Unique behavior of marine conditions in the Java Sea reconstructed from a 70 yr coral δ¹⁸O and Sr/Ca record from the Seribu Islands, Indonesia

Abstract

The Indonesian Throughflow (ITF) plays an important role in the heat flux and water budget between the Pacific and Indian oceans and may modulate climate variability. During the boreal winter monsoon, low-salinity, buoyant water carried from the Java Sea to the Southern Makassar Strait retards the sea-surface transport of the ITF, which may affect the Asian monsoon and climate change. However, observation records are inadequate to elucidate the marine environment around the Indonesian Seas. We analyzed coral Sr/Ca and δ¹⁸O from the Seribu Islands, Java Sea, and reconstructed sea-surface temperature (SST) and sea-surface salinity (SSS) for 1931–2002. The SST data indicate abrupt warming in the mid-1950s and, almost simultaneously, a rapid SSS shift to saline conditions. The relationships between SST around the Seribu Islands and climate variability in the Pacific and Indian oceans have changed after this abrupt warming events. Before the mid-1950s, during September to November, SST varied with the Indian Ocean Dipole, whereas El Niño–Southern Oscillation also affected SST variation after the mid-1950s. This abrupt change seems to be related to a regime shift in the tropical Pacific and Indian oceans, but there are no clear changes corresponding to other regime shifts such as that of 1970s. SSS variation exhibits no relationship with climatic factors, indicating that the dominant controlling factors of SST and SSS should be considered separately. Marine conditions in the Java Sea that affect the ITF show unique behavior, and further local studies in the Indonesian Seas are crucial to understand ITF behavior.

Introduction

The tropical Indonesian Archipelago, which lies between the Pacific and Indian oceans, is important for understanding climate variability related to ocean–atmosphere interactions such as the El Niño–Southern Oscillation (ENSO) and the Indian Ocean Dipole (IOD). In this area, the Indonesian Throughflow (ITF) runs from the Pacific Ocean to the Indian Ocean through the Indonesian Seas and is a major pathway in global thermohaline circulation (Fig. 1). Because the ITF plays an important role in the heat flux and water budget, and the air–sea heat flux, it may modulate climate variability in the Pacific and Indian oceans (e.g., Sprintall et al., 2014). During the boreal winter monsoon, eastward surface currents carry buoyant, low salinity seawater from the South China Sea to the southern Makassar Strait through the Java Sea (Fig. 1). This water retards transportation of the ITF in the surface layer of the Makassar Strait and is called the “freshwater plug” (Gordon et al., 2003). In contrast, during the boreal summer monsoon, the sea-surface currents reverse and remove the freshwater plug from the southern Makassar Strait. Thus, the behavior of the freshwater plug has been suggested to influence the Asian monsoon and climate change (Gordon et al., 2003). Although understanding marine conditions and their variability in the Indonesian Seas is critical to elucidating regional and global climate change, instrumental and observed records around this area are inadequate. For example, Sprintall et al. (2003) reported in situ observed data of sea-surface temperature (SST) and sea-surface salinity (SSS) in the Lombok Strait, the Ombai Strait, and the Timor passage, which are the outlets of the ITF; however, according to the data presented by Cahyarini et al. (2014), the SST data obtained by Sprintall et al. (2003) differ by approximately 1°C from satellite data.

Fig. 1.

Map of (a) SST and (b) SSS around the Indonesian Seas in January 1997. Maps of SST and SSS were obtained from IGOSS nmc Reyn_SmithOlv2 (http://iridl.ldeo.columbia.edu/SOURCES/.IGOSS/.nmc/.Reyn_SmithOIv2/.monthly/) and CARTON-GIESE SODA 2.1.6 (http://iridl.ldeo.columbia.edu/SOURCES/.CARTON-GIESE/.SODA/.v2p1p6/.salt/). Black arrows in (b) represent the ITF pathway and blue solid and dushed arrows show seasonal surface flow during the boreal winter and summer monsoons. The studied coral sample was collected from the Seribu Islands; an enlarged map showing the sampling site is embedded in (a). (c) Comparison of mean SSS for Bunaken (1.25°N, 124.25°E), the Makassar Strait (5.25°S, 117.75°E), the Lombok Strait (8.25°S, 115.25°E), Timor (10.25°S, 123.75°E), and the Seribu Islands (5.75°S, 106.75°E). The mean SSS data are from CARTON-GIESE SODA 2.1.6 for the period of 1960–1989. (d) Coral δ¹⁸O in Bunaken (Charles et al., 2003), the Makassar Strait (Linsley et al., 2017; Murty et al., 2017), the Lombok Strait (Charles et al., 2003; Murty et al., 2018), Timor (Cahyarini et al., 2014) and the Seribu Islands (this study). All plotted data is the average for 1960–1989.

