The El Niño-Southern Oscillation (ENSO) is a climate pattern characterized by a seesaw pattern of sea surface temperature (SST) and atmospheric pressure between the eastern and western tropical Pacific. It significantly influences global climate. However, the response of ENSO variations to increased greenhouse gases and the influence of different tropical Pacific climatic mean states in the pre-industrial period remain unclear. Understanding the natural variability of ENSO and its response to these factors is crucial for validating climate models and predicting future behavior under global warming. Existing observational data, spanning only about 150 years, are insufficient to investigate long-term changes in the mean state. Paleoclimate datasets and climate models offer insights, but existing reconstructions using Southern Oscillation Index (SOI), SST records, and zonal SST gradients have yielded conflicting results regarding changes in the tropical Pacific mean state during the Medieval Warm Period (MWP) and the Little Ice Age (LIA). Some studies suggest La Niña-like conditions during the MWP and El Niño-like conditions during the LIA, while others show the opposite. These discrepancies highlight the need for additional, robust reconstructions, particularly those that minimize the influence of factors such as the Intertropical Convergence Zone (ITCZ) shifts, which can confound interpretations of the mean state. This study leverages the unique teleconnection between East Asian precipitation and tropical Pacific SST anomalies to construct a high-resolution, 1800-year record of the tropical Pacific mean state, providing a more complete picture of its long-term evolution.

Previous research on the mean state of the tropical Pacific during the MWP and LIA has relied on various proxies, including SST reconstructions from corals and foraminifera, and SOI reconstructions from hydroclimate records. These studies have presented conflicting results, with some suggesting a prevalence of La Niña-like conditions during the MWP and El Niño-like conditions during the LIA, and vice versa. Discrepancies have been attributed to the complex interplay between ENSO and the ITCZ. The influence of ITCZ movements on equatorial Pacific hydroclimate complicates the use of these records as strict indicators of the tropical Pacific's mean state. The lack of consistent findings highlights the limitations of existing datasets and the need for alternative approaches. Climate models also offer insights into past ENSO behavior, but different models produce varying results, further emphasizing the need for more comprehensive and robust paleoclimate reconstructions. The study presented here aims to overcome these limitations by using East Asian precipitation records as a robust indicator of tropical Pacific mean state, minimizing the influence of ITCZ variations.

This study utilizes precipitation reconstructions from two East Asian lakes: Lake Nvshan (Jiang-Huai region, JH) and Lake Gonghai (North China region, NC). These regions are outside the direct influence of the ITCZ, making their precipitation records less susceptible to ITCZ-related biases. The precipitation reconstruction from Lake Nvshan is based on sediment redness (a*/L* ratio), a proxy sensitive to changes in lake water level, reflecting precipitation patterns. The Lake Gonghai reconstruction uses pollen assemblages, converted into precipitation proxies based on correlations with modern pollen distributions. Twenty precipitation records from East Asia spanning the last millennium were analyzed to establish the spatial pattern of precipitation changes during the MWP and LIA. These records were classified according to dry-wet climate grades. The spatial pattern of precipitation variability in East Asia demonstrates a tripole pattern, linked to ENSO events. The authors then construct an 1800-year index of the tropical Pacific mean-state by taking the difference between the normalized precipitation records from Lake Nvshan and Lake Gonghai. A positive index indicates El Niño-like conditions, while a negative index suggests La Niña-like conditions. The reliability of the index is evaluated by comparing it to other existing tropical Pacific mean-state reconstructions, including those derived from zonal SST gradients, coral δ¹⁸O values, and stalagmite δ¹⁸O records. Statistical analyses, including Pearson correlation, are used to assess the relationships between the constructed index and various forcing mechanisms such as total solar irradiance (TSI), effective radiative forcing (ERF), global mean temperature (GMT), and Northern Hemisphere temperature (NHT), and greenhouse gas concentrations. Monte Carlo simulations were used to estimate uncertainties in the index due to chronological and analytical uncertainties. The variance of the tropical Pacific mean-state was assessed using a 50-year Lanczos high-pass filter and a 100-year running biweight variance.

