Introduction

The Asian Summer Monsoon (ASM) originates in the tropical ocean and plays a key role in global heat and water vapor transport, with significant impacts on the wellbeing of ~ 60% of the world’s population1. Understanding the evolution of the ASM and its forcing mechanism on orbital timescales has long been of major scientific interest. Multiple proxy indicators derived from diverse geological archives have been developed, some of which are regarded as reliable indicators of past changes of the ASM2,3,4,5,6. However, the interpretation and periodicities of ASM changes based on these proxies, particularly in the case of the loess deposits and speleothems, have been found to be inconsistent, resulting in the so-called “Chinese 100-kyr problem”7. Well-dated speleothem δ18O records from south-central China are dominated by precessional cycles (~ 23 kyr) over the past 640 ka6, indicating that the ASM intensity was driven by local summer insolation4,5. In contrast, most ASM proxy records from loess deposits in northern China, represented by the magnetic susceptibility (MS), show a dominant 100-kyr cyclicity which coincides with the glacial-interglacial cycles of the benthic foraminiferal δ18O record8,9. However, in several recent proxy records from the Chinese Loess Plateau (CLP), the ASM variations are not forced solely by the 100-kyr eccentricity cycle; for example, the Sr/Ca ratio of microcodium also shows the 41-kyr obliquity cycle10; 10Be flux shows composite 23-kyr and 100-kyr cycles5; loess carbonate δ13C shows composite 23-kyr, 41-kyr and 100-kyr cycles11; and MS flux shows a dominant precessional cyclicity12. These findings complicate the interpretation of the evolution and forcings of the ASM on orbital timescales.

We suspect that these discrepancies stem from the uncertainty and complexity inherent in the climatic interpretation of the currently-used proxies. Indeed, the precise interpretation of speleothem δ18O records remains debated, with various explanations proposed, including precipitation amount, moisture composition and source, and pathway effects13. Similarly, loess proxy records are influenced by multiple factors such as precipitation and temperature variations11, pedogenic smoothing14, variations in sedimentation rate15, and mineral phase transformations16. The operation of these additional factors would significantly impact the sensitivity of loess monsoon proxies in responding to changes in climate compositions, resulting in the dilution of periodicities signals within loess records.

The origin of most existing monsoon proxies is based on specific physicochemical processes that are used to extract information about individual climatic factors such as temperature, and the amount or isotopic composition of precipitation. However, these physiochemical processes are often influenced by other environmental variables, greatly increasing the difficulty of interpreting these monsoon proxy records. By contrast, biological proxies can respond to climate change more directly and sensitively; furthermore, the use of biological proxies eliminates the need to isolate individual climatic variables in complex natural environments, thus enhancing their ability to capture changes in overall monsoon intensity. This is because organisms, whose eco-physiological behaviors are often controlled by various environmental factors17, typically respond to changes in the overall environment rather than to individual climatic factors. In addition, on geological timescales, the response of organisms to paleoclimatic fluctuations is grounded in evolutionary theory18,19. Thus, organisms operate autonomously based on physical and chemical laws (e.g., biological laws) that determine the evolutionary trajectory of species under external pressures18,19. Therefore, a biological proxy synchronously records shifts in paleoclimatic conditions in a way that is mechanistically independent of most abiotic proxies. This facilitates the process of corroboration and cross-verification among diverse proxies, thus facilitating the dependable interpretation of past monsoon variations.

In this study, we selected land snails in Chinese loess deposits as a bio-archive and utilized the variation of their shell morphology as a potential proxy of changes in monsoon patterns, based on the following observations: (1) As an “indicator animal” in loess-based paleoenvironmental research, land snails are highly sensitive to environmental changes, and their fossil remains are widely and well-preserved in Quaternary loess-paleosol sequences20. (2) The shell is deposited in situ after the snail’s death and is highly resistant to pedogenesis and post-depositional processes9, thus retaining pristine climatic information. (3) Shell shape has the potential to serve as a novel climate proxy: as a fundamental biological trait, shell shape regulates various essential biological functions, such as protection against predation and drought21. Previous studies have shown that snail shell shape plays a crucial role in the resilience of land snails to desiccation stress22,23,24,25. Specifically, the aperture shape of the shell may be an adaptive response to desiccation stress as land snails primarily lose body water through this opening23,25. Thus, there is the possibility of using variations in land snail shell morphology as a proxy for paleoclimate reconstruction, particularly in semiarid regions like the CLP, which is characterized by distinct dry/wet fluctuations during glacial-interglacial cycles.

