5-6年运气频带相位漂移的\”归因份额\”分解:EEMD 分离慢速分量前后的多站 meta 方案

Written by

in

Language

Evidence level: C

一个”5-6年节律”到底漂了多少相位,得先问一句:漂移是频带内真的在漂,还是慢速大周期”拖”出来的表观假象?本方案给出一个可测的拆解办法。

正文

运气学说把气候与疾病节律绑在”五运”(约5年)与”六气”(约6年)两套周期上,历代医家据此推演岁气。从现代气候科学看,5-6年恰好落在 ENSO 为主的年际变率频带内,并非凭空捏造的尺度。问题在于:即便某气象站的序列在 5-6 年频带确实有信号,它的相位(节律峰谷出现的时刻)逐年相对理论历表漂移——这个漂移本身需要”归因”。

相位漂移至少有两种来源。第一种是真实漂移:频带内振荡本身在变速、变频,例如 ENSO 事件的爆发时刻本身不严格守时。第二种是表观漂移:序列里混着长期趋势和年代际大周期(如太平洋年代际振荡),这些慢速分量没有分离干净时,会把低频的”弯”误算成频带内的相位移动,制造出本不存在的漂移份额。

方法设计分四步。第一步,对多站长序列气象数据做集合经验模态分解(EEMD)——这是专门处理非线性、非平稳序列的自适应分解工具,已被用于湖泊浮游植物动态与全球气候振荡的时频分析(PMID 30802701),以及地表温度多时间尺度模式的提取(PMID 30366325),能把趋势、年代际分量与年际频带分开。第二步,分别在”分离慢速分量前”和”分离后”两个版本序列上,配对测量 5-6 年频带相对运气历表的相位漂移量。第三步,计算缩减比例:分离后漂移量减少的部分记为”表观漂移占比”,剩下的记为”残余真实漂移占比”——这就是本文要的归因份额。第四步,多站结果用随机效应 meta 汇总,并与红噪声零分布对比:随机游走类过程本身就能大量伪造出”像模像样”的准周期,爱尔兰泥炭档案的研究已经证明,与太阳周期相似的多尺度周期在纯随机模拟里同样常见(PMID 27045989)。若运气频带的相位漂移量不显著超过红噪声零分布,那么任何”漂移”都先默认是噪声。

对易学的含义:运气历法从干支推演,天然假设相位是严格守时的。而真实气候系统的准周期相位从来都是松散的。若 meta 结果显示表观漂移占大头,说明部分”历法不准”的批评其实错怪了历法——是分析方法的低频泄漏制造了假漂移;若残余真实漂移占大头,则说明运气周期至多是”相位松散的气候吸引子”,按严格历表推演必然有系统误差。两个结果都比”对/错”的二值判断更有信息量。

证据与局限

已验证: EEMD 作为非线性非平稳序列的分解工具,在气候与生态长序列的时频分析中被反复使用且有效(PMID 30802701、PMID 30366325);红噪声/随机游走可系统性伪造准周期信号,周期检测必须对照随机零假设(PMID 27045989)。

部分验证: 5-6 年频带与 ENSO 年际变率的对应关系在气候学界是成熟共识,但”运气历表”与该频带的定量对齐研究仍很少。

待验证: 本文提出的归因份额分解本身——表观漂移占比与残余真实漂移占比的实测分布目前没有任何站点数据;运气历法与气候/疫病序列的因果链也未建立。本方案是方法学提案,不是已成立的结论。

相关阅读

证据等级: C
参考文献: PMID 30802701, PMID 30366325, PMID 27045989

L1 网络状态:该主题尚未进入 L1 对应网络。经查 CORRESPONDENCE_NETWORK.md 与 EDGE_TABLE_science-yixue.md(2026-11-11),”运气/EEMD/相位漂移”均无对应边记录,本文按 C 级封顶处理,缺口已登记至 science-yixue-l1-gaps.md。


免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。

Attribution Share Decomposition of 5-6 Year Luck Band Phase Drift: Multi-station Meta Scheme Before and After EEMD Separation of Slow Components

Evidence level: C

A “5–6-year rhythm” must first be interrogated: does the observed phase drift represent genuine movement within the frequency band, or is it an apparent artifact dragged out by a slow, large-scale cycle? This protocol provides a measurable method for such decomposition.

Methodological Framework

Component Function
EEMD (Ensemble Empirical Mode Decomposition) Separates slow components from the original signal to isolate true phase drift.
Multi-station meta-analysis Aggregates data across multiple observation points to enhance statistical robustness.
Attribution share decomposition Quantifies the proportion of observed drift attributable to each identified component.

Step-by-Step Procedure

  1. Data Collection: Gather multi-station observational data over the 5–6 year period.
  2. EEMD Decomposition: Apply Ensemble Empirical Mode Decomposition to separate the slow component from the original signal.
  3. Phase Drift Analysis: Calculate the phase drift within the isolated frequency band.
  4. Attribution Share Calculation: Determine the proportion of observed drift attributable to each identified component.
  5. Meta-Analysis: Aggregate results across all stations to produce a comprehensive meta-analysis.

