Heinrich (H) events are significant millennial climatic events that occurred during the last glacial period. Their footprints exited most parts of the global world, including the Asian monsoon (AM) region. However, the internal hydroclimatic characteristics of the AM system and their link to the external forcing during the H stadials are not fully understood. This study reconstructed monsoon precipitation changes during 33,000-12,500 a BP using grain size characteristics, elementary ratios, and pollen percentages from sediments of a crater lake (Tianchi Lake) in northeastern China. The H1 ~ H3 events are identified by high values of median grain size (MGS) and low values of Fe/Mn ratios and pollen percentages. By comparison with other records in the AM region, we found that H events in monsoon precipitation records were more obvious in the relatively higher-latitude region than in the lower-latitude sites. These inconsistent responses within the AM region during the H stadials might be attributed to both the high- and low-latitude climate processes.

During the last glacial period, a series of millennial-scale climate variabilities have been characterized by abrupt transitions between cold (stadial) and warm (interstadial) states, known as the Dansgaard-Oeschger (D/O) cycles [1, 2]. Several of these cold periods are related to extreme ice rafting episodes named the Heinrich events (H events) [3], which are generally attributed to large freshwater discharges from the Laurentide ice sheet into the North Atlantic Ocean occurring irregularly throughout the ice age [4]. However, the responses to H events in different regions of the world should be further studied.

The Asian monsoon (AM) region is a unique geographical area that encompasses a larger domain where the summer monsoon can extend as far north as 55°N [5], and such region could be influenced by a combination of climate forcing from the high-latitude North Hemisphere and the tropical regions [6, 7]. The climate changes coincident with H events have also occurred in the AM region from the records of lacustrine sediments [8, 9], loess [6, 10], and stalagmite [11, 12], etc. However, there are still controversies about the hydroclimatic characteristics in the AM region during the H stadials: some studies have shown that monsoon precipitation decreased significantly during the H stadials [9, 13], whereas other studies have concluded that monsoon precipitation increased obviously or was characterized with non-obvious oscillation during the H stadials [14, 15]. The discrepancy could be attributed to the lack of H events records with high-precise chronology and incomprehensive comparison of spatial patterns. Therefore, we investigated the hydroclimatic history at a relatively higher-latitude area in the AM region, where records covered the last glacial periods are scarce, to study the responses of monsoon precipitation to H events.

Here, we reconstructed a succession of hydrological records (33,000-12,500 a BP) from the sediment profile of a crater lake (Tianchi Lake) in northeastern China. In this study, three drought H events (i.e., H1, H2, and H3) have been identified from the multiproxy analysis, including median grain size (MGS), Fe/Mn ratios, and pollen percentages, at the time ranges of 16,200-15,000, 24,800-23,600, and 30,700-29,500 a BP, respectively. Furthermore, this work reports a comprehensive comparison of hydroclimatic records from lower-latitude to higher-latitude areas in the AM region. The detailed structure of these H event signals will provide a better understanding of the hydroclimatic spatial features in the AM system and the possible mechanisms during the H stadials.

Tianchi Lake (126°E, 48°44N, Figure 1(a)), a hydrological-closed crater lake with a diameter of ~400 m at the summit of Nangelaqiu Hill (Figure 1(b)), is located in the Wudalianchi Geopark, Heilongjiang Province in northeastern China. The Nangelaqiu Hill (149.9 m high and 596.9 m above mean sea level) is a typical conical volcano erupted at ~0.8 Myr and ~0.46 Myr ago [16, 17]. The area is characterized by a typical temperate monsoon climate, where the modern annual mean temperature is 0 ~ 0.5°C and the annual mean precipitation is ~470 mm (approximate 80% of rainfall occurred from June to September, Figure 1(c)). The modern regional vegetation in the catchment is mainly composed of temperate mixed conifer and broad-leaved forest. Tianchi Lake is primarily fed by summer rainfall without direct inflows and outflows and is subject to a relatively limited human impact. Therefore, this lake is an ideal site for preserving paleoclimate archives.

One sediment core, labeled as TC2 (~500 cm length), was collected from Tianchi Lake using a 55 mm diameter percussion corer in November 2012. The core was sectioned at 0.5 cm intervals in the lab resulting in a total of 984 subsamples. AMS 14C dating was carried out on 33 samples using plant residues, seeds, and organic matter (details for AMS 14C dating are available in Table 2). All the results were calibrated to calendar ages by the calibration curve IntCal20 [20] and the age-depth model was established using the Bayesian method with the Bacon. R program (version 2.2, [21]).

315 samples from the parts of core TC2 (267-495.5 cm) were used for grain size distribution analysis. The samples were treated with 10-20 mL H2O2 (30%) to remove organic matter and mixed with 10 mL of HCL (10%) to remove calcareous cement and shell materials. Then, all samples were washed and dispersed using 10 mL (NaPO3)6 (10%). Finally, the grain size distributions were measured using a Mastersizer 2000 laser diffraction particle size analyzer at the Soil and Environmental Change laboratory, Taishan University, and the relative error of the measurements was less than 2%.

