Daily Rainfall-runoff Model Research Articles

Improving the simulations of the hydrological model in the karst catchment by integrating the conceptual model with machine learning models

Hydrological modelling can be complex in nonhomogeneous catchments with diverse geological, climatic, and topographic conditions. In this study, an integrated conceptual model including the snow module with machine learning modelling approaches was implemented for daily rainfall-runoff modelling in mostly karst Ljubljanica catchment, Slovenia, which has heterogeneous characteristics and is potentially exposed to extreme events that make the modelling process more challenging and crucial. In this regard, the conceptual model CemaNeige Génie Rural à 6 paramètres Journalier (CemaNeige GR6J) was combined with machine learning models, namely wavelet-based support vector regression (WSVR) and wavelet-based multivariate adaptive regression spline (WMARS) to enhance modelling performance. In this study, the performance of the models was comprehensively investigated, considering their ability to forecast daily extreme runoff. Although CemaNeige GR6J yielded a very good performance, it overestimated low flows. The WSVR and WMARS models yielded poorer performance than the conceptual and hybrid models. The hybrid model approach improved the performance of the machine learning models and the conceptual model by revealing the linkage between variables and runoff in the conceptual model, which provided more accurate results for extreme flows. Accordingly, the hybrid models improved the forecasting performance of the maximum flows up to 40 % and 61 %, and minimum flows up to 73 % and 72 % compared to the CemaNeige GR6J and stand-alone machine learning models. In this regard, the hybrid model approach can enhance the daily rainfall-runoff modelling performance in nonhomogeneous and karst catchments where the hydrological process can be more complicated.

Satellite-derived spatiotemporal data on imperviousness for improved hydrological modelling of urbanised catchments

• Performance of both lumped and distributed conceptual rainfall-runoff models were improved dramatically when temporal variations in impervious surface fraction (ISF) were used as input to these models. • Both SIMHYD and AWBM could be used for assessing urbanisation effect on streamflow of urbanized catchments. • Both lumped and distributed conceptual rainfall-runoff models showed similar performance in simulation of daily streamflows for the small urban catchment tested. • Given the observed rainfall for the study catchment, one percent increase in impervious surface fraction would lead to 6.9–7.3 mm (2.9–3%) increase in the annual streamflow during 1991–2018. Conceptual rainfall-runoff models have been used extensively to simulate streamflow in urban catchments. However, spatiotemporal changes in the impervious surface area, as an important land use indicator for urban environment, has not been used explicitly as input into conceptual rainfall-runoff models. In this study, spatiotemporal impervious surface fraction (ISF) data were incorporated into two lumped and distributed (1 km) models, namely SIMHYD and AWBM in an urbanized catchment in southeast Queensland, Australia. A sub-pixel classification technique was used to derive ISF from 6 LandSat images between 1988 and 2015. Daily rainfall-runoff models were calibrated with ISF as an input variable. To evaluate model performance for different simulation periods, four model runs including M1 (lumped models with constant ISF), M2 (lumped models with temporally varying ISF), M3 (distributed models with fixed spatial ISF) and M4 (distributed models with spatiotemporal ISF and climate data). Results showed that (1) the sub-pixel classification algorithm can be used to accurately derive ISF data with a mean absolute error (MAE) ranging from 0.07 to 0.12 for 6 images, (2) ISF in the catchment has increased gradually from 12.2% to 34.6% from 1988 to 2015, (3) the performance of SIMHYD and AWBM were not improved noticeably when distributed climate and ISF data were used, (4) simulation capability of lumped and distributed SIMHYD and AWBM were improved (e.g., increase in NSE from 0.68 to 0.73, and decrease in PBIAS from −30.2% to 1.2% for lumped SIMHYD) when temporal (M2) and/or spatiotemporal (M4) ISF data (5-year) were used, (5) difference in model performance is small and insignificantly whether annual or 5-yearly ISF data were used and (6) annual streamflow would increase by 6.9–7.3 mm (equal to 2.9–3%) in response to every 1% increase in ISF for the catchment tested. The methodology developed in this study can be applied for assessment of the effect of urbanisation on regional water balance.

