Daily River Flow Research Articles

Improved higher lead time river flow forecasts using sequential neural network with error updating

Abstract This paper presents a novel framework to use artificial neural network (ANN) for accurate forecasting of river flows at higher lead times. The proposed model, termed as sequential ANN (SANN), is based on the heuristic that a mechanism that provides an accurate representation of physical condition of the basin at the time of forecast, in terms of input information to ANNs at higher lead time, helps improve the forecast accuracy. In SANN, a series of ANNs are connected sequentially to extend the lead time of forecast, each of them taking a forecast value from an immediate preceding network as input. The output of each network is modified by adding an expected value of error so that the residual variance of the forecast series is minimized. The applicability of SANN in hydrological forecasting is illustrated through three case examples: a hypothetical time series, daily river flow forecasting of Kentucky River, USA and hourly river flow forecasting of Kolar River, India. The results demonstrate that SANN is capable of providing accurate forecasts up to 8 steps ahead. A very close fit (>94% efficiency) was obtained between computed and observed flows up to 1 hour in advance for all the cases, and the deterioration in fit was not significant as the forecast lead time increased (92% at 8 steps ahead). The results show that SANN performs much better than traditional ANN models in extending the forecast lead time, suggesting that it can be effectively employed in developing flood management measures.

Features of artificial neural networks technique implementation to runoff calculations problems in ungauged basins

In the last decade scientific questions related to the river runoff calculations in ungauged basins are well-discussed by the wide range of hydrologists and caused to intensive development of modern scientific pathways and state-of-art techniques within this field. 323 watersheds located in the usa were selected as the experimental geographical objects. In the article we provide solution of optimum model parameters searching problem using the modern technique based on artificial neural networks. These model parameters required for a daily runoff calculations in ungauged basins with implementation by the land-surface hydrological model swap. In the article step-by-step genetic evolution of developed technique was presented. Quality of model parameters estimation by using our technique was analyzed on every stage of its development. River runoff calculations efficiency with implementation of found parameters was analyzed as well. Model parameters range expansion based on full accounting of its automated calibration results, and input information multiplying method implementation, lead to satisfactory results of daily river flow calculations in ungauged basins for whole investigated sample.

Deterministic chaotic dynamics of Raba River flow (Polish Carpathian Mountains)

Is the underlying dynamics of river flow random or deterministic? If it is deterministic, is it deterministic chaotic? This issue is still controversial. The application of several independent methods, techniques and tools for studying daily river flow data gives consistent, reliable and clear-cut results to the question. The outcomes point out that the investigated discharge dynamics is not random but deterministic. Moreover, the results completely confirm the nonlinear deterministic chaotic nature of the studied process. The research was conducted on daily discharge from two selected gauging stations of the mountain river in southern Poland, the Raba River.

Impact of multi-resolution analysis of artificial intelligence models inputs on multi-step ahead river flow forecasting

Summary In this paper an attempt is made to show that the performance of daily river flow forecasting is improved when data-preprocessing techniques are used with computational intelligence methods. Especially for forecasting longer lead-times, one of the inherent problems in all forecasting methods is that the reliability of forecasting decreases with increasing the lead-time. Therefore wavelet multi-resolution analysis is coupled with artificial neural networks (ANN) and adaptive neuro-fuzzy interface system (ANFIS) which are two promising methods for forecasting nonlinear and non-stationary time series. Different models with a combination of the different input data sets are developed for 1, 2, 3, 4 and 5 days ahead forecasting in Harvey River, Western Australia. Daubechies and Symlet wavelets are used to decompose river flow time series to different levels. Hybrid wavelet neural networks (WNN) models are trained using Levenberg–Marquart (LM) algorithm and the wavelet neuro-fuzzy (WNF) models with subtractive clustering method to find the optimum number of fuzzy rules. Comparing the results with those of the original ANN and ANFIS models indicates that the hybrid models produce significantly better results, especially for the peak values and longer lead-times.

