fig. S8b); alternatively, northwestward advection of anomalies by the climatological currents through the southern boundary is probably the leading driver of the observed SEIO freshening and cooling trends on isopycnal surfaces of 26.5–27.4 kg m−3.
(a) Annual potential temperature anomalies averaged between 26.5 and 27.4 kg m−3 isopycnal surfaces in the SEIO derived from IAP (blue), NCEI (red), Ishii (yellow), EN4 (purple), and LICOM3 (green). (b) As in (a), but for salinity anomalies. (c) Salinity trends (color shading; psu decade−1) and annual-mean climatological current (vectors; m s−1) averaged between 26.5 and 27.4 kg m−3 isopycnal surfaces for 1960–2010 based on LICOM3.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
(a) Annual potential temperature anomalies averaged between 26.5 and 27.4 kg m−3 isopycnal surfaces in the SEIO derived from IAP (blue), NCEI (red), Ishii (yellow), EN4 (purple), and LICOM3 (green). (b) As in (a), but for salinity anomalies. (c) Salinity trends (color shading; psu decade−1) and annual-mean climatological current (vectors; m s−1) averaged between 26.5 and 27.4 kg m−3 isopycnal surfaces for 1960–2010 based on LICOM3.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
(a) Annual potential temperature anomalies averaged between 26.5 and 27.4 kg m−3 isopycnal surfaces in the SEIO derived from IAP (blue), NCEI (red), Ishii (yellow), EN4 (purple), and LICOM3 (green). (b) As in (a), but for salinity anomalies. (c) Salinity trends (color shading; psu decade−1) and annual-mean climatological current (vectors; m s−1) averaged between 26.5 and 27.4 kg m−3 isopycnal surfaces for 1960–2010 based on LICOM3.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Existing studies have reported the broad-scale freshening trends of SAMW and AAIW in the Indian and Pacific Oceans during the past several decades based on both in situ ocean observations (e.g., Wong et al. 1999; Bindoff and McDougall 2000; Durack and Wijffels 2010; Durack 2015; Swart et al. 2018) and climate model simulations (Banks et al. 2000; Swart et al. 2018). These studies have attributed these water-mass changes to increased freshwater fluxes in their formation region, namely the midlatitude Southern Ocean, which is linked to the amplification of the global water cycle under anthropogenic warming (e.g., Wong et al.1999; Durack and Wijffels 2010; Durack et al. 2012; Durack 2015). Among others, Durack and Wijffels (2010) further demonstrated that the poleward migration of wintertime outcropping of isopycnal lines owing to ocean surface warming makes a comparable contribution to the increased freshwater fluxes; Haumann et al. (2016) suggested that the increased northward sea ice transport from Antarctica also favors the freshening of the AAIW, in addition to precipitation. The freshening of SAMW and AAIW is transported to the SEIO by the circulation gyres and eddies (e.g., Rintoul 2018) during the past half-century, which drives the persistent HSL rise and also makes a considerable contribution to the regional sea level rise in the SEIO. These results underscore the key role played by ventilated water masses in conveying climate change signatures into regional oceans, with regional sea level changes as one of the remarkable consequences.
b. Warming of the 0–400-m layer
The term Res in Eq. (11) is the residual term representing unresolved processes such as vertical and lateral diffusion and computational errors.
We apply the budget analysis to four model-based datasets: LICOM3, SODA2.2.4, ORA-S4, and GECCO2. These datasets can realistically represent the temperature change of 0–400 m in the SEIO, especially the accelerated warming after 1990 (Fig. S9). Here, all the terms are shown in the manner of 1990–2010 minus 1960–90 (Fig. 7). The results from four datasets overall suggest substantial uncertainties in both AdvT and HFF. Among the four datasets, the assimilation approach of SODA2.2.4 and ORA-S4 directly nudges ocean temperature toward the observed value. As such, the heat of the two datasets is not conserved in these two datasets. Alternatively, GECCO2 allows for the correction of surface heat fluxes to achieve a better matching between the modeled and observed temperatures. Therefore, we are not able to close the heat budget in GECCO2 with NCEP–NCAR fields, although NCEP–NCAR is used to force the OGCM of GECCO2. Only LICOM3 is a real heat-conserving system. Yet, we used monthly products to compute AdvT in an offline manner, which will involve a large error in the estimation owing to the nonlinearity of the advection term. In addition, the nonlinearity is likely to be strong because the SEIO is a region with enhanced eddy activity (e.g., Jia et al. 2011; Guo et al. 2020b). Besides, there are unresolved processes in the heat budget, such as vertical and lateral diffusion, which cannot be estimated using the model outputs. These limitations make it difficult to achieve an accurate estimation of the heat budget.
