This page has only limited features, please log in for full access.
We systematically explore the effect of calibration data length on the performance of a conceptual hydrological model, GR4H, in comparison to two Artificial Neural Network (ANN) architectures: Long Short-Term Memory Networks (LSTM) and Gated Recurrent Units (GRU), which have just recently been introduced to the field of hydrology. We implemented a case study for six river basins across the contiguous United States, with 25 years of meteorological and discharge data. Nine years were reserved for independent validation; two years were used as a warm-up period, one year for each of the calibration and validation periods, respectively; from the remaining 14 years, we sampled increasing amounts of data for model calibration, and found pronounced differences in model performance. While GR4H required less data to converge, LSTM and GRU caught up at a remarkable rate, considering their number of parameters. Also, LSTM and GRU exhibited the higher calibration instability in comparison to GR4H. These findings confirm the potential of modern deep-learning architectures in rainfall-runoff modelling, but also highlight the noticeable differences between them in regard to the effect of calibration data length.
Georgy Ayzel; Maik Heistermann. The effect of calibration data length on the performance of a conceptual hydrological model versus LSTM and GRU: A case study for six basins from the CAMELS dataset. Computers & Geosciences 2021, 149, 104708 .
AMA StyleGeorgy Ayzel, Maik Heistermann. The effect of calibration data length on the performance of a conceptual hydrological model versus LSTM and GRU: A case study for six basins from the CAMELS dataset. Computers & Geosciences. 2021; 149 ():104708.
Chicago/Turabian StyleGeorgy Ayzel; Maik Heistermann. 2021. "The effect of calibration data length on the performance of a conceptual hydrological model versus LSTM and GRU: A case study for six basins from the CAMELS dataset." Computers & Geosciences 149, no. : 104708.
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.
Georgy Ayzel; Liubov Kurochkina; Dmitriy Abramov; Sergei Zhuravlev. Development of a Regional Gridded Runoff Dataset Using Long Short-Term Memory (LSTM) Networks. Hydrology 2021, 8, 6 .
AMA StyleGeorgy Ayzel, Liubov Kurochkina, Dmitriy Abramov, Sergei Zhuravlev. Development of a Regional Gridded Runoff Dataset Using Long Short-Term Memory (LSTM) Networks. Hydrology. 2021; 8 (1):6.
Chicago/Turabian StyleGeorgy Ayzel; Liubov Kurochkina; Dmitriy Abramov; Sergei Zhuravlev. 2021. "Development of a Regional Gridded Runoff Dataset Using Long Short-Term Memory (LSTM) Networks." Hydrology 8, no. 1: 6.
Operational national-scale hydrological forecasting systems are widely used in many countries for flood early warning systems and water management. However, this kind of system has never been implemented in Russia. OpenForecast v2—the first national-scale operational runoff forecasting system in Russia—has been developed and deployed to fill this gap. OpenForecast v2 delivers 7 day-ahead streamflow forecasts for 843 gauges across Russia. The verification study has been carried out using 244 gauges for which operational streamflow data were openly available and quality-controlled for the entire verification period (14 March–6 July 2020). The results showed that the developed system provides reliable and skillful runoff forecasts for up to one week. The benchmark testing against climatology and persistence forecasts showed that the system provides skillful predictions for most analyzed basins. OpenForecast v2 is in operational use and is openly available on the Internet.
Georgy Ayzel. OpenForecast v2: Development and Benchmarking of the First National-Scale Operational Runoff Forecasting System in Russia. Hydrology 2021, 8, 3 .
AMA StyleGeorgy Ayzel. OpenForecast v2: Development and Benchmarking of the First National-Scale Operational Runoff Forecasting System in Russia. Hydrology. 2021; 8 (1):3.
Chicago/Turabian StyleGeorgy Ayzel. 2021. "OpenForecast v2: Development and Benchmarking of the First National-Scale Operational Runoff Forecasting System in Russia." Hydrology 8, no. 1: 3.
