Rainfall-runoff Model Research Articles

Using normalised difference infrared index patterns to constrain semi-distributed rainfall–runoff models in tropical nested catchments

Abstract. A parsimonious semi-distributed rainfall–runoff model has been developed for flow prediction. In distribution, attention is paid to both the timing of the runoff and the heterogeneity of moisture storage capacities within sub-catchments. This model is based on the lumped FLEXL model structure, which has proven its value in a wide range of catchments. To test the value of distribution, the gauged upper Ping catchment in Thailand has been divided into 32 sub-catchments, which can be grouped into five gauged sub-catchments at which internal performance is evaluated. To test the effect of timing, first the excess rainfall was calculated for each sub-catchment, using the model structure of FLEXL. The excess rainfall was then routed to its outlet using the lag time from the storm to peak flow (TlagF) and the lag time of recharge from the root zone to the groundwater (TlagS), as a function of catchment size. Subsequently, the Muskingum equation was used to route sub-catchment runoff to the downstream sub-catchment, with the delay time parameter of the Muskingum equation being a function of channel length. Other model parameters of this semi-distributed FLEX-SD model were kept the same as in the calibrated FLEXL model of the entire upper Ping River basin (UPRB), controlled by station P.1 located at the centre of Chiang Mai province. The outcome of FLEX-SD was compared to the (1) observations at the internal stations, (2) calibrated FLEXL model, and (3) the semi-distributed URBS model – another established semi-distributed rainfall–runoff model. FLEX-SD showed better or similar performance during calibration and especially in validation. Subsequently, we tried to distribute the moisture storage capacity by constraining FLEX-SD on patterns of the NDII (normalised difference infrared index). The readily available NDII appears to be a good proxy for moisture stress in the root zone during dry periods. The maximum moisture-holding capacity in the root zone is assumed to be a function of the maximum seasonal range of NDII values and the annual average NDII values to construct two alternative models, namely FLEX-SD-NDIIMaxMin and FLEX-SD-NDIIAvg. The additional constraint on the moisture-holding capacity (Sumax) by NDII, particularly in FLEX-SD-NDIIAvg, improved both the model performance and the realism of its distribution across the UPRB, which corresponds linearly to the percentage of evergreen forests (R2=0.69). To check how well the models represents simulated root zone soil moisture (Sui), the performance of the FLEX-SD-NDII models was compared to the time series of the soil water index (SWI). The correlation between the Sui and the daily SWI appeared to be very good and was even better than the correlation with the NDII, which does not provide good estimates during wet periods. The SWI, which is model-based, was not used for calibration but appeared to be an appropriate index for validation.

Future changes in M-day minimum and maximum discharges in basins of Slovakia

AbstractThis study evaluates future changes in M-day minimum and maximum discharges in selected river basins of Slovakia, which have been divided into four groups, i.e., western, central, northern, and eastern Slovakia. Four types of data were available for the analysis, i.e., observed mean daily discharges, modeled mean daily discharges using the rainfall-runoff model, and simulated mean daily discharges according to the climate scenarios. The Indicators of Hydrological Alteration program was used to estimate the M-day discharges.The results revealed an increase in the M-day minimum discharges in northern Slovakia. Eastern and Western Slovakia show a decrease in M-day minimum discharges and an increase in m-daily maximum discharges by 2100.

Predicting streamflow with LSTM networks using global datasets

Streamflow predictions remain a challenge for poorly gauged and ungauged catchments. Recent research has shown that deep learning methods based on Long Short-Term Memory (LSTM) cells outperform process-based hydrological models for rainfall-runoff modeling, opening new possibilities for prediction in ungauged basins (PUB). These studies usually feature local datasets for model development, while predictions in ungauged basins at a global scale require training on global datasets. In this study, we develop LSTM models for over 500 catchments from the CAMELS-US data base using global ERA5 meteorological forcing and global catchment characteristics retrieved with the HydroMT tool. Comparison against an LSTM trained with local datasets shows that, while the latter generally yields superior performances due to the higher spatial resolution meteorological forcing (overall median daily NSE 0.54 vs. 0.71), training with ERA5 results in higher NSE in most catchments of Western and North-Western US (median daily NSE of 0.83 vs. 0.78). No significant changes in performance occur when substituting local with global data sources for deriving the catchment characteristics. These results encourage further research to develop LSTM models for worldwide predictions of streamflow in ungauged basins using available global datasets. Promising directions include training the models with streamflow data from different regions of the world and with higher quality meteorological forcing.

