Runoff Time Series Research Articles

Influence of Precipitation Characteristics and Vegetation on Runoff and Sediment: A Case on the Basin in the Three Gorges Reservoir Region

Sediment is the main carrier of pollutants in river channels. This study analyzed the distribution characteristics of precipitation, runoff, and sediment and their response characteristics in the Daning River basin. Based on daily precipitation (1979–2017), runoff (1989–2017), and sediment (1997–2017) time series, the Gini concentration index, precipitation concentration index (PCI), precipitation concentration degree (PCD), and precipitation concentration period were applied to assess the concentration characteristics of precipitation, runoff, and sediment on the daily, monthly, and seasonal scales. At each intensity level, precipitation was negatively correlated to the PCI and PCD. The normalized difference vegetation index (NDVI) values had strong negative correlations with rainy days with light precipitation (0.1–9.9 mm). The degrees of concentration were in the same order for the multiscale analysis: runoff < precipitation < sediment. Although the amount of daily precipitation of more than 25 mm displayed a significant increasing trend, suggesting an increased risk of flood and soil erosion, the significantly improved vegetation cover reduced the sediment-carrying capacity of the surface runoff, with significant decreases in the total amount and multiscale concentration degrees of sediment being observed. The results of the study provide a reference for the improvement of the potable water safety and ecological environment in the Three Gorges Reservoir region.

Seasonal sub-basin-scale runoff predictions: A regional hydrometeorological Ensemble Kalman Filter framework using global datasets

Study regionThe São Francisco River Basin (SFRB) in Brazil Study focusIn semi-arid regions, interannual variability of seasonal rainfall and climate change is expected to stress water availability and increase the recurrence and intensity of extreme events such as droughts or floods. Local decision makers therefore need reliable long-term hydro-meteorological forecasts to support the seasonal management of water resources, reservoir operations and agriculture. In this context, an Ensemble Kalman Filter framework is applied to predict sub-basin-scale runoff employing global freely available datasets of reanalysis precipitation (ERA5-Land) as well as bias-corrected and spatially disaggregated seasonal forecasts (SEAS5-BCSD). Runoff is estimated using least squares predictions, exploiting the covariance structures between runoff and precipitation. The performance of the assimilation framework was assessed using different ensemble skill scores. New hydrological insights for the regionOur results show that the quality of runoff predictions are closely linked to the performance of the rainfall seasonal predictions and allows skillful predictions up to two months ahead in most sub-basins. The anthropogenic conditions such as in the Western Bahia state, however, must be taken under consideration, since non-stationary runoff time-series have poorer skill as such unnatural variations can not be captured by long-term covariances. In sub-basins which are dominated by little anthropogenic influence, the presented framework provides a promising and easily transferable approach for skillful operational seasonal runoff predictions on sub-basin scale.

A Hybrid Model Integrating Elman Neural Network with Variational Mode Decomposition and Box–Cox Transformation for Monthly Runoff Time Series Prediction

Precise and reliable monthly runoff prediction plays a vital role in the optimal management of water resources, but the nonstationarity and skewness of monthly runoff time series can pose major challenges for developing appropriate prediction models. To address these issues, this paper proposes a novel hybrid prediction model by introducing variational mode decomposition (VMD) and Box–Cox transformation (BC) into the Elman neural network (Elman), named the VMD-BC-Elman model. First, the observed runoff is decomposed into sub-time series using VMD for better frequency resolution. Second, the input datasets are transformed into a normal distribution using Box–Cox, and as a result, skewedness in the data is removed, and the correlation between the input and output variables is enhanced. The proposed VMD-BC preprocessing technology is expected to overcome the problems arising from nonstationary and skewed runoff data. Finally, Elman is used to simulate the respective sub-time series. The proposed model is evaluated using monthly runoff time series at Zhangjiashan, Zhuangtou and Huaxian hydrological stations in the Wei River Basin in China. The model performances are compared with those of single models (SVM, Elman), decomposition-based (VMD-SVM, VMD-Elman et al.) and BC-based models (BC-SVM and BC-Elman) by employing four metrics. The results show that the hybrid models outperform single models, and the VMD-BC-Elman model performs best in all considered hybrid models with an NSE greater than 0.95, R greater than 0.98, NMSE less than 4.7%, and PBIAS less than 0.4% in both the training and testing periods. The study indicates that the VMD-BC-Elman model is a satisfactory data-driven approach to predict nonstationary and skewed monthly runoff time series, representing an effective tool for predicting monthly runoff series.

