Journal of Hydrologic Engineering

Design storms guarantee uniformity regarding quality and operation standards of engineering projects and have been employed widely in urban drainage system design. Commonly used urban design storms, such as the Chicago (K-C) storm and Pilgrim and Cordery (P-C) storm, are calculated using prescribed or historical hyetographs. A prescribed hyetograph is unsuitable for a particular urban region in most cases, and a historical hyetograph takes no account of the spatiotemporal variations between the rainfall pattern at the rain station and that within the calculated region. Additionally, neither method can make adaptive adjustments for climate change. To obtain a more practical design storm with consideration of spatiotemporal variations and climatic changes, this study introduced proper orthogonal decomposition (POD) and dynamic mode decomposition (DMD) to urban design storm study. To demonstrate the feasibility and advantages of the proposed methodology, four cities (London, New York, Sydney, Wuhan) in different continents with different climatic characteristics were selected as case studies. The principal results are as follows. (1) Breaking the assumption of a uniform precipitation distribution, the proposed DMD-POD method is effective in capturing design storms under current climatic conditions and is sufficiently flexible to adapt to climate change. (2) The low-order representation of the rainfall field indicates substantial change in the storm patterns in urban areas. The peak discharge in New York and Wuhan is almost 10% higher than before urbanization, while that in Sydney and London is more than 10% lower on average. The peak time is largely unchanged in New York and Sydney, while it is 4 and 5 h later in London and Wuhan, respectively. (3) Compared with the K-C storm and P-C storm, the flood peak of POD storm increases and appears slightly earlier. The flood peak time in Wuhan simulated with the POD storm is 1 h (2 h) earlier than that simulated with the K-C storm (P-C storm). The peak flow obtained by the POD storm is 9.55% (25.05%) greater than that obtained by the K-C storm (P-C storm), which means that the POD design storm demands a higher level of safety for an engineering project under the same return period.

Drainage facilities are key infrastructure elements of urban stormwater drainage system (USDSs) for flood mitigation during short-duration extreme rainstorms. To obtain the optimal operation rules of drainage facilities within the shortest response time, a novel framework and approach was developed in this study. In the proposed approach, a complex USDS based on Storm Water Management Model (SWMM) was developed to simulate rainfall runoff and provide hydraulic characteristics, and was linked to the particle swarm optimization (PSO) algorithm to design short-term operation rules. Also, to overcome computational inefficiencies of the PSO algorithm to meet the requirements of response time, multisurrogate models in combination with radial basis function network (RBFN), kriging model, and artificial neural network (ANN) were coupled in the optimization search process to substantially reduce the number of evaluations. A practical case in Yinchuan city of China was then taken for the application. The proposed approach demonstrated high performance in flood mitigation compared with the current operation rules for pumping stations and water gates, with the core area inundations reduced to almost 0 (10 years), and the reduction rate being 71.22% (30 years) and 55.16% (50 years), respectively. In addition, the number of pump switches and the water level of the storage tanks were limited within the allowable range to ensure the reasonable and safe operation of the pumping stations. The new method required only 120 evaluations to obtain the optimal solution, which could provide sufficient response time for operators and thus serve as a practical and highly effective way to reduce urban inundation.

In many cases, estimation of the peak discharge is the primary goal of hydrologic modeling. We employed a dataset of 1,648 rainfall-runoff events in 104 watersheds in Texas to explore the peak rate factor (PRF) of 2.08 recommended by the Natural Resources Conservation Service (NRCS) for determining peak discharges with the unit hydrograph method, as well as the PRF’s dependency on watershed and storm characteristics. For each event, a unit hydrograph—assumed to follow a two-parameter gamma distribution—was determined by deconvolving the direct runoff hydrograph with the excess rainfall hyetograph. Results showed PRFs reaching up to 9.75 in Houston watersheds with a median of 0.58 and PRFs up to 11.02 with a median of 1.42 for the rest of Texas. It was also found that the recommended PRF of 2.08 falls between the 75th and 90th percentile of the 1,043 events analyzed in all regions except Houston, and is above the 95th percentile in all 605 events analyzed in Houston. Finally, statistical analysis showed that the PRF is primarily dependent on the watershed’s geomorphology but not on its slope, as was suggested earlier. This dependency, however, although statistically significant, explains only marginally the PRF variability.