Based on the foregoing we analyzed the changes in the shell morphology of the land snail Cathaica orithyia (Martens, 1879), in the Xifeng section, on the CLP in northern China, over the last 470 kyr. Our objectives were to explore the orbital-scale fluctuations in shell morphology and to assess its potential as a new ASM proxy, and hence to expect to obtain a new understanding of the ASM evolution.

Materials and methods

Sampling and chronology

A total of 472 samples for snail fossil analysis were collected from the upper 43.2 m of the Xifeng Sect. (35°45’N, 107°49’E), where over 70% of the mean annual precipitation (MAP: 528 mm) is contributed by the ASM (Fig. 1). Soil blocks with a cross section of 15 cm × 40 cm and spaced at intervals of 10 cm (~ 1 kyr) were excavated from which snail fossils were extracted by washing. Each sample contained an average of 2 fossils, with a range between 1 and more than 10. Considering the impact of the genetic background on shell morphology, we examined the shell shape variation within a single species to eliminate interspecies differences, with a total of 95 fully mature fossil C. orithyia shells being selected for analysis. The chronology of the section was established through cross-correlation of its magnetic susceptibility record with the magnetic susceptibility time series presented by Sun et al.26. Specifically, using AnalySeries software (version 2.0.5.2; https://github.com/PaleoIPSL/Analyseries)27, the magnetic susceptibility record of Sun et al.26. was used as the target curve for our magnetic susceptibility record. The tie points selected for the initial correlation of these two magnetic-susceptibility curves are often the midpoint between the start and end of an interval of rapid change in the respective magnetic susceptibility proxy records (diamond labels in Fig. S1). The fundamental requirement for selecting tie points was to ensure that the sedimentation rates did not change excessively. More detailed information about the section, sampling and the confirmation of the reliability of chronology can be found in Bao et al.9. , as both studies utilized snail fossil samples obtained during the same field sampling expedition.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

(A) Map showing the location of the study area in Asia (red dot) and other paleoclimate records mentioned in the text (black dots). The northern boundary of the modern ASM (constituted by EASM and ISM) is represented by the green dashed line, while the grey arrows depict the present-day summer mean 700 hPa streamline (modified from Chen et al.28). The gray shading denotes areas above 3000 m a.s.l. and the yellow shading represents the Chinese Loess Plateau. EASM: East Asian summer monsoon; ISM: Indian summer monsoon. (B) Seasonal variation of modern climatic para

Geometric morphometric analysis

Geometric morphometric analysis (GMA) was employed to examine the variations in shell shape of C. orithyia in apertural view (Fig. 2, Fig. S2). Morphometric studies commonly utilize the arrangements of morphological landmarks (LMs) and semi-landmarks (semi-LMs) as the primary data source, from which shape information is extracted through Procrustes superimposition29. Bookstein30defines landmarks as loci with both names (such as ‘bridge of the nose’ or ‘tip of the chin’) and Cartesian coordinates. These names are intended to imply biological homology among forms, which must be expressed through a geometric homology—that is, the correspondence between points, curves, or surfaces31. However, the use of LMs faces the difficulty of accurately representing curves and surfaces due to the inability to homologize landmark positions along the curve or surface across different individuals. Therefore, semi-LMs are necessary in morphometric studies as they are used to establish the geometric homology between corresponding semi-LMs across the sample, enabling the representation of homologous curves and surfaces through sets of points. The configuration of LMs and semi-LMs of the fossil shells of C. orithyia is presented in Fig. 2 and Table S1. In this study, 9 LMs were utilized to capture the overall shape of fossil C. orithyia, while 23 semi-LMs were utilized to accurately depict the contour shape of its body whorl.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Apertural view of C. orithyia, shell structure terms, and the configuration of landmarks (LMs, solid circles) and semi-landmarks (open circles).

To analyse the shape variations in fossil C. orithyia, we first calculated the average shape of the shell using the Cartesian coordinates of LMs and semi-LMs defined in Fig. 2. The shell shape of each specimen was then aligned to this average shape using a Generalized Procrustes analysis32,33, which normalized the Cartesian shape coordinates into Procrustes shape coordinates. Subsequently, the α parameter in TpsRelw software34was set to zero to analyse relative warps (RWs), which involves a principal component analysis (PCA) of the Procrustes shape coordinates30,35,36. The RWs are quantitative indices of shell shape variations that are independent of the spatial scale; thus, they enable the quantitative comparison of fossil C. orithyia shell shapes and paleoclimatic factors.