The difference between 1990–2010 and 1960–90 (1990–2010 minus 1960–90) of heat advection (Adv) terms for the upper layer (0–400 m) of the SEIO based on (a) LICOM3, (b) SODA2.2.4, (c) ORA-S4, and (d) GECCO2. The letters E, W, N, S, and B denote the advection components across the eastern, western, northern, southern, and bottom (i.e., 400 m) boundaries, respectively. (e)–(h) As in (a)–(d), but for surface heat flux forcing term derived from ERA-20C, NOAA-20CR, ECMWF, and NCEP–NCAR, respectively; Qnet denotes the surface net heat fluxes, and Qlhf, Qshf, Qsw, and Qlw denote latent heat flux, sensible heat flux, shortwave radiation, and longwave radiation, respectively, all of which are integrated over the SEIO region and transformed into the forcing on ocean temperature [see Eqs. (11) and (13)]. Positive values denote downward heat fluxes into the ocean.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
The difference between 1990–2010 and 1960–90 (1990–2010 minus 1960–90) of heat advection (Adv) terms for the upper layer (0–400 m) of the SEIO based on (a) LICOM3, (b) SODA2.2.4, (c) ORA-S4, and (d) GECCO2. The letters E, W, N, S, and B denote the advection components across the eastern, western, northern, southern, and bottom (i.e., 400 m) boundaries, respectively. (e)–(h) As in (a)–(d), but for surface heat flux forcing term derived from ERA-20C, NOAA-20CR, ECMWF, and NCEP–NCAR, respectively; Qnet denotes the surface net heat fluxes, and Qlhf, Qshf, Qsw, and Qlw denote latent heat flux, sensible heat flux, shortwave radiation, and longwave radiation, respectively, all of which are integrated over the SEIO region and transformed into the forcing on ocean temperature [see Eqs. (11) and (13)]. Positive values denote downward heat fluxes into the ocean.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
The difference between 1990–2010 and 1960–90 (1990–2010 minus 1960–90) of heat advection (Adv) terms for the upper layer (0–400 m) of the SEIO based on (a) LICOM3, (b) SODA2.2.4, (c) ORA-S4, and (d) GECCO2. The letters E, W, N, S, and B denote the advection components across the eastern, western, northern, southern, and bottom (i.e., 400 m) boundaries, respectively. (e)–(h) As in (a)–(d), but for surface heat flux forcing term derived from ERA-20C, NOAA-20CR, ECMWF, and NCEP–NCAR, respectively; Qnet denotes the surface net heat fluxes, and Qlhf, Qshf, Qsw, and Qlw denote latent heat flux, sensible heat flux, shortwave radiation, and longwave radiation, respectively, all of which are integrated over the SEIO region and transformed into the forcing on ocean temperature [see Eqs. (11) and (13)]. Positive values denote downward heat fluxes into the ocean.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
The AdvT change is determined by a delicate imbalance of the advective fluxes on the five boundaries (Figs. 7a–d). The total AdvT change is positive (warming effect) in LICOM3, SODA2.2.4, and GECCO2 but negative (cooling effect) in ORA-S4. In other words, changes in ocean circulation act to drive the warming in three of the four datasets. Despite substantial interdataset differences, a consistent feature that emerges from all datasets is the enhanced southward heat advection across the northern boundary. The meridional heat transport in the upper layer of the southern Indian Ocean is predominantly through the wind-driven shallow meridional overturning circulation (specifically, the subtropical cells; e.g., McCreary et al. 1993; Schott et al. 2002; Miyama et al. 2003). The surface branch of the subtropical cells is mainly via Ekman transport of Indian Ocean trade winds, and its strengthening indicates the important role played by local wind forcing in the Indian Ocean. In SODA2.2.4 and GECCO2, the increased heat input from the north is discharged by the South Equatorial Current across the western boundary, while in LICOM3 and ORA-S4 the outflow on the southern boundary (probably through the Leeuwin Current) also contributes to the heat discharge.
HFF acts to damp the SEIO warming in three of the four atmospheric datasets, except for the ECMWF data (Figs. 8e–h). The four datasets suggest a consistent damping effect of the warming by enhanced latent heat release (Qlhf). They also uniformly show a warming effect by longwave radiation Qlw in the SEIO. Note that in LICOM3 and GECCO2, the cooling by HFF is much larger than the warming by AdvT, which likely explains the underestimation of the cooling and TSL rising trends in LICOM3 (see Fig. S10f). Biases in the simulated ocean circulation changes and errors in surface atmospheric forcing fields may be mutually responsible. Here, we need to state that in all of the four datasets, the sum of AdvT and HFF is not of a proper value to explain the observed upper SEIO warming. These results highlight the complexity of the time-varying ocean thermodynamics in the SEIO, which is challenging for model simulation. Nevertheless, comparisons of Fig. 8 reach consensus in suggesting two potential processes that may drive the accelerated warming since 1990, namely the enhanced meridional heat transport from the north and the increased surface longwave radiation, which deserve further discussion.
