Floods can cause significant problems for humans and can damage the economy. Implementing a reliable flood monitoring warning system in risk areas can help to reduce the negative impacts of these natural disasters. Artificial intelligence algorithms and statistical approaches are employed by researchers to enhance flood forecasting. In this study, a dataset was created using unique features measured by sensors along the Hunza River in Pakistan over the past 31 years. The dataset was used for classification and regression problems. Two types of machine learning algorithms were tested for classification: classical algorithms (Random Forest, RF and Support Vector Classifier, SVC) and deep learning algorithms (Multi-Layer Perceptron, MLP). For the regression problem, the result of MLP and Support Vector Regression (SVR) algorithms were compared based on their mean square, root mean square and mean absolute errors. The results obtained show that the accuracy of the RF classifier is 0.99, while the accuracies of the SVC and MLP methods are 0.98; moreover, in the case of flood prediction, the SVR algorithm outperforms the MLP approach.
The Danube and its tributaries have been crossing mountains and plains in their almost unchanged riverbeds for thousands of years, regardless of national and administrative boundaries. Nevertheless, even decades ago, several countries provided access to only limited data and information concerning the water level and flood protection status of their rivers. In recent years, information was exchanged mainly on the basis of bilateral agreements and on successful activities of basin-wide organizations, but for others, information could only be obtained by browsing the Internet, which is sometimes rather complicated and definitely time-consuming. The EU Strategy for the Danube Region Environmental Risks Priority Area initiated a project aimed at developing the Danube Hydrological Information System, which was supported by the International Commission for the Protection of the Danube River. A comprehensive overview of the complex national flood and ice forecasting systems, identification of the shortcomings of the existing forecasting practices as well as an improvement of the exchange and availability of hydrological and meteorological data between the involved countries constituted crucial fields of interests for the project. Hence the main aim of the article is to present and discuss key data and functionalities of the system. The key findings show that all authorized meteorological and hydrological data of the Danube River are stored in a central database and made available online to all licensed hydrological and flood protection institutions for further processing in virtually real time. At this moment 12 countries of the Danube have joined forces to work out the proposals that are essential for the future, for safer Danube.
PL
Dunaj i jego dopływy przecinają góry i równiny w swoich prawie niezmienionych korytach od tysięcy lat, nie zważając na granice państwowe i administracyjne. Mimo to jeszcze kilkadziesiąt lat temu kilka krajów udostępniało jedynie ograniczone dane i informacje na temat poziomu wód i stanu ochrony przeciwpowodziowej swoich rzek. W ostatnich latach informacje przekazywane były głównie na podstawie umów dwustronnych i pomyślnie realizowanych działań organizacji działających w całym dorzeczu, ale dla innych informacje można było uzyskać jedynie poprzez przeglądanie Internetu, czasem w dość skomplikowany sposób i z pewnością wymagający czasu. Strategia UE dla Obszaru Priorytetowego Ryzyko Środowiskowe Regionu Dunaju zainicjowała projekt mający na celu rozwój Systemu Informacji Hydrologicznej Dunaju, który był wspierany przez Międzynarodową Komisję Ochrony Rzeki Dunaj. Kompleksowy przegląd złożonych krajowych systemów prognozowania powodzi i oblodzenia, identyfikacja braków w istniejących praktykach prognozowania, jak również poprawa wymiany i dostępności danych hydrologicznych i meteorologicznych pomiędzy krajami uczestniczącymi w projekcie stanowiły kluczowe obszary zainteresowania projektu. Dlatego też głównym celem artykułu jest przedstawienie i omówienie kluczowych danych i funkcjonalności systemu. Z najważniejszych ustaleń wynika, że wszystkie autoryzowane dane meteorologiczne i hydrologiczne dotyczące rzeki Dunaj są przechowywane w centralnej bazie danych i udostępniane online wszystkim licencjonowanym instytucjom hydrologicznym i ochrony przeciwpowodziowej do dalszego przetwarzania w czasie praktycznie rzeczywistym. W tej chwili 12 krajów naddunajskich połączyło siły, aby wypracować propozycje, które są niezbędne dla przyszłości, dla bezpieczniejszego Dunaju.
