This study proposes a novel method, which is a U-shaped convolutional neural network that combines non-local attention mechanisms, Res2net residual modules, and terrain information (UNR-Net). The original U-Net method and the linear regression (LR) method are conducted as benchmarks. Generally, the UNR-Net has demonstrated promise in performing a 10× downscaling for daily 2-m temperature over North China with lead times of 1–7 days and shows superiority to the U-Net and LR methods. To be specific, U-Net and UNR-Net demonstrate higher Nash-Sutcliffe Efficiency coefficient values compared to LR by 0.052 and 0.077, respectively. The corresponding improvements in pattern correlation coefficient are 0.013 and 0.016, while the root mean square error values are higher by 0.22 and 0.338, respectively. Additionally, the structural similarity index metric is higher by 0.033 and lower by 0.015. Furthermore, regions with significant errors are primarily distributed in complex terrain areas such as the Taihang Mountains, where UNR-Net exhibits noticeable improvements. In addition, the 12 components-based error decomposition method is proposed to analyze the error source of different models. Generally, the smallest errors are observed during the summer season and the sequence error component is proven to be the main source error of 2-m temperature forecasts. Furthermore, UNR-Net consistently demonstrates the lowest errors among all 12 error components. Therefore, combining the numerical weather prediction model and deep learning method is very promising in downscaling temperature forecasts and can be applied to routine forecasting of other atmospheric variables in the future.

中文翻译：

---

深度学习每日 2 米温度降尺度

本研究提出了一种新颖的方法，即结合非局部注意力机制、Res2net残差模块和地形信息（UNR-Net）的U形卷积神经网络。以原始U-Net方法和线性回归（LR）方法作为基准。总体而言，UNR-Net 在对华北地区每日 2 米温度进行 10 倍降尺度方面表现出了良好的前景，交付时间为 1-7 天，并且显示出优于 U-Net 和 LR 方法的优越性。具体来说，与 LR 相比，U-Net 和 UNR-Net 的 Nash-Sutcliffe 效率系数值分别高出 0.052 和 0.077。模式相关系数的相应改进为 0.013 和 0.016，而均方根误差值分别提高了 0.22 和 0.338。此外，结构相似性指数指标较高 0.033，较低 0.015。此外，误差较大的区域主要分布在太行山等复杂地形地区，UNR-Net在这些区域表现出明显的改善。此外，提出了基于12分量的误差分解方法来分析不同模型的误差源。一般来说，夏季观测到的误差最小，序列误差分量被证明是2米气温预报的主要误差源。此外，UNR-Net 始终表现出所有 12 个错误分量中最低的错误。因此，将数值天气预报模型与深度学习方法相结合，在降尺度气温预报方面非常有前景，并可应用于未来其他大气变量的常规预报。