Multi-model evaluation of catchment- and global-scale hydrological model simulations of drought characteristics across eight large river catchments

Although global- and catchment-scale hydrological models are often shown to accurately simulate long-term runoff time-series, far less is known about their suitability for capturing hydrological extremes, such as droughts. Here we evaluated simulations of hydrological droughts from nine catchment scale hydrological models (CHMs) and eight global scale hydrological models (GHMs) for eight large catchments: Upper Amazon, Lena, Upper Mississippi, Upper Niger, Rhine, Tagus, Upper Yangtze and Upper Yellow. The simulations were conducted within the framework of phase 2a of the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP2a). We evaluated the ability of the CHMs, GHMs and their respective ensemble means (Ens-CHM and Ens-GHM) to simulate observed hydrological droughts of at least one month duration, over 31 years (1971–2001). Hydrological drought events were identified from runoff-deficits and the Standardised Runoff Index (SRI). In all catchments, the CHMs performed relatively better than the GHMs, for simulating monthly runoff-deficits. The number of drought events identified under different drought categories (i.e. SRI values of -1 to -1.49, -1.5 to -1.99, and ≤-2) varied significantly between models. All the models, as well as the two ensemble means, have limited abilities to accurately simulate drought events in all eight catchments, in terms of their occurrence and magnitude. Overall, there are opportunities to improve both CHMs and GHMs for better characterisation of hydrological droughts.

Dynamic parameterization of soil surface characteristics for hydrological models in agricultural catchments

The detrimental impacts of surface runoff and soil erosion, particularly in cultivated areas, call for the use of distributed runoff and soil erosion models with a view to supporting adapted catchment management strategies. However, runoff model parameterization remains challenging in agricultural catchments due to the high spatial and seasonal variability of soil properties. Data acquisition is demanding and may not always be feasible. Therefore, model parameterization in such environments have been the subject of numerous research efforts. The combined analysis of land use management and soil surface state was proposed in literature to address this issue and demonstrated its potential for runoff analysis and modelling. However, these research findings were related to specific rainfall sequences and/or soil surface state. In this study, existing knowledge on soil surface state and its application to runoff model parameterization were synthetized and included in an easy-to-use parameterization software (PREMACHE), providing a framework for modelers lacking of means and/or data for modelling complex agricultural catchments.To develop and evaluate the software, a dataset was acquired over 9 years on more than 110 plots in a 1045 ha agricultural catchment, including crop types, soil surface state, rainfall and runoff time series. Soil surface state dynamics was modeled based on crop types and daily rainfall. It was evaluated in the experimental catchment and validated in a nearby catchment. Soil hydrodynamic properties (e.g. infiltration capacity) were deduced from this framework and literature data at a daily time step, for each plots. Moreover, runoff events were measured when the modeled infiltration capacity was low, indicating that the parametrization adequately captured its temporal dynamics. The software developed in this study, as well as setup values deduced from the monitoring campaigns are provided with the manuscript for application in other ungauged catchments and explore their impact on agricultural catchment hydrological dynamics.