The concentrations of elements (Fe, Mn, Al, and Ti) were measured with an X-ray Fluorescence (XRF) spectrometer (AxiosmAX-Minerals, PANalytical B.V.) using the fusion method. The air-dried sediments and volcanic rocks (i.e., LH and GQ: Laohei hill and Nangelaqiu hill) were grounded and passed through a 75 μm sieve. Subsequently, the samples mixed with lithium borate were fused into a glass disc and then analyzed using a spectrometer (Rh-anode as X-ray tube) with a relative error of 0.1 ~ 1.0%. Element analysis was performed in the Institute of Geology and Geophysics, Chinese Academy of Science.

4.1. Lithology and Chronology

As observed in the lithology field (Figure 2(a)), the color of different sediment layers in the core TC2 varies significantly. The upper 140 cm sediments are grey-yellow and sediments at depth 140-345 cm are black, all of which are rich in organic residues. The parts at depth 345-460 cm are dark grey clay, and the parts below 460 cm are clay in color with greyish red, all of which contain volcanic clastic materials.

The results of AMS 14C dating (Table 2) suggest that the ages are generally in stratigraphic order and increase steadily with depth (Figure 2(b)), indicating a continuous deposition and less influence of ‘reservoir effect’ in the lake sediments. The resultant age-depth model suggests that the average sedimentation rate of the profile is 14.99 cm·ka-1 during the past 33,000 years.

4.2. Paleoclimate and H Events Recorded in Tianchi Lake

For different areas and specific depositional environments, investigating the source of materials contributed to sediments is essential for reconstructing the paleoclimate history. Tianchi Lake is a typically closed crater lake without direct inflows and outflows around the catchment, indicating uncomplicated source materials deposited into the lake. The Al and Ti elements are considered to be relatively stable during chemical weathering and diagenesis and are influenced weakly by the biogeochemical processes compared to other ones [22, 23]. Most of the values of Al2O3 and TiO2 from the core TC2 are scattered between the volcanic debris (LH and GQ) and Xifeng loess (Figure 3(a)), which suggests that the mixture of terrestrial volcanic clastic surrounding the lake and eolian dust are the dominant source materials to the sediments of Tianchi Lake.

The variations of grain size in lake sediments can reflect the expansion and retreat of lake levels, which indicates the humidity condition around the area [29, 30]. Due to the small surface of Tianchi Lake, terrestrial debris can easily enter into the lake sediments in the center of the lake. Hence, coarse fractions originating from the volcanic clastic will easily deposit into the lake when the lake level retreats with less precipitation. On the contrary, the sample site is far away from the lakeshore when the lake level expands in the humid period, which makes it difficult to accumulate with coarse particles [31]. In general, the Fe/Mn ratios in lake sediments are controlled by both the bottom-water redox conditions and the elemental components from source materials [32]. The comparison of Tianchi Lake sediments and two types of potential sources shows that Fe/Mn ratios in the sediments from Tianchi Lake are not located between, but higher than those from the volcanic clastic and eolian dust (Figure 3(b)), indicating that bottom-water redox conditions in the lake could be the primary factor of the change in Fe/Mn ratios. Generally, higher Fe/Mn ratios suggest the reduced mode in the water column under the condition of increased rainfall with high lake levels, whereas lower ratios indicate the oxic mode following decreased precipitation with low lake levels [33, 34].

The MGS (median grain size) of the core TC2 (Figure 4(a)) ranges greatly between 7.64 and 15.87 μm, whereas the Fe/Mn ratios (Figure 4(b)) vary smoothly between 86.1 and 108.5 with a few largely fluctuations. In the sediments of Tianchi Lake, the MGS and Fe/Mn ratios showed a strong correlation with each other and exhibited similar trends. Furthermore, we can identify three drought events during the 16,200-15,000, 24,800-23,600, and 30,700-29,500 a BP, respectively, from the time-series analysis of records, which corresponded well with the H events recorded from North Atlantic ice-rafted debris (IRD) stack (Figure 4(d), [35]) and North Greenland ice core (Figure 4(e), [36]). In addition, the broadleaf tree pollen percentages in Tianchi Lake (Figure 4(c), [37]) supported these findings, which also indicated less moisture during the aforementioned periods (i.e., H1, H2, and H3). It is noteworthy that the Fe/Mn ratios and pollen percentages do not show obviously consistent changes with the grain-size record during the H events, which could be related to the differences in sensitivity of proxies and time resolution of the records. Additionally, the differences of elemental components from source materials, though not the main influencing factor, could also affect the Fe/Mn ratios in the sediments. However, in general, the reconstructed paleoclimate from multiple proxies in Tianchi Lake has recorded the signals of H events obviously, which is helpful for further detailed research about the H events characteristics in the AM region.