Coral geochemistry has been applied to improve our understanding of the behavior of the ITF and related marine environments (Moore et al., 1997; Charles et al., 2003; Cahyarini et al., 2014, 2016; Linsley et al., 2017). For example, Murty et al. (2017) reconstructed SSS on the basis of coral skeletons from the southern Makassar Strait. They reported that the surface water around the southern part of the strait is carried from the Luzon Strait and the Java Sea, and their mixing ratio is related to the wind strength of the East Asian winter monsoon. Charles et al. (2003) collected two coral samples from Bunaken and Bali and measured coral δ¹⁸O (δ¹⁸O_c) to examine the relationship between monsoon and ocean interaction. Coral record of Bunaken which is located Pacific Ocean side, showed the effect of monsoon and ENSO, and that of Bali which is near the Indian Ocean, also showed an interannual variation related with ENSO. The association with ENSO has also been reported in previous studies using corals in the Makassar Strait. Moore et al. (1997) reported that δ¹⁸O_c of Langaki located in the Makassar Strait is affected by monsoon in seasonal scale, while the drought condition during the ENSO warm phase is found in time-series variation of δ¹⁸O_c. In addition, the interannual variability of the composite δ¹⁸O_c data from the southern Makassar Strait was reported to have relation with Niño 3.4 SST anomaly (Linsley et al., 2017). Cahyarini et al. (2014) used coral Sr/Ca and δ¹⁸O_c data from Timor Island to reconstruct SST and SSS and reported that there are strong correlations between SST/SSS and IOD on a decadal cycle, whereas on an interannual cycle, there are no strong correlations between coral reconstructed indices such as SST and SSS and climate phenomena such as ENSO and IOD. Although our understanding of marine environments under the pathway of ITF and its relation to climate conditions such as monsoons, ENSO, and IOD are improving based on coral studies and in situ observations, additional basic data, especially over long-time scales, are more required to understand marine environments in the complex setting of the Indonesian Seas.

We obtained Sr/Ca and δ¹⁸O_c data from Lanchau Island, one of the Seribu Islands, to reconstruct SST and seawater δ¹⁸O (δ¹⁸O_sw), which is related to the surface-ocean balance between evaporation and precipitation and approximate to SSS (Gagan et al., 1998; Cahyarini et al., 2008, 2014). Our study differs from a previous work (Cahyarini et al., 2016) that measured Sr/Ca from inshore and offshore areas of the Seribu Islands to reconstruct air temperature and SST in that we also attempt to reconstruct SSS (which is thought to markedly influence the ITF) by combining Sr/Ca and δ¹⁸O_c data. The results of this study, together with those of previous studies using a coral from the Indonesian Seas, will help to elucidate the relationship between the ITF, the complex marine environments in the Indonesian Seas, and climate change.

