[论文解读] Deep Learning for Day Forecasts from Sparse Observations

Marcin Andrychowicz, Lasse Espeholt+6 more|arXiv (Cornell University)|Jun 6, 2023

Meteorological Phenomena and Simulations被引用 40

一句话总结

MetNet-3 是一个基于观测的神经天气模型，通过学习密集与稀疏观测并将输出密化，在高空间（1 km）与高时间分辨率（2 分钟）预测降水、风、温度和露点，预测时长可达 24 小时前。

ABSTRACT

Deep neural networks offer an alternative paradigm for modeling weather conditions. The ability of neural models to make a prediction in less than a second once the data is available and to do so with very high temporal and spatial resolution, and the ability to learn directly from atmospheric observations, are just some of these models' unique advantages. Neural models trained using atmospheric observations, the highest fidelity and lowest latency data, have to date achieved good performance only up to twelve hours of lead time when compared with state-of-the-art probabilistic Numerical Weather Prediction models and only for the sole variable of precipitation. In this paper, we present MetNet-3 that extends significantly both the lead time range and the variables that an observation based neural model can predict well. MetNet-3 learns from both dense and sparse data sensors and makes predictions up to 24 hours ahead for precipitation, wind, temperature and dew point. MetNet-3 introduces a key densification technique that implicitly captures data assimilation and produces spatially dense forecasts in spite of the network training on extremely sparse targets. MetNet-3 has a high temporal and spatial resolution of, respectively, up to 2 minutes and 1 km as well as a low operational latency. We find that MetNet-3 is able to outperform the best single- and multi-member NWPs such as HRRR and ENS over the CONUS region for up to 24 hours ahead setting a new performance milestone for observation based neural models. MetNet-3 is operational and its forecasts are served in Google Search in conjunction with other models.

研究动机与目标

• 证明一个基于观测的神经网络可以在高分辨率下预测多种天气变量达到 24 小时前走
• 开发一种密化方法，将稀疏观测转换为密集预测，而不依赖于域内的密集 Ground Truth
• 在 CONUS 范围内将 MetNet-3 与最先进的数值天气预报基线（ENS、HREF、HRRR、HRES）进行比较
• 展示模型在延迟低、并能在高空间和时间分辨率下利用稀疏数据源实现扩展性

提出的方法

• 使用具备地形嵌入的三部分式 MetNet-3 架构（顶层嵌入、U-Net 主干、MaxViT Transformer）以捕捉局部与长程相互作用
• 在训练时引入对输入观测站的掩蔽，并使用空间参数共享输出密集的预测
• 通过对降水和地表变量的 Softmax 输出预测概率分布，从而实现 CRPS 和 CSI 评估，同时对同化状态通道使用 MSE
• 将前瞻时间编码为一 Hot 编码，并对输入和隐状态进行加法和乘法应用以对前瞻时间进行条件化
• 以交叉熵损失对降水和地表变量进行训练，并对同化状态使用 MSE，采用梯度重缩放以平衡多输出损失
• 在 16 个 TPU 核上进行模型并行以处理大尺度时空输入

Figure 1 : Abstract depiction of densification aspects. (a) During training a fraction of the weather stations are masked out from the input, while kept in the target. (b) To evaluate generalization to untrained locations, a set of weather stations represented by squares is never trained on and only

实验结果

研究问题

• RQ1 一个基于观测的神经网络是否能够在高分辨率下使用稀疏观测预测多种大气变量至 24 小时前吗？
• RQ2 学习的密化过程是否能从稀疏观测的位置得到密集场预测？
• RQ3 MetNet-3 相对于多成员集合（ENS、HREF 等）以及确定性 NWP 基线在前瞻时间和变量上的表现如何？
• RQ4 前瞻时间条件化与训练策略对长时程预测精度有何影响？

主要发现

• MetNet-3 在 CRPS 指标上优于多成员基线 ENS 与 HREF，直至大约 19 小时的瞬时降水和直至 18 小时的中等降水，当以 CSI 阈值评估时。
• MetNet-3 将对前瞻时间的优势从此前观测基模型（MetNet-2）的 12 小时扩展到 24 小时，覆盖多种变量。
• MetNet-3 提供降水和表面变量的概率预测（完整分布），在各前瞻时间段的 CRPS 指标均优于 ENS。
• 密化使得在 1 km 的降水和 4 km 的表面变量上实现高分辨率预测，使用 CONUS 分布的 942 个观测站的稀疏观测。
• 模型实现高时间分辨率（2 分钟）且运行时延接近 1 秒/前瞻时间，在 Google Search 中提供预测。
• 训练采用对 Ground Stations 的随机掩蔽策略及特定的前瞻时间采样方案以提升长前瞻性能。

Figure 2 : Case study for Sat Apr 23 2022 12:00 UTC featuring the Rocky Mountains of Colorado showing the mean of the ENS and MetNet-3 6 hour wind speed forecasts (top, left and center) along with the OMO stations ground truth (top, right) and the error of ENS and MetNet-3 on the individual weather