Although GMA is a valuable tool for shape quantification it has several unavoidable drawbacks. The physical meaning of the GMA result (RWs) is highly abstract, which originates from the characteristics of the PCA. To explore the climatic significance of shell shape variations, we selected several areas that displayed notable morphological changes as the subjects of our simplified investigation. This approach facilitates the provision of cogent biological interpretations for different RWs. To further clarify the relationships between specific shell shapes and RWs, several relative area parameters (RAs) have been defined to examine their correlations with RWs: RAap/sh represents the relative area of the aperture in relation to the whole shell projection; RAtri/ap denotes the relative area of the triangular region surrounding LM 4–6 and semi-LM 28–32 in relation to the aperture (as shown in Fig. 2); and RAcs/sh indicates the relative area of the cross-section of a shell whorl compared to its whole shell projection, where the cross-section comprises both the triangular region and the aperture. Additionally, the spire index is the relative length ratio of shell height to width (Fig. 2).

Results and discussion

Shell morphology variations of the fossil shells of C. orithyia

The first four RWs (RW1 to RW4) of the GMA results contain the majority of the shell shape variation information for C. orithyia, accounting for 39.88% (RW1), 20.09% (RW2), 9.61% (RW3), and 9.22% (RW4), respectively, and each exhibits distinct shape characteristics (Fig. 3). Overall, the shell shape variations of fossil C. orithyia reflected by all RWs correspond to a normal distribution (Fig. 3a-d). The shell shape comprises two aspects: (1) the spire index, and (2) the upper lip of the aperture (UPA; Fig. 3i). Specifically, the position of LM 1 in RW1, RW2, and RW4 shows obvious differences, which indicate the variations of the spire index, while LM 5–6 and semi-LM 28–30 in RW1 and RW2 indicate significant shape variations of the upper lip of the aperture.

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

Results of GMA analysis of the fossil shells of C. orithyia from the Xifeng section. (a-d) GMA results of the first four shell shape indices (RW1–4), including the overall distribution for each RW. (e-h) The distributions of each RW in different loess-paleosol units. (i-m) The extreme shapes of each RW. The upper lip of the aperture refers to a partial outline of the aperture, encompassing LM 5, LM 6, and semi-LMs 28–30.

To further explore the biological significance of the shell shape variations, we analyzed the correlation between the RWs and common land snail shell shape indices. Both RW1 and RW2 are strongly correlated with the relative aperture area (Fig. 4a) but only weakly correlated with the spire index (Fig. 4d). This indicates that the major shell shape changes in RW1 and RW2 pertain to the aperture. Although both RW3 and RW4 are related to the spire index, their correlation with the spire index differs significantly, with a weak correlation observed for RW3 and a strong correlation for RW4 (Fig. 4d).

Although both RW1 and RW2 reflect the shape variation of the aperture, in practice they reflect completely different changes in the shape of the upper lip of the aperture. Visually, the shape change of the upper lip is longitudinal in RW1 (Fig. 3i) and lateral in RW2 (Fig. 3j). This observation is supported by the relationships between the RWs and RAs. For RW1, we found that it has a robust correlation with RAap/sh (Fig. 4a), which represents the relative size of the effective aperture area related to the whole shell projection and indicates the direct channel of moisture exchange between the shell interior and exterior. Additionally, the strong correlation between RW1 and RAtri/ap (Fig. 4b) further confirms that RW1 represents the longitudinal variation of the upper lip; this is because only the longitudinal change of the upper lip of the shell can cause significant changes in RAtri/ap. By contrast, the lateral change of the shell upper lip often causes the area of both the triangular region and the aperture to increase simultaneously, which means that RAtri/ap would not change with the lateral variation of the shell upper lip, and thus it has no significant effect on RAtri/ap. Hence, the observed robust correlations of RW2 with RAap/sh and RAcs/sh (Fig. 4a and c) suggest that RW2 primarily reflects the lateral variation of the upper lip.