Materials and Methods

Drill-core SER03-05 was collected in 2003 from a massive Porites coral at a water depth of 2–3 m at Lanchau Island (5.9°S, 106.6°E) located in the Java Sea (Fig. 1a). The SST record from the Integrated Global Ocean Services System (IGOSS, 5.5°S, 106.5°E) shows a bimodal seasonal cycle, with SST maxima in May and November and minima in February and August, respectively. Over the Indonesian Seas, monsoonal winds have been considered to be the dominant influence on SST variations (Qu et al., 2005), and the highest SST each year is observed when the monsoonal winds die down and change. Warm SST anomalies were detected by means of satellite data around Lanchau Island during the 1991/92 and 1987/88 strong El Niño events, whereas during 1982/83 and 1997/98 strong El Niño events with positive IODs, the SST changed from cool anomaly to warm anomaly. In 1994, when there was a positive IOD with moderate El Niño event, cool SST anomaly was detected. The SSS record from the CRATON-GIESE Simple Ocean Data Assimilation 2.1.6 (SODA, 5.75°S, 106.75°E) shift to fresh condition from February to April and saline condition from October to December. In the southern Indonesian region including the Java Sea, the amount of rainfall increases during the northwest monsoon (Aldrian and Susanto, 2003), and it seems that the rainfall affects the SSS of coastal area through river runoff. During the northwest monsoon, upwelling occurs in the central part of north side of Java Island to Lombok Island, while there is no strong upwelling in the western part of the Java Sea (Siswanto and Suratno, 2008). Then effects of upwelling on SST/SSS seem to be negligible around our sampling site.

The drilled core was cut into slabs for X-radiography to observe the annual density bands (Supplementary Fig. S1). The growth rate of the coral was approximately 11–31 mm/year (average 21 mm/year). We analyzed the upper 148 cm of the core along sampling transects that followed the growth axis. Powder samples were collected at 0.4-mm intervals by milling along a ledge cut into the coral slabs, as described by Gagan et al. (1994). Measurements of coral Sr/Ca and δ¹⁸O_c were obtained from every fourth sample (1.6-mm intervals), corresponding to a temporal resolution of approximately one month.

Sr/Ca measurements were performed by Agilent 720 inductively coupled plasma optical emission spectrometry at Okayama University (Okayama city, Okayama, Japan). Briefly, 100 μg of powder was dissolved in 5 ml of 2% HNO₃. A standard solution for Sr/Ca was prepared from JCp-1, provided by the Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology (AIST; Tsukuba city, Ibaraki, Japan) (Okai et al., 2002). The relative standard deviation was 0.3% (n = 314, 1σ).

Measurements of δ¹⁸O were obtained by using an online system employing an IsoPrime isotope-ratio mass spectrometer (GV Instruments Ltd.) coupled to a Multicarb automatic sample treatment system at the Center for Advanced Marine Core Research at Kochi University (CMCR; Nankoku city, Kochi, Japan) and AIST, and an automated individual-carbonate reaction (Kiel) device coupled with a Finnigan MAT251 mass spectrometer at the Research School of Earth Sciences, Australian National University (ANU; Canberra, ACT, Australia). The upper part of the core (~419.2 mm) was analyzed at ANU and the rest was mainly and complementarily analyzed at CMCR and AIST, respectively. All δ¹⁸O_c data were normalized to the Vienna Pee Dee Belemnite (V-PDB) scale using the international standard NBS-19 (δ¹⁸O = –2.2‰) of the National Institute of Standards and Technology (Gaithersburg, Maryland, USA). The IAEA-603 standard (δ¹⁸O = –2.37‰ ± 0.04‰) of the International Atomic Energy Agency (Vienna, Australia) was partially used for the measurements at CMCR. The standard deviation for replicate measurements of δ¹⁸O on NBS-19 within the mass spectrometer runs was 0.06‰ at CMCR (n = 143), 0.04‰ (n = 49) at AIST, and 0.03‰ at ANU (n = 53).

The age model for the SER03-05 coral record was based on counting of annual density bands (Supplementary Fig. S1) and the seasonal Sr/Ca cycle. First, we estimated the age of coral growth by counting the annual density bands visible in X-radiographs of the coral slabs backward from 2003. We then improved the precision of the coral chronology by anchoring the two maxima (minima) of Sr/Ca to the lowest (highest) SSTs in February/August (May/November) within each year, interpolating linearly between maxima/minima.

