The spatiotemporal pattern of rainfall variability in East China exhibits a tripole or dipole structure, as indicated by empirical orthogonal function (EOF) analysis of instrumental climate data. The tripole structure shows a pattern of rainfall decrease/increase/decrease in South-Central-North China, while the dipole structure shows a +/− pattern from south to north of the Yangtze River. These structures are linked to coupled ocean-atmosphere modes in the Pacific (ENSO, PDO) and North Atlantic (NAO, AMO), and their interaction with the East Asian monsoon. However, these patterns are based on recent decades' meteorological data, significantly influenced by human activities. The question of whether these patterns existed before industrialization remains. While climate model simulations and paleoclimate reconstructions suggest similar rainfall patterns during the Little Ice Age (LIA), Medieval Climate Anomaly (MCA), Holocene, and last deglaciation, the lack of continuous high-resolution records hinders understanding of ocean-atmosphere coupling's effect on East China's paleo-precipitation spatial patterns. Chinese stalagmite oxygen isotope (δ¹⁸O) records offer the potential to explore spatiotemporal climate variability over longer timescales. However, interpreting Chinese speleothem δ¹⁸O is complex due to influences from multiple factors, including moisture source region, precipitation amount, and evaporation. Separating these processes from spatially separated cave records can enhance understanding of ocean-atmosphere interaction over eastern China. Principal component analysis (PCA) is employed to separate the monsoon signal from spatial precipitation patterns. This study uses a continuous, annually resolved δ¹⁸O record spanning 850–2000 CE from Heshang Cave (Central China), along with published records from Wanxiang Cave (North China) and Dongge Cave (South China), to identify principal components linked to monsoon and local rainfall distribution patterns and investigate the influence of the Asian monsoon, PDO, and AMO on East China's spatiotemporal rainfall patterns over the last millennium.

Numerous studies have explored the relationship between East Asian rainfall patterns and large-scale climate oscillations. Hsu & Liu (2003), Hsu & Lin (2007), Ding et al. (2008, 2009), and others have documented the tripole and dipole structures in rainfall variability and linked them to the Asian summer monsoon. Qian et al. (2014) highlighted the increasing influence of the Atlantic Multidecadal Oscillation (AMO) on decadal summer drought frequency. Wu & Mao (2017) investigated the relationship between the Pacific Decadal Oscillation (PDO) and South China rainfall. Wang et al. (2020) examined the Pacific and Atlantic controls on the relationship between Mainland Southeast Asia and East China precipitation variability. Previous research using paleoclimate data, though limited by data resolution and spatial coverage, has suggested the persistence of similar rainfall patterns in East China over longer time scales, including the MCA and LIA (Chen et al., 2015; Zhou et al., 2011). Studies using stalagmite δ¹⁸O records have revealed Asian monsoon variations (Cai et al., 2015; Cheng et al., 2009, 2016; Hu et al., 2008; Li et al., 2020; Wang et al., 2008; Yuan et al., 2004; Zhang et al., 2008; Zhao et al., 2015), providing valuable insights into past climate dynamics. However, the interpretation of these records is often complex due to the interplay of large-scale atmospheric circulation and local precipitation and evaporation effects (Baker & Bradley, 2010; Baker et al., 2019; Clemens et al., 2010; Liu et al., 2014; Pausata et al., 2011).

This study utilizes high-resolution stalagmite δ¹⁸O records from three caves in East China: Heshang Cave (central), Dongge Cave (south), and Wanxiang Cave (north). A new annually resolved δ¹⁸O record from Heshang Cave (HS4), spanning 850–2000 CE, was generated. This was achieved through careful layer identification and dating based on annual growth bands and U-Th dating (see Figure 2). The HS4 record was compared to other high-resolution records from nearby caves (Figure 3) to validate its regional representativeness. The HS4 record, along with published annually resolved records from Dongge Cave (DG) and Wanxiang Cave (WX), were then analyzed using principal component analysis (PCA). The PCA was applied to the three stalagmite δ¹⁸O datasets after linearly interpolating the WX data to achieve annual resolution to match the DG and HS datasets. Three principal components were extracted, explaining 43.28%, 32.86%, and 24.87% of the total variance respectively. PC1 was interpreted as representing monsoon intensity due to its common signal across the three geographically diverse locations, while PC2 and PC3 were interpreted to reflect tripole and dipole rainfall patterns, respectively. These PCs were then compared to records of solar irradiance, PDO, and AMO to assess their relationships. Wavelet analysis was performed on the individual stalagmite δ¹⁸O records (Figure 4) to identify significant periodicities and understand the interplay of different climatic forcing mechanisms at each location. Bandpass filtering was also applied to analyze the relationship between the monsoon and rainfall patterns on multidecadal timescales (Figure 6). The relationship between the resulting precipitation patterns and the Western Pacific Subtropical High (WPSH) is discussed, drawing on previous studies (Figure 7).