Evidence and Limitations (证据与局限)

  • Evidence Level: A (empirical observation), B (theoretical derivation), C (statistical analysis).
  • Limitations: The method assumes linearity in the decomposition process, which may not hold for all real-world scenarios.
  • Traditional Consensus: Aligns with classical Yijing Studies frameworks regarding cyclical patterns and phase shifts.
  • Structural Analogy: Comparable to modern signal processing techniques used in physics and engineering.

This structured approach ensures that the decomposition of attribution shares is both rigorous and aligned with traditional metaphysical concepts, while maintaining a clear distinction between empirical evidence and theoretical constructs.

Body

Yijing Studies (a framework of pattern and timing) links climate and disease rhythms to the “Five Stages” (a ~5-year cycle) and the “Six Qi” (a ~6-year cycle), upon which traditional physicians have historically deduced annual climate patterns. From the perspective of modern climatology, the 5–6 year scale coincides with the interannual variability band dominated by ENSO, rather than being an arbitrary construct. The critical issue is that even if a meteorological station’s record exhibits signal within this 5–6 year band, its phase (the timing of the rhythm’s peaks and troughs) drifts relative to the theoretical Yijing calendar year by year—this drift itself requires attribution.

At least two sources account for this phase drift. The first is true drift: the oscillation within the band itself accelerates, decelerates, or changes frequency; for example, the onset of ENSO events does not strictly adhere to a fixed clock. The second is apparent drift: the record contains long-term trends and decadal-scale cycles (such as the Pacific Decadal Oscillation). When these slow components are not fully separated, they can be misinterpreted as low-frequency “bends” within the band, artificially inflating the apparent phase shift.

Methodological design proceeds in four steps. First, we apply Ensemble Empirical Mode Decomposition (EEMD)—an adaptive tool specifically for nonlinear, non-stationary sequences—to multi-station long-term meteorological data. EEMD has been utilized for time-frequency analysis of lake phytoplankton dynamics and global climate oscillations (PMID 30802701), as well as for extracting multi-timescale surface temperature patterns (PMID 30366325), thereby separating trends, decadal components, and the interannual band. Second, we pair-measure the phase drift relative to the Yijing calendar on two versions of the sequence: one before and one after separating the slow components. Third, we calculate the reduction ratio: the portion of drift eliminated after separation is recorded as “apparent drift proportion,” while the remainder is noted as “residual true drift proportion”—these constitute the attribution shares sought in this study. Fourth, results from multiple stations are aggregated using random-effects meta-analysis and compared against the red-noise null distribution. Random-walk-like processes can themselves generate substantial quantities of quasi-cycles that appear structured; research on Irish peat archives has demonstrated that multi-scale cycles similar to the solar cycle are equally common in pure random simulations (PMID 27045989). If the phase drift within the Yijing band does not significantly exceed the red-noise null distribution, any observed “drift” is initially attributed to noise.

Implications for Yijing Studies: The Yijing calendar, derived from the Heavenly Stems and Earthly Branches, inherently assumes strict phase adherence. In contrast, the quasi-cycles of real climate systems exhibit loose phases. If meta-analysis reveals that apparent drift constitutes the majority, it suggests that criticisms of “inaccurate calendars” may be misplaced—the low-frequency leakage in analysis methods created false drift. Conversely, if residual true drift dominates, it implies that Yijing cycles represent at most “loose-phase climate attractors,” meaning strict calendar-based deduction inevitably incurs systematic error. Both outcomes provide more nuanced information than a binary “correct/incorrect” judgment.

Evidence and Limitations

Verified: EEMD, as a decomposition tool for nonlinear non-stationary sequences, has been repeatedly employed and proven effective in time-frequency analyses of long-term climate and ecological records (PMID 30802701, PMID 30366325). However, red noise/random walks can systematically mimic quasi-periodic signals; therefore, cycle detection must be benchmarked against a random null hypothesis (PMID 27045989).

Partially Verified: The correspondence between the 5–6 year frequency band and ENSO interannual variability is a mature consensus within the climate science community. Nevertheless, quantitative alignment studies linking “luck calendars” to this specific frequency band remain scarce.

Pending Verification: The attribution decomposition proposed herein—specifically the measured distributions of apparent drift versus residual true drift—lacks any site-specific data. Furthermore, the causal chain connecting luck calendars to climate or epidemic sequences has not yet been established. This proposal constitutes a methodological framework rather than an established conclusion.

Related Reading

Evidence Level: C
References: PMID 30802701, PMID 30366325, PMID 27045989

L1 Network Status: This topic has not yet entered the corresponding L1 network. According to CORRESPONDENCE_NETWORK.md and EDGE_TABLE_science-yixue.md (2026-11-11), no edge records exist for “luckiness/EEMD/phase drift.” This article is capped at Level C, and the gap has been registered in science-yixue-l1-gaps.md.


Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *