发表机构
Nanjing University of Information Science and Technology; Nanyang Technological University; Harbin Institute of Technology; Princeton University; Karolinska Institutet（南京信息工程大学; 南洋理工大学; 哈尔滨工业大学; 普林斯顿大学; 卡罗林斯卡学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对现有HEALPix网格天气预报方法在映射重建误差和跨面通信问题，提出球面螺旋扫描网络S$^3$N，通过L2Proj双向映射和AQSS四螺旋状态空间扫描实现连续信息传播，在4、7、10天提前期取得更优结果并减缓长期误差增长。
AI 中文摘要
基于机器学习的天气预报（MLWP）在全球天气预报中已取得强劲性能。近期基于分层等面积等纬度像素化（HEALPix）的方法采用HEALPix（HP）网格，以避免传统经纬度（LL）网格在极点附近的面积畸变。然而，现有的基于HP的方法通常使用逐点映射方法，并在独立的基面或局部窗口内处理HP像素。因此，映射可能引入重建误差，且跨面通信依赖于手工设计的边界处理或移位窗口。我们提出球面螺旋扫描网络（S$^3$N）以解决这两个局限性。首先，L2Proj通过其连续有限元表示的$L^2$投影，提供了在LL和HP网格之间映射大气场的双向方法。其次，注意力引导的四螺旋状态空间扫描（AQSS）块使用跨纬度注意力，沿四条全局极点到极点的螺旋路径引导选择性状态空间更新。该设计使得信息能够在HP基面边界上连续传播，而无需额外的边界处理机制。实验表明，S$^3$N在4天、7天和10天的提前期上取得了更好的结果，并且在长期预报中表现出更慢的误差增长。
英文摘要
Machine learning-based weather prediction (MLWP) has achieved strong performance in global weather forecasting. Recent Hierarchical Equal Area isoLatitude Pixelation (HEALPix)-based methods use the HEALPix (HP) grid to avoid area distortion near the poles of conventional latitude-longitude (LL) grids. However, existing HP-based approaches often use pointwise mapping methods and process HP pixels within separate base faces or local windows. Consequently, the mapping may introduce reconstruction errors and cross-face communication depends on handcrafted boundary handling or shifted windows. We propose the Spherical Spiral Scanning Network (S$^3$N) to address both limitations. First, L2Proj provides a bidirectional method for mapping atmospheric fields between the LL and HP grids through an $L^2$ projection of their continuous finite-element representations. Second, the Attention-Guided Quad-Spiral State-Space Scanning (AQSS) block uses cross-latitude attention to guide selective state-space updates along four global pole-to-pole spiral paths. This design enables continuous information propagation across HP base-face boundaries without additional boundary-processing mechanisms. Experiments show that S$^3$N achieves better results at 4-, 7-, and 10-day lead times, and exhibits slower error growth in long-range forecasting.