Scenario-based multi-objective optimization of reservoirs in silt-laden rivers: A case study in the Lower Yellow River

Severe sedimentation often takes place in the river channel of silt-laden rivers, which is often mitigated through water-sediment regulation of the reservoirs. However, watersediment regulation is often competitive with other objectives of reservoirs, like water supply and hydropower generation; on the other hand, the reduction of channel sedimentation is often achieved at the expense of reservoir sedimentation, which reduces the service life of reservoirs. The Yellow River used to be the river with largest sediment transport over the world, but has experienced significant declination of runoff and sediment in recent years. This study presents a scenario-based multi-objective optimization operation model for the Xiaolangdi reservoir considering hydropower generation, reservoir sedimentation and channel sedimentation, with a generalized linear model coupled to calculate channel sedimentation based on runoff and sediment time series. A stochastic model that can reproduce both spatial correlations and low frequency attributes of the data series is adopted to generate two different scenarios based on different periods of observation and the performance of the multi-objective operation model under different scenarios is tested. The results indicate that: (1) the proposed optimization model can generate different schemes of reservoir operation and enhance operation performance; (2) the generalized linear model can well fit the relationship between daily channel sedimentation and runoff-sediment factors, but tends to overestimate the erosion efficiency after 2005; (3) the reservoir sedimentation and channel sedimentation show linear competitive relation, i.e., an average increase of 1 ton in reservoir sedimentation would result in declination of channel sedimentation from 0.455 to 0.488 tons, while the competitive relationship between hydropower generation and reservoir sedimentation is non-linear and weak; (4) the increase in the proportion of non-flood sediment load to the total sediment load makes it more difficult to prevent the reservoir from silting up.

Application of improved seasonal GM(1,1) model based on HP filter for runoff prediction in Xiangjiang River.

Runoff forecasting is essential for the reasonable use of regional water resources, flood prevention, and mitigation, as well as the development of ecological civilization. Runoff is influenced by the intersection of many factors, and the change process is extremely complex, showing significant stochasticity, nonlinearity, and heterogeneity, making traditional prediction models less adaptable. The Hodrick-Prescott filter (HP filter) is a well-established signal separation method. The traditional GM(1,1) model is capable of portraying the growth trend of the time series but cannot capture the periodic characteristics of the time series. Therefore, a novel coupled prediction model-HPF-GM(1,1) model is proposed in this study and applied to the runoff prediction of the Zhuzhou section of Xiangjiang River in Hunan Province. This model enables to separate seasonal factors from non-seasonal factors in the runoff time series, and introduce seasonal factors based on the traditional GM(1,1) model, which solves the challenge that the traditional GM(1,1) model is unable to predict seasonal time series. The results show that the HPF-GM(1,1) model has a mean relative error of 4.82%, a root mean square error of 7.44, and a Nash efficiency coefficient of 0.93, which is better than the traditional GM(1,1) model, the DGGM(1,1) model and the SGM(1,1) model of prediction accuracy. Obviously, the HP filter provides a new approach to data pre-processing of runoff series and the proposed HPF-GM(1,1)-coupled model extends new ideas for nonlinear runoff prediction.

Runoff forecasting model based on CEEMD and combination model: a case study in the Manasi River, China

Abstract Accurate forecasting of runoff is necessary for water resources management. However, the runoff time series consists of complex nonlinear and non-stationary characteristics, which makes forecasting difficult. To contribute towards improved forecasting accuracy, a novel combination model based on complementary ensemble empirical mode decomposition (CEEMD) for runoff forecasting is proposed and applied in this paper. Firstly, the original runoff series is decomposed into a limited number of intrinsic mode functions (IMFs) and one residual based on CEEMD, which makes the runoff time series stationary. Then, approximate entropy is introduced to judge the complexity of each IMF and residual. According to the calculation results of approximate entropy, the high complexity components are predicted by Gaussian process regression (GPR), the medium complexity components are predicted by support vector machine (SVM), and the low complexity components are predicted by autoregressive integrated moving average model (ARIMA). The advantages of each forecasting model are used to forecast the appropriate components. In order to solve the problem that the forecasting performance of GPR and SVM is affected by their parameters, an improved fireworks algorithm (IFWA) is proposed to optimize the parameters of two models. Finally, the final forecasting result is obtained by adding the forecasted values of each component. The runoff data collected from the Manasi River, China is chosen as the research object. Compared with some state-of-the-art forecasting models, the comparison result curve between the forecasted value and actual value of runoff, the forecasting error, the histogram of the forecasting error distribution, the performance indicators and related statistical indicators show that the developed forecasting model has higher prediction accuracy and is able to reflect the change laws of runoff correctly.