Briefly, both RW1 and RW2 reflect the variation of the shape of the shell upper lip, with the former reflecting longitudinal changes, and the latter latitudinal changes. Compared to RW3, which has an ambiguous relationship with the shell morphological variation, including the spire index, RW4 is more strongly related to the variation of the spire index.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Relationships between RWs, RAs and the spire index. Only statistically significant results (P < 0.05) are shown. (a) RAap/sh: relative area of the aperture (red) to the entire shell projection (green). (b) RAtri/ap: relative area of the triangular region (red) to the aperture (green). (c) RAcs/sh: relative area of the whorl cross-section (red) to the shell projection (green). (d) Spire index: relative length of shell height (red) to shell width (green).

Environmental significance of variations in land snail aperture shape

Due to the scarcity of modern relatives of C. orithyia, it is difficult to establish a modern reference in the study area to explore the environmental significance of shell morphology. Fortunately, the correlation between shell morphology (although from other species) and climatic/environmental parameters has been well explored elsewhere, on temporal and spatial scales and with various research protocol (field observation, common garden and controlled laboratory experiments)37,38,39,40,41. From the perspective of variation in shell aperture shape represented by RW1 and RW2, these studies generally consider that the aperture area of land snail shells is generally greater in moister environments41,42. For example, a study from Poland found that larger shell apertures of snails are developed in higher humidity41; however, a small aperture, usually accompanied by a reduction in the area of exposed surface, could be advantages in drier conditions with higher water stress21,39. Specifically, they found that temporal variation in relative humidity showed positive relationships with aperture width and aperture height. Another study demonstrated that shell volume and aperture area tend to be larger in areas with higher annual rainfall level, altitude, and mean average temperature by using snails (Helix pomatia var. Banatica) from two distinct sites characterized by different climatic conditions42. Also, for snail fossils, several studies found that the shell size was related to climate condition of interglacial-glacial phases, with smaller sizes in glacial intervals43,44,45.

Seen from the above, most studies emphasize the key role of environmental aridity/humidity in regulating the shell morphology variations, especially the shell aperture shape. Evaporation through the aperture is the main mechanism by which land snails lose body water23,25. As an animal that prefers warm and humid environments, snails become inactive and cease growth at temperatures below 10 °C or when the environment becomes excessively dry46. The modern climate in Xifeng (Fig. 1B), influenced by the East Asian monsoon system, generally shows a cold-dry winter and a warm-humid summer, with significant intra-annual variations in temperature (−4.2 to 21.4 °C, mean: 9.2 °C), precipitation (3.7 to 111.4 mm, mean: 44.0 mm) and relative humidity (52 to 76%, mean: 62%). Also, field observation found that C. fasciola in the CLP typically enters hibernation from October-November to early March-April and resumes activity once temperatures rise above ~ 10 °C from late spring to early autumn47. Thus, it is reasonable to assume that the life style of our snails is divided into two very different stages—active and resting—according to different water loss contexts. In the active stage, water directly evaporates through the open aperture; and in the resting stage, evaporation from the animal is effectively reduced as a result of its tight attachment to the resting substrate and the epiphragm that seals the aperture. By their close attachment to the substrate and the epiphragm, the ability of snails to resist drought is enhanced by resting and a decrease in the relative virtual area of the aperture (RAap/sh, Fig. 4).

In RW1, the longitudinal shift of the UPA (Fig. 2i) indicates the downward deflection of the aperture (DDA; Fig. S3a), which is a common phenomenon observed during the final stage of shell growth in many land snails21. The DDA reduces the relative aperture area, which enhances water retention. In other words, DDA can regulate the drought resistance ability of C. orithyia by altering the relative aperture area. As noted above, for land snails in the resting stage, the relative aperture area has only a limited ability to resist arid conditions, while in the active stage, land snails tend to reduce evaporation by developing a narrower aperture. Overall, we infer that the RW1 of C. orithyia reflects its ability to resist drought during the active period, with a smaller relative aperture area corresponding to greater drought resistance, while conversely a larger relative aperture area corresponds to weaker drought resistance.

In RW2, the lateral shift of the UPA (Fig. 2j) corresponds to the forward extent of the UPA (Fig. S3b), which can induce a leftward shift in LM 5, thereby augmenting both the whorl cross-sectional and aperture areas. However, the changes in the parietal wall (Fig. 1b) result in very minor changes in the effective aperture area, and thus the virtual aperture area responsible for evaporation/water loss is not increased by the forward extension of the upper lip, so that RW2 barely influences the evaporation rate. Overall, this indicates that the variations in the relative aperture area represented by RW2 are unlikely to significantly impact the drought resistance of C. orithyia. However, due to the forward extension, the UPA enables a closer contact with the substrate surface, which increases the ability of the resting animal to reduce water loss and hence survive the resting stage. From this perspective, RW2 is a possible index of the snail’s ability to resist drought during the dormant or resting stage.