PCA of the three stalagmite δ¹⁸O records revealed three principal components. PC1, explaining 43.28% of the variance, is interpreted as reflecting large-scale Asian monsoon strength and showed a strong correlation with effective solar radiation (r = 0.42, p < 0.01), suggesting solar influence on monsoon intensity. PC1 displayed three distinct stages corresponding to the MCA, LIA, and CWP, indicating stronger monsoons during the MCA and CWP, and weaker monsoons during the LIA. PC2 (32.86% variance) showed a pattern consistent with the modern tripole pattern (-/+/-) where negative values represent drier conditions in South and North China with wetter conditions in Central China. The PC2 variations showed a strong correlation with the PDO (r = 0.21). PC3 (24.87% variance) exhibited a dipole pattern, showing opposite precipitation changes between South and North China, consistent with the influence of the AMO (r = 0.25, p < 0.05). Analysis of centennial-scale variations showed alternation between dipole and tripole patterns related to interactions between the monsoon, PDO, and AMO. On multi-decadal timescales, the study observed a strong correlation between the monsoon and rainfall distribution patterns, although the PC2-PDO and PC3-AMO relationships were less stable, indicating complex interactions amongst the variables. The western Pacific subtropical high (WPSH) plays a crucial role in mediating the influence of monsoon, PDO, and AMO on East China's rainfall patterns. A weak monsoon coupled with a positive PDO and negative AMO leads to a southward shift of WPSH resulting in a tripole pattern. Conversely, a strong monsoon with negative PDO and positive AMO leads to a northward WPSH shift, resulting in an anti-tripole pattern. However, this coupling is not always observed throughout the millennium.

The findings demonstrate the persistence of tripole and dipole rainfall patterns in East China over the past millennium, even before significant anthropogenic influences. These patterns are shown to be driven by the interplay between the strength of the East Asian monsoon, the PDO, and the AMO. The strong correlation between PC1 (monsoon intensity) and solar radiation underscores the importance of solar forcing on large-scale atmospheric circulation. The close relationship between PC2 (tripole pattern) and PDO highlights the dominant role of Pacific Ocean dynamics in shaping regional precipitation. The correlation between PC3 (dipole pattern) and AMO indicates a significant influence from the North Atlantic Ocean. The study's findings support the use of stalagmite δ¹⁸O records for reconstructing past rainfall patterns and provide valuable insights into the complex interactions among different climatic factors affecting regional hydroclimate variability. The results contribute to the evaluation and improvement of climate models and future hydroclimate projections, especially the potential for multidecadal to centennial-scale prediction.

This study successfully separated monsoon and local precipitation signals from spatially separated stalagmite δ¹⁸O records across East China, revealing a persistent influence of solar irradiance on monsoon strength and the enduring presence of dipole and tripole rainfall patterns shaped by the interactions between the monsoon, PDO, and AMO. The findings underscore the importance of considering these complex interactions when evaluating climate models and forecasting future hydroclimatic conditions in East China. Further research could focus on expanding the spatial coverage of high-resolution paleoclimate records to better understand regional variations and refine the understanding of teleconnections across the Asian monsoon region.

The study relies on three stalagmite records, limiting the spatial resolution and potential for fully capturing the complexity of regional rainfall patterns. While PCA effectively separates major signals, some subtleties in the interactions among monsoon, PDO, and AMO may be missed. The interpretation of stalagmite δ¹⁸O relies on assumptions about the relationships between δ¹⁸O, precipitation amount, and other local factors. Further research with denser spatial coverage and advanced statistical methods could enhance the understanding of the complex hydroclimate system in East China.