Catchment scale runoff time-series generation and validation using statistical models for the Continental United States

We developed statistical models to generate runoff time-series at National Hydrography Dataset Plus Version 2 (NHDPlusV2) catchment scale for the Continental United States (CONUS). The models use Normalized Difference Vegetation Index (NDVI) based Curve Number (CN) to generate initial runoff time-series which then is corrected using statistical models to improve accuracy. We used the North American Land Data Assimilation System 2 (NLDAS-2) catchment scale runoff time-series as the reference data for model training and validation. We used 17 years of 16-day, 250-m resolution NDVI data as a proxy for hydrologic conditions during a representative year to calculate 23 NDVI based-CN (NDVI-CN) values for each of 2.65 million NHDPlusV2 catchments for the Contiguous U.S. To maximize predictive accuracy while avoiding optimistically biased model validation results, we developed a spatio-temporal cross-validation framework for estimating, selecting, and validating the statistical correction models. We found that in many of the physiographic sections comprising CONUS, even simple linear regression models were highly effective at correcting NDVI-CN runoff to achieve Nash-Sutcliffe Efficiency values above 0.5. However, all models showed poor performance in physiographic sections that experience significant snow accumulation.

Predicting runoff series in ungauged basins of the Brazilian Cerrado biome

Information concerning water availability in a basin can be key to trustworthy and robust decisions, and reduce disputes over water among its multiple users. The Cerrado region, however, lacks basic hydrological information concerning water availability in many of its basins. In this study, two regionalization frameworks based on the donation of parameter sets from a hydrological model calibrated in gauged catchments were assessed. These approaches were evaluated using a leave-one-out cross-validation for the gauged catchments. Parameters donation by spatial proximity led to KGE and rNSE of 0.58 and 0.56, while by attributes proximity led to KGE and rNSE of 0.56 and 0.45, respectively. A regional-sample data set for the Cerrado (HydroCerrado) was made available with information compiled for the 411 gauged catchments used in this study and runoff time series simulated for 4531 level 5 ottobasins by donating parameter sets by spatial proximity.

Long-term modelling of the Arkun multipurpose reservoir

The Arkun hydropower and dam project is located on the Coruh River in the North-west of Turkey. The project was completed in 2014. The reservoir is mainly used for energy generation. For this purpose, a main powerhouse with 225 MW and an environmental powerhouse with 12 MW were constructed. Additionally, the reservoir and operation scheme shall contribute to mitigate floods. A minimum ecological flow is mandatory in the Coruh River between the river section reaching from the eco powerhouse which is located downstream of the CFSGD dam and the main powerhouse which is connected to the reservoir via a 14 km headrace tunnel. During the feasibility stage the hydropower projects in Turkey frequently applied simplified methods for the estimation of the energy generation potential. This approach led to an overestimation of the energy general in many cases. Therefore, a reservoir operation model was created which considered monthly runoff data from 1963 to 2005. This model considered the operation of the two powerhouses, evaporation, potential irrigation uses, and spilling. In order to provide a reliable annual generation prognosis, the future development status was considered by the adaptation of the run-off time series. The results of this modelling confirm both the estimated power production during feasibility stage and the actual power production during the operation period.

Runoff responses to human activities and climate change in an arid watershed of central Iran

ABSTRACT In this study, contributions of climate change and human activity to streamflow changes were estimated to enable decision makers to develop adaptation strategies for management of regional water resources. Flow trends and climate variables were analysed with the Mann-Kendall method. The year 1986 was selected to perform the Pettitt test to identify the change point of the runoff time series. Then, three methods were used for impact differentiation: climatic elasticity, least-squares support-vector machine (LS-SVM), and the soil and water assessment tool (SWAT). The results showed that climate change (38–67%) and human activities (33–62%) influence runoff reduction. Thus three management scenarios are introduced to reduce the effects of climate change and human activity: (1) adjusting wheat and barley cultivation levels; (2) maintaining wheat and barley cultivation levels and replacing other crops with potatoes; (3) increasing irrigation efficiency. All scenarios showed an increase in runoff, but the first scenario had the most impact.