Water constraints on European power supply under climate change: impacts on electricity prices

Recent warm, dry summers showed the vulnerability of the European power sector to low water availability and high river temperatures. Climate change is likely to impact electricity supply, in terms of both water availability for hydropower generation and cooling water usage for thermoelectric power production. Here, we show the impacts of climate change and changes in water availability and water temperature on European electricity production and prices. Using simulations of daily river flows and water temperatures under future climate (2031–2060) in power production models, we show declines in both thermoelectric and hydropower generating potential for most parts of Europe, except for the most northern countries. Based on changes in power production potentials, we assess the cost-optimal use of power plants for each European country by taking electricity import and export constraints into account. Higher wholesale prices are projected on a mean annual basis for most European countries (except for Sweden and Norway), with strongest increases for Slovenia (12–15%), Bulgaria (21–23%) and Romania (31–32% for 2031–2060), where limitations in water availability mainly affect power plants with low production costs. Considering the long design life of power plant infrastructures, short-term adaptation strategies are highly recommended to prevent undesired distributional and allocative effects.

Hydrological assessment for mini hydropower potential at Sungai Pahang @ Temerloh

Sg Pahang at Temerloh was considered for assessment of hydropower potential using hydrological analysis method and hydrological model. The available data related to topography, soil, land use, weather and discharge pertaining to the study catchment were used to characterize the catchment. The characterization was required for water resources hence hydropower assessment. The hydrology of the study catchment was simulated through the model. This hydrological study is required due to the proposed mini hydroelectric power plant at Pulau Temerloh. It is essential to evaluate the existing river flow characteristic and to model the environmental flow assessment of the river. Two rainfalll stations, JPS Temerloh and Pintu Kawalan Paya Kertam Station are selected to develop the Rainfall Intensity Duration frequency (RIDF) Curve to determine the rainfall intensity of the area. Daily river flow were recorded at Sg Pahang at Temerloh and Sg Pahang at Lubok Paku were used to develop the Flow Duration Curve (FDC) to study the characteristic of Sungai Pahang flow. The 7 days low flow with 10 years return period (7Q10 low flow) was obtained using both Gumbel Method and Log Pearson Type III Method. The results from FDC shows that 50% percentage of time the Sg Pahang @ Temerloh is exceeded over a historical period is 400 m3/s and 50% percentage of time the Sg Pahang @ Lubok Paku is exceeded over a historical period is 650 m3/s. The required environmental flow are set to be 7Q10 low flow which is 64.215 m3/s for Sg Pahang at Temerloh and 79.24 m3/s for Sg Pahang at Lubok Paku. The results show the water resources are abundant and hence boost the mini hydropower potentiality at Sg Pahang.

M5 model tree application in daily river flow forecasting in Sohu Stream, Turkey

This study investigate the potential of M5 model tree in predicting daily stream flows in Sohu river located within the municipal borders of Ankara, Turkey. The results of the M5 model tree was compared with support vector machines. Both modelling approaches were used to forecast up to 7-day ahead stream flow. A comparison of correlation coefficient and root mean square value indicates that M5 model tree approach works equally well to the SVM for same day discharge prediction. The M5 model tree also works well up to 7-day ahead discharge forecasting in comparison of SVM with this data set. An advantage of using M5 model tree approach is the availability of simple linear models to predict the discharge as well use of less computational time.

Future Flows Hydrology: an ensemble of daily river flow and monthly groundwater levels for use for climate change impact assessment across Great Britain