Urban Rainfall-Runoff Modeling Using HEC-HMS and Artificial Neural Networks: A Case Study

Urban flooding nowadays becomes common throughout the world. The main reason for these floods is rapid urban development and climate change. During the monsoon, the flows in the urban drains will be high and the main reason for these high flows is the existence of a combined network system (i.e. drainage and stormwater). Further, the flow in the city (under study) drainage network was very high and some areas of the network exceeds more than discharge carrying capacity. Hence, this may result in overflow from the manholes and create an overland flood problem. Rainfall-Runoff modeling in these situations in the urban catchment will be essential and required to understand the flow pattern that helps in flood management. Therefore, the current study chose Hydrologic Modeling System (HEC-HMS) and Artificial Neural Network (ANN) for rainfall-runoff modeling at an hourly period for the Kukataplly (zone-12) watershed of Hyderabad city, Telangana State in India. This zone-12 watershed was one of the most affected hydraulic zones of Greater Hyderabad Municipal Corporation (GHMC) during the monsoon period in the past 21 years. The present study focuses on a comparative study between HEC-HMS and ANN has been carried out to comprehend the flood scenario in the study area. Finally, the performance of the model is checked with statistical indices such as Nash-Sutcliff Efficiency (NSE), and Coefficient of Determination (R2). HEC-HMS yielded good results (NSE = 0.74 and R2 = 0.76) when it has taken care of the maximum possible nonlinear complex data to be analysed.

Prospective analysis of low flows in a climate change context in West Africa: case of the N’zo-Sassandra watershed (West of Côte d’Ivoire)

The objective of this study is to analyze the prospective of the low water flows of the N'zo-Sassandra River located in the West of Ivory Coast. The hydroclimatic data used (rainfall, ETP, flow) extend on the one hand over the historical period (1961-2017) and on the other hand over the future period 2071-2100 (CORDEX data). The methodological approach adopted is based on the analysis of past low flows, the climate prospective analysis and the hydrological prospective analysis. The hydrological impact variable used is the annual minimum monthly mean flow (MAMF). Thus, a pessimistic climate forecast scenario was used (RCP 8.5.). This climate scenario was then fed into a rainfall-runoff hydrological model (GR2M model) calibrated to the 1961-1990 reference period and validated over the 1991-2017 period. In response to the climate forcing, future monthly low-flow rates (MAMF) were simulated. An increase in the number of flows in the class ]0-4] m3/s was observed, reflecting low flows in the 2071–2100-time frame. The best statistical law retained following the frequency analysis of the low-water flows is the lognormal law. By 2071-2100, a decrease in rainfall (-19.2%) and an increase in temperature (+3.9°C) are predicted in the N'zo-Sassandra watershed. Frequent low-water flows estimated from the lognormal law have shown a bias of -50.38%, which translates into a decrease of half of the frequent low-water flows by the end of the 21st century compared to the current period. This decrease in flows will have consequences on the availability of water resources for irrigation, fishing, breeding, drinking water supply, navigation, etc. This is why we propose as alternative resources for agricultural needs, groundwater which is less impacted by global changes.

In Defense of Metrics: Metrics Sufficiently Encode Typical Human Preferences Regarding Hydrological Model Performance.

Building accurate rainfall-runoff models is an integral part of hydrological science and practice. The variety of modeling goals and applications have led to a large suite of evaluation metrics for these models. Yet, hydrologists still put considerable trust into visual judgment, although it is unclear whether such judgment agrees or disagrees with existing quantitative metrics. In this study, we tasked 622 experts to compare and judge more than 14,000 pairs of hydrographs from 13 different models. Our results show that expert opinion broadly agrees with quantitative metrics and results in a clear preference for a Machine Learning model over traditional hydrological models. The expert opinions are, however, subject to significant amounts of inconsistency. Nevertheless, where experts agree, we can predict their opinion purely from quantitative metrics, which indicates that the metrics sufficiently encode human preferences in a small set of numbers. While there remains room for improvement of quantitative metrics, we suggest that the hydrologic community should reinforce their benchmarking efforts and put more trust in these metrics.

