运气周期相位漂移的真相:表观漂移与真实漂移的科学分离

Written by

in

Language

Evidence level: B

运气周期(5年/6年)被观测到的”相位漂移”,很可能是慢速气候振荡(AMO/PDO)与增暖趋势叠加造成的”表观漂移”,而非周期本身在变——现代信号分离方法可以把两者切开。

正文

运气学说认为气候存在五运六气的节律:岁运以五年为一小周期(五运轮转),六气以六年为一大周期(六气司天在泉),两相叠加形成三十年一纪、六十年一周的复合节律。这套以天文-历法为骨架的周期模型,暗含一个前提:周期本身是稳定的,如同钟摆。钟摆的周期若发生漂移,是值得警惕的大事——那意味着自然的节律变了。

但问题来了:当我们用现代气象站长序列去检验运气周期时,常常发现”相位对不上”——本该在某年出现的运气特征提前或推迟了。这是运气周期在真实漂移,还是我们看错了?

气候科学在过去十年给出了一条关键的思路:观测到的”周期漂移”,很多是慢速外强迫叠加在周期信号上造成的表观效应,而非周期本身的属性变化。

一个教科书级的教训:AMO 之争

大西洋多年代际振荡(AMO)曾被认为是气候系统内部自发的 40-70 年振荡:1950-1960 年代变冷、之后变暖、再之后又变冷,看起来就是一个完美的慢速钟摆。但 Mann 等人在 2020 年用 24 个气候模式的对照模拟(无外强迫的”控制运行”)证明:控制模拟中不存在任何可与噪声区分的多年代内部振荡,观测到的 40-50 年谱峰,实际上是 1950-1970 年代硫酸盐气溶胶的降温与之后温室气体增暖的”表观振荡”(PMID 31900412)。随后他们又在 2021 年用过去千年模拟证明,前工业时期那些”多年代振荡”同样是火山脉冲在 50-70 年频带的投影假象(PMID 33674487)。

这个教训翻译成运气学的语言就是:我们站在一条正在被外力推动的曲线上观测周期,曲线本身的弯曲会被误读成周期的相位移动。全球变暖的长期趋势(相当于运气学说里”太过”的长期化)叠加在 5-6 年的运气周期上,趋势的斜率变化会系统性地把周期的峰值点向后或向前推移——这就是”表观漂移”。

为什么趋势会伪造相位漂移

设想一条正弦曲线(运气周期)加一个单调上升的直线(增暖趋势)。当直线变陡时,正弦的上升半周期被”拉长”,下降半周期被”压缩”,峰值出现的年份就会系统性偏移。Tung 和 Zhou 2013 年在《PNAS》上系统展示了如何用数据归因变暖与变冷事件:不先把趋势和周期分离,任何对周期相位/振幅的估计都会被污染(PMID 23345448)。这正是我们在运气检验中反复遇到”相位对不上”的头号嫌疑犯——不是天干地支算错了,是信号分解没做对。

AMO/PDO 对中国气候的真实调制

更重要的是,AMO 与太平洋年代际振荡(PDO)并非”纸上谈兵”的统计构造,它们实实在在地调制中国东部降水与气温的年代际格局:PDO 暖位相对应华北干旱化与长江流域多雨,AMO 则起加强或削弱作用,二者位相组合决定”南涝北旱”或”南旱北涝”的分布。把这种 10-70 年的慢速调制与 5-6 年的运气周期叠在一起,恰恰落在运气检验最敏感的时间尺度窗口内——运气周期约 2.5 个周期(约 15 年)的检验长度,正好覆盖半个 AMO 子谐波。若不分离,AMO 的位相切换会被误认为运气周期自身的相位漂移。

把”表观”和”真实”切开的三步法

第一,连续小波谱 + 趋势-周期分离:先对站长序列做去趋势或 EMD/低通分离,把慢速分量(趋势+AMO/PDO)从 5-6 年频带中切出去,再提取运气频带的相位角。

第二,交叉小波相干:把运气频带相位与 AMO/PDO 指数做交叉小波分析,检验运气频带的”漂移”是否与慢速振荡指数显著相干——若相干显著,说明漂移是被慢速振荡”牵着走”的表观漂移;若不显著,才可能是运气周期自身的真实变化(PMID 41392677 已在中国站级数据上示范了多元小波相干的可行性)。