Although both RW1 and RW2 represent morphological changes in the shell aperture, they represent two distinct aspects of a snail’s adaptative survival strategy. Specifically, RW1 corresponds to the snail’s ability to resist desiccation during active periods, while RW2 reflects its capacity for water retention during dormancy.

Periodicities of the aperture shape of fossil C. orithyia (represented by RW1 and RW2).

To identify periodic components in the time series of major aperture shape, we grouped the samples into 10-kyr intervals and then performed spectral analysis on the RWs, using the REDFIT program in Past3 software48. As shown in Figs. 5 and 6, the dominant Earth orbital cycle in the RW1 record is ~ 23 kyr, which is also the case for speleothem δ18O records from China6. Wavelet analysis also show some degree of precession signal for RW1 between 150 and 250 ka (Fig. S4). More importantly, cross-spectral analysis shows that RW1 were coherent with speleothem δ18O within the precessional band, although the correlation between the two does not pass the significance test due to some phase offsets (Fig. S5, Fig. S6a, b). The RW1 score generally increases with increasing speleothem δ18O, indicating that a reduction in shell aperture size corresponds to a weakening of ASM intensity. As snails with smaller apertures have a greater resistance to drought, the increasing proportion of C. orithyia with a small aperture in the population suggests that the snail’s living habitat is experiencing drought stress. Therefore, our results suggest a tele-connection between the speleothem δ18O record in southern China and monsoon proxy records from northern China, demonstrating that a decrease in ASM intensity was accompanied by aridification in northern China. As noted in the introduction, the ~ 20-kyr periodicity in speleothem δ18O has also been identified in loess records from northern China, such as in the δ13C record of inorganic carbonate11and a recently reported loess record that spliced by MS for interglacials and the inverse sand content for glacials49. From another perspective, these results from northern China, together with our RW1 record, suggest that the speleothem δ18O record is more indicative of the overall ASM intensity, or the large-scale monsoon circulation, rather than the local precipitation amount.

In contrast to RW1, the RW2 record demonstrates intervals with obliquity-like cycles in addition to precession cycles (Fig. 6b, Fig. S4). On the orbital scale, the RW2 record closely resembles the record of summer monsoonal precipitation reconstructed from the Sr/Ca ratio of loess microcodium10 (Fig. 5d, Fig. S6c), and the two show moderate coherence within the obliquity band (Fig. S5). An obliquity signal of monsoonal precipitation has also been found in a loess carbonate δ13C record11. These observations suggest that, in addition to precession, obliquity forcing may be an additional critical factor in regulating aridity/humidity cycles over northern China, and thus in driving ASM variations. Additionally, as RW2 reflects the snail’s capacity for water retention during its dormant period, our RW2 record may indicate that certain monsoon proxies, particularly trace elements in microcodium, are biased towards precipitation variations during arid periods in northern China, thus reflecting drought severity. Consequently, arid conditions exert selective pressure on the water-retention capacity of snail shells during dormant periods, resulting in the pronounced obliquity signal in RW2.

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

RW records, published ASM records, and Earth orbital parameters over the last ~ 470 kyr. The gray shading denotes the marine isotope stages that correspond to loess and paleosol layers. (a, b) Precipitation reconstructions based on the δ13C of loess inorganic carbonate11 (δ13CIC) and loess 10Be flux5, respectively. (c) RW1 and the composite Chinese speleothem δ18O record6. (d) RW2 and paleoprecipitation reconstructed from the Sr/Ca ratio of loess microcodium10. (e) RW3 and land surface paleotemperatures reconstructed from brGDGTs (bacterial biomarkers) in loess50. (f) Magnetic susceptibility (MS) record of the Xifeng Sect26; paleosol units are labeled as Si. (g) RW4 and Earth orbital eccentricity51. (h) LR04 stacked benthic δ18O record52; numbers indicate marine isotope stages.

