Choosing between post-processing precipitation forecasts or chaining several uncertainty quantification tools in hydrological forecasting systems

This study aims to decipher the interactions of a precipitation post-processor and several other tools for uncertainty quantification implemented in a hydrometeorological forecasting chain. We make use of four hydrometeorological forecasting systems that differ by how uncertainties are estimated and propagated. They consider the following sources of uncertainty: system A, forcing, system B, forcing and initial conditions, system C, forcing and model structure, and system D, forcing, initial conditions, and model structure. For each system's configuration, we investigate the reliability and accuracy of post-processed precipitation forecasts in order to evaluate their ability to improve streamflow forecasts for up to 7 d of forecast horizon. The evaluation is carried out across 30 catchments in the province of Quebec (Canada) and over the 2011–2016 period. Results are compared using a multicriteria approach, and the analysis is performed as a function of lead time and catchment size. The results indicate that the precipitation post-processor resulted in large improvements in the quality of forecasts with regard to the raw precipitation forecasts. This was especially the case when evaluating relative bias and reliability. However, its effectiveness in terms of improving the quality of hydrological forecasts varied according to the configuration of the forecasting system, the forecast attribute, the forecast lead time, and the catchment size. The combination of the precipitation post-processor and the quantification of uncertainty from initial conditions showed the best results. When all sources of uncertainty were quantified, the contribution of the precipitation post-processor to provide better streamflow forecasts was not remarkable, and in some cases, it even deteriorated the overall performance of the hydrometeorological forecasting system. Our study provides an in-depth investigation of how improvements brought by a precipitation post-processor to the quality of the inputs to a hydrological forecasting model can be cancelled along the forecasting chain, depending on how the hydrometeorological forecasting system is configured and on how the other sources of hydrological forecasting uncertainty (initial conditions and model structure) are considered and accounted for. This has implications for the choices users might make when designing new or enhancing existing hydrometeorological ensemble forecasting systems.

Improvement of methods of hydrological forecasting using geoinformation technologies

This article discusses the possibility of improving hydrological forecasting methods based on a neural network. The hydrological series, its importance and forecasting features are considered. For hydrological forecasting using the MapInfoProfessional geoinformation system, an electronic map has been developed containing information about the rivers of Russia, as well as gauging stations on the Ob River. The electronic map is the basis for creating a module for short-term hydrological forecasting based on an artificial neural network. The features of a neural network, methods of its training and implementation are considered. The developed artificial neural network is a layer of neurons with a linear activation function and a delay line at the input. To predict the levels of hydrological series, real water levels at gauging stations of the Ob River in the Novosibirsk region will be used. The developed module and its capabilities have been tested. The study was carried out on the basis of models of hydrological series, as well as on the basis of levels of real hydrological series. Based on the study, dependence of the root-mean-square error on the number of previous values of series was revealed. The study also shows that it is possible to use a neural network for the current one-step forecasting of levels of hydrological series in conditions of insufficient information about the runoff region and its characteristics.

Choosing between post-processing precipitation forecasts or chaining several uncertainty quantification tools in hydrological forecasting systems

This study aims to decipher the interactions of a precipitation post-processor and several other tools for uncertainty quantification implemented in a hydrometeorological forecasting chain. We make use of four hydrometeorological forecasting systems that differ by how uncertainties are estimated and propagated. They consider the following sources of uncertainty: A) forcing, B) forcing and initial conditions, C) forcing and model structure, and D) forcing, initial conditions, and model structure. For each system's configuration, we investigate the reliability and accuracy of post-processed precipitation forecasts in order to evaluate their ability to improve streamflow forecasts for up to seven days of forecast horizon. The evaluation is carried out across 30 catchments in the Province of Quebec (Canada) and over the 2011–2016 period. Results are compared using a multicriteria approach, and the analysis is performed as a function of lead time and catchment size. The results indicate that the precipitation post-processor resulted in large improvements in the quality of forecasts with regard to the raw precipitation forecasts. This was especially the case when evaluating relative bias and reliability. However, its effectiveness in terms of improving the quality of hydrological forecasts varied according to the configuration of the forecasting system, the forecast lead time, and the catchment size. The combination of the precipitation post-processor and the quantification of uncertainty from initial conditions showed the best results. When all sources of uncertainty were quantified, the contribution of the precipitation post-processor to provide better streamflow forecasts was not remarkable and, in some cases, it even deteriorated the overall performance of the hydrometeorological forecasting system. Our study provides an in-depth investigation on how improvements brought by a precipitation post-processor to the quality of the inputs to a hydrological forecasting model can be cancelled along the forecasting chain, depending on how the hydrometeorological forecasting system is configured and on how the other sources of hydrological forecasting uncertainty (initial conditions and model structure) are considered and accounted for. This has implications for the choices users might make when designing new or enhancing existing hydrometeorological ensemble forecasting systems.

