Journal of Hydrologic Engineering

This paper studies the frequency analysis of droughts using a copula with the application of regionalization in the context of a bivariate homogeneity analysis. Drought events indicated by severity and duration were extracted from monthly flow averages. A K-means clustering algorithm was used to form initial regions. A fuzzy C-means algorithm was used to form the final groups of sites that meet the criteria of bivariate discordancy, bivariate homogeneity, and size. The application of the Gumbel, Clayton, and Frank copulas for bivariate drought frequency analysis was studied. Results show the importance of a clear definition of drought in every scenario since, in our example, the longest drought does not necessarily correspond to the most severe one. Another important observation of this study was that, given the average annual rainfall of a catchment, droughts seem to occur in almost all regions, humid or arid. However, areas with higher annual rainfall can experience shorter but more severe drought. The procedures of this study are applicable for flood frequency analysis as well. Furthermore, ungauged sites can be integrated in the procedure of regionalization.

Assuming time-averaged velocity as a random variable, a one-dimensional velocity distribution for wide open channels was derived by maximizing the Tsallis entropy, subject to mass conservation. The derived distribution was tested by using field and experimental data and was also compared with other well-known velocity distributions. The Tsallis entropy-based velocity distribution was found to predict the velocity near the bed well.

Flood routing of seasonal waterways with no base flow involves considerable uncertainties due to transmission losses. This case would be more complex when water flows into network and branching rivers. A finite-difference scheme is developed in this study to solve Saint-Venant equations for unsteady flow through irregular branching streams. An alternative technique (staggered method) was employed for solving linearized equations. Water elevation at fluvial connections is determined at any time step implicitly using the influence line method. The accuracy of the model and its ability to simulate flow and losses is examined for steady and unsteady flow estates. This is done for simple and complex systems of channels (in situ and artificial systems) with and without consideration of seepage losses. Due to uncertainties involved in practice, results of the model verification showed an acceptable range of error when estimating seepage losses in complex systems. The results of this study reveal the considerable effect of seepage loss on estimation of discharge and flow direction in river branches. Using Lane’s hydrograph and characteristics of Hughes Wash River, Muskat’s equation predicted peak flood and transmission losses more accurately than other seepage equations.

Applying a lumped runoff model in ungauged basins requires the relationships between the parameter characteristics of a model with watershed characteristics. The authors of this study designed various methods to evaluate the parameter characteristics of the storage function model using the kinematic wave model. They applied genetic algorithms to 101 observed events at 17 dam basins in Korea. The results confirmed that there is a limit to developing an empirical equation for parameter estimation on the basis of observed data. As part of the measures taken to overcome the limit, a runoff simulation was implemented that considered 35,000 various conditions by configuring virtual basins, and the parameter of the storage function model was evaluated based on the results. Hence, the authors discovered that a lag time of the storage function method is primarily determined to be closely connected with the characteristics of the longest stream, and the storage coefficient has high relevancy with the basic characteristics.

A new permeameter is proposed for performing laboratory hydraulic conductivity tests on gravels with hydraulic conductivity values ranging from 0.1 to $1\mathrm{m}/\mathrm{s}$. A small diameter riser is connected to a large diameter cylinder, which holds the coarse-grained specimen saturated in a water bath. The release of pneumatic pressure, applied to the free surface in the riser, induces an underdamped oscillatory response of the water level in the riser, similar to an underdamped in situ slug test response in monitoring wells. A closed-form model is used to analyze the measured oscillatory hydraulic head data to calibrate the minor losses in the permeameter and the hydraulic conductivity of the specimen by performing tests without and with a specimen. The average model error of calibrated pressure head values in the riser for the tests considered are about 5% of the initial displacement of about 2 cm. The hydraulic conductivity values are calibrated considering the replicate tests, the tests of different specimen lengths, and the different time periods within a test to verify that the results reflect the hydraulic conductivity of the specimen alone. The Kozeny-Carman equation, which considers the specific surface area of the tested material, gave a hydraulic conductivity value within 5% of the measured value for the marbles, which is a good comparison because the uniform marbles have a known specific surface area. For all the various tests performed on each specimen, most of the hydraulic conductivity values were within 10% of the average, while the specimens with hydraulic conductivity greater than $1\mathrm{m}/\mathrm{s}$ were within 10 to 20% of the average.