Hydrological Response of the Kunhar River Basin in Pakistan to Climate Change and Anthropogenic Impacts on Runoff Characteristics

Pakistan is amongst the most water-stressed countries in the world, with changes in the frequency of extreme events, notably droughts, under climate change expected to further increase water scarcity. This study examines the impacts of climate change and anthropogenic activities on the runoff of the Kunhar River Basin (KRB) in Pakistan. The Mann Kendall (MK) test detected statistically significant increasing trends in both precipitation and evapotranspiration during the period 1971–2010 over the basin, but with the lack of a statistically significant trend in runoff over the same time-period. Then, a change-point analysis identified changes in the temporal behavior of the annual runoff time series in 1996. Hence, the time series was divided into two time periods, i.e., prior to and after that change: 1971–1996 and 1997–2010, respectively. For the time-period prior to the change point, the analysis revealed a statistically significant increasing trend in precipitation, which is also reflected in the runoff time series, and a decreasing trend in evapotranspiration, albeit lacking statistical significance, was observed. After 1996, however, increasing trends in precipitation and runoff were detected, but the former lacked statistical significance, while no trend in evapotranspiration was noted. Through a hydrological modelling approach reconstructing the natural runoff of the KRB, a 16.1 m3/s (or 15.3%) reduction in the mean flow in the KRB was simulated for the period 1997–2010 in comparison to the period 1971–1996. The trend analyses and modeling study suggest the importance of anthropogenic activities on the variability of runoff over KRB since 1996. The changes in streamflow caused by irrigation, urbanization, and recreational activities, in addition to climate change, have influenced the regional water resources, and there is consequently an urgent need to adapt existing practices for the water requirements of the domestic, agricultural and energy sector to continue being met in the future.

Climate-driven Model Based on Long Short-Term Memory and Bayesian Optimization for Multi-day-ahead Daily Streamflow Forecasting

Many previous studies have developed decomposition and ensemble models to improve runoff forecasting performance. However, these decomposition-based models usually introduce large decomposition errors into the modeling process. Since the variation in runoff time series is greatly driven by climate change, many previous studies considering climate change focused on only rainfall-runoff modeling, with few meteorological factors as input. Therefore, a climate-driven streamflow forecasting (CDSF) framework was proposed to improve the runoff forecasting accuracy. This framework is realized by using principal component analysis (PCA), long short-term memory (LSTM) and Bayesian optimization (BO), referred to as PCA-LSTM-BO. To validate the effectiveness and superiority of the PCA-LSTM-BO method along with one autoregressive LSTM model and two other CDSF models based on PCA, BO, and either support vector regression (SVR) or gradient boosting regression trees (GBRT), namely, PCA-SVR-BO and PCA-GBRT-BO, respectively, were compared. A generalization performance index based on the Nash-Sutcliffe efficiency (NSE), called the GI(NSE) value, is proposed to evaluate the generalizability of the model. The results show that (1) the proposed model is significantly better than the other benchmark models in terms of the mean square error (MSE<=185.782), NSE>=0.819, and GI(NSE) <=0.223 for all the forecasting scenarios; (2) the PCA in the CDSF framework can improve the forecasting capacity and generalizability; (3) the CDSF framework is superior to the autoregressive LSTM models for all the forecasting scenarios; and (4) the GI(NSE) value is demonstrated to be effective in selecting the optimal model with better generalizability.

