Stretching Near and Far: Unstructured Multiscale Lagrangian Sea Ice Deformation and Dynamic Regimes

Arctic sea ice is becoming thinner, younger, and more mobile, making it increasingly important to understand how ice motion and deformation are organized and evolving across scales. Because of its brittle response to external forcing, sea ice deformation is highly localized in space and time, producing a heterogeneous ice cover in which fractured and compact regions coexist. Existing deformation diagnostics capture part of this picture: Eulerian velocity-gradient metrics are instantaneous and local, whereas Lagrangian diagnostics are typically connected to a specific scale or collapse the flow into a single domain-averaged rate. Pointwise diagnostic comparisons are also sensitive to small spatial offsets between otherwise similar flow features. Here, we introduce an objective Lagrangian framework to overcome point-to-point comparison sensitivities and treat deformation as an evolving multiscale process. We develop techniques for comparing pairwise separation matrices across time periods and drift products, and for studying deformation time series at specific spatial scales. We demonstrate the framework with 45-day pan-Arctic integrations from 2000 to 2020 and 4-day integrations in the Beaufort Gyre region during spring 2017, identifying drift anomalies and regime transitions consistent with independent observations and documented events. When evaluating both neXtSIM and OSI SAF drift products, we find they broadly agree on their timing but differ in how stretching intensity is distributed across scales. Requiring only drift fields as input, the framework offers a systematic way to detect multiscale changes in ice drift and to quantitatively compare sea ice dynamics in different drift products or time periods.