Changes in the hydrological and sediment regimes of two neighboring catchments in the past sixty years

Despite advances in watershed hydrology research, little is known about the hydrologic differences and similarities between neighboring catchments. Two neighboring catchments, the Nanliu River catchment (NLRC) and the Qinjiang River catchment (QJRC), were studied to see how they differed in annual precipitation, runoff, and sediment load in the last 60 years. Four hydrological signatures (river regime, periodicity, change trends, and precipitation-runoff-sediment correlations) estimated from a wide variety of catchment classification and corresponding indices, such as the runoff ratio (RQP) and baseflow index (BFI). The findings revealed that the two catchments shared similar general features, including similar runoff patterns (RQP= 0.45 and 0.43, and BFI = 2.83% and 2.12% in NLRC and QJRC respectively), minor differences in periodicities, and similar annual precipitation, runoff, and sediment load changes. The annual sediment load has been steadily decreasing due to human activity. The simple precipitation-runoff regression model developed in one catchment can be transferred to another. Nash–Sutcliffe efficiency (NSE) coefficients were 0.79 and 0.62, respectively. The runoff-sediment load regression model, on the other hand, showed poor performance (NSE of −0.34 and −0.99, respectively), despite capturing the main patterns with correlation coefficients of 0.82 (p < 0.001) and 0.74 (p < 0.001), respectively. These findings suggested that (1) annual runoff ratios are influenced more by climate (particularly precipitation) than by surface and subsurface characteristics of the earth. The rainfall-runoff model for the neighboring catchment can recreate runoff dynamics at yearly scales. Human activities, such as dam construction and sand mining, significantly impacted the annual sediment yield. (2) The two neighboring catchments had similar hydrologic ages because of similar climate, geology, tectonics, and human disturbance. These findings provided a solid evidence for predicting runoff and sediment in ungauged basins.

Assessing CFSR climate data for rainfall-runoff modeling over an ungauged basin between Iraq and Iran

In most developing countries, networks of gauge observations are frequently sparse and may have significant spatial and temporal gaps in the meteorological records. Reanalysis products are extensively used as alternative climate inputs for watershed modeling that provide continuous data worldwide. However, the direct utilization of such products is highly doubtful at different basin scales. This article aims to evaluate the accuracy of the Climate Forecast System Reanalysis (CFSR) product for streamflow simulation using the Soil and Water Assessment Tool (SWAT) model in the northeastern part of the Little Zab River Basin (LZRB), a mountainous basin between Iraq and Iran. The quality of CFSR was assessed from two viewpoints: (1) the precipitation aspect, an analysis of the CFSR precipitation compared to observational meteorological data (OMD) from 2004 to 2013 on a daily time scale, and (2) the hydrology aspect, a comparison of the SWAT model performance based on OMD (scenario I) and the CFSR product (scenario II) for monthly streamflow simulations. Results showed that the CFSR precipitation had a satisfactory correlation with rain gauge data ranging from 0.55 to 0.81. The performance of OMD in SWAT-based streamflow simulation achieved a ‘very good’ rating in calibration (2006–2010) and validation (2011–2013). Meanwhile, the CFSR-based model demonstrated a ‘very good’ hydrologic prediction skill in calibration. In contrast, CFSR revealed a lower performance in the validation period as it considerably overestimated base and peak flows in LZRB. Nevertheless, model uncertainty analysis after calibration showed that more than 88% of measured streamflow values at the basin outlet were bracketed within the 95 percent prediction uncertainty (95PPU) band for OMD and CFSR. These findings revealed that the CFSR product, after proper bias correction, could be a reliable alternative climate input for hydrologic modeling and climatological studies in semi-arid basins where gauge data is limited.

Why do our rainfall–runoff models keep underestimating the peak flows?