Advances in Hydrological Forecasting

Hydrological forecasting is of primary importance to better inform decision-making on flood management, drought mitigation, water system operations, water resources planning, and hydropower generation, among others [...]

How Well Can Machine Learning Models Perform without Hydrologists? Application of Rational Feature Selection to Improve Hydrological Forecasting

With more machine learning methods being involved in social and environmental research activities, we are addressing the role of available information for model training in model performance. We tested the abilities of several machine learning models for short-term hydrological forecasting by inferring linkages with all available predictors or only with those pre-selected by a hydrologist. The models used in this study were multivariate linear regression, the M5 model tree, multilayer perceptron (MLP) artificial neural network, and the long short-term memory (LSTM) model. We used two river catchments in contrasting runoff generation conditions to try to infer the ability of different model structures to automatically select the best predictor set from all those available in the dataset and compared models’ performance with that of a model operating on predictors prescribed by a hydrologist. Additionally, we tested how shuffling of the initial dataset improved model performance. We can conclude that in rainfall-driven catchments, the models performed generally better on a dataset prescribed by a hydrologist, while in mixed-snowmelt and baseflow-driven catchments, the automatic selection of predictors was preferable.

Application of the Hargreaves method for calculating the potential evapotranspiration in the hydrological modelling

To improve the results of the Slovenian Environment Agency’s hydrological forecasting system, especially in the river basins with lower specific runoff (Pomurje) and during high water events following a long dry period, we decided to find a good method for calculating daily values of the potential evapotranspiration (PET). We were deciding between several temperature-based methods for the daily reference evapotranspiration (ET0) values calculation. For selected meteorological stations we calculated ET0 using three different methods and then compared them to the ET0 values calculated using the much more complex Penman-Monteith method. Among the tested temperature methods the results given by the Hargreaves method fitted best to the results of the Penman-Monteith method. The reason for this may lie in the fact that the Hargreaves method, besides the mean daily air temperature as other temperature-based methods, considers the daily temperature range as well. Afterwards, considering the ground cover factor, we calculated the PET values from the ET0 values and then applied them in the hydrological modelling. The model setups for the Sava, Soča, and Mura Rivers were reanalysed twice, considering firstly the climatologic monthly PET values that were already used in the hydrological forecasting system of the Slovenian Environment Agency for many years, and, secondly the daily PET values calculated according to the Hargreaves method and using hourly air temperature 2 m above the ground, originating from the short-term weather forecasting model ALADIN or the INCA/AT meteorological system. At all selected calculation points, the model setups using daily PET values showed better performance over the model setups using climatological monthly values.

Tracking the Uncertainty Propagation Process Between Hydrological Forecasting and Reservoir Real-Time Optimal Operation

The inherent uncertainty in hydrological forecasting poses a challenge for reservoir real-time optimal operation. In this paper, a stochastic framework is proposed to track the uncertainty propagation process between hydrological forecasting and reservoir operation. The framework simulates the comprehensive uncertainty of hydrological forecasts in the form of ensemble forecasts and scenario trees. Based on the derived analytic relationship between the performance metric Nash-Sutcliffe efficiency coefficient (NSE) and forecast uncertainty probability distribution, we use three methods (two are commonly used classical methods and one is the Gaussian copula method) simultaneously to generate inflow forecast ensembles. Compared with the two classical methods, the Gaussian copula method additionally takes into account the temporal correlation of reservoir inflows. Then, the neural gas method is employed to transform the generated ensembles into a scenario tree, which is further used as an input for reservoir stochastic optimization. To improve the adaptability to uncertainties in inflow forecasts, we establish a stochastic optimization model that optimizes the expectation of objective values over all scenarios. Meanwhile, we propose a parallel differential evolution (DE) algorithm based on parallel computing techniques for solving the stochastic optimization model efficiently. Risk assessment is performed to capture the uncertainty and corresponding risk associated with the reservoir optimal decision. The proposed framework is demonstrated in a flood control reservoir system in China. Furthermore, we conduct several numerical experiments to explore the effect of forecast uncertainty level and temporal correlation on reservoir real-time optimal operation. The results indicate that the temporal correlation of inflows must be considered in inflow stochastic simulation and reservoir stochastic optimization, otherwise the operational risk is likely to be overestimated or underestimated, thus leading to operation failures. Based on the risk simulation surface, reservoir operators can examine the robustness of operational decisions and thus make more reliable final decisions.