Coral Sr/Ca ratios were calibrated using IGOSS SST data for the 1 × 1 degree grid area surrounding Lanchau Island for 1982–2002. The linear best fit of the data was established using the maximum Sr/Ca values, which corresponded to the austral winter (July to September) and the minimum Sr/Ca values, which corresponded to the austral summer (April to June). Calibration was performed as follows (Supplementary Fig. S2a);

   Sr/Ca (mmol/mol) = (–0.044 ± 0.005) × T (°C) + (9.880 ± 0.155)    (1)   

Eq. (1) exhibits a high correlation (r² = 0.63) between Sr/Ca and SST from IGOSS data. Published values of the regression slope, which corresponds to the temperature dependence of coral Sr/Ca ratios from the Indonesian Seas, range from –0.04 to –0.07 mmol/mol/°C (Gagan et al., 1998; Cahyarini et al., 2014, 2016; Krawczyk et al., 2020). The calibration in this study falls at the lower end of the range.

The values of δ¹⁸O_sw were estimated by subtracting SST components from δ¹⁸O_c (Cahyarini et al., 2008).

   Δδ¹⁸O_sw = (δ¹⁸O_c – δ¹⁸O_{c, mean}) – γ₁/β₁ (Sr/Ca – Sr/Ca_mean)    (2)   

δ¹⁸O_{c, mean} and Sr/Ca_mean are the mean values of δ¹⁸O_c and Sr/Ca, respectively. γ₁ indicates the slope of δ¹⁸O_c–IGOSS SST; and β₁ indicates the slope of Sr/Ca–IGOSS SST. During calculation of Eq. (2), we applied values of –0.22‰/°C for γ₁ and –0.044 mmol/mol/°C for β₁ (Supplementary Fig. S2b).

Results and Discussion

We analyzed Sr/Ca and δ¹⁸O_c with monthly resolution for the period 1931–2002. The range of Sr/Ca ratios was 8.48 to 8.78 mmol/mol, with a mean value of 8.61 mmol/mol, while that of δ¹⁸O_c was –5.2‰ to –7‰, with a mean value of –6‰. Although δ¹⁸O_c reflects both SST and δ¹⁸O_sw (Gagan et al., 1998; Cahyarini et al., 2008, 2014), the difference in mean annual SST across the region is small (~0.6°C); thus, the differences in average δ¹⁸O_c correspond to variations of δ¹⁸O_sw (SSS). As mentioned above, the distribution and variation of surface salinity are key to understanding ITF behavior, and the Java Sea exhibits extremely low salinity, mainly resulting from heavy precipitation and transport of lower-salinity water from the South China Sea (Fig. 1b). A comparison of averaged δ¹⁸O_c values from several places in the Indonesian Seas indeed showed the lowest value in Seribu coral (Fig. 1d). The distribution of δ¹⁸O_c in the Indonesian Seas matches that of salinity data (Fig. 1c, d). Because the averaged water temperature among previously studied sites is almost identical (>28°C, Fig. 1a), δ¹⁸O_c recorded in Seribu coral probably reflects the lower-salinity conditions around the Seribu Islands.

SST reconstructed from Sr/Ca ratios exhibit a seasonal cycle with a range of 1–4°C, and the variability of reconstructed SST is almost the same with the satellite data (Supplementary Fig. S2c). Reconstructed Δδ¹⁸O_sw (SSS) demonstrates fresh condition during the northwest monsoon, while saline condition during the southeast monsoon. This tendency corresponds with seasonal pattern of rainfall in the southern Indonesia and SODA SSS (Supplementary Fig. S2d) although there are several discrepancies between Δδ¹⁸O_sw and SODA SSS. These local and seasonal marine condition at the sampling site would be mainly controlled by the Asian monsoon and its related surface current. Both δ¹⁸O_c and Sr/Ca of the coral mostly reflect such a local condition, although relatively large and dynamic variations of SST and SSS on interannual to decadal scale found in the long-term record cannot be attribute to only the monsoon variation.

For the long-term trend of reconstructed SST, we found a warming trend of +0.74°C/70 year (Fig. 2a). Dividing a whole period into 1931–1954, 1955–1982 and 1983–2002 in which SST and/or SSS changed largely as discussed below, the mean reconstructed SST has changed from 28.6°C, 29.1°C to 29.2°C while Δδ¹⁸O_sw has changed from –0.11‰, 0.17‰ to –0.11‰ at each period. The SST data of this study exhibit a shift to warm conditions in the mid-1950s and the Δδ¹⁸O_sw record displays a rapid shift to saline conditions at almost the same time, which continued until 1982 (Fig. 2b). Stress bands could not be identified in the X-radiograph of the coral skeleton and indicating that there was no abnormality in growth around that time (Supplementary Fig. S1). Also, coral Mg/Ca which has a potential for proxy of skeletal growth rate showed no abnormal variation during this period (Supplementary Fig. S3).