Assessing climate change effects on declining groundwater levels using wavelet entropy (case study of Khorramabad city)

Abstract Changing global climate suggests a warmer future which may alter the hydrological cycle, affecting surface water as well as groundwater resources. The wavelet entropy (WE) criterion is a new indicator for analyzing time series fluctuations. In this study, effective factors decreasing the groundwater level in Khorramabad city during the years 2005–2018 were evaluated by the use of WE criterion. In general, it can be said that the decreasing WE criterion or time series complexity of a phenomenon shows the time series decrease of natural fluctuation, which leads to an unfavorable trend. In this regard, in order to identify the factors affecting groundwater level decrease in Khorramabad, the groundwater level was divided into 4 time periods, and after being investigated, the monthly time series of runoff, temperature, and precipitation of this city were also divided into 4 periods. Each of these subsets were decomposed into several other subsets at different time scales under the wavelet transform, and finally, after calculation of the normalized wavelet energy for this subset, the WE criterion was calculated for each period. Investigation of WE complexity shows a 21.3% decrease in groundwater level in the second period, but in the third and fourth periods, it increased by 145 and 272%, respectively. Also, according to the results of analysis of WE changes for the precipitation time series, 35.2, 32.8, and 10.06% decrease in the second, third, and fourth periods were shown, respectively. The air temperature time series complexity decreased by 26.8% only in the third time period and in the second and fourth period, it shows an increase of 29.65 and 34.7%, respectively. However, the runoff time series did not show any reduction complexity according to the WE criterion. These results indicate that the impact of climatic factors has been more effective than human factors in reducing the groundwater level of Khorramabad.

Detection of abrupt shift and non-parametric analyses of trends in runoff time series in the Dez river basin

Abstract The present study aims to investigate the homogeneity of runoff time series and also to review the existence of trends in Tale Zang hydrometric station (the hydrometric station at the inflow into the Dez reservoir) runoff, using 61 years (1956–2016) daily recorded data. The Pettit test, which is a common method in investigating the homogeneity of time series, was used to identify change points. Both Mann-Kendall and auto-correlated Mann-Kendall tests were applied to analyze the existence or non-existence of trends in each annual, seasonal, and monthly time series observed in a runoff. In time series, significant trends at the 95% level of confidence were recognized, upper and lower limit values were presented for Sen's slope and it was tested for the increasing or decreasing trends in nature. Based on the results of this study, the significant change point at the 95% level of confidence was recognized in the annual, spring, summer, autumn, March, May, June, July, August, September, and October data sets in 1997,1997, 1999, 1997, 1999, 1999, 1999, 1997, 2000, 2007 and 2008, respectively. Analyzing the existence of a trend at the 95% level of confidence indicated that in spring, and in March, September, and October, for both Mann-Kendall and auto-correlated Mann-Kendall tests, the trend is significant and additive until the change point.

LamaH-CE: LArge-SaMple DAta for Hydrology and Environmental Sciences for Central Europe

Abstract. Very large and comprehensive datasets are increasingly used in the field of hydrology. Large-sample studies provide insights into the hydrological cycle that might not be available with small-scale studies. LamaH-CE (LArge-SaMple DAta for Hydrology and Environmental Sciences for Central Europe, LamaH for short; the geographical extension “-CE” is omitted in the text and the dataset) is a new dataset for large-sample studies and comparative hydrology in Central Europe. It covers the entire upper Danube to the state border of Austria–Slovakia, as well as all other Austrian catchments including their foreign upstream areas. LamaH covers an area of about 170 000 km2 in nine countries, ranging from lowland regions characterized by a continental climate to high alpine zones dominated by snow and ice. Consequently, a wide diversity of properties is present in the individual catchments. We represent this variability in 859 gauged catchments with over 60 catchment attributes, covering topography, climatology, hydrology, land cover, vegetation, soil and geological properties. LamaH further contains a collection of runoff time series as well as meteorological time series. These time series are provided with a daily and hourly resolution. All meteorological and the majority of runoff time series cover a span of over 35 years, which enables long-term analyses with a high temporal resolution. The runoff time series are classified by over 20 attributes including information about human impacts and indicators for data quality and completeness. The structure of LamaH is based on the well-known CAMELS (Catchment Attributes and MEteorology for Large-sample Studies) datasets. In contrast, however, LamaH does not only consider independent basins, covering the full upstream area. Intermediate catchments are covered as well, which allows together with novel attributes the considering of the hydrological network and river topology in applications. We not only describe the basic datasets used and methodology of data preparation but also focus on possible limitations and uncertainties. LamaH contains additionally results of a conceptual hydrological baseline model for checking plausibility of the inputs as well as benchmarking. Potential applications of LamaH are outlined as well, since it is intended to serve as a uniform data basis for further research. LamaH is available at https://doi.org/10.5281/zenodo.4525244 (Klingler et al., 2021).