(a) Annual-mean climatological surface wind stress (vectors; N m−2) and wind stress curl (color shading; 10−6 N m−3) over 1990–2010 derived from ERA-20C. (b) As in (a), but showing their linear trends for 1990–2010.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
(a) Annual-mean climatological surface wind stress (vectors; N m−2) and wind stress curl (color shading; 10−6 N m−3) over 1990–2010 derived from ERA-20C. (b) As in (a), but showing their linear trends for 1990–2010.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
(a) Annual-mean climatological surface wind stress (vectors; N m−2) and wind stress curl (color shading; 10−6 N m−3) over 1990–2010 derived from ERA-20C. (b) As in (a), but showing their linear trends for 1990–2010.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
In climatology, the SEIO is under the influence of the southeasterly trade winds, characterized by positive WSCs in its subtropical sector that favor upper-ocean heat convergence (Fig. 8a). Since 1990, this climatological wind pattern has strengthened over the entire southern Indian Ocean (Fig. 8b); particularly, most areas of the SEIO are covered by positive WSC trends. These changes in surface winds act to enhance local Ekman downwelling and depression of the isopycnal surfaces associated with the heaving mode. The downwelling in the SEIO corresponds to upper-ocean heat convergence of heat, including the enhanced southward heat transport from the equatorial Indian Ocean.
The warming effect exerted by the reduced upward Qlw is the manifestation of global warming under anthropogenic greenhouse gas forcing. Figure 9a shows the 1990–2010 trend of the 0–400-m temperature over the global ocean, which shows prevailing warming trends except for a cooling of the eastern Pacific owing to the transition toward the La Niña–like condition (e.g., England et al. 2014). Figure 9b shows the difference between 1990–2010 and 1960–90 in Qlw, which shows quasi-uniform increasing trends over the global oceans, consistent with basic features of anthropogenic greenhouse gas forcing. These results support the combined effect of anthropogenic radiation forcing and wind-driven heat convergence on the accelerated upper-ocean warming and TSL rise since ∼1990. As a result, the total SSL trend of the SEIO is 20.9 ± 9.1 mm decade−1 according to IAP data during 1990–2010 (Fig. 9c), which is close to the trend from 1990 to 2018 (20.3 ± 7.0 mm decade−1; Fig. 2e) and is significantly greater than the global-mean SSL trend of 10.4 ± 1.2 mm decade−1. In addition to warming, the freshening above 400 m presents also contributes to enhanced SSL rise after 1990, which arises mainly from decadal variations (Figs. 3e–h). We leave the discussion of 0–400-m salinity change in section 6 through an upper-ocean freshwater budget.
(a) Trend map of 0–400-m averaged temperature during 1990–2010 derived from IAP. (b) The difference between 1990–2010 and 1960–90 in Qlw derived from ERA-20C data. (c) As in (a), but for SSL.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
(a) Trend map of 0–400-m averaged temperature during 1990–2010 derived from IAP. (b) The difference between 1990–2010 and 1960–90 in Qlw derived from ERA-20C data. (c) As in (a), but for SSL.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
(a) Trend map of 0–400-m averaged temperature during 1990–2010 derived from IAP. (b) The difference between 1990–2010 and 1960–90 in Qlw derived from ERA-20C data. (c) As in (a), but for SSL.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
6. Summary and discussion
Existing studies based on Argo data have proposed the essential role played by ocean salinity in the sea level change of the SEIO since 2005 (e.g., Llovel and Lee 2015; Jyoti et al. 2019; Huang et al. 2020). By analyzing four ocean observational datasets, this study investigates multidecadal sea level rise since the mid-twentieth century and confirms the considerable effect of ocean salinity change. We revealed a complex mechanism with contributions of temperature and salinity varying with time. Dramatically different results for trends starting from 1960 to 1990 are obtained, as specified in Fig. 10 and summarized below.
Summary of this study. (a) Trends of SSL, TSL, and HSL for 1960–2018 and the contributions of different layers and processes. The percentage of spicing or heaving mode is the contribution of the corresponding mode to the total temperature or salinity changes on different depths. (b) As in (a), but for the 1990–2018 trends. Results are derived from IAP data.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Summary of this study. (a) Trends of SSL, TSL, and HSL for 1960–2018 and the contributions of different layers and processes. The percentage of spicing or heaving mode is the contribution of the corresponding mode to the total temperature or salinity changes on different depths. (b) As in (a), but for the 1990–2018 trends. Results are derived from IAP data.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Summary of this study. (a) Trends of SSL, TSL, and HSL for 1960–2018 and the contributions of different layers and processes. The percentage of spicing or heaving mode is the contribution of the corresponding mode to the total temperature or salinity changes on different depths. (b) As in (a), but for the 1990–2018 trends. Results are derived from IAP data.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
The long-term sea level change over the 1960–2018 period (Fig. 10a) is superimposed upon prominent decadal variability, showing an SSL trend of 9.1 ± 2.6 mm decade−1. The decadal variability is predominantly induced by the thermosteric component associated with strong upper-ocean temperature variability, while its long-term trend contains a contribution by the halosteric component (∼42%) comparable to the thermosteric component (∼58%). During this period, the HSL shows a persistent rising of 3.9 ± 1.0 mm decade−1 and weak short-term fluctuations. The essential effect of salinity change on sea level change arises mainly from a persistent freshening trend over 400–1000 m (with a contribution of ∼59%). Further analysis demonstrates that this freshening trend is mainly associated with density-compensated temperature and salinity changes of the spicing mode, reflecting property alterations of the SAMW and AAIW. These changes originate from the Southern Ocean and spread into the SEIO, likely via the subsurface ocean circulation as passive tracers, which has driven a persistent HSL rise since 1960.