Fig. 2.

Time-series of (a) reconstructed SST using coral Sr/Ca and (b) Δδ¹⁸O_sw reconstructed from combined Sr/Ca and δ¹⁸O_c data with monthly resolution. The y-axis for Δδ¹⁸O_sw is reversed.

To examine which global climate factors had the greatest effects on the temporal variations of SST and SSS around the Java Sea, monthly data were compared to ENSO and IOD variations. Comparison of coral data with ENSO variability shows that a warm SST anomaly accompanying lower SSS conditions was associated with very strong El Niño events (>2 on the Ocean Niño index) that occurred in 1983 and 1998 (Supplementary Fig. S4). However, other El Niño events seem to have not affected marine conditions around the Java Sea, suggesting ENSO variability does not predominantly affect the Java Sea. Because positive IODs mostly occurred with El Niño events, independently determining the impact of the IOD on the Java Sea is difficult. Although Cahyarini et al. (2016) suggested that seasonal SST variations at the Seribu Islands co-vary with that in the southeastern Indian Ocean, which is correlated with IOD variation, the reconstructed SST in this study does not consistently reflect IOD variation (Supplementary Fig. S5).

For further insight into the SST and SSS variations, we compare seasonally aggregated data with variations of ENSO and IOD. SST for the September to November (SON) and December to February (DJF) periods, when IOD and ENSO are maturely developed, varied widely compared to that in other seasons (Fig. 3, Supplementary Figs. S6, S7). The abrupt warming that we detected in the mid-1950s is also apparent in the SON and DJF data. Focusing on the SON period, corresponding to the mature phase of the IOD and the onset season of El Niño, the variations of both SST and SSS are similar to those of the IOD with a negative correlation prior to the abrupt warming occurred in the mid-1950s (Fig. 3). This trend is also visible in other seasons (Supplementary Fig. S7), so we hypothesize that SST and SSS in the Java Sea were mainly controlled by climate factors in the Indian Ocean rather than those in the Pacific Ocean at least from 1931 to the mid-1950s. The peak of the abrupt warming occurred in 1955–1956 and the moderate La Niña event during 1954–1956 could have contributed to this warming (Fig. 3). In contrast, Δδ¹⁸O_sw shifted to saline conditions at the same time (Fig. 2), but this trend generally seems to have followed the IOD variation. This trend in which Δδ¹⁸O_sw followed the IOD variation especially before the mid-1950s are found in all seasons, as is apparent in the data for June to August (JJA) and SON (Fig. 3, Supplementary Fig. S7).

Fig. 3.

The comparison of variability of Niño 3.4 index with the anomalies of (a) reconstructed SST and (c) Δδ¹⁸O_sw, and dipole mode index (DMI) with the anomalies of (b) reconstructed SST and (d) Δδ¹⁸O_sw during September–October–November. The peak of rapid warming in the mid-1950s is shown with a yellow bar in left panels. The correlation coefficients with climate factors (Niño 3.4 index or DMI) are represented in all figures. For the comparison of SST (a, b), R values were calculated before and after the peak of rapid warming (1932–1954 and 1957–2002), while those for the Δδ¹⁸O_sw (c, d) were obtained for all periods (1932–2002). We acquire monthly Niño 3.4 index (based on NOAA ERSSTv5) and DMI (based on NOAA ERSSTv5) from KNMI climate explorer (https://climexp.knmi.nl/).