第三,置换检验:用合成时间序列(含趋势+已知相位漂移)标定效应量,回答”漂移多大才算显著”,避免把红噪声背景下的随机抖动当漂移。

对运气研究的含义

这一步分离一旦做对,运气检验的结论会分叉成两种截然不同的解读:若相位漂移全部可由慢速振荡解释,说明运气周期本身依然稳定,变的是”背景”——运气学说作为稳定节律框架的假设反而被保全;若分离后仍有残余漂移,那才是真正需要解释的异常。换句话说,先证伪”表观漂移”,再谈”真实漂移”,是运气周期检验走向可证伪科学的第一步。

证据与局限

已验证: AMO 观测到的多年代振荡在气候模式对照模拟中不显著,表观谱峰可由气溶胶与温室气体外强迫解释(PMID 31900412);前工业多年代振荡是火山脉冲的频带投影假象(PMID 33674487);趋势与周期不分离会污染相位/振幅估计(PMID 23345448);AMO/PDO 确实调制中国东部降水的年代际格局。

部分验证: 运气学说气象变量相关性有中国直接证据(风疹与 3 年前气象变量的关联 PMID 30467697;手足口病与运气因子 PMID 38543836),但均未做趋势-周期分离。

待验证: ① 中国多站长序列去趋势后,5年/6年运气频带相位与 AMO/PDO 指数的交叉小波相干是否显著(区分表观/真实漂移的直接检验);② 分离后残余相位漂移的存在性与量级;③ 运气周期在”控制模拟无外强迫”框架下的稳定性检验。

相关阅读

证据等级: B
参考文献: PMID 31900412, PMID 33674487, PMID 23345448, PMID 41392677, PMID 30467697


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。

The Truth About Luck Cycle Phase Drift: Scientific Separation of Apparent and Real Drift

Evidence level: B

The observed “phase drift” of the Yunqi cycle (5-year/6-year) is likely an “apparent drift” caused by the superposition of slow climate oscillations (AMO/PDO) and warming trends, rather than an actual change in the cycle itself—modern signal separation methods can disentangle the two.

Main Text

Yunqi theory posits that climate operates on a rhythm of Five Transitions and Six Qi: the Sui Yun (Annual Cycle) operates on a five-year sub-cycle (Five Transitions rotate), while the Liu Qi operates on a six-year super-cycle (Six Qi govern Heaven and Earth), combining to form composite rhythms of thirty years per Ji and sixty years per Zhou. This cycle model, built on an astronomical-calendrical skeleton, implicitly assumes a premise: the cycle itself is stable, like a pendulum. If the period of a pendulum were to drift, it would be a cause for alarm—implying that the natural rhythm has changed.

However, a problem arises: when we test Yunqi cycles against modern long-term meteorological station data, we often find “phase misalignment”—Yunqi characteristics that should appear in a certain year occur earlier or later. Is the Yunqi cycle truly drifting, or are we misinterpreting the signal?

Climate science has offered a key insight over the past decade: “observed cycle drift” is often an apparent effect caused by slow external forcing superimposed on the cycle signal, not an intrinsic property change of the cycle.

A Textbook Lesson: The AMO Controversy

The Atlantic Multidecadal Oscillation (AMO) was once considered a spontaneous internal oscillation of the climate system with a period of 40–70 years: cooling in the 1950s, warming thereafter, then cooling again, appearing as a perfect slow pendulum. However, Mann et al. (2020) demonstrated using 24 climate model control simulations (no external forcing) that no multidecadal internal oscillation distinguishable from noise exists in control simulations; the observed spectral peaks at 40–50 years are actually an “apparent oscillation” resulting from sulfate aerosol cooling in the 1950s–1970s followed by greenhouse warming (PMID 31900412). Subsequently, they demonstrated in 2021 that pre-industrial “multidecadal oscillations” were similarly projection artifacts of volcanic pulses projected into the 50–70 year band (PMID 33674487).

Translated into Yunqi terminology, this lesson means: we are observing a cycle on a curve being pushed by external forces; the curvature of the curve itself is misread as phase movement of the cycle. The long-term warming trend (equivalent to the “excessive” prolongation in Yunqi theory) superimposed on the 5–6 year Yunqi cycle causes systematic shifts in peak points forward or backward due to changes in the trend slope—this constitutes “apparent drift”.