Two integrated conceptual–wavelet-based data-driven model approaches for daily rainfall–runoff modelling

Abstract Rainfall–runoff modelling is crucial for enhancing the effectiveness and sustainability of water resources. Conceptual models can have difficulties, such as coping with nonlinearity and needing more data, whereas data-driven models can be deprived of reflecting the physical process of the basin. In this regard, two hybrid model approaches, namely Génie Rural à 4 paramètres Journalier (GR4 J)–wavelet-based data-driven models (i.e., wavelet-based genetic algorithm–artificial neural network (WGANN); GR4 J–WGANN1 and GR4 J–WGANN2), were implemented to improve daily rainfall–runoff modelling. The novel GR4 J–WGANN1 hybrid model includes the outflow (QR) and direct flow (QD) obtained from the GR4 J model, and the GR4 J–WGANN2 hybrid model includes the soil moisture index (SMI) obtained from the GR4 J model as input data. In hybrid models, wavelet analysis and the Boruta algorithm were implemented to decompose input data and select wavelet components. Four gauging stations in the Eastern Black Sea and Kızılırmak basins in Turkey were used to observe modelling performance. The GR4 J model exhibited poor performance for extreme flow forecasting. The novel GR4 J–WGANN1 approach performed better than the GR4 J–WGANN2 model, and the hybrid models improved modelling performance up to 40% compared to the GR4 J model. In this regard, integrated conceptual–wavelet-based data-driven models can be useful for improving the conceptual model performance, especially regarding extreme flow forecasting.

New hybrid GR6J-wavelet-based genetic algorithm-artificial neural network (GR6J-WGANN) conceptual-data-driven model approaches for daily rainfall–runoff modelling

New hybrid GR6J-wavelet-based genetic algorithm-artificial neural network (GR6J-WGANN) conceptual-data-driven model approaches for daily rainfall–runoff modelling

A comparison between lumped and distributed hydrological models for daily rainfall-runoff simulation

In Uruguay, the Santa Lucía Chico watershed has been studied in several hydrologic/hydraulic works due to its economic and social importance. However, few studies have been focused on water balance computation in this watershed. In this work, two daily rainfall-runoff models, a distributed (SWAT) and a lumped one (GR4J), were implemented at two subbasins of the Santa Lucía Chico watershed, with the aim of providing a thorough comparison for simulating daily hydrographs and identify possible scenarios in which each approach is more suitable than the other. Results showed that a distributed and complex model like SWAT performs better in watersheds characterized by anthropic interventions such as dams, which can be explicitly represented. On the other hand, for watersheds with no significant reservoirs, the use of a complex model may not be justified due to the higher effort required in modeling design, implementation, and computational cost, which is not reflected in a significant improvement of model performance.

An Approach toward Integrated Management for Wadi Haseeb Watershed using Remote Sensing GIS and Hydrological modeling System, East Khartoum, Sudan

This study has been conducted in Wadi Haseeb Watershed, which originates from the upland in South East of Khartoum State. It descends in the North West direction to the confluence with the Blue Nile. The objective of this study was to specify an integrated management strategy to harness the potentiality of the watershed for sustainable agricultural purposes. An integrated approach of remote sensing (RS), geographic information system (GIS), hydrologic engineering center geospatial hydrologic modeling extension (HEC-GeoHMS) and hydrologic modeling system (HMS4.0) was performed to develop a daily rainfall-runoff model. The watershed was delineated and divided into four basins yielding an average water volume of 23 million cubic meters month-1. The results indicated that the first scenario needed 48 MCM month-1 whereas watershed discharge was 23 MCM month-1 with a deficit of 25 MCM month-1. The second scenario needed 32 MCM month-1, leading to a deficit of 9 MCM month-1. The third scenario needed 4 MCM month-1 with a water surplus of 19 MCM month-1. However, it was not economically and technically feasible to construct a canalization system for this small area. The questionnaire analysis indicated that high flood levels increased the cultivated area of some farms from 0.8 to 22 ha and the yield of sorghum from 400 to 2500 kg ha-1. The effect of the flood on the cultivated area and crop yield was highly significant (P ≤ 0.01) during high flood periods and significant (P ≤ 0.05) during the low floods.