High suspended sediment concentrations (SSCs) from natural and anthropogenic sources are responsible for biological impairments of many streams, rivers, lakes, and estuaries, but techniques to estimate sediment concentrations or loads accurately at the daily temporal resolution are not available. This paper presents a time series separation and reconstruction technique (TSSR) that separates time series data (e.g., streamflow or SSC series) into two components: magnitude or duration curve and time sequence or simply sequence as referred throughout this paper. This technique uses only magnitude to estimate sediment rating curves (SRCs) and sediment duration curves (SDCs) and uses both components to estimate daily SSC series for gauged and ungauged sites. Nash-Sutcliffe model efficiency between observed and TSSR estimated SDCs for study watersheds ranged from 0.89 to 0.99. For two selected gauged sites, model efficiencies between observed and TSSR estimated SSC series were 0.96 and 0.98, and for two ungauged sites model efficiencies were 0.65 and 0.88. TSSR’s improved performance is attributed to its ability to separate time series into two components, build individual SRCs, complete SDCs from the magnitude, and convert the estimated SDCs to SSC series using stored sequence to reshuffle the SDCs. For gauged sites, sequence embedded in the observed SSC series of the site itself is used, but for ungauged sites, sequence is obtained from nearby gauged sites.

This paper examines a linked simulation-optimization procedure based on the combined application of an artificial neural network (ANN) and genetic algorithm (GA) with the aim of developing an efficient model for the multiobjective management of groundwater lenses in small islands. The simulation-optimization methodology is applied to a real aquifer in Kish Island of the Persian Gulf to determine the optimal groundwater-extraction while protecting the freshwater lens from seawater intrusion. The initial simulations are based on the application of SUTRA, a variable-density groundwater numerical model. The numerical model parameters are calibrated through automated parameter estimation. To make the optimization process computationally feasible, the numerical model is subsequently replaced by a trained ANN model as an approximate simulator. Even with a moderate number of input data sets based on the numerical simulations, the ANN metamodel can be efficiently trained. The ANN model is subsequently linked with GA to identify the nondominated or Pareto-optimal solutions. To provide flexibility in the implementation of the management plan, the model is built upon optimizing extraction from a number of zones instead of point-well locations. Two issues are of particular interest to the research reported in this paper are: (1) how the general idea of minimizing seawater intrusion can be effectively represented by objective functions within the framework of the simulation-optimization paradigm, and (2) the implications of applying the methodology to a real-world small-island groundwater lens. Four different models have been compared within the framework of multiobjective optimization, including (1) minimization of maximum salinity at observation wells, (2) minimization of the root mean square (RMS) change in concentrations over the planning period, (3) minimization of the arithmetic mean, and (4) minimization of the trimmed arithmetic mean of concentration in the observation wells. The latter model can provide a more effective framework to incorporate the general objective of minimizing seawater intrusion. This paper shows that integration of the latest innovative tools can provide the ability to solve complex real-world optimization problems in an effective way.

This paper presents a relationship between the runoff curve number (CN) and mean potential evapotranspiration (PET) using long-term daily rainfall-runoff data. To this end, CNs were derived for different durations from rainfall-runoff data of eight different agro-climatic river basins in India, and these were correlated with PET derived for respective watersheds using the standard Penman-Monteith method. The resulting high coefficient of determination ($R^2$) values, viz, 0.99 for Hemawati and Mohegoan; 0.98 for Haridanagar, Kalu, and Seonath; 0.96 for Manot, Ghodahadho, and Ramganga strongly support the versatility of the derived relationship. This relationship also invokes determination of PET from the available published CN values, and therefore may be quite useful in field applications.

The subdivision of catchments into appropriate topography-based hydrologic units is an essential step in rainfall-runoff modeling, with the hillslope serving as a common fundamental unit for this purpose. Hillslope-based modeling approaches can utilize, for instance, the hillslope width function as a one-dimensional representation of three-dimensional landscapes by introducing profile curvatures and plan shapes. In this work, an algorithm was developed to delineate and extract hillslopes and hillslope width functions based on a new approach to calculate average profile curvatures and plan shapes from digital terrain data. The proposed method uses fuzzy logic rules and provides a quick and reliable assessment of hillslope characteristics, classifying hillslopes according to nine elementary landscapes (the so-called Dikau shapes). The algorithm was first tested on two contrasting (flat and steep) catchments in Quebec, Canada. The hillslope width functions obtained with the proposed method were able to preserve the modeled hillslope surface areas within approximately 1% while preserving monotonicity. The algorithm was then applied to the Plynlimon catchments in the United Kingdom for comparison with a previously published scheme. Despite fundamental differences in slope metrics calculations, the proposed method produced qualitatively similar large-scale features.

