RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting

Abstract

In this study, we present RainNet, a deep convolutional neural network for radar-based precipitation nowcasting.Its design was inspired by the U-Net and SegNet families of deep learning models which were originally designed for binary segmentation tasks. RainNet was trained to predict continuous precipitation intensities at a lead time of five minutes, using several years of quality-controlled weather radar composites provided by the German Weather Service (DWD). That data set covers Germany with a spatial domain of 900×900 km, and has a resolution of 1 km in space and 5 minutes in time. 5There are many nowcasting systems which work operationally all around the world to provide precipitation nowcasts (Reyniers, 2008; Wilson et al., 1998). These systems, in their core, utilize a two-step procedure that was originally suggested by Austin and Bellon (1974), consisting of tracking and extrapolation. In the tracking step, a velocity is obtained from a series of consecutive radar images. In the extrapolation step, that velocity is used to propagate the most recent precipitation observation into the future. Various flavors and variations of this fundamental idea have been developed and operationalized over the past 40 decades, and provide value to users of corresponding products. Still, the fundamental approach to nowcasting has not changed much over the recent years -a situation that might change with the increasing popularity of deep learning in various scientific disciplines.Deep learning refers to machine-learning methods for artificial neural networks with "deep" architectures. Rather than relying on engineered features, deep learning derives low-level image features on the lowest layers of a hierarchical network, 45 and increasingly abstract features on the high-level network layers, as part of the solution of an optimization problem based on training data . Deep learning took its rise from the field of computer science when it started to dramatically outperform reference methods in image classification (Krizhevsky et al., 2012), machine translation (Sutskever et al., 2014), followed by speech recognition . Three main reasons caused this substantial breakthrough in predictive efficacy: the availability of "big data" for model training, the development of activation functions and network architectures 50 that result in numerically stable gradients across many network layers (Dahl et al., 2013), and the ability to scale the learning process massively by parallelization on graphics processing units (GPUs). Today, deep learning is rapidly spreading in many data-rich scientific disciplines, and complements researchers' toolboxes with efficient predictive models, including the field of geosciences (Reichstein et al., 2019).But while expectations in atmospheric sciences are high (see e.g., Dueben and Bauer, 2018;Gentine et al., 2018), the 55 investigation of deep learning in radar-based precipitation nowcasting is still in its infancy, and universal solutions are not yet available. Shi et al. (2015) were the first to ...

Quantifying the Location Error of Precipitation Nowcasts

“…While similar conclusions have already been drawn by using spatially sensitive verification measures such as the Fractions Skill Score (see, e.g., [6]), our framework allows us to isolate the location error for specific models and situations, to better understand the factors that govern these errors, and hence to use that knowledge in order to specifically improve the extrapolation of motion patterns in existing nowcasting models. As an example, we have demonstrated how the use of the sinuosity index can help us to better understand the predictive skill and hence the uncertainty of our models in specific situations.…”

“…Was it our prediction of the future location of a precipitation feature, or was it how precipitation intensity changed over time? Some verification scores, such as the Fractions Skill Score [6], apply a metric over spatial windows of increasing size in order to examine how the forecast performance depends on the spatial scale. Yet, we still lack the ability to explicitly isolate and quantify the location error.…”

A Precipitation Nowcasting Mechanism for Real-World Data Based on Machine Learning

“…In addition to the RNN-based method, Ayzel et al presented the RainNet [5] inspired by U-Net [6] to predict shortterm precipitation. e model followed an encoder-decoder architecture in which the encoder downscaled the spatial resolution and the decoder upscaled the learned patterns to a higher spatial resolution.…”

“…In addition to RNN-based models, Ayzel et al proposed RainNet [5] inspired by U-Net [6], a deep convolutional neural network for radar-based precipitation nowcasting.…”

SOPNet Method for the Fine-Grained Measurement and Prediction of Precipitation Intensity Using Outdoor Surveillance Cameras

“…Moreover, satellite images are commonly used, and various CNNs have been applied to satellite images to identify and forecast precipitation intensity. For example, variants of popular CNNs, such as ResNet [6], U-Net [7], and GAN [8], have been used to analyze satellite and infrared images. Although standard CNNs can employ a 2D convolutional operator to extract and generate features of a series of images, standard CNNs cannot identify time-series relationships between images.…”