In precipitation nowcasting, it is common to track the motion of precipitation in a sequence of weather radar images and to extrapolate this motion into the future. The total error of such a prediction consists of an error in the predicted location of a precipitation feature and an error in the change of precipitation intensity over lead time. So far, verification measures did not allow isolating the extent of location errors, making it difficult to specifically improve nowcast models with regard to location prediction. In this paper, we introduce a framework to directly quantify the location error. To that end, we detect and track scale-invariant precipitation features (corners) in radar images. We then consider these observed tracks as the true reference in order to evaluate the performance (or, inversely, the error) of any model that aims to predict the future location of a precipitation feature. Hence, the location error of a forecast at any lead time Δt ahead of the forecast time t corresponds to the Euclidean distance between the observed and the predicted feature locations at t + Δt. Based on this framework, we carried out a benchmarking case study using one year worth of weather radar composites of the German Weather Service. We evaluated the performance of four extrapolation models, two of which are based on the linear extrapolation of corner motion from t − 1 to t (LK-Lin1) and t − 4 to t (LK-Lin4) and the other two are based on the Dense Inverse Search (DIS) method: motion vectors obtained from DIS are used to predict feature locations by linear (DIS-Lin1) and Semi-Lagrangian extrapolation (DIS-Rot1). Of those four models, DIS-Lin1 and LK-Lin4 turned out to be the most skillful with regard to the prediction of feature location, while we also found that the model skill dramatically depends on the sinuosity of the observed tracks. The dataset of 376,125 detected feature tracks in 2016 is openly available to foster the improvement of location prediction in extrapolation-based nowcasting models.
Arthur Costa Tomaz De Souza; Georgy Ayzel; Maik Heistermann. Quantifying the Location Error of Precipitation Nowcasts. Advances in Meteorology 2020, 2020, 1 -12.
AMA StyleArthur Costa Tomaz De Souza, Georgy Ayzel, Maik Heistermann. Quantifying the Location Error of Precipitation Nowcasts. Advances in Meteorology. 2020; 2020 ():1-12.
Chicago/Turabian StyleArthur Costa Tomaz De Souza; Georgy Ayzel; Maik Heistermann. 2020. "Quantifying the Location Error of Precipitation Nowcasts." Advances in Meteorology 2020, no. : 1-12.
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.
Georgy Ayzel; Tobias Scheffer; Maik Heistermann. RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting. Geoscientific Model Development 2020, 13, 2631 -2644.
AMA StyleGeorgy Ayzel, Tobias Scheffer, Maik Heistermann. RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting. Geoscientific Model Development. 2020; 13 (6):2631-2644.
Chicago/Turabian StyleGeorgy Ayzel; Tobias Scheffer; Maik Heistermann. 2020. "RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting." Geoscientific Model Development 13, no. 6: 2631-2644.
The focus of this is to reveal the value of making national data archives available for scientific research by showing the specific example from the field of regional runoff reconstruction. Thus, for northwest Russia, we developed two gridded datasets of monthly runoff reconstruction: for the first dataset (BASE), we used only the freely available data from the Global Runoff Data Centre (GRDC), while, for the second dataset (SOTA), we complemented the GRDC data with digitized runoff records from Russian national observational runoff archives (R5). The accuracy of developed datasets in terms of monthly runoff prediction was assessed using the Nash-Sutcliffe efficiency (NSE) for a wide range of river basins. The results show that accounting for R5 data for runoff reconstruction underpins a substantial gain in NSE of SOTA over the BASE dataset. Moreover, both datasets, on average, outperform 10 state-of-the-art global hydrological models and one European-scale regional hydrological model.
Georgy Ayzel; Liubov Kurochkina; Sergei Zhuravlev. The influence of regional hydrometric data incorporation on the accuracy of gridded reconstruction of monthly runoff. Hydrological Sciences Journal 2020, 1 -12.
AMA StyleGeorgy Ayzel, Liubov Kurochkina, Sergei Zhuravlev. The influence of regional hydrometric data incorporation on the accuracy of gridded reconstruction of monthly runoff. Hydrological Sciences Journal. 2020; ():1-12.
