基于卫星图像与大气场的潜式整流流热带气旋预报
Tropical Cyclone Forecasting via Latent Rectified Flow using Satellite Imagery and Atmospheric Fields
- Khulna University of Engineering & Technology（库尔纳工程技术大学）
机构由 AI 辅助整理，请以论文原文为准。
中文总结 AI 辅助
本研究提出单步潜式整流流模型，联合预报热带气旋的卫星图像与大气场，采样速度快、预报精度高，路径误差较基线降低，为热带气旋预报提供高效方案。
中文摘要 AI 辅助
在气候变化背景下，热带气旋的破坏性不断增强，高效预报其结构与路径已成为必要需求。深度生成模型有望替代计算成本高昂的数值天气预报（NWP），但现有系统仅能生成卫星图像或大气场中的一类，无法同时生成两类；且需要大量采样步骤，普通硬件难以支撑；此外其风暴路径来自回归头，与生成的大气无物理关联。本研究提出一种单步模型，可联合预报未来9小时的GRIDSAT-B1红外图像与4种ERA5大气场（U风、V风、气温、地表气压）。该模型采用5通道变分自编码器将每帧5×256×256的图像压缩为4×64×64的潜变量，再通过带分解时间注意力模块的条件整流流UNet，从3帧历史图像、最佳路径坐标及时间戳预测未来3帧；随后通过奖励微调（DRaFT），基于由预测风场经引导流计算得到的可微路径误差进行优化。在2022年留存风暴测试中，该模型的峰值信噪比（PSNR）达16.35 dB，结构相似性（SSIM）达0.759，在所有提前预报时效均优于复现的级联扩散基线（提前9小时时效时高0.84 dB），且采样速度快约30倍（56毫秒 vs 1673毫秒）；提前9小时的路径误差为62.4公里，比基线低15%，奖励微调研究显示，在各类采样预算下路径误差可进一步降低8%-11%。
英文摘要
Tropical cyclones are growing more destructive in a changing climate, and efficient forecasting of their structure and track has become a necessity. Deep generative models promise an alternative to computationally expensive numerical weather prediction (NWP), yet current systems produce either satellite imagery or atmospheric fields, never both; they need many sampling steps, putting them out of reach of modest hardware; and their storm tracks come from regression heads with no physical link to the generated atmosphere. This work presents a single-pass model that jointly forecasts GRIDSAT-B1 infrared imagery and four ERA5 atmospheric fields (U-wind, V-wind, air temperature, and surface pressure) out to nine hours. A five-channel variational autoencoder compresses each 5 x 256 x 256 frame to a 4 x 64 x 64 latent, and a conditional rectified-flow UNet with a factorized temporal-attention module predicts the next three frames from three past frames, their best-track coordinates, and timestamps. The model is then reward-fine-tuned (DRaFT) against a differentiable track error derived from the predicted winds through a steering-flow calculation. On held-out 2022 storms the model reaches 16.35 dB PSNR and 0.759 SSIM, ahead of a reproduced cascaded-diffusion baseline at every lead time (+0.84 dB at +9 h) while sampling ~30x faster (56 ms vs. 1673 ms). Track error at +9 h is 62.4 km, 15% below the baseline, and a reward fine-tuning study demonstrates a further 8-11% track-error reduction across sampler budgets.