Abstract. The dataset Future Flows Hydrology was developed as part of the project "Future Flows and Groundwater Levels'' to provide a consistent set of transient daily river flow and monthly groundwater level projections across England, Wales and Scotland to enable the investigation of the role of climate variability on river flow and groundwater levels nationally and how this may change in the future. Future Flows Hydrology is derived from Future Flows Climate, a national ensemble projection derived from the Hadley Centre's ensemble projection HadRM3-PPE to provide a consistent set of climate change projections for the whole of Great Britain at both space and time resolutions appropriate for hydrological applications. Three hydrological models and one groundwater level model were used to derive Future Flows Hydrology, with 30 river sites simulated by two hydrological models to enable assessment of hydrological modelling uncertainty in studying the impact of climate change on the hydrology. Future Flows Hydrology contains an 11-member ensemble of transient projections from January 1951 to December 2098, each associated with a single realisation from a different variant of HadRM3 and a single hydrological model. Daily river flows are provided for 281 river catchments and monthly groundwater levels at 24 boreholes as .csv files containing all 11 ensemble members. When separate simulations are done with two hydrological models, two separate .csv files are provided. Because of potential biases in the climate–hydrology modelling chain, catchment fact sheets are associated with each ensemble. These contain information on the uncertainty associated with the hydrological modelling when driven using observed climate and Future Flows Climate for a period representative of the reference time slice 1961–1990 as described by key hydrological statistics. Graphs of projected changes for selected hydrological indicators are also provided for the 2050s time slice. Limitations associated with the dataset are provided, along with practical recommendation of use. Future Flows Hydrology is freely available for non-commercial use under certain licensing conditions. For each study site, catchment averages of daily precipitation and monthly potential evapotranspiration, used to drive the hydrological models, are made available, so that hydrological modelling uncertainty under climate change conditions can be explored further. doi:10.5285/f3723162-4fed-4d9d-92c6-dd17412fa37b

Future Flows Climate: an ensemble of 1-km climate change projections for hydrological application in Great Britain

Abstract. The dataset Future Flows Climate was developed as part of the project ''Future Flows and Groundwater Levels'' to provide a consistent set of climate change projections for the whole of Great Britain at both space and time resolutions appropriate for hydrological applications, and to enable climate change uncertainty and climate variability to be accounted for in the assessment of their possible impacts on the environment. Future Flows Climate is derived from the Hadley Centre's ensemble projection HadRM3-PPE that is part of the basis of UKCP09 and includes projections in available precipitation (water available to hydrological processes after snow and ice storages have been accounted for) and potential evapotranspiration. It corresponds to an 11-member ensemble of transient projections from January 1950 to December 2098, each a single realisation from a different variant of HadRM3. Data are provided on a 1-km grid over the HadRM3 land areas at a daily (available precipitation) and monthly (PE) time step as netCDF files. Because systematic biases in temperature and precipitation were found between HadRM3-PPE and gridded temperature and precipitation observations for the 1962–1991 period, a monthly bias correction procedure was undertaken, based on a linear correction for temperature and a quantile-mapping correction (using the gamma distribution) for precipitation followed by a spatial downscaling. Available precipitation was derived from the bias-corrected precipitation and temperature time series using a simple elevation-dependant snow-melt model. Potential evapotranspiration time series were calculated for each month using the FAO-56 Penman-Monteith equations and bias-corrected temperature, cloud cover, relative humidity and wind speed from HadRM3-PPE along with latitude of the grid and the day of the year. Future Flows Climate is freely available for non-commercial use under certain licensing conditions. It is the dataset used to generate Future Flows Hydrology, an ensemble of transient projections of daily river flow and monthly groundwater time series for representative river basins and boreholes in Great Britain. doi:10.5285/bad1514f-119e-44a4-8e1e-442735bb9797.

Wavelet neural network model for river flow time series

A new hybrid model that combines wavelets and an artificial neural network (ANN) called the wavelet neural network (WNN) model is proposed and applied for time series modelling of river flow. Time series of daily river flow of the Malaprabha River basin (Karnataka state, India) were analysed by the WNN model. The observed time series are decomposed into sub-series using a discrete wavelet transform and then an appropriate sub-series is used as input to the neural network for forecasting hydrological variables. The hybrid model (WNN) was compared with the standard ANN and auto-regressive (AR) models. The WNN model was able to provide a good fit with the observed data, especially the peak values during testing. The benchmark results from WNN model applications show that the hybrid model produces better results than the ANN and AR models in estimating hydrograph properties.