4.3. Spatial Characteristics of H Events in the AM Region and the Possible Forcing Mechanisms

The increased coarse particles of core TC2 in Tianchi Lake (Figure 5(e)) during H1 ~ H3 stadials are temporally consistent with the findings of other records from Qinghai Lake and Sihailongwan Lake (Figures 5(f) and 5(g), Table 1)), indicating clear signals of H events in the relatively higher-latitude area of AM region. Unlike the higher-latitude area, there are no significant signals of H events performed in the lower-latitude area from the records of the South China Sea, Tengchong Qinghai Lake, Bay of Bengal, and Huguangyan Maar Lake (Figures 5(h)–5(k), Table 1)). It is emphasized that the climate proxies in the AM region discussed above are indicative of humidity or hydroclimatic changes because various indicators may respond inconsistently to climate variations. In summary, it can be concluded that the higher-latitude area was obviously subjected to strong climate oscillations during the H stadials, in contrast to the lower-latitude area in the AM region.

The AMOC was suggested to play a key role in driving hydroclimate changes in the AM region during the H stadials as indicated by the integration of paleoclimate records and modeling results [10]. Global northward ocean heat transport would decrease when the AMOC slows down during the H stadials [39], and the temperature in the North Hemisphere would decrease with areas of sea-ice expansion in both the North Pacific and North Atlantic [10]. The increase of ice volume might affect Asian summer monsoon (ASM) precipitation via some high-latitude processes [4345]. On the other hand, the ASM circulation is dynamically linked to the latitudinal position of the ITCZ [38, 46, 47], which indicates that the southward movement of ITCZ following the AMOC slows down or even collapses would result in a weak monsoon circulation in the AM region [46, 48].

The records from the higher-latitude area of AM region matched well with the records of temperature changes from Greenland ice core (Figure 5(a)) and 231Pa/230Th ratios (Figure 5(b)), a strength index of the Atlantic meridional overturning circulation (AMOC), of the North Atlantic deep sediments during the H stadials. This consistency, together with the insignificant signals in lower-latitude records, likely suggests that the climate oscillations in the high-latitude North Hemisphere have a vigorous influence on the hydroclimatic variation in the AM region through high-latitude processes. However, the oxygen isotopic records of Hulu and Xiaobailong Caves (Figures 5(c) and 5(d)) significantly performed the signal of H events, indicating that high-latitude processes are not the only factors. The δ18O records in speleothems throughout Asia are often interpreted as an index for the ‘strength’ or ‘intensity’ of the AM circulation, a low-latitude process [11, 49], and one notable feature of these δ18O records is their coherent temporal variability across the entire AM region beyond the latitudes [11, 12, 50, 51]. Hence, the positive changes of stalagmite oxygen isotopes in Hulu and Xiaobailong Caves during the H stadials represented a weak ASM circulation induced by the southward shift of the ITCZ, which is also supported by the model simulations [52, 53].

However, the insignificant signals of H events in lower-latitude records were inconsistent with the obvious weakening of the ASM, likely indicating the forcing of other factors on lower-latitude precipitation during the H stadials. The sites in higher latitudes are located near the margin of the monsoon region, where rainfall changes are highly sensitive to the strengthening or weakening of the monsoon circulation [54, 55]. Therefore, the H events performed obviously in these records. Meanwhile, the sites in lower latitudes were generally located in the core monsoon region, where precipitation might not change largely during the H stadials.

Previous studies suggest that the spatial/temporal features of monsoon rainfall change signs within the AM region on different timescales [56, 57]. In this study, our records and comprehensive comparison with others revealed that both the high- and low-latitude processes might play important roles in influencing hydrology in the AM region during the H stadials. However, the detailed mechanisms between each climate system (like AMOC and AM system) still remain unclear [58]. Therefore, more precise chronology, high-resolution constrained records, and model simulations about the H events in the AM region are essential for further detailed exploration.

In the present study, we reconstructed a succession of monsoon precipitation records between 33,000 and 12,500 a BP from sediments of Tianchi Lake in the Wudalianchi of Northeast China. The results have identified three drought H events (i.e., H1, H2, and H3) from the time-series of median grain size, elementary ratios (Fe/Mn), and pollen percentages, which are consistent with the records of North Atlantic IRD and Greenland ice core. The asynchronous responses of the paleoclimate records to H events indicate that the higher-latitude area was subjected to stronger climate oscillation during H stadials compared to the lower-latitude area in the AM region, which could be attributed to both the high- and low-latitude climate processes. This work has the potential to provide a better understanding of hydroclimatic characteristics associated with zonal distributions in the AM region during the H stadials from a new perspective, but the physical mechanisms and detailed linkage between each climate system should be further investigated in the future.

The data used to support the findings of this study can be accessed by contacting the corresponding author.

The authors declared that they have no conflicts of interest.

This study was jointly supported by the National Natural Science Foundation of China (grants 41822707, 41888101, and 41721002), the Strategic Priority Research Program of the Chinese Academy of Sciences (grant XDB26000000), and the Youth Innovation Promotion Association CAS (grant 2018498).

Exclusive Licensee GeoScienceWorld. Distributed under a Creative Commons Attribution License (CC BY 4.0).