Управление информационными процессами в системах моделирования природных объектов

[Departement_IRSTEA]Eaux [TR1_IRSTEA]ARCEAU

Flood forecasting is becoming increasingly important across the world. The exposure of people and property to flooding is increasing and society is demanding improved management of flood risk. At the same time, technological and data advances are enabling improvements in forecasting capabilities. One area where flood forecasting is seeing technical developments is in the use of probabilistic forecasts—these provide a range of possible forecast outcomes that indicate the probability or chance of a flood occurring. While probabilistic forecasts have some distinct benefits, they pose an additional decision-making challenge to those that use them: with a range of forecasts to pick from, which one is right? (or rather, which one(s) can enable me to make the correct decision?). This paper describes an innovative and transferable approach for aiding decision-making with probabilistic forecasts. The proposed risk-based decision-support framework has been tested in a range of flood risk environments: from coastal surge to fluvial catchments to urban storm water scales. The outputs have been designed to be practical and proportionate to the level of flood risk at any location and to be easy to apply in an operational flood forecasting and warning context. The benefits of employing a benefit-cost inspired decision-support framework are that flood forecasting decision-making can be undertaken objectively, with confidence and an understanding of uncertainty, and can save unnecessary effort on flood incident actions. The method described is flexible such that it can be used for a wide range of flood environments with multiple flood incident management actions. It uses a risk-based approach taking into account both the probability and the level of impact of a flood event. A key feature of the framework is that it is based on a full assessment of the flood-related risk, taking into account both the probability and the level of impact of a flood event. A recommendation for action may be triggered by either a higher probability of a lower impact flood or a low probability of a very severe flood. Hence, it is highly innovative as it is the first application of such a risk-based method for flood forecasting and warning purposes. A final benefit is that it is considered to be transferrable to other countries.

It is a recognised problem that the structure and parameters of hydrological models are often not transferable in time. This may reduce model performance, with negative implications for applications in water resources modelling or flood forecasting. We hypothesise that integration of time-varying model parameters has the potential to better account for different active meteorological or hydrological processes that lead to different flood event types. To enhance flood forecasting the parameter sets should therefore be weighted dynamically according to the event type. In addition, ensemble approaches can be employed to reduce parameter uncertainty and to increase simulations accuracy for both continuous and event-based flood forecast models. Here, we introduce a machine learning based flood-type specific dynamic parameter weighting for ensemble flood forecasting. The potential of this method is demonstrated for an example case study application, employing five different machine learning approaches. Four types of flood events were considered, heavy-rainfall floods, long-duration-rainfall floods, sequence-of-rain floods and minor events with no clear generating processes. The results show that applying flood-type-specific parameter sets outperforms the other methods. Moreover, the performance consistently matched or exceeded the benchmark models. This improved performance is particularly relevant for practical applications in operational flood forecasting.

Flood forecasting has proved to be an effective and helpful non-structure method which can effectively reduce the impact of floods by providing early warnings ahead of time. Among many flood forecasting models, Rainfall-Runoff Correlation Diagram (RRCD) gives more accurate prediction by making use of large amount of history data and acting as experience expect, and thus is widely applied in operational flood forecasting. This paper proposed a new method to draw the RRCD through an equation instead of subjectivity. The method is then applied to Zhang River Reservoir with 126 flood events from historical meteorological and hydrological data from 1971 to 2000. Results show that this method gives performance as good as experience in more objective aspect and can be easily applied to consider the influence of rainfall hours cast on RRCD.

Operational flood forecasting and risk management requires high-resolution, spatially and temporally explicit predictions of inundation dynamics alongside robust uncertainty quantification. In practice, forecast skill is strongly constrained by uncertainties in meteorological forcing, boundary conditions, and human controls (e.g., reservoir releases) as well as inherent model uncertainties. While physics-based 2D hydrodynamic models provide physically consistent inundation dynamics, their computational cost make it impractical to generate several of ensemble members needed for uncertainty quantification and the real-time exploration of “what-if” intervention scenarios.We present DeepWaive, a physics-informed Foundation Model, that translates precipitation- or discharge-driven boundary conditions and static geospatial input into transient, spatially explicit 2D inundation dynamics within seconds. By leveraging deep-learning architectures trained on synthetic 2D hydrodynamic simulations, DeepWaive achieves zero-shot transferability to previously unseen basins without the need for domain-specific re-training. Crucially, the model architecture maintains the flexibility for optional site-specific fine-tuning, allowing for further optimization using either regional hydrodynamic models or in-situ sensor data to meet localized precision requirements. Benchmarking against classical numerical solvers demonstrates high predictive fidelity, with R² values ranging from 0.85 to 0.97, achieved alongside acceleration factors of 105–106. The model maintains scalability for domains up to 40,000 km2 and event durations exceeding 24 hours.Building on this capability, we develop an ensemble-to-probability workflow that propagates meteorological and hydrological forecast ensembles, and alternative reservoir release scenarios, through DeepWaive to generate probabilistic inundation products (e.g., spatial exceedance probabilities for depth and velocity thresholds) and impact-relevant summary metrics.Within the Indo-German FLAIR project (Flood Forecasting using AI for Regional Sustainability, funded by BMBF), DeepWaive provides the fast dynamic core required to (i) quantify and communicate forecast uncertainty, (ii) support rapid sensitivity analyses of key uncertainty sources, and (iii) enable tight coupling to consortium modules on EO-derived flood variables, data assimilation, and reservoir operation optimization.