Estimating suspended sediment concentration using neural differential evolution (NDE), multi layer perceptron (MLP) and radial basis function (RBF) models

Sediment transport has attracted the attention of engineers from various aspects and different methods have been used for its estimation. So, several experimental equations have been submitted by experts. Though the results of these methods have considerable differences with each other and with experimental observations, because the sediment measures have some limits, these equations can be used in estimating sediment load. With regard to the fact that Givichay River has high sediment production in the region, it is chosen as the study area. This river is one of the Qezeuzan River branches and through this river it joins to the Sefidrud River. Sefidrud dam is one of the most sediment receiver dams in the world and now more than half of the dam capacity has been filled with sediment. With regard to the fact that there are not enough sediment measure stations in this region, therefore different methods have been used. In this study, neural differential evolution (NDE) models are proposed to estimate suspended sediment concentration. NDE models are improved by combining two methods, neural networks and differential evolution. In the first part of the study, NDE model is trained using daily river flow and suspended sediment data belonging to Givi Chay River in northwest of Iran and various combinations of current daily stream flow and past daily stream flow, suspended sediment data are used as inputs to the NDE model so as to estimate current suspended sediment. In the second part of the study, the suspended sediment estimations provided by NDE model are compared with multi layer perceptron (MLP), radial basis function (RBF) and sediment rating curves (SRC) results. The Root mean squared error (RMSE) and the determination coefficient (R2) are used as comparison criteria. Obtained results demonstrate that NDE are in good agreement with the observed suspended sediment concentration; while they depict better results than RBF, MLP and SRC methods. For example, in Givi Chay River station, the determination coefficient (R2) is 0.9621 for NDE model, while it is 0.9114, 0.90 and 0.8447 for RBF, MLP and SRC models, respectively. However, for the estimation of maximum sediment peak, the NDE was mostly found to be better than the RBF and the other techniques. The results also indicate that the NDE may provide better performance than the RBF, MLP and SRC in the estimation of the total sediment load (Re = -26%). Key words: Givi Chay River, neural differential evolution, multi-layer perceptron model,radial basis function, sediment rating curves.

Incorporating land-use changes and surface–groundwater interactions in a simple catchment water yield model

Pressure on limited water resources and the environment requires better understanding of how landscape change impacts river flow. Rainfall-runoff models have traditionally focused on estimating total river flows with less emphasis on modelling the groundwater component or the consequences of different land-use change scenarios. In this paper, we present the GWlag model, a water-generation model that predicts river flows with explicit accounting of the impacts of catchment land-use change and surface–groundwater interactions. The paper firstly describes the theory that underpins the model and its calibration then presents a case study application in the Tarcutta Creek catchment of the Murray–Darling Basin, Australia. The case study aims at: (i) demonstrating the ability of the model to predict daily river flows; (ii) modelling the impacts of hypothetical plantation forestry expansions on river flows; and (iii) showing the impacts of reduced recharge on the low-flow regime using three indices, namely, Q90/Q50 (where Qn refers to nth percentile flow), slope of low-flow part of flow duration curve, and % of zero-flow days. Results showed that predicted flows agreed favourably to those observed at the gauge especially during low-flow conditions. The hypothetical plantation expansion from 32% to 87% of the catchment area has resulted in reductions of 48% and 32%, in Q50 and Q20, respectively. The low-flow indices demonstrated the great sensitivity of low flow to reductions in recharge with the trend of the low-flow response changing to non-linear for recharge reductions beyond 10%. GWlag daily river flow predictions compared favourably to those obtained from four other rainfall-runoff models in terms of the Nash–Sutcliffe model efficiency (E). However, GWlag produced the highest E-value for log-transformed flows thus highlighting the model's superior predictive capability during low-flow conditions.