Rainfall-Runoff Short-Term Forecasting Method Based on LSTM

A rainfall-runoff forecasting method based on Long Short Term Memory (LSTM) was proposed in this study, which can extract the trend characteristics of runoff time series data through introducing daily rainfall data collected at related upstream stations and making use of the advantages of LSTM in saving long-term sequence feature information and avoid vanishing gradient, and identify the nonlinear mapping relationship in between, thus establishing a short-term runoff forecasting model. In this study, 24-hour and 5-day short-term forecasting models were established based on the runoff data collected at Danba Hydrologic Station in Dadu River Basin and historical rainfall collected at three upstream stations (Xiaojin, Dawei, and Fubian River). The experimental results showed that the forecasting models performed well during the inspection period. In 24-hour forecasting, RMSE was 98.016 and MAE was 45.709, which were 73.993 and 32.699, respectively in 5-day forecasting, indicating better performance and increased forecasting accuracy than simple LSTM model.

Runoff Prediction Based on Deep Belief Networks

&#x0D; &#x0D; &#x0D; Runoff prediction has recently become an essential task with respect to assessing the impact of climate change to people’s livelihoods and production. However, the runoff time series always exhibits nonlinear and non-stationary features, which makes it very difficult to be accurately predicted. Machine learning have been recently proved to be a powerful tool in helping society adapt to a changing climate and its subfield, deep learning, showed the power in approximate nonlinear functions. In this study, we propose a method based on deep belief networks (DBN) for runoff prediction. In order to evaluate the proposed method, we collected runoff datasets from Srepok and Dak Nong rivers located in mountain regions of the Central Highland of Vietnam in the periods of 2001-2007 at Dak Nong hydrology station and 1990-2011 at Buon Don hydrology station, respectively. Experimental results show that DBN outperforms, respectively, LSTM, BiLSTM, multi-layer perceptron (MLP) trained by particle swarm optimization (PSO) and MLP trained by stochastic gradient descent (SGD) in which gradients are computed using the backpropagation (BP) procedure. The results also confirm that DBN is suitable to employ for the task of runoff prediction.&#x0D; &#x0D; &#x0D;

Application of ANN and HEC-RAS model for flood inundation mapping in lower Baro Akobo River Basin, Ethiopia

Study regionLower Baro River, Ethiopia. Study focusThis paper presents the novelty of ANN and HEC-RAS model for flood inundation mapping in lower Baro Akobo Basin River, Ethiopia. ANN and HEC-RAS model is applied and successfully improves the accuracy of prediction and flood inundation in the region. This study uses 14 meteorological stations on a daily basis for 1999−2005 and 2006−2008 periods, and Topographical Wetness Index (TWI) to the train and test the model respectively. The runoff time series obtained in ANN model is linked to HEC-RAS and the flood depths were generated. The flood inundation generated in HEC-RAS model result was calibrated and validated in Normal Difference Water Index (NDWI). New hydrological insights for the regionAs the inundation map generated from the runoff values of ANN model reveals, the lower Baro river forms huge inundation depth up to 250 cm. The performance the ANN model was evaluated using Nash-Sutcliffe Efficiency (NSE = 0.86), PBIAS = 8.2 % and R2 = 0.91 and NSE = 0.88, PBIAS = 8.5 % and R2 = 0.93 during the training and testing periods respectively. The generated inundation areas in HEC-RAS and the water bodies delineated in NDWI were covered with 94.6 % and 96 % as overlapping areas during the calibration and validation periods respectively. Therefore, it is concluded that the integration of the ANN approach with the HEC-RAS model has improved the prediction accuracy in traditional flood forecasting methods.