Tree-ring Inferred Winter Temperature Change Over The Past Three Centuries In The Hengduan Mountains Of Northwestern Yunnan

Situated on the southeastern Tibetan Plateau,the Hengduan Mountains of northwestern Yunnan is a sensitive area of climate change.Under the background of global warming,this area has undergone an apparent temperature rise,which has led to an evident natural landscape change and threatened the valuable and vulnerable ecological resources.Paleoclimatology study of the region can provide a background of long-term climatic variability for assessing the current warming and testing hypotheses about the climate forcing,and therefore contribute to the evaluation,prediction and adaptation for climate change impacts.However,a lack of long-term,high-resolution proxy records containing cold season temperature signals has limited our understanding of the regional climate change.To this end,we introduced Pinus yunnanensis Franch into dendroclimatic research,and chose the winter temperature variations as the target for tree-ring width based reconstruction and relative analyses.In this study,we built a robust 296-year tree-ring width chronology of Pinus yunnanensis Franch from two high-altitude sites in Shangri-La,northwestern Yunnan and checked the tree growth-climate relationships.It turned out that before 1992,winter temperature was the limiting factor of tree growth,when the chronology contains clear and steady winter temperature signals.However,the temperature sensitivity of tree-ring index decreased sharply after 1992,which is an obvious manifestation of the"divergence problem".We explored the cause of the "divergence problem" from various perspectives,such as the validity of instrumental climate data,the methods of chronology standardisation,and the statistical techniques used in tree growth-climate response analyses.As a result,we ruled out the some common causes of divergence like the inhomogeneity of climatic data,the failure of detrending,the aging of trees,and the microenvironment-scale disturbance.Although the exact cause of the divergence in this research is still hard to pinpoint at the current stage,we present a reasonable hypothesis that after 1992,winter temperature is no longer the dominant limiting factor due to its rapid increase,and the tree radial growth is controlled by several interwoven environmental factors,while May-June precipitation has the potential to become the new dominant limiting factor.Based on the calibration period from 1959 to 1992,we reconstructed the December-February mean temperature for the past three centuries using a linear regression model.The reconstruction passed all standard calibration-verification schemes and explained nearly 73%of the variance of the instrumental temperature series.The winter temperature of study area may have reached the highest level of the past three centuries during the 21st century according to the comparison between instrumental and reconstructed data.Spectral analysis revealed significant quasi-periodicities of 23.2a,4.9a,4.0a,3.3a and 2.3a.Spatial correlation analysis shows that this reconstruction represents large-scale temperature variations in southwest China and the eastern Tibetan Plateau.Most of the warm and cold periods in our reconstruction matched well with other tree-ring based reconstructions from nearby regions.Comparisons were made between our reconstructed December-February mean temperature and a yearly sunspot number series,a large volcanic eruption chronology and the Atlantic Multidecadal Oscillation(AMO)series to explore the possible driving forces of temperature change.The results of series comparison and cross wavelet transform analysis suggest that winter temperature variations in the study area have an unsteady phase relation with solar activity in decadal and multidecadal scales.Comparison between our reconstruction and the chronology of volcanic eruptions with Vocano Eruption Index>5 since 1718 suggest that extremely cold winters are supposed to appear after large volcanic eruptions.During the past three centuries,the winter temperature shows a persistent close association with AMO,with warm(cold)periods coinciding with the positive(negative)phases of the AMO,suggesting that the AMO may have been a key driver of multidecadal winter temperature variations on the southeastern Tibetan Plateau.