Holocene and Deglacial Sea Surface Temperatures in Sundaland

A research by Dhani Irwanto, 6 October 2025

Abstract

We present a regional synthesis of sea surface temperature (SST) evolution across Sundaland—the now-drowned continental shelf of Southeast Asia—using the Osman et al. (2021) LGMR global proxy–model assimilation. Two metrics were derived: a boundary-wide Sundaland mean and an inner-tropical mean (±6° latitude), both averaged at 100-year intervals from 22.5 ka BP to the present. The SST record shows a pronounced deglacial warming, with the coldest conditions centered at ≈ 19.7–19.0 ka BP rather than ≈ 21 ka BP, a locally expressed Younger Dryas-type slowdown between 14.1 and 12.1 ka BP, and a delayed Holocene thermal maximum centered at ≈ 5–3 ka BP. These phase offsets reflect tropical oceanic leads and lags relative to global benchmarks, shaped by monsoon feedbacks, shelf flooding, and smoothing inherent to LGMR data assimilation. The Sundaland series thus refines our understanding of Indo-Pacific thermal evolution and highlights the nuanced regional phasing of post-glacial climate recovery.

Keywords: Sundaland, Sea Surface Temperature, Holocene, Deglaciation, Osman 2021, LGMR, Tropical Climate, Younger Dryas

1. Introduction

During the Last Glacial Maximum (LGM), when global sea levels stood more than 120 m below their present level, the continental shelf connecting modern Indonesia, Malaysia, and surrounding seas formed a vast subcontinent known as Sundaland. Its low-latitude position at the heart of the Indo-Pacific Warm Pool (IPWP) made it a key region for both ocean–atmosphere interaction and early human dispersal. Reconstructing sea surface temperature (SST) variations across Sundaland is therefore crucial for understanding how post-glacial warming, monsoon variability, and sea-level rise transformed this once-emergent landscape.

Osman et al. (2021) introduced the Last Glacial Maximum Reanalysis (LGMR), a globally resolved temperature reconstruction that assimilates more than 700 paleoclimate proxy records—including marine sediments, ice cores, and terrestrial archives—into a climate-model framework. The data assimilation technique combines proxy constraints with model physics to produce spatio-temporally consistent fields of surface temperature and isotopic composition from 24 ka BP to the present. The LGMR achieves near-global coverage at approximately 2° spatial resolution and 120 time steps, validated against modern instrumental records and independent proxies. It therefore provides an unprecedented foundation for analyzing regional climate evolution within a globally coherent context.

Building on this dataset, the present analysis focuses on Sundaland’s SST trends within two complementary spatial masks: the full Sundaland boundary and a restricted inner-tropical belt (±6° latitude). This dual perspective allows evaluation of both regional mean conditions and tropical-core behavior, testing whether Sundaland’s thermal evolution followed global trajectories or exhibited unique Indo-Pacific dynamics.

2. Data and Methods

The analysis utilizes the LGMR (Osman et al., 2021) gridded sea-surface-temperature field (variable sst). The Sundaland boundary was delineated using a geographic shapefile representing the shelf area presently submerged under the Java, South China, and Sulu Seas. Two spatial subsets were defined: (1) all grid cells within the boundary and (2) those confined to ±6° latitude to represent the equatorial core. For each of the 120 chronological steps (spanning 24 ka BP → 0 ka BP), SST values were averaged using a simple arithmetic mean. Temporal aggregation at 100-year intervals reduced small-scale variability while preserving long-term structure.

Key climatic benchmarks were annotated according to established chronologies: the LGM (~21 ka BP), the Younger Dryas (12.9–11.7 ka BP), the Early Holocene warming (~11 ka BP), and the Mid-Holocene Thermal Maximum (8–6 ka BP). No area weighting was applied, as the objective was to maintain transparency and comparability with previous Sundaland-scale studies. Visualization employed a simple time-series overlay between the two means, emphasizing contrasts in amplitude and timing.

3. Results

The Sundaland-wide and inner-tropical SST series both display a strong deglacial warming trend from the Last Glacial Maximum through the early Holocene. The lowest mean SSTs occur at 19.7–19.0 ka BP, about 2 kyr later than the canonical global LGM, indicating a slightly delayed tropical temperature minimum. A marked warming followed after 18 ka BP, punctuated by a subdued but distinct slowdown between 14.1 and 12.1 ka BP—interpreted as a regional expression of the Younger Dryas event. SSTs then stabilized at elevated levels through the Holocene, reaching a thermal maximum at ≈ 5–3 ka BP, later than most Indo-Pacific records. Thereafter, a gradual decline persisted toward modern values, consistent with orbital forcing and monsoon realignment during the late Holocene.

This overall trajectory, encompassing early deglacial warming and a prolonged Holocene optimum, mirrors the large-scale evolution of tropical ocean systems. The inner-tropical (±6°) mean remains consistently warmer than the whole-region mean throughout the sequence, differing by roughly 0.4–0.6 °C on average. This offset reflects the latitudinal SST gradient within the Sundaland domain and confirms the relative thermal stability of the equatorial core. Both curves reproduce the timing of key deglacial transitions documented in coral proxy records (Gagan et al., 2004) and global temperature stacks (Shakun et al., 2012; Marcott et al., 2013).