The Red River of the North basin (RRB) is vulnerable to spring snowmelt flooding because of its flat terrain, low permeability soils, and the presence of river ice jams resulting from the river's northward flow direction. The onset and magnitude of major flood events in the RRB have been very difficult to forecast, in part due to limited field observations of snow water equivalent (SWE). Coarse-resolution (25-km) passive microwave observations from satellite instruments are well suited for the monitoring of SWE. Despite routine use in the Earth sciences community to document the impacts of climate change, the use of passive microwave observations in operational flood forecasting is rare. This paper compares daily satellite passive microwave SWE observations from special sensor microwave/imager (SSM/I) and special sensor microwave imager/sounder (SSMIS), advanced microwave scanning radiometer for earth observing system (AMSR-E), and advanced microwave scanning radiometer 2 (AMSR2) from 2003 to 2016 to modeled output from the SNOw Data Assimilation System (SNODAS) and Global Snow Monitoring for Climate Research -2 (GlobSnow-2) in the RRB to determine the differences between the remotely sensed SWE estimates and the model products currently used in flood forecasting. Results show statistically significant differences between the satellite observations and SNODAS in the northern region of the basin that were not evident in the southern region. Satellite estimates of peak SWE in the forecast subbasins in the northern region were 40–125% higher than the model results which points to the lack of ground observations used to constrain the model simulations. This paper recommends that satellite SWE observations should be considered for improving operational snowmelt forecasting in the RRB.

The importance of the development of systems for on-going (short-range) prediction of high water and spring floods on rivers and their systems is substantiated. In this respect, the need arises for implementation of measures for flood protection and reduction in losses sustained from these events. Another critical problem is the prediction of inflow to reservoirs and control of the passage of water through hydroprojects in periods of above-normal precipitation. Adescription is given for the general structure of forecasting systems. Moreover, studies conducted by a large group of countries within the European Community, which were performed successively in two programs 1999 – 2003, 2003 to the present) have been briefly reported on. Studies conducted throughout the second program are directed toward development of a single European forecasting system. This paper makes note of the fact that in our country there is no single comprehensive program of studies for development of this type of system.

Flood disasters occur frequently in China, resulting in property damage and fatalities. To reduce these risks, a high-resolution flood forecasting and monitoring system was developed based on a simple, distributed hydrological model: CREST. There are mainly four parts in this flood forecasting system, which are the data pre-processing and input module, the distributed hydrological model, the results output and post-processing module, as well as the results verification module. The data inputted in this system are high-resolution CMORPH satellite remote sensing precipitation and surface condition data, including ASTER DEM, HWSD soil types and MODIS vegetation and land cover types, or so. Those data were interpolated to 1 km spatial resolution and then used to drive the hydrological model to run. The CREST model was developed to provide rapid online prediction of large area which may be continental scale or even global scale, while it is also applicable at small scales, such as small basins. It can simulate the spatiotemporal variation of water and energy fluxes as long as storages on a regular grid with the grid cell resolution being user-defined, which ranges from meters to kilometers. The results outputted from the forecasting system are hydrological variables like soil moisture, potential and actual evapotranspiration, surface and river runoffs, etc. Results were then applied to GIS system and then translated into flood disaster warning information in the data output and post-processing module of the forecasting system. Both hydrological variables and flood disaster warning information will be tested and verified in the results verification module. With the 1 km horizontal spatial resolution, the forecasting system successfully simulated basic hydrological variables and processes such as actual evapotranspiration, soil moisture, and surface runoff according to the analysis and comparison with observation data. System hindcasted flood disaster information of several flood events were carefully analyzed. The results showed that the hydrographs of several stations hindcasted by the forecasting system demonstrated that the forecasting system can approximately hindcast river runoffs and streamflows. The performance of the forecasting system was tested by analyzing several real cases of flooding in different places in China, and results indicated that the system hindcasted flood information, including flood timing and spatial extent, were useful for disaster preventions and mitigations.

<scp>RMetS M</scp>eteorological <scp>M</scp>asterclass <scp>S</scp>eries: Anticipating floods, droughts and heatwaves

Ensemble predictions and perceptions of risk, uncertainty, and error in flood forecasting

System for hydrological forecasting and alert running in an operational way are important tools for floods impacts reduction. The present study describes the development and results evaluation of an operational discharge forecasting system of the upper Uruguay River basin, sited in Southern Brazil. Developed system was operated every day to provide experimental forecasts with special interest for Barra Grande and Campos Novos hydroelectric power plants reservoirs inflow, with 10 days in advance. We present results of inflow forecasted for floods occurred between July 2013 to July 2016, the period which the system was operated. Forecasts results by visual and performance metrics analysis showed a good fit with observations in most cases, with possibility of floods occurrence being well predicted with antecedence of 2 to 3 days. Comparing the locations, it was noted that the sub-basin of Campos Novos, being slower in rainfall-runoff transformation, is easier forecasted. The difference in predictability between the two basins can be observed by the coefficient of persistence, which is positive from 12h in Barra Grande and from 24h to Campos Novos. These coefficient values also show the value of the rainfall-runoff modeling for forecast horizons of more than one day in the basins.

