서브메뉴
검색
Understanding Shear Localizations in Glaciers Using High-Performance Computing and Machine Learning
Understanding Shear Localizations in Glaciers Using High-Performance Computing and Machine Learning
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202105613
- ISBN
- 9798265429414
- DDC
- 000
- 서명/저자
- Understanding Shear Localizations in Glaciers Using High-Performance Computing and Machine Learning
- 발행사항
- [Sl] : Stanford University, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 187 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-05, Section: B.
- 주기사항
- Advisor: Suckale, Jenny.
- 학위논문주기
- Thesis (Ph.D.)--Stanford University, 2025.
- 초록/해제
- 요약Accurately projecting the future evolution of ice sheets requires a deep understanding of the complex, multiscale processes that govern ice deformation. While most of the ice sheet deforms slowly through distributed flow, fast-flowing regions account for the majority of ice discharge and thus dominate mass loss. The transition between slow and fast flow is usually confined to narrow, highly dynamic zones where multiple interacting processes, including basal conditions, internal deformation, and microstructural evolution, govern the onset and localization of rapid flow.I investigate two types of localized deformation zones in glaciers: the transition from internal ice deformation to basal sliding, and shear margins. At the transition from flow to sliding, I find that basal topography can induce highly localized deformation regions, identified as internal shear bands within the ice flow, connecting topographic highs. The power-law exponent in ice rheology amplifies the feedback between shear heating and shear localization, resulting in the spontaneous formation of these internal shear bands, which can induce flow separation within the ice. I develop a regime diagram summarizing conditions under which topography of specific amplitude and wavelength leads to shear band formation for a given ice rheology.I further demonstrate that ice microstructural features, particularly grain size and ice fabric orientations, significantly amplify localization. By constructing a two-way coupled modeling framework that integrates microscale ice evolution with macroscale glacier dynamics using a Fourier Neural Operator, I illustrate how grain-scale softening and anisotropy profoundly influence flow dynamics. I also found that this feedback between micro- and macroscale physics is not monotonic. Instead, it can change dramatically in different flow regimes. When ice flows slowly, thermal effect dominates over strain induced recrystallization processes. This leads to large grain size formation and weak, slow ice anisotropy development in the basal region, which reduces the ice flow through rheology. When ice flows fast, strain induced recrystallization processes dominates over thermal effects. This leads to new grains forming and strong, fast ice anisotropy development, which accelerates the ice flow.Applying this multiscale modeling approach to the eastern shear margin of Thwaites Glacier, I show that ice within the shear margin exhibits substantial anisotropy, with weakening localized within a very narrow zone, aligning closely with radar observations. However, comparisons between model predictions and observational data reveal limitations in current fabric models and emphasize the critical need to incorporate brittle processes and the historical context of past shearing.Finally, recognizing that ice-sheet-scale models often remain computationally intensive despite parallelization efforts, I enhance computational efficiency by integrating temporal parallelization into the ice sheet model BISICLES. This advancement enables more scalable and efficient simulations, thereby facilitating improved long-term climate projections.
- 일반주제명
- Pressure distribution
- 일반주제명
- Eigenvalues
- 일반주제명
- Grain size
- 일반주제명
- Viscosity
- 일반주제명
- Deformation
- 일반주제명
- Rheology
- 일반주제명
- Shear strain
- 일반주제명
- Geomorphology
- 일반주제명
- Physics
- 기타저자
- Stanford University.
- 기본자료저록
- Dissertations Abstracts International. 87-05B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017360747
■00520260202105613
■006m o d
■007cr#unu||||||||
■020 ▼a9798265429414
■035 ▼a(MiAaPQ)AAI32316430
■035 ▼a(MiAaPQ)Stanfordrs703xb0989
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a000
■1001 ▼aLiu, Emma Weijia.
■24510▼aUnderstanding Shear Localizations in Glaciers Using High-Performance Computing and Machine Learning
■260 ▼a[Sl]▼bStanford University▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a187 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-05, Section: B.
■500 ▼aAdvisor: Suckale, Jenny.
■5021 ▼aThesis (Ph.D.)--Stanford University, 2025.
■520 ▼aAccurately projecting the future evolution of ice sheets requires a deep understanding of the complex, multiscale processes that govern ice deformation. While most of the ice sheet deforms slowly through distributed flow, fast-flowing regions account for the majority of ice discharge and thus dominate mass loss. The transition between slow and fast flow is usually confined to narrow, highly dynamic zones where multiple interacting processes, including basal conditions, internal deformation, and microstructural evolution, govern the onset and localization of rapid flow.I investigate two types of localized deformation zones in glaciers: the transition from internal ice deformation to basal sliding, and shear margins. At the transition from flow to sliding, I find that basal topography can induce highly localized deformation regions, identified as internal shear bands within the ice flow, connecting topographic highs. The power-law exponent in ice rheology amplifies the feedback between shear heating and shear localization, resulting in the spontaneous formation of these internal shear bands, which can induce flow separation within the ice. I develop a regime diagram summarizing conditions under which topography of specific amplitude and wavelength leads to shear band formation for a given ice rheology.I further demonstrate that ice microstructural features, particularly grain size and ice fabric orientations, significantly amplify localization. By constructing a two-way coupled modeling framework that integrates microscale ice evolution with macroscale glacier dynamics using a Fourier Neural Operator, I illustrate how grain-scale softening and anisotropy profoundly influence flow dynamics. I also found that this feedback between micro- and macroscale physics is not monotonic. Instead, it can change dramatically in different flow regimes. When ice flows slowly, thermal effect dominates over strain induced recrystallization processes. This leads to large grain size formation and weak, slow ice anisotropy development in the basal region, which reduces the ice flow through rheology. When ice flows fast, strain induced recrystallization processes dominates over thermal effects. This leads to new grains forming and strong, fast ice anisotropy development, which accelerates the ice flow.Applying this multiscale modeling approach to the eastern shear margin of Thwaites Glacier, I show that ice within the shear margin exhibits substantial anisotropy, with weakening localized within a very narrow zone, aligning closely with radar observations. However, comparisons between model predictions and observational data reveal limitations in current fabric models and emphasize the critical need to incorporate brittle processes and the historical context of past shearing.Finally, recognizing that ice-sheet-scale models often remain computationally intensive despite parallelization efforts, I enhance computational efficiency by integrating temporal parallelization into the ice sheet model BISICLES. This advancement enables more scalable and efficient simulations, thereby facilitating improved long-term climate projections.
■590 ▼aSchool code: 0212.
■650 4▼aPressure distribution
■650 4▼aEigenvalues
■650 4▼aGrain size
■650 4▼aViscosity
■650 4▼aDeformation
■650 4▼aRheology
■650 4▼aShear strain
■650 4▼aGeomorphology
■650 4▼aPhysics
■690 ▼a0800
■690 ▼a0484
■690 ▼a0605
■71020▼aStanford University.
■7730 ▼tDissertations Abstracts International▼g87-05B.
■790 ▼a0212
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17360747▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