Assessing the expression of large-scale climatic fluctuations in the hydrological variability of daily Seine river flow (France) between 1950 and 2008 using Hilbert–Huang Transform

Summary Daily Seine river flow variability from 1950 to 2008 was analyzed and compared to the winter-months North-Atlantic Oscillation (NAO) using Hilbert–Huang Transform (HHT). For the last 10 years, HHT has proven its efficiency for the analysis of transient oscillatory signals. HHT provides an interesting alternative to other techniques for time-frequency or time scale analysis of non-stationary signals. In this study we aimed at delineating the different components characterizing daily flow of the Seine river, on the short-term, intra-seasonal, annual and interannual time scales and eventually interpret them in the context of regional North-Atlantic climate regime fluctuations. HHT results highlighted the existence of similar scales of variability beard by internal components of each NAO or river flow signal at interannual scales. Hypotheses on a possible link between the Madden–Julian Oscillation pattern and intra-seasonal variability of river flow could be also proposed, which would highlight linkages between river flow variability and global climate oscillations. Finally, all oscillating components were found to increase in amplitude in both climatic and hydrological signals in the end of the 1950–2008 period of study, with a first step in the late 1960s–early 1970s and a second step in the early 1990s, which demonstrated the capabilities of HHT to handle non-stationarity of natural processes and to help interpreting hydrological variability in a context of climate changes.

An appraisal of the performance of data-infilling methods for application to daily mean river flow records in the UK

River flow records are fundamental for the sustainable management of water resources and even very short gaps can severely compromise their utility. Suitably-flagged flow estimates, derived via judicious infilling, are potentially highly beneficial to data users. The UK National River Flow Archive provides stewardship of, and access to, UK river flow records. While many datasets held on the archive are complete, gaps remain across a wide range of flow records. A comprehensive assessment of existing techniques for infilling these gaps is currently lacking. This paper therefore assesses 15 simple infilling techniques (including regression, scaling and equipercentile approaches), each relying upon data transfer from hydrologically-similar donor stations, to generate estimates of flow at 26 representative gauging stations. Results reveal the overall superiority of equipercentile and multiple regression techniques compared to the poorer capability of catchment area scaling. Donor station choice has a strong influence on technique performance. Modifying datasets to improve homogeneity, by seasonally grouping flows or excluding certain periods, offers improved performance. These findings provide a foundation upon which guidance on infilling river flow records can be based in future, allowing hydrometric practitioners and data end-users alike to adopt a consistent and auditable approach towards infilling.

A fuzzy-autoregressive model of daily river flows

A model for the identification of daily river flows has been developed, consisting of the combination of an autoregressive model with a fuzzy inference system. The AR model is devoted to the identification of base flow, supposed to be described by linear laws. The fuzzy model identifies the surface runoff, by applying a small set of linguistic statements, deriving from the knowledge of the physical features of the nonlinear rainfall–runoff transformation, to the inflow entering the river basin. The model has been applied to the identification of the daily flow series of river Volturno at Cancello-Arnone (Southern Italy), with a drainage basin of around 5560km2, observed between 1970 and 1974. The inflow was estimated on the basis of daily precipitations registered during the same years at six rain gauges located throughout the basin. The first two years were used for model training, the remaining three for the validation. The obtained results show that the proposed model provides good predictions of either low river flows or high floods, although the analysis of residuals, which do not turn out to be a white noise, indicates that the cause and effect relationship between rainfall and runoff is not completely identified by the model.

Daily Discharge Forecasting Using SupportVector Machine

 Abstract—Support Vector Machine (SVM) is a kind of learning machine for simulation or prediction. In this paper, Support Vector Machine (SVM) is used to forecast daily river flow and the results of these models are compared with observed daily values. Daily river flow data On Ghara-soo river in north of Iran are used in this study. The daily flow and rain data of station on Ghara-soo as exit discharge and three station of this Catchment Names: shast kalate, ziyarat and Kurd kuy are used to train and test the developed models. The observed data that are used in this study start from 1992 to 2010 in 18 year's period (6550 days). 75% of the whole data set are used for training the models and 25% of the whole data set are used for testing step. In this regard, five kind of different input data that affect the river flow has been identified and based on this method, the river flow is predicted. In this research, predicted data are compared with actual data through the RMSE index. For checking of proposed model performance in river flow forecasting, the information of Ghara-soo River has been used. The comparison shows that the proposed method yields a high accuracy in the prediction.

