器测和树轮资料对厄尔尼诺-南方涛动指数的响应特征

摘要

El Ni(n)o-Southern Oscillation (ENSO) plays the most conspicuous role in shaping the interannual (0～ 10 years) climate variability of the globe,which is a key concern for studies on climate change for both of the modern and historical periods.Different indices have been developed to characterize its features and evolutions.The indices are defined for different oceanic regions or including different oceanic and atmospheric components.However,the instrumental data to infer changes of ENSO are limited to the industrial era.This limits our ability to fully comprehend the regimes of ENSO in the historical period,when both anthropogenic activities play a less important role on climate change.Tree rings are to our knowledge the most frequently used data to reconstruct the ENSO variability of the past millennia,because they are climate sensitive,accurately dated,highly resolved and have a large and dense spatial coverage.Although numerous studies have been conducted on the linkages between ENSO and climate changes in both the instrumental and historical periods,it is still unknown about the responses of instrumental and tree-ring data to different indices of ENSO.This study presents a systematic survey on the linkages between different ENSO indices and the instrumental temperature,precipitation and the tree-ring data for different timescales in eastern Asia and North America.This can improve our understandings on the linkages between regional climate and different ENSO indices and can also shed lights on whether there are systematic biases for treering based ENSO reconstructions.This study employed the instrumental temperature and precipitation data,as well as a tree-ring network of 335 tree-ring chronologies,including 196 chronologies for eastern Asia and 139 chronologies for North America.All these chronologies are over 400 years and have been widely used for climate reconstructions including ENSO.The ENSO indices employed herein include the definition based on sea surface temperature (SST) in eastern and central Pacific Ocean,i.e.the Ni(n)ol+2 and Ni(n)o3.4,and the SST gradients in this area,i.e.the Trans-Ni(n)o Index (TNI),and the SST differences between central and eastern and western Pacific Ocean,i.e.the ENSO Modoki Index (EMI).In addition,we also employed the atmospheric component of the ENSO,i.e.the Southern Oscillation (SO) and the index with the most comprehensive oceanic and atmospheric components,i.e.the Multivariate ENSO Index (MEI).ENSO indices have higher correlations with temperature and tree rings than at interannual scales than at interdecadal scales.Both the climate and the tree-ring data have the highest correlations with MEI among the ENSO indices.This indicates that the coupled indices incorporating both oceanic and atmospheric components have closer linkages with climate and thus tree rings than individual oceanic and atmospheric components.Both climate data and tree rings have the lowest correlation with TNI,suggesting that the temperature gradients in central and eastern tropical Pacific Ocean have least impacts on large-scale climate anomalies.EMI has low correlations with climate data but have high correlations with tree rings.This may be because that the EMI can cause common changes of many climate variables,which are similar as the combination of the limiting climate variables for tree growths.ENSO indices have the highest correlations with temperature and tree-ring data in southeastern Asia,Indian subcontinent and the western bank of Canada.Both precipitation and tree-ring data have significant correlations in central Asia and northwestern USA.There are also mismatches among the correlations of climate and tree-ring data with ENSO indices.Tree-ring data have high correlations with many tree-ring chronologies from wet regions,such as the southern Himalaya,where no significant correlation is observed between climate data and the ENSO.Instead of identifying the relationships between ENSO and tree rings before an ENSO reconstruction,our study highlights the necessity to examine the linkages among tree rings,regional climate variables and ENSO.%厄尔尼诺-南方涛动是影响全球年际(1～10年)气候变化最显著的海气因子,是现代气候和古气候的热点研究对象.目前有多个指数能够刻画厄尔尼诺-南方涛动特征与演化,树轮资料是重建其过去千年以来演化历史所使用的最为广泛的代用资料,然而系统对这些指数与器测温度、降水和树轮资料在不同时间尺度上关联的研究尚未开展.本文研究了东亚和北美的器测与树轮资料对6种厄尔尼诺-南方涛动指数的响应差异,发现器测和树轮资料对描述多气候变量耦合变化的多变量厄尔尼诺-南方涛动指数(MEI)的响应均最高.厄尔尼诺-南方涛动Modoki指数(EMI)与器测资料的相关偏低,但是该指数与树轮资料的相关偏高,这可能是由于该指数和树轮所关联的多个气候因子的协同变化相似.不同厄尔尼诺-南方涛动指数与东南亚和印度次大陆以及加拿大中西部的温度相关显著,在年际尺度上更明显;这些指数与中亚和美国西北部的降水的相关最显著,但在年际尺度上的相关提升不明显.厄尔尼诺-南方涛动指数除了与上述区域的树轮资料显著相关以外,该指数还与部分湿润地区(如喜马拉雅南麓)的树轮资料显著相关,这可能是这些湿热地区的树轮往往受到多个气候因子共同影响,而这些气候因子的协同变化与厄尔尼诺-南方涛动指数的关联显著.研究指出,进行厄尔尼诺-南方涛动指数重建时,不仅要考虑指数与树轮之间的关系,还需要详细建立树轮、区域气候因子和厄尔尼诺-南方涛动这三者之间的关联.