Figure 1. Mean SST time series for Sundaland (whole boundary) and the inner tropics (±6°), with key climatic intervals highlighted

4. Discussion

The Sundaland SST evolution broadly parallels the global deglacial pattern yet reveals distinctive tropical phasing and amplitude. The coldest interval occurs around ≈ 19.4 ka BP—about two millennia later than the canonical global LGM—suggesting that tropical oceans reached their temperature minima slightly after maximum ice volume, possibly due to delayed deep-ocean mixing and greenhouse gas rise. The ensuing warming accelerated after 18 ka BP, interrupted by a modest slowdown between 14.1 and 12.1 ka BP that corresponds to a regionally expressed Younger Dryas-type event. Although muted compared with high-latitude signals, this episode marks the tropical imprint of global circulation perturbations transmitted through the Indo-Pacific Warm Pool (IPWP).

The mid- to late-Holocene evolution likewise departs subtly from global reconstructions. The thermal maximum appears at ≈ 5–3 ka BP rather than the canonical 8–6 ka BP, likely reflecting continued shelf flooding, monsoon realignment, and prolonged heat retention across the newly inundated Sunda shelf. Comparison with Osman et al. (2021) global composites indicates that Sundaland warmed broadly in phase with other tropical basins but maintained slightly higher absolute SSTs throughout the Holocene, consistent with its shallow-shelf setting and strong ocean–land coupling. The agreement with coral records from the western Pacific (Gagan et al., 2004) further demonstrates that the LGMR framework captures Indo-Pacific thermal evolution with realistic regional detail, reaffirming Sundaland’s role as a dynamically sensitive yet climatologically buffered component of the IPWP.

1. LGM phase (~19.4 ka BP). The SST minimum appearing at ≈ 19.4 ka BP, slightly younger than the canonical ≈ 21 ka BP, is consistent with other tropical reconstructions showing that Indo-Pacific surface waters began to warm earlier than the global ice-volume maximum. This phase lead likely reflects tropical sensitivity to rising greenhouse gases and orbital precession, initiating equatorial convection before full glacial retreat.
2. Regional expression of the Younger Dryas. The subdued warming between 14.1 and 12.1 ka BP represents a local manifestation of the Younger Dryas, shifted earlier by about one to two millennia. Such displacement may stem from regional feedbacks in the Indo-Pacific Warm Pool (IPWP), where ocean–atmosphere coupling and early resumption of overturning circulation produced a tropical lead relative to Northern Hemisphere cooling. Comparable leads have been reported in tropical SST syntheses (e.g., Tierney et al., 2020).
3. Mid-Holocene peak timing (5–3 ka BP). The delayed maximum SST relative to the global mid-Holocene (8–6 ka BP) can be attributed to continued shelf inundation and regional monsoon asymmetry. As postglacial flooding transformed Sundaland into a mosaic of seas and islands, enhanced heat retention and sustained humidity may have extended warm conditions well into the middle Holocene. Additionally, the LGMR assimilation’s temporal smoothing likely distributed the Holocene thermal maximum over a broader interval, shifting the apparent peak toward later centuries.
4. Chronometric and methodological factors. The Osman et al. (2021) LGMR dataset integrates multiple proxy types with varying age control, producing an estimated uncertainty of ±0.5–1 kyr for regional means. Its Kalman-filter approach dampens abrupt transitions but preserves long-term coherence; when averaged across Sundaland’s broad spatial domain, this smoothing can produce 1–2 kyr apparent offsets in peak or trough timing.

In summary, the phase shifts observed in the Sundaland SST curves do not contradict global reconstructions but rather highlight the spatial heterogeneity and lag–lead behavior of tropical oceans during deglaciation. The slow-warming interval at 14.1–12.1 ka BP likely represents a regional Younger Dryas signature modulated by tropical feedbacks, while the delayed thermal maximum at 5–3 ka BP reflects prolonged warmth associated with monsoon dynamics, shelf inundation, and model assimilation smoothing.

5. Conclusion

Analysis of the Osman et al. (2021) LGMR dataset reveals parallel SST histories for Sundaland’s full extent and its inner-tropical core. Both exhibit canonical deglacial transitions, but with regionally distinct phasing: the LGM minimum near ≈ 19.4 ka BP, an early Younger Dryas-like cooling at 14.1–12.1 ka BP, and a delayed Holocene peak at 5–3 ka BP. These offsets underscore Sundaland’s tropical sensitivity and the asynchronous yet coherent behavior of the Indo-Pacific Warm Pool relative to global climate evolution. They also emphasize how shelf flooding, monsoon feedbacks, and assimilation smoothing influence the apparent timing of climatic events. Together, these findings position Sundaland as a key indicator of tropical ocean variability and as a benchmark region for integrating paleoclimate, sea-level, and archaeological evidence of the late Quaternary transformation of Southeast Asia.

References

1. Gagan, M.K., Hendy, E.J., Haberle, S.G., & Hantoro, W.S. (2004). Post-glacial evolution of the Indo-Pacific Warm Pool and ENSO. Quaternary Science Reviews, 23(7–8), 1227–1243.
2. Kaufman, D.S., et al. (2020). A global database of Holocene paleotemperature records. Scientific Data, 7(115).
3. Marcott, S.A., Shakun, J.D., Clark, P.U., & Mix, A.C. (2013). A reconstruction of regional and global temperature for the past 11,300 years. Science, 339(6124), 1198–1201.
4. Osman, M.B., Tierney, J.E., Zhu, J., et al. (2021). Globally resolved surface temperatures since the Last Glacial Maximum. Nature, 599, 239–244.
5. Shakun, J.D., Clark, P.U., He, F., et al. (2012). Global warming preceded by increasing CO₂ during the last deglaciation. Nature, 484, 49–54.
6. Tierney, J.E., Zhu, J., King, J., et al. (2020). Glacial cooling and climate sensitivity revisited. Nature, 584, 569–573.