Since ∼1990, the sea level rise has been greatly accelerated (Fig. 10b). The SSL trend for the 1990–2010 period is 25.4 ± 6.6 mm decade−1, 3 times greater than that of 1960–2018. This rapid sea level rise is dominated by the thermosteric component (20.2 ± 5.2 mm decade−1; contribution ∼80%), and the role of the halosteric component is secondary (∼20%). The warming of the 0–400-m layer is critical in driving the total SSL rise (∼78%), which is in turn associated with the heaving mode (vertical displacements of isopycnal surfaces). Although we cannot get a robust conclusion through the heat budget analysis of four OGCM-based datasets, two processes stand out as candidate drivers for the enhanced SEIO warming. The upper-ocean heat convergence in response to surface wind trends and reduced upward longwave radiation associated with anthropogenic greenhouse gas warming worked mutually to drive the SEIO warming.
The salinity effect on multidecadal sea level change highlighted here differs from that on shorter time scales. The freshening of the upper ocean has been shown to dominate decadal HSL changes of the SEIO during the Argo era (e.g., Llovel and Lee 2015; Huang et al. 2020). The upper SEIO above 400 m presents prominent interannual-to-decadal variations, with an obvious freshening after 2010 (Fig. 3), making a ∼44% contribution to HSL rise. Freshwater input from rivers is essential on the salinity of the tropical Indian Ocean, particularly in the Bay of Bengal and the eastern Arabian Sea (Han and McCreary 2001). Yet, this effect is supposed to be negligible in the freshwater budget of the SEIO (Du et al. 2015; Hu et al. 2019), given the lack of strong rivers along Western Australian coasts. We performed a freshwater budget analysis for the 0–400-m layer of the SEIO (Fig. 11). The results suggest that the advection term AdvS can explain the majority of salinity tendency, with a correlation of 0.90, and the surface freshwater forcing [evaporation minus precipitation (E − P)] plays a secondary role (Figs. 11a,b). It is discernible that AdvS and E − P show covariances and operate mutually to drive some of the salinity tendencies, because both are closely associated with the tropical Pacific climate (e.g., Du et al. 2015, 2019; Hu et al. 2019; Guo et al. 2021). Furthermore, we analyzed the advection at different boundaries, suggesting that the salinity advection on the eastern (ITF), western (South Equatorial Current), and northern boundaries are also important for the total AdvS, indicating the complicated circulation variability in this region (Figs. 11b,e). These results based on LICOM3 are generally consistent with previous studies based on various datasets (e.g., Du et al. 2015; Hu et al. 2019; Jyoti et al. 2019) and highlight the difference in mechanism between decadal variability of the upper layer and multidecadal trend of the 400–1000-m layer. The decadal variability in the 0–400-m layer is mostly affected by the advection from the north, while the long-term change of 400–1000 m is mostly affected by SAMW and AAIW from the south.
Freshwater budget for the upper layer (0–400 m) of the SEIO: (a) Salinity tendency (black; ∂S/∂t), surface freshwater forcing term (yellow; E − P), horizontal advection term Adv (red), and residual term (purple; RES), derived from monthly LICOM3 dataset and smoothed with a 13-month running mean. (b) Components of Adv: E, W, N, S, and B denote the advection components across the eastern, western, northern, southern, and bottom (i.e., 400 m) boundaries of the SEIO region, respectively. (c) As in (b), but for surface forcing terms (E − P), with shown evaporation (E; red) and precipitation (−P; yellow). (d)–(f) The standard deviations of (a)–(c), respectively.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Freshwater budget for the upper layer (0–400 m) of the SEIO: (a) Salinity tendency (black; ∂S/∂t), surface freshwater forcing term (yellow; E − P), horizontal advection term Adv (red), and residual term (purple; RES), derived from monthly LICOM3 dataset and smoothed with a 13-month running mean. (b) Components of Adv: E, W, N, S, and B denote the advection components across the eastern, western, northern, southern, and bottom (i.e., 400 m) boundaries of the SEIO region, respectively. (c) As in (b), but for surface forcing terms (E − P), with shown evaporation (E; red) and precipitation (−P; yellow). (d)–(f) The standard deviations of (a)–(c), respectively.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
Freshwater budget for the upper layer (0–400 m) of the SEIO: (a) Salinity tendency (black; ∂S/∂t), surface freshwater forcing term (yellow; E − P), horizontal advection term Adv (red), and residual term (purple; RES), derived from monthly LICOM3 dataset and smoothed with a 13-month running mean. (b) Components of Adv: E, W, N, S, and B denote the advection components across the eastern, western, northern, southern, and bottom (i.e., 400 m) boundaries of the SEIO region, respectively. (c) As in (b), but for surface forcing terms (E − P), with shown evaporation (E; red) and precipitation (−P; yellow). (d)–(f) The standard deviations of (a)–(c), respectively.
