发表于《Nature》的研究（2025-12-24）指出：约9,000年前，在早全新世这一段比今天更暖的时期，东南极冰盖（德龙宁·莫德地一带）的部分浮冰架曾以“异常迅速”的速度崩解并触发冰体加速入海。研究团队在日本昭和站附近的吕措-霍尔姆湾海底获取沉积岩心，利用稀有铍同位素与微体化石等证据，把冰架破裂的时间约束在9,000年前，并将其与海洋变暖联系起来。

从量级看，南极冰量若完全融化足以使全球海平面上升约190英尺，因此关键在于机制与速率。研究强调“环极深层水”这一相对温暖、含盐的深层海流（在海面下数百英尺）曾反复涌上大陆架，侵蚀并削弱漂浮冰架的支撑作用；冰架一旦碎裂，后方陆上冰就失去“支撑”而更快流向海洋。融水还使表层海水淡化、分层更强、抑制上下混合与散热，从而让暖深水更易向岸推进，形成自我强化的正反馈循环。

作者将这一全新世模式与当代观测对照：西南极的思韦茨冰川在暖海水入侵下变薄，松岛等出口冰川在深层暖水可触及搁浅线时也表现出快速退缩。传统上被视为更稳定的东南极也被最新重建与卫星/重力资料提示存在沿岸持续失冰（如托滕、登曼等）。研究认为，若模型忽略“融水—海洋”反馈，可能低估冰架破裂与内陆冰释放的速度；即便本世纪海平面仅上升数英尺，也足以显著放大风暴潮、频繁内涝与盐水入侵等风险，而未来几十年的排放与海洋吸热将深刻影响这一轨迹。

A Nature study dated Dec 24, 2025 reports that a sector of the East Antarctic Ice Sheet in Dronning Maud Land lost its buttressing ice shelves rapidly about 9,000 years ago during the early Holocene, when regional temperatures were above today’s. Sediment cores from Lützow-Holm Bay near Syowa Station, analyzed with rare beryllium isotopes and microfossils, constrain the timing of breakup and link it to ocean warming.

Antarctica stores enough ice to raise global sea level ~190 ft if fully melted, so process rates matter. The trigger was relatively warm, salty circumpolar deep water (flowing hundreds of feet below the surface) surging onto the continental shelf, melting shelves from below and removing their “brace,” which let inland ice accelerate seaward. Meltwater then freshened the surface, increased stratification, reduced cooling/mixing, and promoted repeated pulses of warm water—a positive feedback.

Modern observations in West Antarctica show similar ocean-driven thinning at Thwaites and rapid retreat at outlets like Pine Island where deep water reaches grounding zones. Reconstructions and satellites/gravity data indicate parts of East Antarctica (e.g., Totten, Denman) are also losing ice, challenging the idea of inherent stability. The study argues models that omit meltwater–ocean feedbacks may understate the risk of faster-than-projected sea-level rise; even a few feet this century would greatly amplify coastal flooding and storm surge.