一項發表於《自然》期刊的新研究指出，地球對極端太陽風暴的反應可能比科學家過去認為的更為劇烈。長期以來，研究人員相信地磁反應存在一個天然的「飽和」上限，即當太陽風強度達到極高值時，地球磁場不再按比例回應。然而，這項研究表明，這種飽和現象可能只是測量分析方式所造成的統計假象，而非真實的物理機制。

問題的核心在於一個被稱為「回歸均值」的統計現象。科學家主要依賴位於拉格朗日L1點的衛星來測量太陽風強度，但該處的觀測值與太陽風實際與地球磁場交互作用時的狀態並不完全一致。當某次測量值異常極端時，其所代表的真實值在統計上往往較為溫和，這種不確定性在數千次重複分析中會系統性地低估地球對強烈太陽風的反應，從而製造出飽和的假象。

研究團隊開發了一個納入主要不確定性來源的統計模型，並運用「回歸校準」技術修正測量偏差後，所謂的飽和效應幾乎完全消失，太陽風強度與地磁反應之間重新呈現線性關係。這意味著若類似1859年卡林頓事件等級的極端太陽風暴再次發生，其地磁衝擊可能是傳統模型估算的兩倍，對現代電網、衛星通訊及導航系統構成遠超預期的威脅。

A new study published in Nature challenges the long-held belief that Earth's geomagnetic response to extreme solar storms reaches a natural saturation point. For years, scientists assumed that beyond a certain solar wind intensity, the magnetosphere stops reacting proportionally. However, the researchers demonstrate that this apparent saturation is likely a statistical illusion arising from how measurements are analyzed, rather than a genuine physical limitation of the Earth's magnetic field.

The key issue lies in the statistical phenomenon of regression toward the mean, compounded by the fact that solar wind measurements are taken at the L1 Lagrange point, roughly 1.5 million kilometers from Earth, where conditions differ from those at the actual point of interaction with the magnetosphere. Exceptionally high readings at L1 tend to overestimate the true solar wind intensity upon arrival, and when these inflated measurements are correlated with Earth's observed response thousands of times, they create a systematic bias that mimics saturation.

After developing a statistical model incorporating sources of uncertainty and applying regression calibration to correct the bias, the researchers found that the saturation effect nearly vanishes, revealing a fundamentally linear relationship between solar wind intensity and geomagnetic response. This implies that a Carrington-class event could produce roughly twice the geomagnetic impact predicted by traditional models, posing a far greater danger to modern electrical grids, satellite communications, and navigation systems than previously estimated. (Key numbers: 1859)