A comparison of particle swarm optimization and genetic algorithm for daily rainfall-runoff modelling: a case study for Southeast Queensland, Australia

Real-time and short-term prediction of river flow is essential for efficient flood management. To obtain accurate flow predictions, a reliable rainfall-runoff model must be used. This study proposes the application of two evolutionary algorithms, particle swarm optimization (PSO) and genetic algorithm (GA), to train the artificial neural network (ANN) parameters in order to overcome the ANN drawbacks, such as slow learning speed and frequent trapping at local optimum. These hybrid ANN-PSO and ANN-GA approaches were validated to equip natural hazard decision makers with a robust tool for forecasting real-time streamflow as a function of combinations of different lagged rainfall and streamflow in a small catchment in Southeast Queensland, Australia. Different input combinations of lagged rainfall and streamflow (delays of one, two and three days) were tested to investigate the sensitivity of the model to the number of delayed days, and to identify the effective model input combinations for the accurate prediction of real-time streamflow, which has not yet been recognized in other studies. The results indicated that the ANN-PSO model significantly outperformed the ANN-GA model in terms of convergence speed, accuracy, and fitness function evaluation. Additionally, it was found that the rainfall and streamflow with 3-day lag time had less impact on the predicted streamflow of the studied basin, confirming that the flow of the studied river is significantly correlated with only 2-day lagged rainfall and streamflow.

A robust approach for calibrating a daily rainfall-runoff model to monthly streamflow data

Development of robust approaches for calibrating daily rainfall-runoff models to monthly streamflow data is of major practical interest. Such approaches would enable widely used hydrological modelling platforms that operate at daily time step to be applied in practical situations where precipitation is available at the daily scale, but observed streamflow is available only at the monthly scale (e.g. predicting inflows into large dams). This study compares the performance of a hydrological model running at daily time step (GR4J) that is calibrated against daily and monthly streamflow data using a wide range of metrics: fit of the daily and monthly flow duration curve, daily and monthly pattern metrics, and long-term bias. The comparison is undertaken for 508 Australian catchments, two evaluation periods and four objective functions (including sum-of-squared-errors of Box-Cox transformed streamflow and the Kling-Gupta efficiency). Monthly calibration performs similar or better than daily calibration in a majority of sites and periods in terms of bias and fit of the flow duration curve. This result holds even when the flow duration curve is computed at the daily time step, which constitutes a major finding of this study. However, performance of monthly calibration is worse than daily calibration for daily pattern metrics such as Nash-Sutcliffe efficiency in a majority of sites and periods. This performance loss can be reduced significantly by using regionalised values for the flow-timing parameter of GR4J. Similar results are obtained for other pattern metrics and all objective functions. These findings suggest that monthly calibration of rainfall-runoff models to daily-rainfall/monthly-streamflow is a viable alternative to daily calibration when no daily streamflow data are available.

Curve Number Calibrations on 204 Australian Catchments

AbstractThe Australian water balance model (AWBM) is automatically calibrated to 204 data sets from the main stream-gauging network in Australia. This daily rainfall-runoff model calculates surface...

Dependence of regionalization methods on the complexity of hydrological models in multiple climatic regions