A flood forecasting system commonly consists of at least two essential components, that is, a numerical weather prediction (NWP) model to provide rainfall forecasts and a hydrological/hydraulic model to predict the hydrological response. While being widely used for flood forecasting, hydrological models only provide a simplified representation of the physical processes of flooding due to negligence of strict momentum conservation. They cannot reliably predict the highly transient flooding process from intense rainfall, in which case a fully 2‐D hydrodynamic model is required. Due to high computational demand, hydrodynamic models have not been exploited to support real‐time flood forecasting across a large catchment at sufficiently high resolution. To fill the current research and practical gaps, this work develops a new forecasting system by coupling a graphics processing unit (GPU) accelerated hydrodynamic model with NWP products to provide high‐resolution, catchment‐scale forecasting of rainfall‐runoff and flooding processes induced by intense rainfall. The performance of this new forecasting system is tested and confirmed by applying it to “forecast” an extreme flood event across a 2,500‐km2 catchment at 10‐m resolution. Quantitative comparisons are made between the numerical predictions and field measurements in terms of water level and flood extent. To produce simulation results comparing well with the observations, the new flood forecasting system provides 34 hr of lead time when the weather forecasts are available 36 hr beforehand. Numerical experiments further confirm that uncertainties from the rainfall inputs are not amplified by the hydrodynamic model toward the final flood forecasting outputs in this case.

Comparison of the Hydro-Climatological Characteristics for the Extra-Ordinary Flood Induced by Tropical Cyclone in the Selected River Basins

In recent years, several regions in Germany experienced devastating floods caused by heavy precipitation events, often associated with severe convective storms. To improve the prediction of such events and corresponding warning strategies, Deutscher Wetterdienst (DWD) is intensifying its collaboration with Germany’s flood forecasting authorities. DWD has significantly enhanced and diversified its forecasting strategies through a range of new model systems. With a focus on seamless and probabilistic prediction in combination with more frequent initializations, DWD’s novel Seamless Integrated Forecasting System (SINFONY) marks a major step forward in predicting severe summertime convective events and heavy precipitation within a lead time of minutes to approximately 12 hours. Additional advancements include the development of 500 m high-resolution NWP with ICON as well as the establishment of an AI center and the integration of AI-based forecasting methods– all together paving the way for next-generation weather forecast systems. Beyond the improvement of forecasting techniques, DWD is placing strong emphasis on collaboration and communication with regional flood forecasting centers, including coordinated outreach to local disaster management authorities. To that end, the joint “Co-Design Project” was launched in 2023 together with regional German Flood Forecasting Centers, aiming to strengthen the hydrometeorological value chain. It is part of DWD’s new binational research initiative “Italia–Deutschland science-4-services network in weather and climate (IDEA-S4S)”. The project focuses on identifying user needs, creating a shared knowledge base for forecast evaluation, developing tailored forecast and warning tools, and supporting decision-making in the face of diverse and uncertain weather predictions. Its four main activities include: a user-oriented evaluation of DWD’s precipitation forecasts the establishment of standardized hydrological verification and thereupon analysis of (new) DWD forecasts within operational flood forecasting models the tailoring of DWD’s new warning system to meet requirements of flood forecasting centers a serious game and E-Learning initiative to improve the communication along the entire warning chain of rainfall and flood forecasts for better decision-making. This contribution provides an overview of the “Co-Design Project”. We will share first results and progress, and look forward to exchanging experiences with other initiatives at the operational intersection of meteorology and hydrology.

Cyber-Physical Systems (CPS) are being developed to provide useful interactions between physical systems and environments and cyber world for a variety of applications. CPS are designed with a set of software and interconnected distributed hardware components that are linked with physical elements to provide advanced monitoring and control mechanisms geared towards enhancing the targeted physical system or environment. These components function seamlessly to offer specific functionalities that help enhance human lives, physical system operations and environments. While CPS can offer many smart enhancements for improving physical processes, the development of such complex systems composed of many distributed and heterogeneous components is extremely difficult. This is due to the many communication, computing, and networking challenges. Using an appropriate middleware that provides a framework to support developing and operating diverse CPS applications is a novel method to address these challenges. The availability of advanced middleware services and platforms can provide effective approaches for enhancing CPS application development processes as well as provide more robust environments for operating CPS applications. Such middleware can significantly reduce the time needed to design, build, test, and operate robust CPS applications. However, designing a common middleware platform for diverse types of CPS applications is not trivial. This paper investigates the middleware challenges for CPS, based on the different types of CPS applications being developed and their specific challenges. In addition, the paper discusses the current efforts of developing middleware platforms for CPS and the open research issues in the field.