Determination of unit hydrographs using observed rainfall-runoff events can be accomplished by minimizing the deviations between the derived and the observed direct runoff hydrograph ordinates, and accurate estimation of loss rate parameters. Linear programming is an optimization tool that can be used to determine the optimal composite unit hydrograph for a watershed from observed multiple rainfall-runoff data, but requires infiltration losses prior to using the model. Nonlinear programming models overcome this problem and can be used to simultaneously determine the unit hydrograph and loss rate parameters. In fact, several known rainfall-runoff events can be used simultaneously to develop a composite unit hydrograph based upon all those events. Such a model has been developed and applied to multiple storms for watersheds near Buckeye, Arizona; Fort Worth, Texas; and San Antonio, Texas. The nonlinear optimization procedures used for this investigation are the general algebraic modeling system (GAMS) and the generalized reduced gradient (GRG) in a spreadsheet application. Results of the two procedures are comparable.

This article explains the application of search algorithms to determine the parameters for Clark’s unit hydrograph. The golden section search algorithm and discrete horizontal line search method were applied to a composite unit hydrograph developed for a watershed. The methodology developed here was applied to watersheds near Buckeye, Arizona, Fort Worth, Texas, and San Antonio, Texas. The only difference in determining the Clark’s unit hydrograph parameters for these watersheds was the time-area relationships. This relationship for Texas watersheds is different than that for Arizona watersheds. Resulting ordinates of Clark’s unit hydrograph are comparable to the derived composite unit hydrograph.

Insufficient unregulated streamflow observation is a practical challenge and limits the effective utilization and management of water resources. In this paper, the recently introduced univariate map-correlation method to select reference stream gauges and estimate daily streamflows at ungauged catchments was evaluated. The concept was further developed into a multivariate map-correlation using ordinary cokriging. The characteristics of the hydrological response correlations among 27 unregulated catchments were used as the primary sources of information for the analyses. In addition, the lack of information about the hydrological response correlations between catchments was supplemented by extracting correlation information from topographic and climatic covariables. The study was applied to investigate regional hydrological response patterns and select reference stream gauges for ungauged catchments. The analyses revealed that distinct regional hydrological response patterns exist in the case study area. Furthermore, the presence of the hydrological response patterns influenced the location of the best-correlated reference stream gauges to the ungauged catchments in the study area. The analyses results also showed that geographic proximity could not entirely explain the catchment hydrological response correlations, and the nearest stream gauge was not necessarily the best-correlated stream gauge. The multivariate map-correlation approach applied in this study added a physical understanding to the hydrological response correlations between the ungauged catchments and reference stream gauges.

Predicting soil water movement and the transport of pollutants requires a good understanding of soil hydraulic properties. One of the most important hydraulic properties of unsaturated soil is the soil water diffusivity. However, because this parameter exhibits significant spatial variation, monitoring the soil water diffusivity over a large area of land would require the acquisition of an impractically large number of measurements in order to avoid introducing significant errors. To address this issue, the soil water diffusivity was measured in 56 soil samples gathered from across the Nanxiaohegou basin. These experimental data were then used to derive single- and double-logarithmic models for predicting the soil moisture diffusivity for the basin as a whole and for specific soil and terrain types within the basin. A linear relationship was observed between the two parameters of the single-logarithmic model, making it possible to establish a third series of single-parameter models. All three model types yielded estimates that were in good agreement with the experimental results. In addition, analyses of the single-parameter models revealed that the parameter varied according to the nature of the sampled soil and terrain, suggesting that it may provide useful information about the spatial variation of the soil moisture diffusivity.

Interbasin water-transfer projects are constructed to solve regional water-shortage problems. This study attempts to develop a set of water-diversion strategies to assist in the decision-making process of effectively diverting the water. The reservoir operation rule curves are improved by considering water-diversion rule curves that consist of hydrological-stage and water-level factors. A set of water-diversion strategies consisting of diversion rule curves, diversion flows, and supply rules is then developed to guide the actual water diversion. To minimize total water supply shortages of the urban, agricultural, and ecological water demands, the optimization model is constructed and the variables of diversion strategy are optimized by genetic algorithm. Furthermore, three scenarios including scenarios of no water diversion, water diversion with constant flow, and diversion with strategy are used to better understand the advantages of applying diversion strategies. The authors present a case study for water-diversion operation of the Biliu River reservoir in China. The results demonstrate that the proposed approach will help the reservoir’s managers to effectively divert the water.