Hydrological models have been widely used to predict runoff in regions with observed discharge data, and regionalization methods have been extensively discussed for providing runoff predictions in ungauged basins (PUB), especially during the PUB decade (2003–2012). Great progress has been achieved in the field of regionalization in previous studies, in which different hydrological models have been coupled with various regionalization methods. However, different conclusions have been drawn due to the use of different hydrological models, regionalization methods, and study regions. In this study, we assessed the performance of the five most widely used regionalization methods (spatial proximity with parameter averaging option (SP-par), spatial proximity with output averaging option (SP-out), physical similarity with parameter averaging option (Phy-par), physical similarity with output averaging option (Phy-out), and regression methods (PCR)) and four daily rainfall-runoff models (GR4J, WASMOD, HBV and XAJ, with 6, 8, 13, and 17 parameters, respectively) at the same time. Our aim was to evaluate how the performance of the regionalization methods depends on (a) the selection of hydrological models, (b) nonstationary climate conditions, and (c) different climatic regions. This investigation used data from 86 independent catchments evenly distributed throughout Norway, covering three different climate zones (oceanic, continental and polar tundra) according to the Köppen-Geiger classification. The results showed that (a) the SP-out and Phy-out methods performed better than the SP-par and Phy-par for all the hydrological models, and the regression method performed worst in most cases; (b) the difference between the parameter averaging option and the output averaging option is positively related to the number of hydrological model parameters, i.e. the greater the number of parameters, the larger the difference between the two options; (c) the XAJ model with the greatest number of parameters produced the best results in most cases, and models with fewer parameters tend to produce similar performance for the different regionalization methods; (d) models with more parameters displayed larger declines in performance than those with fewer parameters for nonstationary conditions; and (e) clear differences in the performance of the regionalization methods exist among the three climatic regions. This study provides insight into the relationship between the complexity of hydrological models and regionalization methods in cold and seasonally snow-covered regions.

Exploring parsimonious daily rainfall-runoff model structure using the hyperbolic tangent function and Tank model

Mathematical models with small numbers of parameters are useful for hydrological analyses and water resources planning as they are easier to prepare, use, and interpret compared to more complex models. Many simple model structures and simulation strategies have been developed, but there is still substantial room for improvement in the efficiency of daily rainfall-runoff modeling. This study proposes daily rainfall-runoff models that require only two to four parameters yet are capable of predicting peak discharge and baseflow. Parsimonious model structures were obtained using the hyperbolic tangent function and combining it with one subsurface layer and two parameters of a Tank model. The daily rainfall-runoff models proposed based on the Two-Parameter Hyperbolic Model (TPHM) and Tank model were applied to 41 watersheds throughout South Korea. Model accuracy was evaluated by comparing observed and simulated watershed streamflow and compared with that of a nine-parameter Tank model. The original TPHM and its variants yielded slightly better efficiency with a smaller number of parameters than the Tank model when reproducing high flows on a daily basis. The accuracy of low flow and flow duration curve predictions provided by the TPHM-based models was higher in watersheds with relatively large baseflow temporal variability. The results suggested that the TPHM-based structures be alternative parsimonious daily rainfall-runoff simulation models. This study demonstrates that daily high streamflow can be efficiently simulated using a two-parameter hydrological model, and daily baseflow simulation can be improved by adding two additional parameters to the model.

Machine Learning Models Coupled with Variational Mode Decomposition: A New Approach for Modeling Daily Rainfall-Runoff

Accurate modeling for nonlinear and nonstationary rainfall-runoff processes is essential for performing hydrologic practices effectively. This paper proposes two hybrid machine learning models (MLMs) coupled with variational mode decomposition (VMD) to enhance the accuracy for daily rainfall-runoff modeling. These hybrid MLMs consist of VMD-based extreme learning machine (VMD-ELM) and VMD-based least squares support vector regression (VMD-LSSVR). The VMD is employed to decompose original input and target time series into sub-time series called intrinsic mode functions (IMFs). The ELM and LSSVR models are selected for developing daily rainfall-runoff models utilizing the IMFs as inputs. The performances of VMD-ELM and VMD-LSSVR models are evaluated utilizing efficiency and effectiveness indices. Their performances are also compared with those of VMD-based artificial neural network (VMD-ANN), discrete wavelet transform (DWT)-based MLMs (DWT-ELM, DWT-LSSVR, and DWT-ANN) and single MLMs (ELM, LSSVR, and ANN). As a result, the VMD-based MLMs provide better accuracy compared with the single MLMs and yield slightly better performance than the DWT-based MLMs. Among all models, the VMD-ELM and VMD-LSSVR models achieve the best performance in daily rainfall-runoff modeling with respect to efficiency and effectiveness. Therefore, the VMD-ELM and VMD-LSSVR models can be an alternative tool for reliable and accurate daily rainfall-runoff modeling.