Abstract. In this paper, the question of how the interpolation of precipitation in space by using various spatial gauge densities affects the rainfall–runoff model discharge if all other input variables are kept constant is investigated. The main focus was on the peak flows. This was done by using a physically based model as the reference with a reconstructed spatially variable precipitation model and a conceptual model calibrated to match the reference model's output as closely as possible. Both models were run with distributed and lumped inputs. Results showed that all considered interpolation methods resulted in the underestimation of the total precipitation volume and that the underestimation was directly proportional to the precipitation amount. More importantly, the underestimation of peaks was very severe for low observation densities and disappeared only for very high-density precipitation observation networks. This result was confirmed by using observed precipitation with different observation densities. Model runoffs showed worse performance for their highest discharges. Using lumped inputs for the models showed deteriorating performance for peak flows as well, even when using simulated precipitation.

Integrated SUSTAIN-SWMM-MCDM Approach for Optimal Selection of LID Practices in Urban Stormwater Systems

Rapid urbanization has increased impervious areas, leading to a higher flood hazard across cities worldwide. Low Impact Development (LID) practices have shown efficacy in reducing urban runoff; nevertheless, choosing the best combinations in terms of implementation cost and performance is of great importance. The present study introduces a framework based on green infrastructure, multi-objective optimization, and decision support tools to determine the most cost-effective LID solutions. The Storm Water Management Model (SWMM) was employed for rainfall-runoff and hydraulic modeling in Region 1, District 11 of Tehran, Iran. Six scenarios of different combinations of LID practices were developed. The system for Urban Stormwater Treatment and Analysis Integration (SUSTAIN) was used to optimize and evaluate each scenario. The selected solutions were imported to the SWMM to evaluate the stormwater system performance. Then, two multi criteria decision making (MCDM) models, including TOPSIS and COPRAS, were employed to rank the scenarios based on four technical and economic criteria. Results showed that scenario 4, consisting of rain barrels, porous pavements, and vegetated swales, had the best performance under TOPSIS with a 7.68 million USD and reduced the runoff volume and peak flow by 20.77% and 19.2%, respectively. However, Under the COPRAS method, Scenario 2 with a combination of rain barrels, bio-retention cells, and vegetated swales showed higher performance than the other scenarios with 3.25 million USD and led to a 15% reduction in the runoff volume and 4.30% in the peak flow. The COPRAS method was more sensitive to cost weights and chose the most economical scenario as the ideal. However, Scenario 4 concluded to be more feasible due to spatial limitations in the study area. The proposed SWMM—SUSTAIN—MCDM framework could be helpful to decision-makers in the design, performance evaluation, cost estimation, and selection of optimal scenarios.

Assessment and Comparison of Satellite-Based Rainfall Products: Validation by Hydrological Modeling Using ANN in a Semi-Arid Zone

Several satellite precipitation estimates are becoming available globally, offering new possibilities for modeling water resources, especially in regions where data are scarce. This work provides the first validation of four satellite precipitation products, CHIRPS v2, Tamsat, Persiann CDR and TerraClimate data, in a semi-arid region of Essaouira city (Morocco). The precipitation data from different satellites are first compared with the ground observations from 4 rain gauges measurement stations using the different comparison methods, namely: Pearson correlation coefficient (r), Bias, mean square error (RMSE), Nash-Sutcliffe efficiency coefficient and mean absolute error (MAE). Secondly, a rainfall-runoff modeling for a basin of the study area (Ksob Basin S = 1483 km2) was carried out based on artificial neural networks type MLP (Multi Layers Perceptron). This model was -then used to evaluate the best satellite products for estimating the discharge. The results indicate that TerraClimate is the most appropriate product for estimating precipitation (R2 = 0.77 and 0.62 for the training and validation phase, respectively). By using this product in combination with hydrological modeling based on ANN (Artificial Neural Network) approach, the simulations of the monthly flow in the watershed were not very satisfactory. However, a clear improvement of the flow estimations occurred when the ESA-CCI (European Space Agency’s (ESA) Climate Change Initiative (CCI)) soil moisture was added (training phase: R2 = 0.88, validation phase: R2 = 0.69 and Nash ≥ 92%). The results offer interesting prospects for modeling the water resources of the coastal zone watersheds with this data.