Chicago/Turabian StyleGeorgy Ayzel; Liubov Kurochkina; Sergei Zhuravlev. 2020. "The influence of regional hydrometric data incorporation on the accuracy of gridded reconstruction of monthly runoff." Hydrological Sciences Journal , no. : 1-12.
Georgy Ayzel. Responses to reviewers. 2020, 1 .
AMA StyleGeorgy Ayzel. Responses to reviewers. . 2020; ():1.
Chicago/Turabian StyleGeorgy Ayzel. 2020. "Responses to reviewers." , no. : 1.
Streamflow prediction is a vital public service that helps to establish flash-flood early warning systems or assess the impact of projected climate change on water management. However, the availability of streamflow observations limits the utilization of the state-of-the-art streamflow prediction techniques to the basins where hydrometric gauging stations exist. Since the most river basins in the world are ungauged, the development of the specialized techniques for the reliable streamflow prediction in ungauged basins (PUB) is of crucial importance. In recent years, the emerging field of deep learning provides a myriad of new models that can breathe new life into the stagnating PUB methods. In the presented study, we benchmark the streamflow prediction efficiency of Long Short-Term Memory (LSTM) networks against the standard technique of GR4J hydrological model parameters regionalization (HMREG) at 200 basins in Northwest Russia. Results show that the LSTM-based regional hydrological model significantly outperforms the HMREG scheme in terms of median Nash-Sutcliffe efficiency (NSE), which is 0.73 and 0.61 for LSTM and HMREG, respectively. Moreover, LSTM demonstrates the comparable median NSE with that for basin-scale calibration of GR4J (0.75). Therefore, this study underlines the high utilization potential of deep learning for the PUB by demonstrating the new state-of-the-art performance in this field.
Georgy Ayzel; Liubov Kurochkina; Eduard Kazakov; Sergei Zhuravlev. Streamflow prediction in ungauged basins: benchmarking the efficiency of deep learning. E3S Web of Conferences 2020, 163, 01001 .
AMA StyleGeorgy Ayzel, Liubov Kurochkina, Eduard Kazakov, Sergei Zhuravlev. Streamflow prediction in ungauged basins: benchmarking the efficiency of deep learning. E3S Web of Conferences. 2020; 163 ():01001.
Chicago/Turabian StyleGeorgy Ayzel; Liubov Kurochkina; Eduard Kazakov; Sergei Zhuravlev. 2020. "Streamflow prediction in ungauged basins: benchmarking the efficiency of deep learning." E3S Web of Conferences 163, no. : 01001.
For only two out of more than 95 * 10³ glaciers in High Mountain Asia (HMA) a continuous time series of mass balance measurements covering more than 30 years (World Glacier Monitoring Service’s ‘reference glaciers’) is available to date. Considering that both glaciers are located in the Tian Shan Range, i.e. the northernmost part of HMA, and that glacier changes in HMA is known to be heterogeneous in space and time, it is clear that a substantial knowledge gap exists regarding the actual dynamics at individual glaciers and their forcing.
Here, we present a novel data set of transient snowline altitude (TSLA) measurements covering all glaciers > 0.5 km² in HMA (n=28,501) for a time frame from the mid 1980s to late 2019 based on more than 10⁵ Landsat satellite images, allowing for investigations of the characteristics of glacier change at unprecedented spatio-temporal resolution and coverage.
Individual glacier’s total maxima of end-of-season TSLAs for the whole period of observation clearly highlight years with many (i.e. 1994, 2009, 2013, 2015) and few (i.e. 1995, 2003, 2012) maxima. Out of the glaciers that show a significant trend throughout the observation period, 90.8% have a positive trend with a median TSLA rise of 7.0 m/year. These figures increase to 95.8% and 13.8 m/year, when only observations of the last two decades are considered.