Why Trends Fabricate Phase Drift

Imagine adding a monotonic rising line (warming trend) to a sine wave (Yunqi cycle). As the line steepens, the rising half-cycle of the sine wave is “stretched,” while the falling half-cycle is “compressed,” causing systematic shifts in the year of peak occurrence. Tung and Zhou (2013) systematically demonstrated in PNAS how to attribute warming and cooling events using data decomposition: without separating trend from cycle, any estimate of cycle phase or amplitude becomes contaminated (PMID 23345448). This is the primary suspect we repeatedly encounter in Yunqi validation when “phase does not match”—not that the Heavenly Stems and Earthly Branches were calculated incorrectly, but that the signal decomposition was performed incorrectly.

Real Modulation of AMO/PDO on Chinese Climate

More importantly, the AMO and Pacific Decadal Oscillation (PDO) are not mere statistical constructs; they genuinely modulate the interannual patterns of precipitation and temperature in eastern China: PDO warm phases correspond to aridification in North China and increased rainfall in the Yangtze River basin, while AMO acts as an amplifier or dampener; their phase combinations determine the distribution of “southern floods/northern droughts” or vice versa. Superimposing this slow modulation (10–70 years) onto the 5–6 year Yunqi cycle precisely falls within the most sensitive temporal window for Yunqi validation—the validation length of approximately 2.5 cycles (~15 years) covers half an AMO sub-harmonic. Without separation, phase switches of the AMO are misread as intrinsic phase drifts of the Yunqi cycle.

A Three-Step Method to Separate “Apparent” from “Real”

First, Continuous Wavelet Spectra + Trend-Cycle Separation: Perform detrending or EMD/low-pass separation on station series first to excise slow components (trend + AMO/PDO) from the 5–6 year band, then extract phase angles within the Yunqi band.

Second, Cross-Wavelet Coherence: Perform cross-wavelet analysis between Yunqi band phase and AMO/PDO indices to test whether “drift” in the Yunqi band is significantly coherent with slow oscillation indices. If coherence is significant, the drift is an apparent drift “led” by slow oscillations; if not significant, it may represent real changes in the Yunqi cycle itself (PMID 41392677 has demonstrated the feasibility of multivariate wavelet coherence on station-level data).

Third, Permutation Testing: Use synthetic time series (containing trend + known phase drift) to calibrate effect sizes, answering “how much drift constitutes significance,” thereby avoiding treating random jitter under red noise background as drift.

Implications for Yunqi Research

Once this separation is performed correctly, conclusions from Yunqi validation bifurcate into two distinct interpretations: if all phase drifts can be explained by slow oscillations, it indicates the Yunqi cycle itself remains stable while the “background” changes—thereby preserving the assumption of Yunqi theory as a stable rhythmic framework; if residual drift persists after separation, that is the true anomaly requiring explanation. In short, disproving “apparent drift” before discussing “real drift” is the first step toward making Yunqi cycle validation a falsifiable science.

Evidence and Limitations

Verified: Observed multidecadal oscillations in AMO are not significant in climate model control simulations; apparent spectral peaks can be explained by aerosol and greenhouse gas external forcing (PMID 31900412); pre-industrial multidecadal oscillations are projection artifacts of volcanic pulses (PMID 33674487); failure to separate trend from cycle contaminates phase/amplitude estimates (PMID 23345448); AMO/PDO genuinely modulate interannual precipitation patterns in eastern China.

Partially Verified: Correlations between Yunqi theory and meteorological variables have direct evidence in China (association between chickenpox and meteorological variables from 3 years prior PMID 30467697; Hand, Foot, and Mouth disease and Yunqi factors PMID 38543836), but none performed trend-cycle separation.

To Be Verified: ① Whether cross-wavelet coherence between 5-year/6-year Yunqi band phase and AMO/PDO indices is significant after detrending multiple Chinese station series (direct test distinguishing apparent/real drift); ② Existence and magnitude of residual phase drift after separation; ③ Stability testing of the Yunqi cycle within a “control simulation with no external forcing” framework.

Related Reading

Evidence Level: B
References: PMID 31900412, PMID 33674487, PMID 23345448, PMID 41392677, PMID 30467697


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 *