Daily rainfall-runoff modelling by neural networks in semi-arid zone: case of Wadi Ouahrane’s basin

This research work will allow checking efficiency of formal neural networks for flows’ modelling of wadi Ouahrane’s basin from rainfall-runoff relation which is non-linear. Two models of neural networks were optimized through supervised learning and compared in order to achieve this goal, the first model with input rain, and the second one with rain and input ETP. These neuronal models were compared with another overall model, the GR4j model. Then, it has been optimized and compared with the three first models, a third model of neural network with rain, ETP and soil moisture (calculated by the model GR4j) input. The neuronal models were optimized with algorithm of Levenberg Marquarld (LM), while the GR4j model was optimized with SCE-UA method. The Nash criterion (%) and the correlation coefficient of Pearson (R) allowed appreciating performances of these models.

Application of PCA and Clustering Methods in Input Selection of Hybrid Runoff Models

This study has proposed and investigated a novel input variable selection method for nonlinear modelling based on principle component analysis (PCA) and cluster analysis. The proposed approach was applied to daily rainfall-runoff modelling of the Brue catchment of the United Kingdom using wavelet based hybrid forms of two nonlinear models, Artificial Neural Networks (ANNs) and Local Linear Regression (LLR), to identify meaningful wavelet decomposed sub-series. The homogenous group formation capability of cluster analysis and redundancy assessment capability of PCA were applied effectively in this study to solve input selection uncertainties associated with wavelet based hybrid models. Though this concept has been represented in the selection of effective wavelet decomposed subseries in runoff modelling, the application has gotten wider implications in time series modelling with highly redundant and large input space. The study revealed the weakness of conventional forms of cross-correlation analysis and also suggested that input selection could be improved by making sufficient natural clusters (equal to the desired number of input data series) of input space and restricting the search within each cluster according to silhouette or correlation value. The study also highlighted the higher modelling capability of ANN over traditional LLR models in rainfall-runoff modelling of the Brue catchment.

Application of Satellite-Based Precipitation Estimates to Rainfall-Runoff Modelling in a Data-Scarce Semi-Arid Catchment

Rainfall-runoff modelling is a useful tool for water resources management. This study presents a simple daily rainfall-runoff model, based on the water balance equation, which we apply to the 11,630 km2 Lesser Zab catchment in northeast Iraq. The model was forced by either observed daily rain gauge data from four stations in the catchment or satellite-derived rainfall estimates from two TRMM Multi-satellite Precipitation Analysis (TMPA) data products (TMPA-3B42 and 3B42RT) based on the Tropical Rainfall Measuring Mission (TRMM) from 2003 to 2014. As well as using raw TMPA data, we used a bias-correction method to adjust TMPA values based on rain gauge data. The uncorrected TMPA data products underestimated observed mean catchment rainfall by −10.1% and −10.7%. Corrected data also slightly underestimated gauged rainfall by −0.7% and −1.6%, respectively. Nash-Sutcliffe Efficiency (NSE) and Pearson’s Correlation Coefficient (r) for the model fit with the observed hydrograph were 0.75 and 0.87, respectively, for a calibration period (2010–2011) using gauged rainfall data. Model validation performance (2012–2014) was best (highest NSE and r; lowest RMSE and bias) using the corrected 3B42 data product and poorest when driven by uncorrected 3B42RT data. Uncertainty and equifinality were also explored. Our results suggest that TRMM data can be used to drive rainfall-runoff modelling in semi-arid catchments, particularly when corrected using rain gauge data.

Evaluating Regional and Global Hydrological Models against Streamflow and Evapotranspiration Measurements