Based on ERA5 meteorological time series and fundamental physiographic glacier characteristics from the Randolph Glacier Inventory v6, we investigated drivers of the observed TSLA fluctuations. Consistent with expectations, a Random Forest analysis finds temperature to be the dominant meteorological driver of TSLA dynamics throughout all regions of HMA when whole years are considered. Conversely, meteorological forcing regimes are highly heterogeneous for different glaciers in the ablation phase, with wind, air temperature and incoming shortwave radiation being the dominant TSLA drivers for the majority of glaciers in HMA. Considering regional domains, TSLA dynamics are considerably determined by physiographic factors, such as latitude, longitude, hypsographic characteristics, slope and aspect of individual glaciers. A hierarchical clustering analysis shows distinct groups of similar forcing setups exist; Their spatial distribution, however, rather follows specific positions in the topoclimatic system than forming distinct regional clusters or aligning to large-scale gradients.
In summary, our findings indicate that spatial and temporal patterns of glacier change in HMA are considerably more complex than currently known. Multidecadal high-resolution TSLA datasets like the one presented here may inform future research to disentangle the complex topoclimatic process-response systems that control the adaptation of individual glaciers to climate change.
David Loibl; Georgy Ayzel; Fiona Clubb; Inge Grünberg; Jan Nitzbon. Dynamics and drivers of High Mountain Asia’s glacier change from the mid 1980s to late 2019. 2020, 1 .
AMA StyleDavid Loibl, Georgy Ayzel, Fiona Clubb, Inge Grünberg, Jan Nitzbon. Dynamics and drivers of High Mountain Asia’s glacier change from the mid 1980s to late 2019. . 2020; ():1.
Chicago/Turabian StyleDavid Loibl; Georgy Ayzel; Fiona Clubb; Inge Grünberg; Jan Nitzbon. 2020. "Dynamics and drivers of High Mountain Asia’s glacier change from the mid 1980s to late 2019." , no. : 1.
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 by 900 km, and has a resolution of 1 km in space and 5 minutes in time. Independent verification experiments were carried out on eleven summer precipitation events from 2016 to 2017. In order to achieve a lead time of one hour, a recursive approach was implemented by using RainNet predictions at five minutes lead time as model input 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 minutes for the routine verification metrics Mean Absolute Error (MAE) and Critical Success Index (CSI, at intensity thresholds of 0.125, 1, and 5 mm/h). Apart from its superiority in terms of MAE and CSI, an undesirable property of RainNet predictions is, however, the level of spatial smoothing. At a lead time of five minutes, 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 minutes 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 five minutes, 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 unfavourable 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 to such future studies.
Georgy Ayzel; Tobias Scheffer; Maik Heistermann. RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting. 2020, 2020, 1 -20.
AMA StyleGeorgy Ayzel, Tobias Scheffer, Maik Heistermann. RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting. . 2020; 2020 ():1-20.
Chicago/Turabian StyleGeorgy Ayzel; Tobias Scheffer; Maik Heistermann. 2020. "RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting." 2020, no. : 1-20.
Georgy Ayzel; Tobias Scheffer; Maik Heistermann. Supplementary material to "RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting". 2020, 1 .
AMA StyleGeorgy Ayzel, Tobias Scheffer, Maik Heistermann. Supplementary material to "RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting". . 2020; ():1.
Chicago/Turabian StyleGeorgy Ayzel; Tobias Scheffer; Maik Heistermann. 2020. "Supplementary material to "RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting"." , no. : 1.
During the last few decades, the rapid separation of the Small Aral Sea from the isolated basin has changed its hydrological and ecological conditions tremendously. In the present study, we developed and validated the hybrid model for the Syr Darya River basin based on a combination of state-of-the-art hydrological and machine learning models. Climate change impact on freshwater inflow into the Small Aral Sea for the projection period 2007–2099 has been quantified based on the developed hybrid model and bias corrected and downscaled meteorological projections simulated by four General Circulation Models (GCM) for each of three Representative Concentration Pathway scenarios (RCP). The developed hybrid model reliably simulates freshwater inflow for the historical period with a Nash–Sutcliffe efficiency of 0.72 and a Kling–Gupta efficiency of 0.77. Results of the climate change impact assessment showed that the freshwater inflow projections produced by different GCMs are misleading by providing contradictory results for the projection period. However, we identified that the relative runoff changes are expected to be more pronounced in the case of more aggressive RCP scenarios. The simulated projections of freshwater inflow provide a basis for further assessment of climate change impacts on hydrological and ecological conditions of the Small Aral Sea in the 21st Century.
