HydroFirn：面向大规模多维firn水力学的数值模型
计算机视觉与模式识别
2026-04-14 v1
摘要
观测表明，格陵兰冰盖firn中熔水和冰层分布的多维动力学。然而，当前大规模firn水力学模型本质上是一维的，限制了其解释观测数据集的能力，并导致对冰盖表面质量平衡和海平面上升估计的不确定性。本文提出一种大规模、多维、多相及热力学模型，用于firn的地下水力学。该模型由于 novel algorithm 而高度高效：仅在饱和区域求解额外的压力方程。此外，模型能够对不饱和-饱和域施加空间异构边界条件，并允许完全不透水冰层的动态形成。数值结果在一维和二维问题上均与解析解高度吻合，这些问题涉及耦合的不饱和-饱和流动、热力学和相变。我们进一步将模型用于探讨西南格陵兰的现场数据，发现横向异构性对熔水渗透深度和冰层形成具有显著影响。改进对这些局部多维过程的理解将为firn致密化提供基于物理的约束，减少将高度计量变化转化为质量变化的不确定性，并提高温暖气候下对淡水排放到海洋的估计。
引用
@article{arxiv.2604.11487,
title = {NTIRE 2026 Challenge on Robust AI-Generated Image Detection in the Wild},
author = {Aleksandr Gushchin and Khaled Abud and Ekaterina Shumitskaya and Artem Filippov and Georgii Bychkov and Sergey Lavrushkin and Mikhail Erofeev and Anastasia Antsiferova and Changsheng Chen and Shunquan Tan and Radu Timofte and Dmitry Vatolin and Chuanbiao Song and Zijian Yu and Hao Tan and Jun Lan and Zhiqiang Yang and Yongwei Tang and Zhiqiang Wu and Jia Wen Seow and Hong Vin Koay and Haodong Ren and Feng Xu and Shuai Chen and Ruiyang Xia and Qi Zhang and Yaowen Xu and Zhaofan Zou and Hao Sun and Dagong Lu and Mufeng Yao and Xinlei Xu and Fei Wu and Fengjun Guo and Cong Luo and Hardik Sharma and Aashish Negi and Prateek Shaily and Jayant Kumar and Sachin Chaudhary and Akshay Dudhane and Praful Hambarde and Amit Shukla and Zhilin Tu and Fengpeng Li and Jiamin Zhang and Jianwei Fei and Kemou Li and Haiwei Wu and Bilel Benjdira and Anas M. Ali and Wadii Boulila and Chenfan Qu and Junchi Li},
journal= {arXiv preprint arXiv:2604.11487},
year = {2026}
}
备注
CVPR 2026 NTIRE Workshop Paper, Robust AI-Generated Image Detection Technical Report