Citation: Journal of Climate 35, 5; 10.1175/JCLI-D-21-0288.1
It is worth mentioning that while precipitation dominates decadal fluctuations of E − P, the local evaporation shows a long-term increasing trend. The enhanced evaporation in the SEIO region may be linked to the amplification of the global water cycle (e.g., Durack et al. 2012; Durack 2015), given that the SEIO is under an evaporation-dominant regime in climatology. The increasing surface freshwater release to the atmosphere acts to attenuate the freshening trend induced by ocean advection and the rising trend of HSL.
The salinity change is not isolated in dynamics but coupled with temperature change. In some areas of the low- and midlatitude oceans, advection is the dominant mechanism. The compensation of temperature and salinity changes in density gives rise to negatively correlated HSL and TSL changes (e.g., Munk 2003; Köhl 2014). In other regions, usually at high latitudes (Köhl 2014; Levitus et al. 2005), the halosteric and thermosteric effects can be positively correlated and work in concert. In this study, we find a nontrivial positive correlation in the SEIO, where the temperature and salinity changes are driven by different mechanisms. The warming of 0–400 m is mainly caused by the enhanced Ekman downwelling linked to surface winds and radiation forcing linked to global greenhouse gas warming, while the significant subsurface freshening trend arises from water-mass property changes transported from the Southern Ocean. Moreover, the time-varying feature of the relative importance of temperature and salinity effects on the sea level change also indicates the complexity of the ocean’s response to anthropogenic climate change. Analysis of this study should be performed over the global ocean to gain more comprehensive insights into the regional sea level changes under climate change and evaluate the performance of climate models in simulating them, which is our ongoing research.
Acknowledgments.
Comments from three anonymous reviewers are gratefully appreciated. This research is jointly supported by the Strategic Priority Research Program of Chinese Academy of Sciences (Grant XDB40000000), the National Natural Science Foundation of China (NSFC) Grant 41806001 and 41976026, the Shandong Provincial Natural Science Foundation (ZR2020JQ17), the National Key R&D Program of China (2019YFA0606702), and the Key Deployment Project of CAS Centre for Ocean Mega-Science (COMS2019Q07). IAP data are provided by the Chinese Academy of Sciences and can be found on Lijing Cheng’s website (http://www.ocean.iap.ac.cn/). NCEI data are provided by the National Oceanic and Atmospheric Administration and can be downloaded from https://www.nodc.noaa.gov/. EN4 and Ishii are obtained from https://www.metoffice.gov.uk/hadobs/en4/download-en4-2-1.html and https://www.data.jma.go.jp/gmd/kaiyou/english/ohc/ohc_global_en.html. The AVISO altimeter products are obtained from the Copernicus Marine Environment Monitoring Service (CEMES), which was previously provided by CLS/Archiving, Validation and Interpretation of Satellite Oceanographic data and can be downloaded from http://www.aviso.altimetry.fr/duacs/. The Argo data are collected from http://www.argo.ucsd.edu and http://argo.jcommops.org. ORA-S4 are downloaded from http://apdrc.soest.hawaii.edu/datadoc/ecmwf_oras4.php. SODA2.2.4 data are provided by jointly the Texas A&M University and University of Maryland and the Laboratory of Atmospheric Sciences and downloaded from https://iridl.ldeo.columbia.edu/. GECCO2 data are downloaded from https://icdc.cen.uni-hamburg.de/en/gecco2.html. ERA-Interim, ERA-20C, and ERA-40 data are downloaded from ECMWF interface website https://apps.ecmwf.int/datasets/. NOAA-20CR and NCEP–NCAR are obtained from NOAA’s PSD website https://www.esrl.noaa.gov/psd/data/gridded/.
REFERENCES
Ablain, M., A. Cazenave, G. Valladeau, and S. Guinehut, 2009: A new assessment of the error budget of global mean sea level rate estimated by satellite altimetry over 1993–2008. Ocean Sci., 5, 193–201, https://doi.org/10.5194/os-5-193-2009.
Antonov, J. I., S. Levitus, and T. P. Boyer, 2002: Steric sea level variations during 1957–1994: Importance of salinity. J. Geophys. Res., 107, 8013, https://doi.org/10.1029/2001JC000964.
Balmaseda, M. A., K. Mogensen, and A. T. Weaver, 2013: Evaluation of the ECMWF Ocean Reanalysis System ORAS4. Quart. J. Roy. Meteor. Soc., 139, 1132–1161, https://doi.org/10.1002/qj.2063.
Banks, H. T., R. A. Wood, J. M. Gregory, T. C. Johns, and G. S. Jones, 2000: Are observed decadal changes in intermediate water masses a signature of anthropogenic climate change? Geophys. Res. Lett., 27, 2961–2964, https://doi.org/10.1029/2000GL011601.
Bindoff, N. L., and T. J. McDougall, 1994: Diagnosing climate change and ocean ventilation using hydrographic data. J. Phys. Oceanogr., 24, 1137–1152, https://doi.org/10.1175/1520-0485(1994)024<1137:DCCAOV>2.0.CO;2.