Georgy Ayzel; Alexander Izhitskiy. Climate Change Impact Assessment on Freshwater Inflow into the Small Aral Sea. Water 2019, 11, 2377 .
AMA StyleGeorgy Ayzel, Alexander Izhitskiy. Climate Change Impact Assessment on Freshwater Inflow into the Small Aral Sea. Water. 2019; 11 (11):2377.
Chicago/Turabian StyleGeorgy Ayzel; Alexander Izhitskiy. 2019. "Climate Change Impact Assessment on Freshwater Inflow into the Small Aral Sea." Water 11, no. 11: 2377.
The development and deployment of new operational runoff forecasting systems are a strong focus of the scientific community due to the crucial importance of reliable and timely runoff predictions for early warnings of floods and flashfloods for local businesses and communities. OpenForecast, the first operational runoff forecasting system in Russia, open for public use, is presented in this study. We developed OpenForecast based only on open-source software and data—GR4J hydrological model, ERA-Interim meteorological reanalysis, and ICON deterministic short-range meteorological forecasts. Daily forecasts were generated for two basins in the European part of Russia. Simulation results showed a limited efficiency in reproducing the spring flood of 2019. Although the simulations managed to capture the timing of flood peaks, they failed in estimating flood volume. However, further implementation of the parsimonious data assimilation technique significantly alleviates simulation errors. The revealed limitations of the proposed operational runoff forecasting system provided a foundation to outline its further development and improvement.
Georgy Ayzel; Natalia Varentsova; Oxana Erina; Dmitriy Sokolov; Liubov Kurochkina; Vsevolod Moreydo. OpenForecast: The First Open-Source Operational Runoff Forecasting System in Russia. Water 2019, 11, 1546 .
AMA StyleGeorgy Ayzel, Natalia Varentsova, Oxana Erina, Dmitriy Sokolov, Liubov Kurochkina, Vsevolod Moreydo. OpenForecast: The First Open-Source Operational Runoff Forecasting System in Russia. Water. 2019; 11 (8):1546.
Chicago/Turabian StyleGeorgy Ayzel; Natalia Varentsova; Oxana Erina; Dmitriy Sokolov; Liubov Kurochkina; Vsevolod Moreydo. 2019. "OpenForecast: The First Open-Source Operational Runoff Forecasting System in Russia." Water 11, no. 8: 1546.
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.
Georgy Ayzel; Maik Heistermann; Tanja Winterrath. Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1). Geoscientific Model Development 2019, 12, 1387 -1402.
AMA StyleGeorgy Ayzel, Maik Heistermann, Tanja Winterrath. Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1). Geoscientific Model Development. 2019; 12 (4):1387-1402.
Chicago/Turabian StyleGeorgy Ayzel; Maik Heistermann; Tanja Winterrath. 2019. "Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1)." Geoscientific Model Development 12, no. 4: 1387-1402.
Today deep learning is taking its rise in hydrometeorological applications, and it is critical to extensively evaluate its prediction performance and robustness. In our study, we use deep all convolutional neural networks for radar-based precipitation nowcasting, which has a crucial role for early warning of hazardous events at small spatiotemporal scales. Our trial and error study focuses the particular importance of selecting and adopting suitable data preprocessing routine, network structure, and loss function regarding input data features, and, as a result, highlights limited transferability of methods in existing studies. Results show that parsimonious deep learning models can forecast a complex nature of a short-term precipitation field evolution and compete for the state-of-the-art performance with well-established nowcasting models based on optical flow techniques.
G. Ayzel; M. Heistermann; Aleksei Sorokin; Oleg Nikitin; Olga Lukyanova. All convolutional neural networks for radar-based precipitation nowcasting. Procedia Computer Science 2019, 150, 186 -192.
AMA StyleG. Ayzel, M. Heistermann, Aleksei Sorokin, Oleg Nikitin, Olga Lukyanova. All convolutional neural networks for radar-based precipitation nowcasting. Procedia Computer Science. 2019; 150 ():186-192.
