RIFT-TIP: Rates of Ice Fracture and Timing of Tabular Iceberg Production

每年从冰架上崩解的板状冰山约占南极洲融化冰总量的一半。当全层裂缝（称为裂口）在冰架上水平传播时，冰山就形成了。由此产生的冰山可达数千平方公里，可能影响野生动物、航运和海洋环流，并可能改变剩余冰架的形状和稳定性，从而对冰排放和海平面上升产生后续影响。目前，冰裂的时间是不可预测的，而且在一些冰盖模型中只被肤浅地包括在内，例如，当冰达到一定厚度时，就会把冰移走。据观察，裂谷的传播速度非常快，每天可达几公里，或者非常缓慢，停滞数年甚至数十年。虽然冰架崩塌可以导致冰川加速这一事实已经得到充分证实，但最近的观测也表明，适度的冰解事件直接并立即影响冰流和基础融化速率，这表明迫切需要限制这一过程的时间，以及它是否会在未来加速。同时，在工程应用的推动下，裂缝近似方法的发展使得离散裂缝的数值表示成为可能，前提是在小尺度和大尺度上的行为是通过观察来校准的。目前缺乏观测限制了这类模拟在冰川学中的价值。我们的研究结合了对南极洲布伦特冰架裂缝生长的直接观察和对同一块冰样本的实验室实验，当它们联系在一起时，可以在多个尺度上产生前所未有的断裂过程细节。这种水平的细节将应用于裂缝问题，使用一种新的可扩展相场模型，该模型允许在裂缝边界使用弥散界面将微尺度过程映射到低分辨率冰盖尺度网格上。我们将在实验室观察裂缝如何在晶体水平上与冰相互作用，并在现场观察裂缝如何在公里水平上与冰架相互作用，以验证和测试该模型。这将阐明裂谷生长背后的三维机制和控制其速率的物理冰性质。在冰盖模型中如何表示产犊过程的逐步改进对整个地球科学界都有好处，从需要估计冰架支撑应力和产犊对接地线动力学的影响的冰盖建模者，到依赖精确的冰架几何形状来限制淡水通量和海冰形成速率的大规模地球系统建模者。

Calving of tabular icebergs from ice shelves accounts for around half of all the ice lost from Antarctica each year. The icebergs form when full-thickness fractures (known as rifts) propagate horizontally through the ice shelf. The resulting icebergs can be thousands of square kilometres in size, can impact wildlife, shipping and ocean circulation and can modify the shape and stability of the ice shelves which remain, with a subsequent impact on ice discharge and sea level rise. The timing of calving is currently unpredictable and is only included superficially in some ice sheet models, for example by removing ice once a certain thickness is reached. Rifts have been observed to propagate very rapidly, at up to several kilometres per day, or very slowly, stagnating for years or even decades. Whilst it is well established that ice shelf collapse can lead to glacier acceleration, recent observations also show moderate calving events directly and immediately impacting ice flow and basal melt rate, indicating an urgent need to constrain the timing of this process and whether it will accelerate in the future. Simultaneously, developments in fracture approximation methods driven by engineering applications have made it possible to represent discrete fractures numerically, provided the behaviour at the small-scale and large-scale is calibrated with observations. Lack of observations currently limits the value of this type of modelling in glaciology. Our research combines direct observations of rift growth on the Brunt Ice Shelf in Antarctica with laboratory experiments on samples of the same ice, which when linked together produce unprecedented detail on the fracture process across multiple scales. This level of detail will be applied to the fracture problem using a new scalable phase-field model that allows microscale processes to be mapped onto a low-resolution ice-sheet-scale grid using diffuse interfaces at crack boundaries. We will conduct laboratory observations of how cracks interact with ice at the crystal level, and in situ observations of how rifts interact with the ice shelf at the kilometre level to validate and test this model. This will illuminate the three-dimensional mechanism behind rift growth and the physical ice properties that control its rate. A step-change improvement in how the calving process is represented in ice sheet models has benefits across the geoscience community, from ice sheet modellers who need to estimate ice shelf buttressing stress and the impact of calving on grounding line dynamics, to large scale earth-system modellers which rely on accurate ice shelf geometry to constrain freshwater fluxes and rates of sea ice formation.