Bindoff, N. L., and T. J. McDougall, 2000: Decadal changes along an Indian Ocean section at 32°S and their interpretation. J. Phys. Oceanogr., 30, 1207–1222, https://doi.org/10.1175/1520-0485(2000)030<1207:DCAAIO>2.0.CO;2.
Cazenave, A., and W. Llovel, 2010: Contemporary sea level rise. Annu. Rev. Mar. Sci., 2, 145–173, https://doi.org/10.1146/annurev-marine-120308-081105.
Chen, C., G. Wang, Y. Yan, and F. Luo, 2021: Projected sea level rise on the continental shelves of the China seas and the dominance of mass contribution. Environ. Res. Lett., 16, 064040, https://doi.org/10.1088/1748-9326/abfdea.
Chen, X., and Coauthors, 2017: The increasing rate of global mean sea-level rise during 1993–2014. Nat. Climate Change, 7, 492–495, https://doi.org/10.1038/nclimate3325.
Cheng, L., and J. Zhu, 2016: Benefits of CMIP5 multimodel ensemble in reconstructing historical ocean subsurface temperature variations. J. Climate, 29, 5393–5416, https://doi.org/10.1175/JCLI-D-15-0730.1.
Cheng, L., K. Trenberth, J. Fasullo, T. Boyer, J. Abraham, and J. Zhu, 2017: Improved estimates of ocean heat content from 1960 to 2015. Sci. Adv., 3, e1601545, https://doi.org/10.1126/sciadv.1601545.
Cheng, L., and Coauthors, 2020: Improved estimates of changes in upper ocean salinity and the hydrological cycle. J. Climate, 33, 10 357–10 381, https://doi.org/10.1175/JCLI-D-20-0366.1.
Church, J. A., and Coauthors, 2001: Change in sea level. Climate Change 2001: The Scientific Basis, J. T. Houghton et al., Eds., Cambridge University Press, 641–693.
Church, J. A., and Coauthors, 2013: Sea level change. Climate Change 2013: The Physical Science Basis. Cambridge University Press, 1137–1216.
Collins, M., and Coauthors, 2013: Long-term climate change: Projections, commitments and irreversibility. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 1029–1136.
Compo, G. P., and Coauthors, 2011: The Twentieth Century Reanalysis Project. Quart. J. Roy. Meteor. Soc., 137 (654), 1–28, https://doi.org/10.1002/qj.776.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, https://doi.org/10.1002/qj.828.
Douglas, B. C., 2001: Sea level change in the era of the recording tide gauge. Sea Level Rise: History and Consequences, B. C. Douglas, M. S. Kearney, and S. P. Leatherman, Eds., Academic Press, 37–64, https://doi.org/10.1016/S0074-6142(01)80006-1.
Du, Y., Y. Zhang, M. Feng, T. Wang, N. Zhang, and S. Wijffels, 2015: Decadal trends of the upper ocean salinity in the tropical Indo-Pacific since mid-1990s. Sci. Rep., 5, 16050, https://doi.org/10.1038/srep16050.
Du, Y., Y. Zhang, and J. Shi, 2019: Relationship between sea surface salinity and ocean circulation and climate change. Sci. China Earth Sci., 62, 771–782, https://doi.org/10.1007/s11430-018-9276-6.
Durack, P. J., 2015: Ocean salinity and the global water cycle. Oceanography, 28, 20–31, https://doi.org/10.5670/oceanog.2015.03.
Durack, P. J., and S. E. Wijffels, 2010: Fifty-year trends in global ocean salinities and their relationship to broad-scale warming. J. Climate, 23, 4342–4362, https://doi.org/10.1175/2010JCLI3377.1.
Durack, P. J., S. E. Wijffels, and R. J. Matear, 2012: Ocean salinities reveal strong global water cycle intensification during 1950 to 2000. Science, 336, 455–458, https://doi.org/10.1126/science.1212222.
Durack, P. J., S. E. Wijffels, and P. J. Gleckler, 2014: Long-term sea-level change revisited: The role of salinity. Environ. Res. Lett., 9, 114017, https://doi.org/10.1088/1748-9326/9/11/114017.
England, M. H., and Coauthors, 2014: Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus. Nat. Climate Change, 4, 222–227, https://doi.org/10.1038/nclimate2106.
Feng, M., G. Meyers, A. Pearce, and S. Wijffels, 2003: Annual and interannual variations of the Leeuwin Current at 32°S. J. Geophys. Res., 108, 3355, https://doi.org/10.1029/2002JC001763.
Feng, M., A. Biastoch, C. Böning, N. Caputi, and G. Meyers, 2008: Seasonal and interannual variations of upper ocean heat balance off the west coast of Australia. J. Geophys. Res., 113, C12025, https://doi.org/10.1029/2008JC004908.
Feng, M., J. Benthuysen, N. Zhang, and D. Slawinski, 2015a: Freshening anomalies in the Indonesian Throughflow and impacts on the Leeuwin Current during 2010–2011. Geophys. Res. Lett., 42, 8555–8562, https://doi.org/10.1002/2015GL065848.
