Reply to: A critical examination of a newly proposed interhemispheric teleconnection to Southwestern US winter precipitation
Reply to: A critical examination of a newly proposed interhemispheric teleconnection to Southwestern US winter precipitation
复制标题
回复：对新提出的美国西南部冬季降水半球间遥相关的严格检查
DOI：
10.1038/s41467-019-10531-3
复制
发表时间：
2019
影响因子：
16.6
通讯作者：
Foufoula-Georgiou, Efi
中科院分区：
文献类型：
--
作者：
Mamalakis, Antonios;Yu, Jin-Yi;Randerson, James T.;AghaKouchak, Amir;Foufoula-Georgiou, Efi
Gibson等人。1对Mamalakis等人提出的物理机制进行了评论。2(下称M18)，并质疑新提出的半球间遥相关的第一步，即大气桥，即新西兰附近的海表面温度(SST)(称为NZI；新西兰指数2)调制西北太平洋SST的第一步。具体地说，他们认为这些关键区域之间没有直接的因果关系，因为观测到的相应SST异常之间的高度统计相关性在很大程度上可以用当地的SST记忆和厄尔尼诺-南方涛动(ENSO)来解释，而且正如M18所报告的那样，过去40年来相关性的增加可能只是内部变化的结果，而不是历史强迫的结果。Gibson等人。1还认为，温暖的NZI与M18所建议的菲律宾地区东部对流活动和云层减少并不相关。我们感谢有机会辩论这些问题。原则上，我们同意，不止一个因素的组合可以推动西北太平洋气候(和SST)的变化(即，M18并不认为NZI是唯一的驱动因素)。然而，我们不同意Gibson等人的一般建议。1 M18中提出的大气桥不受数据(观测和模型)的支持。在这里，我们提出了证据，证明NZI确实携带着不能被排除的非冗余信息，它不能仅用内部变化来解释，我们还提供了进一步的分析，支持拟议的大气桥的因果机制。我们还指出了Gibson等人分析中的一些复杂的局限性，这可能影响了他们的结论。Gibson等人的论点之一。1在挑战NZI与西北太平洋SST的滞后关系时，这些关键区域之间的统计相关性在考虑(使用部分相关)本地SST记忆或ENSO(具体地说，南方涛动指数-SOI)时减少。然而，他们的分析显示出一些局限性。首先，M18并不认为NZI是西北太平洋海温变化的唯一驱动因素。事实上，M18已经在他们提出的远程连接机制中调用了本地SST内存(参见M18图5中的步骤2)。M18简单地论证了一种新的西太平洋半球间遥相关的出现，在海洋-大气耦合系统中，这种遥相关可以影响北太平洋气候，并最终影响美国西南部(SWU)的降水。澄清了这一点后，有意义的问题不是在考虑其他预测因素/机制时，西北太平洋NZI和SST的相关性是否会降低(这是意料之中的)，而是是否仍然存在其他预测因素没有解释的统计上显著的关系模式。正如Gibson等人S 1自己的结果所表明的那样，在考虑了本地SST记忆和SOI(参见他们的图1E，f)后，统计上显著的(a=0.05时的本地假设检验)相关性的一致模式仍然明显，这意味着NZI不是西北太平洋SST的冗余预测因子。因此，我们不认为Gibson等人的结果。1.对我们的总体建议提出质疑。请注意，尽管在Gibson等人的结果中。1(他们的图1F)存在局部统计显著相关性的整个模式，作者似乎仅使用错误发现率(FDR)方法的结果来评估相关性重要性。这可能具有误导性，因为FDR只控制I类错误的可能性(拒绝…的真正空值
Gibson et al. 1 comment on the physical mechanism suggested by Mamalakis et al. 2 (hereafter referred to as M18), and question the first step of the newly proposed interhemispheric teleconnection, ie, the atmospheric bridge, whereby sea surface temperature (SST) close to New Zealand (termed as NZI; the New Zealand Index 2) modulates the SST in the Northwestern Pacific. Specifically, they suggest that there is no direct causal relationship between these key areas since the observed high statistical correlations between the corresponding SST anomalies can be largely explained by local SST memory and the El Niño-Southern Oscillation (ENSO), and that the increase in the correlations over the past four decades, as reported in M18, is likely the result of internal variability alone, not caused by historical forcings. Gibson et al. 1 also argue that warm NZI is not associated with decreased convective activity and cloud cover over the east of the Philippines region, as suggested by M18. We appreciate the opportunity to debate these issues. In principle, we agree that a combination of more than one contributors can drive climate (and SST) variability in the northwestern Pacific (ie, M18 did not argue that NZI is the only driver). However, we disagree with the general suggestion by Gibson et al. 1 that the atmospheric bridge, as proposed in M18, is not supported by the data (observations and models). Here, we present evidence that indeed NZI carries non-redundant information that cannot be dismissed, it cannot be explained by internal variability alone, and we provide further analysis that supports the causal mechanism of the proposed atmospheric bridge. We also point out some intricate limitations in the analysis of Gibson et al., that might have affected their conclusions. One of the arguments of Gibson et al. 1 in challenging the NZI lagged association with northwestern Pacific SST is that the statistical correlations between these key areas decrease when accounting for (using partial correlation) local SST memory or ENSO (specifically, the Southern Oscillation Index-SOI). However, their analysis exhibits some limitations. Firstly, M18 did not argue that NZI is the only driver of SST variability in the northwestern Pacific. In fact, M18 already invoked the local SST memory in their proposed teleconnection mechanism (see step 2 in Figure 5 of M18). M18 simply argued for the emergence of a new western Pacific interhemispheric teleconnection, which in the ocean-atmosphere coupled system can affect, among other contributors, the north Pacific climate and ultimately the precipitation in the southwestern US (SWUS). Having clarified this, the meaningful question is not whether correlations of NZI and SST in the northwestern Pacific decrease when considering additional predictors/mechanisms (this is to be expected), but whether there are still patterns of statistically significant relations not explained by other predictors. As Gibson et al.’s 1 own results suggest, a consistent pattern of statistically significant (local hypothesis testing at a= 0.05) correlations is still evident after accounting for both local SST memory and SOI (see their Fig. 1e, f), which means that NZI is not a redundant predictor of the SST in the northwestern Pacific. Thus, we do not think the results of Gibson et al. 1 challenge our general suggestion. Note that although in the results of Gibson et al. 1 (their Fig. 1f) there is an entire pattern of local statistically significant correlations, the authors seem to assess correlation significance solely by using the results from the false discovery rate (FDR) method. This can be misleading, because FDR only controls the likelihood of the type I error (rejecting a true null …
用于跟踪热带辐合带的多元概率框架：近期气候学和过去趋势的分析
影响因子：
5.2
作者：
Mamalakis, Antonios;Foufoula‐Georgiou, Efi
通讯作者：
Foufoula‐Georgiou, Efi
对新提出的美国西南部冬季降水半球间遥相关的严格检查
影响因子：
16.6
作者：
P. Gibson;D. Waliser;M. DeFlorio
通讯作者：
M. DeFlorio