Modulation of centennial-scale hydroclimate variations in the middle Yangtze River Valley by the East Asian-Pacific pattern and ENSO over the past two millennia

Meteorological observations reveal an inverse relationship of summer monsoon rainfall between the tropical western North Pacific (WNP) and middle Yangtze River Valley (mid-YRV) on interannual and intraseasonal timescales. This seesaw-like pattern, named the East Asian-Pacific (EAP) pattern, is one of the major teleconnections that affect summer precipitation in the mid-YRV. However, due to the scarcity of long and continuous instrumental data, the behavior and mechanism of the EAP pattern on a long timescale are poorly understood. Here, we present a composite high-resolution summer monsoon rainfall record from 230 BC to 1810 AD, based on 1221 delta O-18 measurements and 22 Th-230 dates from two stalagmites (EB1 and EB5) in Shenqi Cave over the mid-YRV. Our new record shows that monsoon rainfall in the mid-YRV varies at a dominant periodicity of similar to 100 years, probably linked to solar activity and ENSO variability. The multi-centennial fluctuations in the EB record, broadly consistent with the North Hemisphere temperature reconstruction, display strong similarities to two stalagmite delta O-18 records from nearby Heshang and Luoshui caves. To represent the integrated hydroclimate variations over the mid-YRV, we extracted the first principal component (PC1) of these three cave delta O-18 records by using the principal component analysis. We reveal a striking inverse relationship of wet/dry conditions between the mid-YRV and tropical WNP, suggesting that changes in the EAP pattern have an important impact on centennial-scale monsoon precipitation changes. Furthermore, we compile monsoon rainfall records from the mid-YRV and WNP to reconstruct a new EAP index. This EAP index exhibits a strong correlation with the ENSO proxy, confirming that the ENSO variability may play a key role in modulating the monsoon rainfall in the mid-YRV through the EAP teleconnection. (C) 2021 Elsevier B.V. All rights reserved.