Evaluating the Effectiveness of Low Impact Development Practices against Climate Induced Extreme Floods

Short but extreme flooding events have been frequent and severe globally due to climate change and urbanization in recent years. Similarly, researchers, scientists, and water managers are suggesting the application of sustainable flood management strategies such as Low Impact Development (LID) to mitigate the impacts of such extreme flooding events. However, most of these strategies have primarily been evaluated using historical precipitation events, which may not accurately represent the impact of climate-induced flooding events, which are projected to become more extreme. In this context, this study assesses the effectiveness of LIDs in combating climate change-induced flooding events. The North American Regional Climate Change Assessment Program (NARCCAP) climate model was applied in this study to quantify the magnitude of future projected storm depths, which are expected to increase due to climate change. Similarly, Personal Computer Storm Water Management Model (PCSWMM) was used to develop a rainfall-runoff simulation model and to assess the effectiveness of three LID techniques (Permeable Pavement, Green Roof, and Bio-Retention Cell) in reducing surface runoff under various climate scenarios. The results revealed that under the climate change scenarios the future projected design depths are expected to increase by up to 104%. Similarly, peak discharge, and total flooding volume were found to increase by 37.72% and 88.73%, respectively under the most extreme climate change scenario. Furthermore, the study demonstrated that applying LID strategies decreases peak discharge, offering a viable solution to tackle flooding events induced by climate change. The results illustrated the performance of permeable pavement was superior in reducing the peak discharge by up to 28.57%. Similarly, applying green roofs and bioretention cells reduced the peak discharge by up to 19.93% and 14.25%, respectively.

The SWTools R package for SILO data acquisition, homogeneity testing and correction

ABSTRACT Rainfall is a fundamental input to water resources models. Data products are making rainfall data available in application ready formats. There is a temptation to adopt these datasets directly without undertaking thorough quality assurance checks. The SWTools R package is introduced to download, summarise, check and, if necessary, correct rainfall data from a commonly used data product in Australia, the SILO database. A case study is presented where interpolation to extend a dataset after a station closed introduced non-homogeneity when compared to adjacent sites. SWTools functions are used to identify this change and apply corrections to the data. Calibrating a rainfall–runoff model to the datasets with and without corrections had a substantial influence on the simulated water availability in the catchment, with a 30% reduction in average runoff after the rainfall corrections. This case study highlights the importance of undertaking data quality assurance checks, particularly when applying a model outside the calibration period.

Runoff predictions in new-gauged basins using two transformer-based models

In hydrology, runoff predictions are challenging when the data is lacking (e.g., predictions in un-gauged basins (PUB) and predictions with limited data (PLD)). Here, PLD refers to the case that the data is not enough for training or fine-tuning a data-driven model well (e.g., a new-gauged basin). We also name PLD as PNB (predictions in new-gauged basins). The difference between PNB and PUB is that the new-gauged basins can provide some data (e.g., runoff observations) while the un-gauged basins cannot. The long short-term memory (LSTM)-based models have shown good performance in runoff predictions due to their advanced structures. However, those structures have low level of flexibility and two nonadjacent positions cannot communicate directly. For high level of flexibility and better performance, we propose a simple Transformer-based rainfall-runoff model named RRS-Former, and want to show its power and also the power of the previously proposed RR-Former in PNB compared with LSTM-based models. The main part of RRS-Former is attention modules in which two arbitrary positions can be connected directly. Four hydrological units including 241 basins in the Catchment Attributes and Meteorology for Large-Sample Studies (CAMELS) dataset are used to compare the performance of our Transformer-based models and that of LSTM-based models. Besides using the k-fold validation to test performance for PNB, we propose a new way named unit-fold validation, in which we train the models by using the basins in three hydrological units and then test the performance using basins in the rest one hydrological unit. The results based on both k-fold and unit-fold show that our RRS-Former and RR-Former have better performance and are more reliable compared with the LSTM-based models.