Baseflow separation from rainfall and streamflow data is a fundamental problem in applied hydrology, unsolved despite extensive investigations. For example, a part of baseflow is interflow, which results from rainfall infiltration, but almost all of the existing baseflow separation methods, such as graphic and filter methods, have nothing to do with it, which is a serious flaw from the viewpoint of science. The objective of this research is thus to present an innovative baseflow separation method based on the recent general unit hydrograph (UH) model and the classic Green-Ampt infiltration equation. Specifically, we divided a rainfall hyetograph into two parts using the Green-Ampt infiltration equation: one for surface flow that generates direct runoff and the other for subsurface flow that recharges groundwater and generates interflow. As with direct runoff from the surface system, we approximated the subsurface system as a linear system and thus applied the general UH model for interflow in the unsaturated soil zone, where excess infiltration is defined by analogy to excess rainfall. We assumed that groundwater flow could be described by the classic recession curve; we then added interflow and the groundwater flow to obtain the baseflow. We validated the proposed method with six real-world case studies representing four interflow patterns. Particularly, we found that, unlike direct runoff, the interflow UH model parameters are not unique, depending on stream stages. This implies that the subsurface system is a time-variant linear system, which can even make a negative interflow if groundwater is recharged from streams. We expect that this research will provide a better baseflow separation method and thus improve our understanding of watershed processes.

Deriving unit hydrographs from rainfall-runoff data using optimization techniques circumvents the cumbersome requirements of the conventional manual method. In one of the two main ways of formulating the optimization problem, the decision variables are the individual ordinates of the unit hydrograph (UH). Because of their high number of degrees of freedom, though, such formulations can easily lead to overfitting the input hydrographs in the training set, resulting in highly irregular shapes for the derived UHs. These may be optimal for back-constructing the hydrographs in the training set, but as they do not resemble a typical hydrograph, they might not be capable of reproducing other unseen events adequately. In this work, we propose a set of linear constraints that can be included in existing optimization formulations to force the UH to feature a monotonically ascending limb, a single peak, and a monotonically receding limb. The ordinate of the peak, where the shift in monotonicity occurs, is not established in advance, but is rather determined within the optimization process. If required, the constraints can also be applied piecewise to obtain multiple peaks and troughs with monotonic limbs between them. Application to two case studies reveals that the reduction in the degrees of freedom does help in preventing overfitting of the input hydrographs, resulting in better performance when reproducing unseen events.

Abstract

Soil hydraulic properties (SHPs) are commonly determined in soil samples with replicas. Whether these replicas are taken at a same location to represent a specific point or at several locations to represent a larger area, the results should be merged into a final data set to be used in modeling. For this data set to be representative, standard errors and a correlation matrix must be considered in the merging process. We present a method to perform this merging and give an example using stochastic realizations of van Genuchten-Mualem (VGM) parameters generated by Cholesky decomposition to merge the SHP and associated statistics into a merged parameter set. To do so, we used VGM parameters obtained at sample scale in three replicas from a Brazilian savanna soil through inverse modeling of laboratory evaporation experiments. The effectiveness and representativeness of the proposed methodology were evaluated by observing the frequency distribution of different levels of output, comparing individual and merged sample properties. The outputs include VGM parameters, retention and conductivity characteristics, and water balance components stochastically predicted by a hydrological model. The performed stochastic merging correctly represented the variability of the combined replicas, especially with respect to hydrological model outputs of soil water balance components. Using the mean hydraulic property parameter values to deterministically predict water balance components may yield values that are substantially different from the mean values of stochastic realizations. This suggests that the deterministic prediction using mean parameter values in vadose zone hydrological modeling may result in unrepresentative outputs.