SST variation exhibits a negative correlation with ENSO variation after this warming (Fig. 3a, r = –0.35, n = 46 (1956–2002)), whereas there was no obvious relationship between them before the abrupt shift (Fig. 3a, r = –0.09, n = 23 (1932–1954)). Therefore, for SST variation with a particular focus on the SON season, there was an abrupt shift in the mid-1950s, during which control of SST variations in the Java Sea by the Indian Ocean climate changed to control by the Pacific Ocean climate. Although similar variation was not detected in reanalyzed SST data from the Extended Reconstructed SST (ERSST) v5 (Supplementary Fig. S8b, f), a coral record from Timor indicated rapid warming in the late 1940s (Cahyarini et al., 2014, Supplementary Fig. S8a, e). A regime shift in the late 1940s or early 1950s has been reported for the north Pacific, the western north America (Minobe, 1997) and the northern hemisphere (Yasunaka and Hanawa, 2002). The anomaly of annual mean SST in the Indian Ocean–maritime continent region (15°N–15°S, 40°E–160°E), including the Indonesian archipelago, decreased in the late 1940s in associated with the Aleutian low (Minobe, 1997). Although the SST variability around the Seribu Islands and Timor (Cahyarini et al., 2014) in the late 1940s to mid-1950s differs from the results of Minobe (1997), it is possible that SST locally rose in the Indonesian Seas due to the complex ITF marine condition. Although the exact timing and behavior of this climate change could differ among sites, the abrupt shift to warmer and more saline conditions during the mid-1950s detected in this study may possibly be attributed to a regime shift. However, other regime-shift-like changes have not been detected in the Java Sea, whereas several regime shifts, such as the one that occurred in the mid-1970s have been reported (Yasunaka and Hanawa, 2002, 2005; Meehl et al., 2009). The effects of the significant regime shift in the mid-1970s extended globally; thus, the absence of clear changes in the Java Sea other than in the mid-1950s suggests complicated marine conditions of the Java Sea in which interaction between global and regional to local climate conditions may control.

While SST variation showed an abrupt shift of relationship between SST and ENSO and IOD, SSS variations show no clear relationship with either ENSO or IOD after the mid-1950s, indicating that the predominant climatic factor(s) controlling the SST and SSS variations in this region should be considered separately. Although temporal SST variations are relatively simply controlled by ENSO and IOD events with a regime shift-like change during the mid-1950s, SSS variations are governed more by local and/or regional conditions such as local precipitation, the Asian monsoon and surface currents. Reconstructed δ¹⁸O_sw from Seribu has showed that δ¹⁸O_sw shifted towards saline condition after the mid-1950s when rapid warming was found at the Java Sea as discussed above (Supplementary Fig. S8c, g). Also, this saline condition intermittently continued to around 1982, and then shift to fresher condition. Although similar variation was not detected SODA SSS data (Supplementary Fig. S8d, h), a coral record from Timor showed the relatively saline condition from the late 1940s to early 1980s and fresher condition since around 1985 (Supplementary Fig. S8c, g). Thus, it seems sea surface condition in southern part of Indonesia shifted to fresher recently based on coral records. Although a mechanism of salinity change in the past cannot be elucidated, this recent freshening trend is likely consistent to that found in SODA SSS records. Because the ITF is modulated by lower-salinity surface water brought by monsoon winds (Gordon et al., 2003), further careful examination of SSS variation in the Indonesian Seas is crucial.

Conclusions

We reconstructed SST and SSS records for the period 1931–2002 based on Sr/Ca and δ¹⁸O in a coral skeleton collected from the Seribu Islands, Java Sea. The reconstructed SST and Δδ¹⁸O_sw (SSS) records show the seasonal cycle associated with monsoon, especially, light-Δδ¹⁸O_sw-values reflect local precipitation and low salinity water carried by the sea surface current. For the long-time scale, we found a warming trend of +0.74°C/70 years and abrupt shifts towards warming and saline conditions in the mid-1950s. Based on investigations of relationship between SST/SSS and ENSO/IOD during September to November when IOD and ENSO are maturely developed, it is suggested that the climate factors that predominantly control SST in the Java Sea have changed in the mid-1950s from the IOD control to the ENSO control. On the other hand, SSS data show no clear relationship with either ENSO or IOD after the mid-1950s, implying that we should consider SST and SSS variations separately at least in this region.