Performance comparison of IHACRES, random forest and copula-based models in rainfall-runoff simulation

In this study, two models of Random Forest and copula-based simulation were used to evaluate the accuracy and efficiency of the IHACRES rainfall-runoff model in simulating the daily discharge of the Siminehroud River in the south of Lake Urmia basin, Iran. A trivariate copula-based model was created using discharge, rainfall and temperature data on a daily scale in the period 1992–2018. Vine family models and their conditional densities were used to implement the copula-based model. By calibrating the IHACRES model and also selecting the tree sequence in accordance with the data, rainfall-runoff simulations were performed in the study area. The accuracy and efficiency of the studied models were evaluated using RMSE and NSE criteria, and also violin plot and Taylor diagram. The results of comparing the error rate of rainfall-runoff simulation in the study area showed that the vine-based model reduces the RMSE statistics by about 14.5 and 16.5%, respectively, compared to the IHACRES and Random Forest models. According to the presented diagrams, the efficiency and certainty of IHACRES and copula-based simulation models are acceptable. While the Random Forest model does not have acceptable accuracy and efficiency in the study area. The copula-based simulation model has a good performance due to the unique tree sequence as well as involving the marginal distributions fitted to the data. Although the copula-based simulation model has increased the efficiency of the model in simulating the daily discharge by about 5% compared to the IHACRES model, it is not significant compared to the mathematical complexity of the copula-based model.

Optimizing the Use of Meninting Multipurpose Reservoir Water in West Lombok District

The construction of the Meninting Reservoir was planned with the objective of meeting the irrigation and domestic water demands on Lombok Island. It served as a multipurpose reservoir, with a maximum storage volume of 12.18 million m³, mainly for supplying irrigation and domestic water. The reservoir had considerable potential for water availability, which could be used to supply water to the South Lombok region with limited water availability but had agricultural land potential. Therefore, this study aimed to evaluate the potential of Meninting Reservoir water availability and its optimum utilization for irrigation and domestic purposes. The irrigation water demand for 1,559.29 ha and domestic water demand of 150 ls-1 was fulfilled by Meninting Reservoir. Water availability in this Reservoir was estimated with the F.J. Mock method of rainfall-runoff model using 25 years of daily rainfall data from Gunung Sari and Sesaot rain gauge stations. The calibration process of the rainfall-runoff parameters models employed observed discharge data from the Aiknyet water level gauge station. The formula for optimizing reservoir water release was prepared using the linear programming method based on operational water level limits, inflow discharge, irrigation, and non-irrigation water requirements, including domestic water. The optimal average annual cropping intensity was 203.96%, 215.87%, and 241.41% for dry, normal, and wet years, respectively. The service reliability of irrigation and domestic water demands reached 100% for all inflow discharge conditions. The k-factor value met the minimum limit of 0.70 and 0.85 for irrigation and domestic water demands, respectively.&#x0D;

Comparison of Two Hydrological Models, the HEC-HMS and Nash Models, for Runoff Estimation in Michałówka River

Floods are among the most devastating natural disasters in small suburban catchments. These phenomena, causing loss of life and massive property damage, pose a serious threat to the economy. Hydrological modeling is extremely important in terms of climate change, and the use of appropriate modeling can be a useful tool for flood risk prevention and mitigation. Rainfall–runoff modeling requires the selection of an appropriate hydrological model in order to obtain satisfactory results. Hydrological models are used in water resource planning and management to estimate catchment runoff. Small uncontrolled catchments play a particularly important role in hydrological phenomena, since changes in them affect flows in the recipient. Hydrologists are particularly interested in developing hydrological models that can be made with a minimum of data and parameters. Nash models and the Hydrologic Engineering Center-Hydrologic Modeling System (HEC-HMS) are examples of simple and most practical hydrologic models. These models were used in this paper to study geographic and qualitative changes in precipitation runoff due to land cover changes. The modeling was carried out for two spatial aspects relating to the years 1940 and 2018. The model allowed for the simulation of the river flow that can occur under different rainfall probabilities. The analysis of the results was used to evaluate the hydrological models used. The hundred-year flow modeled with the Nash model for 1940 was 13.4 m3∙s−1, whereas the second model gave slightly lower flow values. In addition, modeling the flow for 2018 (after changing the land cover) highlighted the increase in the flow value for both models, where again the flow volume was slightly higher for the Nash model and amounted to about 19 m3∙s−1. The flow differences for individual models were not too large. This made it possible to conclude that the simulated outflow hydrographs are in good agreement, and this means that the models accurately reproduce the flow of the Michałówka River. The study showed that rapid urbanization adversely affects hydrological processes. In addition, the study showed that a well-distributed model can outperform a global flood forecasting model, especially in terms of magnitude, as in the current study example.