Abstract Land surface and global hydrological models are often used to characterize global water and energy fluxes and stores and to model their future trajectories. This study evaluates estimates of streamflow and evapotranspiration (ET) obtained with a priori parameterization from a land surface model [CSIRO Atmosphere Biosphere Land Exchange (CABLE)] and a global hydrological model (H08) against a global dataset of streamflow from 644 largely unregulated catchments and ET from 98 flux towers and benchmarks their performance against two lumped conceptual daily rainfall–runoff models [modèle du Génie Rural à 4 paramètres Journalier (GR4J) and a simplified version of the HYDROLOG model (SIMHYD)]. The results show that all four models perform poorly in simulating the monthly and annual runoff values, with the rainfall–runoff models outperforming both CABLE and H08. The model biases in runoff are generally reflected as a complementary opposite bias in ET. All models can generally reproduce the observed seasonal and interannual runoff variability. The correlations between the modeled and observed runoff time series are reasonable, with the rainfall–runoff models performing slightly better than CABLE and H08 at the monthly time scale and all four models performing similarly at the annual time scale. The results suggest that while the land surface and global hydrological models cannot adequately simulate the actual runoff time series and long-term average volumes, they can reasonably simulate the monthly and interannual runoff variability and trends and can therefore be reliably used for broadscale or comparative regional and global water and energy balance assessments and simulations of future trajectories. They can be improved through validating the models or calibrating some of the more sensitive and less physically based parameters.

Modelling the Climate Change Impact On Monthly Runoff in Central Slovakia

The aim of this paper was to evaluate the possible impact of climate change on the runoff regime of the Hron River basin up to the Banská Bystrica profile until the year 2100. The daily rainfall-runoff model Hron which was developed at the Department of Land and Water Resources Management, SUT Bratislava, was used. The model was calibrated on a period between 1981-1995 and validated on 1996-2010 For the simulation of the possible impact of climate change, the KNMI and MPI climate change scenarios, which illustrate changes in daily precipitation, daily air temperature and the average air humidity in the river basin for future periods, were used. In conclusion the average monthly flows for the future time horizons of 2011-2040, 2041-2070 and 2071-2100 with a reference period (1981-2010) were compared. Based on the results of the two hydrological models and the two different climate change scenarios, we can expect an increase in long-term mean monthly discharges for the winter and spring periods and a decrease for the summer period.

新安江日降雨径流模型在安徽省沿江江南地区应用

当集水面积相差较大时，采用水文比拟法计算设计径流，计算成果可能误差较大，采用此法计算新建水库设计径流时，可能会带来水库设计水位的偏差。通过建立皖东南宣城市广德县粮长门水库新安江日降雨径流模型，计算粮长门水库设计径流，并移用到皖西南安庆市破罡湖流域进行验证，得到较为满意的效果。 When the difference of catchment areas is large, the errors of calculation result may be compara-tively large using the hydrologic analogy method to calculate design runoff. Using this method to calculate design runoff of a new reservoir may bring deviation of the reservoir’s design water level. Through the establishment of the Xin’an River daily rainfall-runoff model of the Liang Changmen Reservoir in Guangde, Xuancheng, southeast of Anhui, runoff of the Liang Changmen Reservoir was calculated; the model was verified by applying it to the Pogang Lake Basin in Anqing, southwest of Anhui, and satisfactory results were got.

Rainfall Runoff Modelling Using Generalized Neural Network and Radial Basis Network

Rainfall runoff study has a wide scope in water resource management. To provide a reliable prediction model is of paramount importance. Runoff prediction is carried out using generalized regression neural network and radial basis neural network. Daily Rainfall runoff model was developed for Nethravathi river basin located at the west coast of Karnataka, India. The comparative study showed Radial basis neural network performed better than generalized neural network during its evaluation by performance indicators

Water resources of Cyprus under changing climatic conditions: Modelling approach, validation and limitations

We examine trends in the water resources of Cyprus by focussing on water flux changes in the important Kouris catchment. Our modelling approach is general and is a synthesis of an adapted conceptual daily rainfall-runoff model, radiation transfer models that use high resolution MODIS satellite climatological data and GCM scenarios for future climatic change. We used climatic data as input to our models, downscaled to the catchment resolution from two climate scenarios: the mild RCP2.6 and the extreme RCP8.5, to estimate water resources by the end of the 21st century. The models show that the present mean annual rainfall resource of 174 Mm3 will be reduced to 162 Mm3 and 132 Mm3, for the mild and extreme scenario, respectively. The present mean discharge of 21.5 Mm3 into the Kouris dam from the catchment will decrease to 16.6 Mm3 and 6.9 Mm3 under the mild and extreme scenario, respectively.