Acknowledgments

We express our deep appreciation to Ministry for Research and Technology, Indonesia (RISTEK), which issued the permit for research, measurement of oceanological parameters, and collection of coral samples in the Seribu Islands, Indonesia. We thank the Indonesian Institute of Sciences (LIPI) scientists who took part in the research and publication activities. This study was performed under the cooperative research program of the Center for Advanced Marine Core Research (CMCR), Kochi University (16B052, 17A014, and 17B014). We thank T. Ishimura and Y. Yoshinaga for their help and support with the preparation and performance of δ¹⁸O measurement at the Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology. We are also grateful to Michael K. Gagan, Joe Cali and Heather Scott-Gagan for assistance with coral milling and δ¹⁸O measurement at the Research School of Earth Sciences, ANU. This study was supported by Japan Society for the Promotion of Science KAKENHI (grant numbers JP15H05329 and JP20K12135 to M. I) and Kurita Water and Environment Foundation (19B063 to M.I.).

References

•  Aldrian,  E. and  Susanto,  R. D. (2003) Identification of three dominant rainfall regions within Indonesia and their relationship to sea surface temperature. Int. J. Climatol. 23, 1435–1452. (https://doi.org/10.1002/joc.950)
•  Cahyarini,  S. Y.,  Pfeiffer,  M.,  Timm,  O.,  Dullo,  W. C. and  Schönberg,  D. G. (2008) Reconstructing seawater δ¹⁸O from paired coral δ¹⁸O and Sr/Ca ratios: Methods, error analysis and problems, with examples from Tahiti (French Polynesia) and Timor (Indonesia). Geochim. Cosmochim. Acta 72, 2841–2853. (https://doi.org/10.1016/j.gca.2008.04.005)
•  Cahyarini,  S. Y.,  Pfeiffer,  M.,  Nurhati,  I. S.,  Aldrian,  E.,  Dullo,  W. C. and  Hetzinger,  S. (2014) Twentieth century sea surface temperature and salinity variations at Timor inferred from paired coral δ¹⁸O and Sr/Ca measurements. J. Geophys. Res. Oceans 119, 4593–4604. (https://doi.org/10.1002/2013JC009594)
•  Cahyarini,  S. Y.,  Zinke,  J.,  Troelstra,  S., Suharsono,  Aldrian,  E. and  Hoeksema,  B. W. (2016) Coral Sr/Ca-based sea surface temperature and air temperature variability from the inshore and offshore corals in the Seribu Islands, Indonesia. Mar. Pollut. Bull. 110, 694–700. (https://doi.org/10.1016/j.marpolbul.2016.04.052)
•  Charles,  C. D.,  Cobb,  K.,  Moore,  M. D. and  Fairbanks,  R. G. (2003) Monsoon-tropical ocean interaction in a network of coral records spanning the 20th century. Mar. Geol. 201, 207–222. (https://doi.org/10.1016/S0025-3227(03)00217-2)
•  Gagan,  M. K.,  Chivas,  A. R. and  Isdale,  P. J. (1994) High-resolution isotopic records from corals using ocean temperature and mass-spawning chronometers. Earth Planet. Sci. Lett. 121, 549–558. (https://doi.org/10.1016/0012-821X(94)90090-6)
•  Gagan,  M. K.,  Ayliffe,  L. K.,  Hopley,  D.,  Cali,  J. A.,  Mortimer,  G. E.,  Chappell,  J.,  McCulloch,  M. T. and  Head,  M. J. (1998) Temperature and surface-ocean water balance of the mid-Holocene tropical Western Pacific. Science 279, 1014–1018. (DOI: 10.1126/science.279.5353.1014)
•  Gordon,  A. L.,  Susanto,  R. D. and  Vranes,  K. (2003) Cool Indonesian throughflow as a consequence of restricted surface layer flow. Nature 425, 824–828. (https://doi.org/10.1038/nature02038)