Feng, M., and Coauthors, 2015b: Decadal increase in Ningaloo Niño since the late 1990s. Geophys. Res. Lett., 42, 104–112, https://doi.org/10.1002/2014GL062509.
Frederikse, T., and Coauthors, 2020: The causes of sea-level rise since 1900. Nature, 584, 393–397, https://doi.org/10.1038/s41586-020-2591-3.
Frölicher, T. L., and Coauthors, 2015: Dominance of the Southern Ocean in anthropogenic carbon and heat uptake in CMIP5 models. J. Climate, 28, 862–886, https://doi.org/10.1175/JCLI-D-14-00117.1.
Frölicher, T. L., E. M. Fischer, and N. Gruber, 2018: Marine heatwaves under global warming. Nature, 560, 360–364, https://doi.org/10.1038/s41586-018-0383-9.
Giese, B. S., and S. Ray, 2011: El Niño variability in simple ocean data assimilation (SODA), 1871–2008. J. Geophys. Res., 116, C02024, https://doi.org/10.1029/2010JC006695.
Good, S. A., M. J. Martin, and N. A. Rayner, 2013: EN4: Quality controlled ocean temperature and salinity profiles and monthly objective analyses with uncertainty estimates. J. Geophys. Res. Oceans, 118, 6704–6716, https://doi.org/10.1002/2013JC009067.
Gregory, J. M., and J. A. Lowe, 2000: Predictions of global and regional sea-level rise using AOGCMs with and without flux adjustment. Geophys. Res. Lett., 27, 3069–3072, https://doi.org/10.1029/1999GL011228.
Guo, Y., Y. Li, F. Wang, Y. Wei, and Z. Rong, 2020a: Processes controlling sea surface temperature variability of Ningaloo Niño. J. Climate, 33, 4369–4389, https://doi.org/10.1175/JCLI-D-19-0698.1.
Guo, Y., Y. Li, F. Wang, Y. Wei, and Q. Xia, 2020b: Importance of resolving mesoscale eddies in the model simulation of Ningaloo Niño. Geophys. Res. Lett., 47, e2020GL087998, https://doi.org/10.1029/2020GL087998.
Guo, Y., Y. Li, F. Wang, and Y. Wei, 2021: Ocean salinity aspects of the Ningaloo Niño. J. Climate, 34, 6141–6161, https://doi.org/10.1175/JCLI-D-20-0890.1.
Häkkinen, S., P. B. Rhines, and D. L. Worthen, 2016: Warming of the global ocean: Spatial structure and water-mass trends. J. Climate, 29, 4949–4963, https://doi.org/10.1175/JCLI-D-15-0607.1.
Hamlington, B. D., and Coauthors, 2020: Investigating the acceleration of regional sea level rise during the satellite altimeter era. Geophys. Res. Lett., 47, e2019GL086528, https://doi.org/10.1029/2019GL086528.
Han, W., and J. P. McCreary, 2001: Modeling salinity distributions in the Indian Ocean. J. Geophys. Res., 106, 859–877, https://doi.org/10.1029/2000JC000316.
Han, W., and Coauthors, 2010: Patterns of Indian Ocean sea-level change in a warming climate. Nat. Geosci., 3, 546–550, https://doi.org/10.1038/ngeo901.
Han, W., and Coauthors, 2014: Intensification of decadal and multi-decadal sea level variability in the western tropical Pacific during recent decades. Climate Dyn., 43, 1357–1379, https://doi.org/10.1007/s00382-013-1951-1.
Han, W., D. Stammer, G. A. Meehl, A. Hu, F. Sienz, and L. Zhang, 2018: Multi-decadal trend and decadal variability of the regional sea level over the Indian Ocean since the 1960s: Roles of climate modes and external forcing. Climate, 6, 51, https://doi.org/10.3390/cli6020051.
Hanawa, K., and L. D. Talley, 2001: Mode waters. Ocean Circulation and Climate: Observing and Modelling the Global Ocean, G. Siedler, J. Church, and J. Gould, Eds., Academic Press, 373–386.
Haumann, F. A., N. Gruber, M. Münnich, I. Frenger, and S. Kern, 2016: Sea-ice transport driving Southern Ocean salinity and its recent trends. Nature, 537, 89–92, https://doi.org/10.1038/nature19101.
Hong, Y., Y. Du, T. Qu, Y. Zhang, and W. Cai, 2020: Variability of the Subantarctic Mode Water volume in the South Indian Ocean during 2004–2018. Geophys. Res. Lett., 47, e2020GL087830, https://doi.org/10.1029/2020GL087830.
Hu, S., and J. Sprintall, 2016: Interannual variability of the Indonesian Throughflow: The salinity effect. J. Geophys. Res. Oceans, 121, 2596–2615, https://doi.org/10.1002/2015JC011495.
