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 5 min, 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 km×900 km and has a resolution of 1 km in space and 5 min in time. Independent verification experiments were carried out on 11 summer precipitation events from 2016 to 2017. In order to achieve a lead time of 1 h, a recursive approach was implemented by using RainNet predictions at 5 min lead times as model inputs for longer lead times. In the verification experiments, trivial Eulerian persistence and a conventional model based on optical flow served as benchmarks. The latter is available in the rainymotion library and had previously been shown to outperform DWD's operational nowcasting model for the same set of verification events. RainNet significantly outperforms the benchmark models at all lead times up to 60 min for the routine verification metrics mean absolute error (MAE) and the critical success index (CSI) at intensity thresholds of 0.125, 1, and 5 mm h⁻¹. However, rainymotion turned out to be superior in predicting the exceedance of higher intensity thresholds (here 10 and 15 mm h⁻¹). The limited ability of RainNet to predict heavy rainfall intensities is an undesirable property which we attribute to a high level of spatial smoothing introduced by the model. At a lead time of 5 min, an analysis of power spectral density confirmed a significant loss of spectral power at length scales of 16 km and below. Obviously, RainNet had learned an optimal level of smoothing to produce a nowcast at 5 min lead time. In that sense, the loss of spectral power at small scales is informative, too, as it reflects the limits of predictability as a function of spatial scale. Beyond the lead time of 5 min, however, the increasing level of smoothing is a mere artifact – an analogue to numerical diffusion – that is not a property of RainNet itself but of its recursive application. In the context of early warning, the smoothing is particularly unfavorable since pronounced features of intense precipitation tend to get lost over longer lead times. Hence, we propose several options to address this issue in prospective research, including an adjustment of the loss function for model training, model training for longer lead times, and the prediction of threshold exceedance in terms of a binary segmentation task. Furthermore, we suggest additional input data that could help to better identify situations with imminent precipitation dynamics. The model code, pretrained weights, and training data are provided in open repositories as an input for such future studies.

Quantitative precipitation nowcasting (QPN) has become an essential technique in various application contexts, such as early warning or urban sewage control. A common heuristic prediction approach is to track the motion of precipitation features from a sequence of weather radar images and then to displace the precipitation field to the imminent future (minutes to hours) based on that motion, assuming that the intensity of the features remains constant (“Lagrangian persistence”). In that context, “optical flow” has become one of the most popular tracking techniques. Yet the present landscape of computational QPN models still struggles with producing open software implementations. Focusing on this gap, we have developed and extensively benchmarked a stack of models based on different optical flow algorithms for the tracking step and a set of parsimonious extrapolation procedures based on image warping and advection. We demonstrate that these models provide skillful predictions comparable with or even superior to state-of-the-art operational software. Our software library (“rainymotion”) for precipitation nowcasting is written in the Python programming language and openly available at GitHub (https://github.com/hydrogo/rainymotion, Ayzel et al., 2019). That way, the library may serve as a tool for providing fast, free, and transparent solutions that could serve as a benchmark for further model development and hypothesis testing – a benchmark that is far more advanced than the conventional benchmark of Eulerian persistence commonly used in QPN verification experiments.

Data-driven models have been recently suggested to surrogate computationally expensive hydrodynamic models to map flood hazards. However, most studies focused on developing models for the same area or the same precipitation event. It is thus not obvious how transferable the models are in space. This study evaluates the performance of a convolutional neural network (CNN) based on the U-Net architecture and the random forest (RF) algorithm to predict flood water depth, the models' transferability in space and performance improvement using transfer learning techniques. We used three study areas in Berlin to train, validate and test the models. The results showed that (1) the RF models outperformed the CNN models for predictions within the training domain, presumable at the cost of overfitting; (2) the CNN models had significantly higher potential than the RF models to generalize beyond the training domain; and (3) the CNN models could better benefit from transfer learning technique to boost their performance outside training domains than RF models.

Extreme precipitation is a weather phenomenon with tremendous damaging potential for property and human life. As the intensity and frequency of such events is projected to increase in a warming climate, there is an urgent need to advance the existing knowledge on extreme precipitation processes, statistics and impacts across scales. To this end, a working group within the Germany-based project, ClimXtreme, has been established to carry out multidisciplinary analyses of high-impact events. In this work, we provide a comprehensive assessment of the 29 June 2017 heavy precipitation event (HPE) affecting the Berlin metropolitan region (Germany), from the meteorological, impacts and climate perspectives, including climate change attribution. Our analysis showed that this event occurred under the influence of a mid-tropospheric trough over western Europe and two shortwave surface lows over Britain and Poland (Rasmund and Rasmund II), inducing relevant low-level wind convergence along the German–Polish border. Over 11 000 convective cells were triggered, starting early morning 29 June, displacing northwards slowly under the influence of a weak tropospheric flow (10 m s⁻¹ at 500 hPa). The quasi-stationary situation led to totals up to 196 mm d⁻¹, making this event the 29 June most severe in the 1951–2021 climatology, ranked by means of a precipitation-based index. Regarding impacts, it incurred the largest insured losses in the period 2002 to 2017 (EUR 60 million) in the greater Berlin area. We provide further insights on flood attributes (inundation, depth, duration) based on a unique household-level survey data set. The major moisture source for this event was the Alpine–Slovenian region (63 % of identified sources) due to recycling of precipitation falling over that region 1 d earlier. Implementing three different generalised extreme value (GEV) models, we quantified the return periods for this case to be above 100 years for daily aggregated precipitation, and up to 100 and 10 years for 8 and 1 h aggregations, respectively. The conditional attribution demonstrated that warming since the pre-industrial era caused a small but significant increase of 4 % in total precipitation and 10 % for extreme intensities. The possibility that not just greenhouse-gas-induced warming, but also anthropogenic aerosols affected the intensity of precipitation is investigated through aerosol sensitivity experiments. Our multi-disciplinary approach allowed us to relate interconnected aspects of extreme precipitation. For instance, the link between the unique meteorological conditions of this case and its very large return periods, or the extent to which it is attributable to already-observed anthropogenic climate change.

Gridded datasets provide spatially and temporally consistent runoff estimates that serve as reliable sources for assessing water resources from regional to global scales. This study presents LSTM-REG, a regional gridded runoff dataset for northwest Russia based on Long Short-Term Memory (LSTM) networks. LSTM-REG covers the period from 1980 to 2016 at a 0.5${}^{°}$ spatial and daily temporal resolution. LSTM-REG has been extensively validated and benchmarked against GR4J-REG, a gridded runoff dataset based on a parsimonious regionalization scheme and the GR4J hydrological model. While both datasets provide runoff estimates with reliable prediction efficiency, LSTM-REG outperforms GR4J-REG for most basins in the independent evaluation set. Thus, the results demonstrate a higher generalization capacity of LSTM-REG than GR4J-REG, which can be attributed to the higher efficiency of the proposed LSTM-based regionalization scheme. The developed datasets are freely available in open repositories to foster further regional hydrology research in northwest Russia.