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

Spectral analysis results for RWs (red lines) and related palaeoclimate records (grey lines). (a) RW1 and the composite Chinese speleothem δ18O record6. (b) RW2 and paleoprecipitation reconstructed from the Sr/Ca ratio of loess microcodium10. (c) RW3 and land surface paleotemperature reconstructed from brGDGTs in loess50. (d) RW4 and Earth orbital eccentricity. In all panels, the yellow lines indicate the 90% confidence level and the labeled dashed lines denote primary orbital frequencies (frequency = 1/period in kyr).

Periodicities of the shell spire index (represented by RW3 and RW4)

The shell spire index is represented by RW4 (Figs. 2m and 3d). This record seems to shows a long periodicity that is similar to the long cycle (405 kyr) of Earth orbital eccentricity51 (Figs. 5g and 6d). We acknowledge that it is somewhat arbitrary to say RW4 reflects long eccentricity cycles due to the length limitations of our records. However, a long cycle of ~ 300 kyr to ~ 400 kyr was also found in the shell δ18O record of C. orithyiafrom the same Sect53, leading us to speculate about the connection between long eccentricity and shell morphology. A 405-kyr eccentricity cycle, which is often found in the oceanic carbon reservoir, can impact low-latitude monsoon rainfall via insolation variations with subsequent Earth’s interior amplification effects54. The monsoon climate will inevitably impact the survival strategy of snails and thus a monsoonal signal can be recorded by the shell structure, which may subsequently involve changes in spire shape represented by RW4, since the changes in spire index are a manifestation of the snail’s adaptability to environmental changes55,56.

The shape characteristics of RW3 are ambiguous (Fig. 3k), making it difficult to link it with a specific shell morphology and hence to elucidate its paleoclimatic significance. Although the correlation between RW3 and brGDGTs-based land surface temperature (LST) is not statistically significant (Fig. S6d), the temporal variation of RW3 is visually similar to that of the brGDGTs-based LST record (Fig. 5e), with both records showing early warming during the last four glacial terminations50 (Fig. 5e). In contrast, the MS record from the same Sect26 (Fig. 5f) shows a different trend, although the temporal variations among physicochemical loess proxies are consistent at the last glacial terminations (Fig. 5a, b, f and h). This implies that RW3 may be affected by temperature variations between glacial-interglacial cycles to some extent.

Linkage between shell morphology and orbital-scale ASM changes

The morphological changes of land snail shells represent their ecological adaptive behavior in response to environmental pressures21,22. Thus, snail morphology can be expected to respond to changes in the snails’ living environment and climatic conditions; more specifically, the changes in the shell morphology of snails in the Asian monsoon regions reflect variations in the environmental aridity/humidity, and thus may capture the information about the ASM variation. The shell morphological changes in Xifeng section can be decomposed using parameters RW1–RW4, with the records of these parameters representing different aspects of shell morphology. For example, RW1 and RW2 reflect changes in shell aperture, while RW3 and RW4 likely reflect changes in the shell spire index. We found that different RW records of shell morphology from the fossil snail shells in the Xifeng section are dominated by different orbital periodicities (i.e. eccentricity, precession, obliquity), indicating that each RW record represents changes in a specific aspect of the ASM system. For example, some RW records may capture information about changes in the large-scale ASM circulation, while others may document changes in environmental humidity or aridity, or even provide insights into temperature variations. However, considering all the RW records together, enables us to capture the overall changes in snail shell morphology, which may record the overall changes in the ASM, implying that ASM variations are simultaneously forced by multiple orbital periodicities. These multiple orbital periodicity signals have been confirmed in recent years by other records from loess deposits, such as the records of the δ13C of soil carbonate and δ18O of microcodium11,57. Mechanismly, the multiple orbital period of ASM can be caused by both astronomical (i.e. summer insolation) and glacial forcing (e.g., changing surface boundary conditions such as ice volume, CO2concentration), which have also been proved by model results58. Further evidence comes from the cross-spectral results, which indicate that the RW1 record is coherence with the summer insolation on the precession band (Fig. S5b), while RW3 and benthic δ18O are plausibly coherent on the 100-kyr band (Fig. S5d).

The shell morphology record obtained in the present study is the first to confirm this ASM behavior from the perspective of biological proxies. A possible explanation for these multiple cycles captured in our shell morphology record is that biological proxies may respond more sensitively to changes in the overall climatic and environmental conditions, recording the overall monsoon climate system composed of various climatic factors including temperature, humidity and precipitation. In contrast, previous proxies based on physicochemical properties may record specific aspects of the EAM system, resulting in inconsistencies in their periodic components7.