Water balance of Gorganrood river basin East of Iran

Recent development of geographic information system (GIS) and remote sensing technology makes it possible to capture and manage a vast amount of spatially distributed hydrological parameters and variables. In this paper, a spatially distributed hydrologic model (WetSpa) is used to estimate daily river flow discharge and to analyze water balance of the Gorganrood river basin, Iran. The model combines topography, landuse and soil maps to predict discharge hydrographs and spatial distribution of hydrological parameters. For each grid cell the model holds water balance in the root zone by determination of soil moisture changes due to continues changes in infiltration of precipitation, runoff, initial absorption, evaporation transpiration, interflow, and percolation to the groundwater zone. The model gives all flow components at any cell, including surface flow, interflow and groundwater flow. The results show that the model is capable in reproducing water balance and its components. They also demonstrate the changes in the water balance during last two decades. Key words: Water balance, distributed hydrological model, WetSpa, Gorganrood.

Analysis of some hydrological parameters of the Nasia River Catchment in the Northern Region of Ghana for its socio-economic development

The Nasia River is the drainage basin of the left bank tributary of the White Volta located in the Northern Region of Ghana. Most years, rivers in the White Volta basin flow only three or four months, causing seasonal deficits across the region. This research therefore studies the hydrology of the Nasia River Basin and the catchment’s conditions so as to determine the potentials for irrigation development. Thirty years of daily river flows and climatic data for Nasia River were obtained from Hydrological Services Department and the Regional Meteorological Agency at Tamale. Soils and geological information about the basin were obtained, whilst topographic maps with scale of 1:50,000 were acquired from the Regional Survey Department at Tamale for the determination of the catchment characteristics. The results obtained in the study suggested that the Nasia river basin has enough surface water resources which could be harnessed for irrigation and other socio-economic development of the area. The seasonal variation in the runoff yield is influenced by the catchment’s physical characteristics. Optimum flow parameters for hydrological design purposes indicate that the guaranteed yield which is exceeded 95% of the time is 0.0438 m3/s, whilst the characteristic drought at 63% exceedance is 0.40 m3/s. Key words: Catchment, hydrology, Nasia River, water balance.

Time Series Modeling of River Flow Using Wavelet Neural Networks

A new hybrid model which combines wavelets and Artificial Neural Network (ANN) called wavelet neural network (WNN) model was proposed in the current study and applied for time series modeling of river flow. The time series of daily river flow of the Malaprabha River basin (Karnataka state, India) were analyzed by the WNN model. The observed time series are decomposed into sub-series using discrete wavelet transform and then appropriate sub-series is used as inputs to the neural network for forecasting hydrological variables. The hybrid model (WNN) was compared with the standard ANN and AR models. The WNN model was able to provide a good fit with the observed data, especially the peak values during the testing period. The benchmark results from WNN model applications showed that the hybrid model produced better results in estimating the hydrograph properties than the latter models (ANN and AR).

Climate impacts on river flow: projections for the Medway catchment, UK, with UKCP09 and CATCHMOD

AbstractThe potential impact of climate change on areas of strategic importance for water resources remains a concern. Here, river flow projections for the River Medway, above Teston in southeast England are presented, which is just such an area of strategic importance. The river flow projections use climate inputs from the Hadley Centre Regional Climate Model (HadRM3) for the time period 1960–2080 (a subset of the early release UKCP09 projections). River flow predictions are calculated using CATCHMOD, the main river flow prediction tool of the Environment Agency (EA) of England and Wales. In order to use this tool in the best way for climate change predictions, model setup and performance are analysed using sensitivity and uncertainty analysis. The model's representation of hydrological processes is discussed and the direct percolation and first linear storage constant parameters are found to strongly affect model results in a complex way, with the former more important for low flows and the latter for high flows. The uncertainty in predictions resulting from the hydrological model parameters is demonstrated and the projections of river flow under future climate are analysed. A clear climate change impact signal is evident in the results with a persistent lowering of mean daily river flows for all months and for all projection time slices. Results indicate that a projection of lower flows under future climate is valid even taking into account the uncertainties considered in this modelling chain exercise. The model parameter uncertainty becomes more significant under future climate as the river flows become lower. This has significant implications for those making policy decisions based on such modelling results. Copyright © 2010 John Wiley & Sons, Ltd.