Hu, S., and Coauthors, 2019: Interannual to decadal variability of upper ocean salinity in the southern Indian Ocean and the role of the Indonesian Throughflow. J. Climate, 32, 6403–6421, https://doi.org/10.1175/JCLI-D-19-0056.1.
Huang, J., W. Zhuang, X. Yan, and Z. Wu, 2020: Impacts of the upper-ocean salinity variations on the decadal sea level change in the southeast Indian Ocean during the Argo era. Acta Oceanol. Sin., 39, 1–10, https://doi.org/10.1007/s13131-020-1574-4.
Huang, R. X., 2020: Heaving, Stretching and Spicing Modes. Springer, 399 pp., https://doi.org/10.1007/978-981-15-2941-2.
Ishii, M., Y. Fukuda, S. Hirahara, S. Yasui, T. Suzuki, and K. Sato, 2017: Accuracy of global upper ocean heat content estimation expected from present observational data sets. SOLA, 13, 163–167, https://doi.org/10.2151/sola.2017-030.
Jia, F., L. Wu, J. Lan, and B. Qiu, 2011: Interannual modulation of eddy kinetic energy in the southeast Indian Ocean by Southern Annular Mode. J. Geophys. Res., 116, C02029, https://doi.org/10.1029/2010JC006699.
Jyoti, J., P. Swapna, R. Krishnan, and C. V. Naidu, 2019: Pacific modulation of accelerated south Indian Ocean sea level rise during the early 21st century. Climate Dyn., 53, 4413–4432, https://doi.org/10.1007/s00382-019-04795-0.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471, https://doi.org/10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
Kirtman, B., and Coauthors, 2013: Near-term climate change: Projections and predictability. Climate Change 2013: The Physical Science Basis, T. F. Stocker et al., Eds., Cambridge University Press, 953–1028.
Koch-Larrouy, A., R. Morrow, T. Penduff, and M. Juza, 2010: Origin and mechanism of Subantarctic Mode Water formation and transformation in the southern Indian Ocean. Ocean Dyn., 60, 563–583, https://doi.org/10.1007/s10236-010-0276-4.
Köhl, A., 2014: Detecting processes contributing to interannual halosteric and thermosteric sea level variability. J. Climate, 27, 2417–2426, https://doi.org/10.1175/JCLI-D-13-00412.1.
Large, W. G., and S. Pond, 1981: Open ocean momentum flux measurements in mode rate to strong winds. J. Phys. Oceanogr., 11, 324–336, https://doi.org/10.1175/1520-0485(1981)011<0324:OOMFMI>2.0.CO;2.
Lee, S. K., W. Park, M. Baringer, A. Gordon, B. Huber, and Y. Liu, 2015: Pacific origin of the abrupt increase in Indian Ocean heat content during the warming hiatus. Nat. Geosci., 8, 445–449, https://doi.org/10.1038/ngeo2438.
Lee, T., I. Fukumori, and B. Tang, 2004: Temperature advection: Internal versus external processes. J. Phys. Oceanogr., 34, 1936–1944, https://doi.org/10.1175/1520-0485(2004)034<1936:TAIVEP>2.0.CO;2.
Levitus, S., J. I. Antonov, T. P. Boyer, H. E. Garcia, and R. A. Locarnini, 2005: Linear trends of zonally averaged thermosteric, halosteric, and total steric sea level for individual ocean basins and the world ocean, (1955–1959)–(1994–1998). Geophys. Res. Lett., 32, L16601, https://doi.org/10.1029/2005GL023761.
Levitus, S., and Coauthors, 2012: World ocean heat content and thermosteric sea level change (0–2000 m), 1955–2010. Geophys. Res. Lett., 39, L10603, https://doi.org/10.1029/2012GL051106.
Li, Y., and F. Wang, 2015: Thermocline spiciness variations in the tropical Indian Ocean observed during 2003–2014. Deep-Sea Res. I, 97, 52–66, https://doi.org/10.1016/j.dsr.2014.12.004.
Li, Y., W. 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, https://doi.org/10.1002/2017GL075050.
Li, Y., W. Han, L. Zhang, and F. Wang, 2019: Decadal SST variability in the southeast Indian Ocean and its impact on regional climate. J. Climate, 32, 6299–6318, https://doi.org/10.1175/JCLI-D-19-0180.1.
Li, Y., and Coauthors, 2020: Eddy-resolving simulation of CAS-LICOM3 for phase 2 of the Ocean Model Intercomparison Project. Adv. Atmos. Sci., 37, 1067–1080, https://doi.org/10.1007/s00376-020-0057-z.
Lin, P., Z. Yu, H. Liu, Y. Yu, and J. Ma, 2020: LICOM model datasets for the CMIP6 Ocean Model Intercomparison Project. Adv. Atmos. Sci., 37, 239–249, https://doi.org/10.1007/s00376-019-9208-5.
Llovel, W., and T. Lee, 2015: Importance and origin of halosteric contribution to sea level change in the southeast Indian Ocean during 2005–2013. Geophys. Res. Lett., 42, 1148–1157, https://doi.org/10.1002/2014GL062611.
Llovel, W., S. Guinehut, and