The 1800-year reconstruction of the tropical Pacific mean-state reveals four distinct centennial-scale phases. The MWP (approximately 1000-1300 CE) is characterized by La Niña-like conditions with low variance, while the LIA (approximately 1400-1850 CE) and the CWP (post-1850 CE) show El Niño-like conditions with high variance. This index shows a strong negative correlation with effective radiative forcing and GMT during the pre-industrial period, supporting the "ocean dynamical thermostat" mechanism. During the pre-industrial era, changes in the mean state closely covary with TSI and ERF, consistent with the ocean dynamical thermostat mechanism, which suggests that increased radiative forcing leads to La Niña-like conditions and vice-versa. However, the persistence of El Niño-like conditions during the CWP, despite continued global warming, cannot be fully explained by this mechanism and is consistent with some GCM simulations showing a weakening of the Pacific Walker circulation under global warming. The analysis of tropical Pacific variability reveals lower variance during the MWP (La Niña-like) and higher variance during the LIA and CWP (El Niño-like). This finding is consistent with other ENSO variance reconstructions. The relationship between variability and mean state suggests a positive feedback, where increased ENSO variance leads to El Niño-like states, and these states, in turn, amplify ENSO variability. The study demonstrates that the responses of the tropical Pacific mean state differ significantly between the pre-industrial and industrial periods. Before the industrial era, natural variability, driven primarily by solar irradiance and volcanic activity, was the dominant factor. However, after 1850, anthropogenic greenhouse gas forcing became the major driver. The study's findings are supported by the close agreement between their mean-state index and other existing reconstructions based on SST patterns and hydroclimate variations, although the results contrast with some SOI reconstructions based on equatorial Pacific hydroclimate.

The findings of this study provide valuable insights into long-term changes in the mean state of the tropical Pacific and their driving mechanisms. The contrasting responses of the tropical Pacific to radiative forcing during the pre-industrial and industrial periods highlight the importance of considering both natural variability and anthropogenic influence. The close agreement between the reconstructed mean-state index and other independent reconstructions, while differing from some SOI reconstructions based on equatorial Pacific hydroclimate, underscores the potential impact of ITCZ variability on the interpretation of hydroclimate proxies. The results reconcile discrepancies in previous studies and provide a more comprehensive understanding of ENSO behavior across different climate periods. The positive feedback loop between ENSO variance and mean state offers implications for future climate projections under continued global warming. The finding that El Niño-like conditions prevailed during the CWP suggests a potential weakening of the Pacific Walker circulation, a phenomenon that needs further investigation.

This study presents a robust 1800-year reconstruction of the tropical Pacific mean state using East Asian precipitation records, revealing distinct La Niña-like and El Niño-like periods corresponding to the MWP and LIA/CWP respectively. The study's methodology successfully mitigates biases associated with ITCZ variability in equatorial Pacific hydroclimate proxies, providing a more nuanced understanding of the complex interplay between ENSO, radiative forcing, and climate change. The results highlight the need for further high-resolution reconstructions from other ENSO-teleconnected regions to refine the understanding of long-term ENSO variability and its response to changing climate conditions. Future research should focus on exploring the mechanisms underlying the positive feedback between ENSO variance and mean state and improving the accuracy of climate model projections of future ENSO behavior.

While this study offers a novel and robust reconstruction of the tropical Pacific mean state, several limitations should be acknowledged. The reliance on East Asian precipitation as an indicator of tropical Pacific conditions relies on the stability and robustness of the teleconnection between these regions. Changes in atmospheric circulation patterns over time could potentially affect the strength of this teleconnection and influence the accuracy of the reconstructed mean state. Furthermore, the proxy records used in the study have inherent uncertainties associated with dating and measurement techniques, which can propagate into the reconstructed index. The study focuses primarily on centennial-scale variability and may not capture shorter-term fluctuations in the tropical Pacific mean state.