Chicago/Turabian StyleG. Ayzel; M. Heistermann; Aleksei Sorokin; Oleg Nikitin; Olga Lukyanova. 2019. "All convolutional neural networks for radar-based precipitation nowcasting." Procedia Computer Science 150, no. : 186-192.
Georgy Ayzel. Responses to reviewers. 2019, 1 .
AMA StyleGeorgy Ayzel. Responses to reviewers. . 2019; ():1.
Chicago/Turabian StyleGeorgy Ayzel. 2019. "Responses to reviewers." , no. : 1.
Three river basins—the Lena, Ganges, and Darling—were selected to study the possibility of reproducing water balance components of river basins, located in different regions of the globe under a wide variety of natural conditions, with the use of the land surface model SWAP and global data sets. Input data including meteorological forcings and land surface parameters were prepared on the basis of the WATCH and ECOCLIMAP global data sets, respectively. Long-term variations of the water balance components of the Lena, Ganges, and Darling river basins were simulated by the SWAP model. The results of simulations were compared with observations. In addition, the natural variability of river runoff caused by the weather noise of atmospheric characteristics was estimated.
E. M. Gusev; O. N. Nasonova; E. E. Kovalev; G. V. Ayzel. Modelling Water Balance Components of River Basins Located in Different Regions of the Globe. Water Resources 2018, 45, 53 -64.
AMA StyleE. M. Gusev, O. N. Nasonova, E. E. Kovalev, G. V. Ayzel. Modelling Water Balance Components of River Basins Located in Different Regions of the Globe. Water Resources. 2018; 45 (2):53-64.
Chicago/Turabian StyleE. M. Gusev; O. N. Nasonova; E. E. Kovalev; G. V. Ayzel. 2018. "Modelling Water Balance Components of River Basins Located in Different Regions of the Globe." Water Resources 45, no. 2: 53-64.
Due to global warming, the problem of assessing water resources and their vulnerability to climate drivers in the Arctic region has become a focus in the recent years. This study is aimed at investigating three lumped hydrological models to predict daily runoff of large-scale Arctic basins in the case of substantial data scarcity. All models were driven only by meteorological forcing reanalysis dataset without any additional information about landscape, soil, or vegetation cover properties of the studied basins. Model parameter regionalization based on transferring the whole parameter set showed good efficiency for predictions in ungauged basins. We run a blind test of the proposed methodology for ensemble runoff predictions on five sub-basins, for which only monthly observations were available, and obtained promising results for current water resources assessment for a broad domain of ungauged basins in the Russian Arctic.
G. V. Ayzel. Runoff Predictions in Ungauged Arctic Basins Using Conceptual Models Forced by Reanalysis Data. Water Resources 2018, 45, 1 -7.
AMA StyleG. V. Ayzel. Runoff Predictions in Ungauged Arctic Basins Using Conceptual Models Forced by Reanalysis Data. Water Resources. 2018; 45 (2):1-7.
Chicago/Turabian StyleG. V. Ayzel. 2018. "Runoff Predictions in Ungauged Arctic Basins Using Conceptual Models Forced by Reanalysis Data." Water Resources 45, no. 2: 1-7.
Georgy Ayzel; Maik Heistermann; Tanja Winterrath. Supplementary material to "Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1)". 2018, 1 .
AMA StyleGeorgy Ayzel, Maik Heistermann, Tanja Winterrath. Supplementary material to "Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1)". . 2018; ():1.
Chicago/Turabian StyleGeorgy Ayzel; Maik Heistermann; Tanja Winterrath. 2018. "Supplementary material to "Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1)"." , no. : 1.
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 extrapolate that motion to the imminent future (minutes to hours), 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 Python programming language, and openly available at GitHub (https://github.com/hydrogo/rainymotion). 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.
Georgy Ayzel; Maik Heistermann; Tanja Winterrath. Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1). 2018, 2018, 1 -23.
AMA StyleGeorgy Ayzel, Maik Heistermann, Tanja Winterrath. Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1). . 2018; 2018 ():1-23.
Chicago/Turabian StyleGeorgy Ayzel; Maik Heistermann; Tanja Winterrath. 2018. "Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1)." 2018, no. : 1-23.