Practical Applications

In vadose zone hydrological modeling, soil hydraulic property functions (retention and conductivity) need to be parameterized. This is commonly done by laboratory measurements in soil samples. For a single location, it is common to take some replicas. To characterize a larger area like a plot or a soil type, samples may be taken at different locations within the area. The analysis usually results in the mean parameter values of the fitted equation, their standard deviation, and the correlation between parameters. The method presented here allows merging this information of all replicas into a single set of mean, standard deviation, and correlations representing all analyzed samples. Subsequently, this allows a stochastic evaluation of soil water balance components (evapotranspiration, runoff, drainage) for the respective soil under specific climatic and management boundary conditions to be performed. The obtained results show the method to be robust. Additionally, we show that using the mean values of the hydraulic parameters to deterministically predict the water balance components may give very different results than the mean values of the stochastic realizations.

Abstract

In this study, the frequency analysis of river flow deficiency (FD) of the Siminehrood River in the south of Lake Urmia located in northwest Iran was investigated with regard to rainfall deficiency (RD) during the period of 1992–2013 using copula functions. The main purpose of this study is to provide a comprehensive method for bivariate simulation and forecasting based on marginal distribution and joint behavior of the studied series. For this purpose, the FD and RD values were extracted using the deficiency value method. By preparing deficiency values, 57 different distribution functions were fitted to the studied values, and the generalized extreme value (GEV) distribution was selected as the best marginal distribution function based on the evaluation criteria. Before fitting the copula function, the correlation between the RD and FD values was examined using Kendall’s tau, and a correlation of 70% was obtained. After selecting the marginal distribution function and examining the correlation, the goodness of fit of seven different copula functions was examined for frequency analysis of RD and FD values in the Siminehrood River at the Dashband station. The results indicated that the Clayton copula had the best performance for creating a joint distribution of RD and FD values. It was also determined from the joint analysis of deficiency values that the FD values can be estimated with high accuracy for RD values of more than 0.68 mm. Also, the results indicated that if rainfall in the study area were less than long-term mean for 10-day and 60-day durations, with different return periods and probabilities, different conditions will occur for FD values, which can be used as typical curves for water resources management and allocation in the basin. Finally, the accuracy of the copula-based model and its conditional density in the two phases of simulation and forecasting were investigated. The accuracy of the copula-based model and its conditional density in the simulation phase was confirmed [$R^2=0.87$, root mean square error $({\mathrm{RMSE})=0.1\mathrm{m}}^3/\mathrm{s}$, and nash-sutcliffe efficiency $(\mathrm{NSE})=0.86$]. In the forecasting phase, the forecasting equation based on the proposed method had a RMSE of ${0.14\mathrm{m}}^3/\mathrm{s}$ and NSE of 0.89. By using the violin plot, the model certainty was also confirmed. According to the proposed equation, FD values can be forecasted affected by RD values for 10-day duration with high certainty and accuracy.

Practical Applications

In this paper, a method for forecasting and simulation of meteorological and hydrological parameters is presented that considers two parameters simultaneously. This study discusses meteorology and hydrology from a different perspective. Given the current climate change, this study uses deficiency values. The proposed forecasting method provides regional results that can be used in the water resources management in each basin specifically. By implementing this method, it is easy to forecast the desired values in the basin with different probabilities and return periods. In this study, river flow deficiency (FD) affected by rainfall deficiency (RD) can be forecasted. The difference between the proposed method and other methods and models of simulation and forecasting is in the connection of two variables with each other, which makes the results more certain and reliable. This method can be used in the field of basin management and water resources allocation and also water resources monitoring.