This study is aimed at evaluating the applicability of Artificial Neural Network (ANN) model technique for river discharge forecasting. Feed-forward multilayer perceptron neural network trained with back-propagation algorithm was employed for model development. Hydro-meteorological data for the Imo River watershed, that was collected from the Anambra-Imo River Basin Development Authority, Owerri – Imo State, South-East, Nigeria, was used to train, validate and test the model. Coefficients of determination results are 0.91, 0.91 and 0.93 for training, validation and testing periods respectively. River discharge forecasts were fitted against actual discharge data for one to five lead days. Model results gave R2 values of 0.95, 0.95, 0.92, 0.96 and 0.94 for first, second, third, fourth, and fifth lead days of forecasts, respectively. It was generally observed that the R2 values decreased with increase in lead days for the model. Generally, this technique proved to be effective in river discharge modelling for flood forecasting for shorter lead-day times, especially in areas with limited data sets.
The typical Mediterranean climate is marked at certain times of the year by sudden torrential rains causing high water flows, which leads to heavy flooding and hydroclimatic fluctuations due to a semi-arid climate. This explains the need for hydrological modeling for water resource management in these contexts. This work concerns the hydrological modeling of the Azzaba catchment area in Haut-Sebou “Morocco”. In the first part of this work, a bibliographic synthesis was carried out to characterize certain factors (physical, geological and climatic), and a hydrological study was carried out by processing rainfall and hydrometric data from the considered time periods. Ultimately, the use of the “ATHYS” platform is beginning to reproduce the flows at the Azzaba outlet. This model is really applicable in the semi-arid context based on several studies carried out on these contexts, since it has to consider the chronological sequence of phenomena on one hand and the influence of the climatic and physicalhydrogeological parameters of the basin (humidity and soil exchange) on the other. Several criteria were used in this study to estimate the model performance; the most common is Nash-Sutcliffe. After observation and analysis of the overall results, it can be concluded that the model reproduces flows in the Azzaba River watershed well, especially in event mode (mean Nash-Sutcliffe value of 0.71). The use of a historical meteorological time series to simulate flow using a daily time step gives average results with a Nash of 0.50, which strengthens the reliability of the ATHYS platform in the Mediterranean climate area.
Accurate and timely flash floods forecasting, especially, in ungauged and poorly gauged basins, is one of the most important and challenging problems to be solved by the international hydrological community. In changing climate and variable anthropogenic impact on river basins, as well as due to low density of surface hydrometeorological network, flash flood forecasting based on “traditional” physically based, or conceptual, or statistical hydrological models often becomes inefficient. Unfortunately, most of river basins in Russia are poorly gauged or ungauged; besides, lack of hydrogeological data is quite typical. However, the developing economy and population safety necessitate issuing warnings based on reliable forecasts. For this purpose, a new hydrological model, MLCM3 (Multi-Layer Conceptual Model, 3rd generation) has been developed in the Russian State Hydrometeorological University. The model showed good results in more than 50 tested basins.
6
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW
This review considers the application of statistical methods and ARIMA (autoregression integrated moving average) models to rainfall-runoff modeling and flood forecasting have been discussed. This is a relatively emerging field of research, characterized by a wide variety of techniques, an amenity of hulk source data, a possibility of intermodel comparisons, determina-tion its adequacy to observable data and also inconsistent reporting of model skin. The paper outlines the basic principles of ARIMA modeling and algorithms used. Literature survey underlines the need for clear guidance in current ARIMA modeling practice, as well as the comparison of ARIMA models with already existing models of rainfall-runoff. Accordingly, a template is proposed in order to assist the construction of future ARIMA rainfall-runoff models.
PL
Przedstawiono zastosowanie metod statystycznych, w tym zwłaszcza modelu ARIMA (autoregresji całkowanej zmiennej średniej), do prognozowania przebiegu sytuacji powodziowych. Omówiono zastosowanie modelu ARIMA do opisu powsta-wania wód powodziowych spowodowanych ulewnymi deszczami oraz spływu tych wód.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.