•  Krawczyk,  H.,  Zinke,  J.,  Browne,  N.,  Struck,  U.,  McIlwain,  J.,  O’Leary,  M. and  Garbe-Schönberg,  D. (2020) Corals reveal ENSO-driven synchrony of climate impacts on both terrestrial and marine ecosystems in northern Borneo. Sci. Rep. 10, 3678. (https://doi.org/10.1038/s41598-020-60525-1)
•  Linsley,  B. K.,  Wu,  H. C.,  Rixen,  T.,  Charles,  C. D.,  Gordon,  A. L. and  Moore,  M. D. (2017) SPCZ zonal events and downstream influence on surface ocean conditions in the Indonesian Throughflow region. Geophys. Res. Lett. 44, 293–303. (https://doi.org/10.1002/2016GL070985)
•  Meehl,  G. A.,  Hu,  A. and  Santer,  B. D. (2009) The mid-1970s climate shift in the Pacific and the relative roles of forced versus inherent decadal variability. J. Climate 22, 780–792. (https://doi.org/10.1175/2008JCLI2552.1)
•  Minobe,  S. (1997) A 50–70-year climatic oscillation over the North Pacific and North America. Geophys. Res. Lett. 24, 683–686. (https://doi.org/10.1029/97GL00504)
•  Moore,  M. D.,  Schrag,  D. P. and  Kashgarian,  M. (1997) Coral radiocarbon constraints on the source of the Indonesian throughflow. J. Geophys. Res. Oceans 102, 12359–12365. (https://doi.org/10.1029/97JC00590)
•  Murty,  S. A.,  Goodkin,  N. F.,  Halide,  H.,  Natawidjaja,  D.,  Suwargadi,  B.,  Suprihanto,  I.,  Prayudi,  D.,  Switzer,  A. D. and  Gordon,  A. L. (2017) Climatic influences on southern Makassar Strait salinity over the past century. Geophys. Res. Lett. 44, 11967–11975. (https://doi.org/10.1002/2017GL075504)
•  Murty,  S. A.,  Goodkin,  N. F.,  Wiguna,  A. A. and  Gordon,  A. L. (2018) Variability in coral-reconstructed sea surface salinity between the northern and southern Lombok Strait linked to East Asian winter monsoon mean state reversals. Paleoceanogr. Paleoclimatol. 33, 1116–1133. (https://doi.org/10.1029/2018PA003387)
•  Okai,  T.,  Suzuki,  A.,  Kawahata,  H.,  Terashima,  S. and  Imai,  N. (2002) Preparation of a new geological survey of Japan geochemical reference material: Coral JCp-1. Geostand. Newsl.: J. Geostand. Geoanal. 26, 95–99. (https://doi.org/10.1111/j.1751-908X.2002.tb00627.x)
•  Qu,  T.,  Du,  Y.,  Strachan,  J.,  Meyers,  G. and  Slingo,  J. (2005) Sea surface temperature and its variability in the Indonesian region. Oceanography 18, 50–61.
• Siswanto and Suratno (2008) Seasonal pattern of wind induced upwelling over Java-Bali sea waters and surrounding area. Int. J. Remote Sens. and Earth Sci. 5, 46–56. (http://dx.doi.org/10.30536/j.ijreses.2008.v5.a1228)
•  Sprintall,  J.,  Potemra,  J. T.,  Hautala,  S. L.,  Bray,  N. A. and  Pandoe,  W. W. (2003) Temperature and salinity variability in the exit passages of the Indonesian Throughflow. Deep Sea Res. Part Ⅱ Top. Stud. Oceanogr. 50, 2183–2204. (https://doi.org/10.1016/S0967-0645(03)00052-3)
•  Sprintall,  J.,  Gordon,  A. L.,  Koch-Larrouy,  A.,  Lee,  T.,  Potemra,  J. T.,  Pujiana,  K. and  Wijffels,  S. E. (2014) The Indonesian seas and their role in the coupled ocean-climate system. Nat. Geosci. 7, 487–492. (https://doi.org/10.1038/ngeo2188)
•  Yasunaka,  S. and  Hanawa,  K. (2002) Regime shifts found in the Northern Hemisphere SST field. J. Meteorol. Soc. Japan 80, 119–135. (https://doi.org/10.2151/jmsj.80.119)
•  Yasunaka,  S. and  Hanawa,  K. (2005) Regime shift in the global sea-surface temperatures: its relation to El Niño–southern oscillation events and dominant variation modes. Int. J. Climatol. 25, 913–930. (https://doi.org/10.1002/joc.1172)