As the largest freshwater lake in China, Poyang Lake has been influenced by the urbanization process since the 1990s. Therefore, nutrient concentration data of Poyang Lake in the past three decades were selected for spatial and temporal heterogeneity analysis. The entropy-based method (EBM) was used to calculate the contribution weight of three independent driving factors (precipitation, agricultural land area, and urban land area) to nutrient concentrations in Poyang Lake in three different periods (1988–1997, 1998–2007, and 2008–2018). The results of this study showed that the concentrations of ${\mathrm{NH}}_3\text{-}\mathrm{N}$ and total nitrogen (TN) have similar trends over time, both of which have been increasing in this century, while the total phosphorus (TP) concentration has experienced a decreasing-increasing process. The nutrient concentrations in the eastern part of Poyang Lake were at the highest level in the whole lake after 1998, and this area has good conditions for the formation of algal blooms, which need special control and management. The effect of precipitation on the nutrient concentrations of the lake in the last 30 years was most prominent. The weight of agriculture’s contribution to nutrient concentrations showed a significant upward trend during the first 20 years, and agriculture and urban construction weights remained stable after experiencing opposite changes in the first 20 years. In the future, precipitation in the Poyang Lake basin needs to be of high concern, and agricultural and urban pollution control measures need to be implemented continuously.

Nova Scotia is a small maritime province with limited capacity to gauge its abundant water resources, but it possesses a diverse geologic setting and surface water distribution. The common engineering practice of using nearest neighbor catchments as hydrologic surrogates may be unreliable in regions such as this. Catchment classification provides a tool to identify and explain variability in hydrologic regimes and inform data transfer across catchments. Here we develop a catchment classification framework using hydrometric, climatic, and landscape data from Nova Scotia, Canada. An inductive classification approach was first used to identify five generalized hydrologic metaclasses based on streamflow signatures derived from 47 long-term streamflow records. We then attempted to replicate this classification using deductive approaches, and identified key physiographic and meteorological variables that could be useful in classifying ungauged catchments. Due to the limited number of gauged catchments, two supervised deductive classification methodologies were applied for comparison: (1) an automated approach often used in more data-rich scenarios (random forests and classification and regression trees); and (2) a nonautomated approach, which involved manual construction and testing of decision trees. The products of the automated approach (random forests), although more robust, may be challenging to apply, while the manually constructed decision tree, which was guided by a combination of local knowledge and theoretical reasoning, could be a useful tool for practitioners. Climate did not emerge as a particularly strong controlling factor in hydrologic variability in this region, but surface water storage had an important role in flow regime across the province. Results demonstrate that this type of hybrid approach can be effective for understanding hydrologic variability and identifying surrogate watersheds in data-limited regions.

The Volga River, as the primary supplier of the Caspian Sea, plays a huge role in its ecosystem sustainability. In this study, we analyze its runoff predictability for different monthly forecast horizons. Additionally, the meteorological and hydrological variables affecting runoff in each month are identified. A wide range of the potential variables was first collected and the Boruta variable preprocessing method was employed to select the important ones. Then the hybrid models were created by combining the selected variables and the data-driven models [i.e., support vector regression (SVR), artificial neural network (ANN), and multiple linear regression (MLR)]. To postprocess the predicted data, the Bayesian model averaging (BMA) method was employed using the combination of the Boruta–artificial neural network (B-ANN) and the Boruta–support vector regression (B-SVR) models. Finally, the Kling-Gupta efficiency (KGE) and the continuous ranked probability skill score (CRPSS) probabilistic evaluation criteria were applied to evaluate the hybrid models. The results showed that the streamflow of the previous steps is the most crucial variable in predicting the streamflow of all next horizons, while its significance decreases as the forecast time horizon increases. Moreover, the temperature variables have the unlike effect on the streamflow prediction and the minimum temperature for winter and spring and the maximum and the average temperatures for summer and autumn are the most effective ones. Correspondingly, BMA-B-ANN-SVR presents the best performance among the hybrid models (for example, the median of 0.8 and 0.91 for CRPSS and KGE in the first horizon, respectively); the reliability of its predicted runoff for different forecast horizons is much better than other hybrid models (B-ANN and B-SVR).