Evaluating climatic change effects on EC and runoff in the near future (2020–2059) and far future (2060–2099) in arid and semi-arid watersheds

This study investigates the effect of climate change on the runoff and electrical conductivity (EC) of the Marun watershed. It used 35 general circulation models (GCMs) and the identification of unit hydrographs and component flows from rainfall, evaporation and streamflow data (IHACRES) rainfall-runoff model for the hydrological simulation. Moreover, a non-parametric regression model based on the multivariate adaptive regression splines (MARS) was utilized to estimate the EC under representative concentration pathway RCP4.5 and RCP8.5 scenarios in the near future F1 (2020–2059) and far future F2 (2060–2099) periods. Also, it used the technique for order of preference by similarity to ideal solution (TOPSIS) method to determine the best GCMs for each region and the k-nearest neighbors (KNN) technique to combine the temperature (Tmean) and precipitation (PCP) outputs and reduce the GCM uncertainty in each cell. According to the results, the highest increase of EC relative to the historical period (1966–2005) that will occur in the F1 period under the RCP4.5 and RCP8.5 scenarios is 17.43% and 15.6%, and for the F2 period is 18.46% and 11.2%, respectively, during autumn. The changes of annual Tmean, PCP, runoff, and EC in F1 period are 8.6%, 2.1%, − 10.7%, and − 11%, respectively, under the RCP4.5 scenario and 10.5%, 5.9%, − 3.5%, and − 12.2%, respectively, under the RCP8.5 scenario. The same values for the F2 period are 12.9%, − 0.1%, − 14.9%, and − 10%, respectively, under the RCP4.5 scenario and 22.6%, 5.2%, 1.2%, and − 12.8%, respectively, under the RCP8.5 scenario relative to the historical period.

Enhancing flood risk assessment through integration of ensemble learning approaches and physical-based hydrological modeling

This study aims to examine three machine learning (ML) techniques, namely random forest (RF), LightGBM, and CatBoost for flooding susceptibility maps (FSMs) in the Vietnamese Vu Gia-Thu Bon (VGTB). The results of ML are compared with those of the rainfall-runoff model, and different training dataset sizes are utilized in the performance assessment. Ten independent factors are assessed. An inventory map with approximately 850 flooding sites is based on several post-flood surveys. The inventory dataset is randomly split between training (70%) and testing (30%). The AUC-ROC results are 97.9%, 99.5%, and 99.5% for CatBoost, LightGBM, and RF, respectively. The FSMs developed by the ML methods show good agreement in terms of an extension with flood inundation maps developed using the rainfall-runoff model. The models’ FSMs showed 10–13% of the total area to be highly susceptible to flooding, consistent with RRI's flood map. The FSMs show that downstream areas (both urbanized and agricultural) are under high and very high levels of susceptibility. Additionally, different sizes of the input datasets are tested to determine the least number of data points having acceptable reliability. The results demonstrate that the ML methods can realistically predict FSMs, regardless of the number of training samples.

Development of simple semi-distributed approaches for modelling complex rainfall–runoff process

ABSTRACT The semi-distributed approaches proposed in the past for simulating rainfall–runoff (RR) still require extensive hydro-meteorological data for their complex calibration process. Therefore, research efforts are needed to develop novel and innovative semi-distributed RR models that require a minimum amount of data and effort. In this study, four semi-distributed RR models are proposed using a simple lumped model in a distributed sense by gradually enforcing spatial distribution in terms of hydro-meteorological and physiographical features in a basin. Only rainfall, runoff, and temperature data from three contrasting basins were employed in developing the proposed models. The results show that semi-distributed models performed better than the lumped model, and the accuracy increased with gradual enforcement of the spatial variations in various data. Further, the results suggest that incorporation of the basin’s physiographic features is the most important aspect in developing efficient semi-distributed RR models followed by hydrological and climatic information in a basin.