The mechanisms and the rate of groundwater recharge were assessed for the uppermost aquifer used for water supply by local settlements in the area surrounding Timbuktu, Republic of Mali, using chloride budgets. The area receives a mean rainfall of 225 mm annually. Three clusters of wells and three soil water profiles were sampled for the study. The well cluster in the vicinity of the River Niger showed effects of induced infiltration via an active wetland fed by the old canal, which once connected Timbuktu with the River Niger. For the well clusters further away from the river, an annual mean recharge of 4 mm was found. A similar recharge, 1–2% of the precipitation, was found by assessing the recharge from the soil water chloride in three pits dug to 2.2-m depth. The recharge occurs preferably in depressions between dunes and due to channeling by soil crusts on the slopes of the dunes. This runoff-runon regime is probably the mechanism that allows any recharge to occur at all. The depressions, called mares in French, have standing water for extended periods after the rainy season, which is mirrored by gley patches in the subsoil. Occasional high nitrate concentrations in soil profiles indicate that conditions for nitrification and nitrate leaching occur after long drought periods followed by onset of heavy rain. This is a witness of the large variability of the precipitation climate in time and space, also mirrored in the variations of the figures for groundwater recharge. The moist conditions give growth to cyano-bacteria found on the soil surface in the mares. N-fixation by cyano-bacteria is probably the reason for occasional very high nitrate contents in soil and groundwater. The water demand by the local settlements has been assessed by using figures for population and livestock density and their water use. The water demand by the local population and their herds is sustainable in relation to the groundwater recharge assessed by this study.

Understanding sources of streamflow and nutrient concentrations are fundamental for the assessment of pollutant loadings that can lead to water quality impairments. The objective of this study was to evaluate the discharge of three main tributaries, draining different land uses with karst features, as well as their combined influences on total nitrogen (TN) and total phosphorus (TP) levels in reservoir-like embayment (R-E) on a stream entering Lake Marion, South Carolina. From 2007–2009, hydrology, TN, and TP data were collected from the 1,555-ha Chapel Branch Creek (CBC) watershed. In general, monthly streamflow in all tributaries was found to be $-10\%$ of rainfall, and as little as 0.1% in the smallest tributary. The third tributary flowed into a cave system and discharged via a cave spring (CS) into the embayment while gaining a sustained groundwater flow from a second cave (GW) system. The CS flow was substantially larger than the flow measured in the other tributaries. The small amount of rainfall that became surface flow and the large flow at the cave spring indicated a significant water loss from the surface watershed to subsurface flow or a groundwater source area substantially larger than the surface watershed. Nutrient concentrations in flows from tributaries draining various land uses were not significantly different ($\alpha =0.05$) for most of the locations. A simple water balance was developed to estimate the R-E outflow to Lake Marion using measured discharges from three tributaries, change in storage computed using a bathymetric survey, daily lake level changes, rainfall, and computed evaporation. Mean monthly TN and TP concentrations in the embayment were substantially lower than the observed means from the two tributary outlets and the CS into the embayment, indicating a loss in the embayment. The second cave system at CS, representing an unknown subsurface drainage area, was the source of nearly 50% of TP loading, over 50% of flow, and over 70% of TN loading to CBC. These results may have implications in water quality management of the CBC watershed.

Ecological water was stolen by industrial and domestic water consumption in the Weihe River, which led to a disquieting decline in riverine ecologic functioning. Based on river regionalization and typical years of hydrologic frequency analysis, a variable eco-environmental water demand (EEWD) calculation method was put forward with a monthly runoff process against its traditional fixed volume. The main stream of the Weihe River in Shaanxi Province was regionalized into seven zones in accordance with the location of gauging stations, boundary of administrative districts, and desirable target of water qualities by water function regionalization; typical years with hydrologic frequencies of 25, 50, 75, and 95% were introduced to take annual variation into account, so as to meet the hydrologic characteristics of the river. In addition, the monthly process was suggested to reflect the seasonal change of EEWD. In this study, EEWD was divided into three parts: water demand for sediment transport, diluting and self-cleaning, and basic ecological water demand according to the major problems with the Weihe River. The results showed that EEWD had obvious annual and seasonal variation at time scale, while it tended to increase from upstream to downstream, with the maximum value being $\mathrm{4,570}\mathrm{M}{\mathrm{m}}^3/\mathrm{a}$. The EEWD showed a strong relationship with runoff, sewage discharge, and sediment erosion and deposition status. The method used herein provided a credible scientific basis for estimating EEWD on the river with serious annual and seasonal hydrologic variation and with great changes along the river, which proposed a significant reference to rivers with similar situations to the Weihe.

Excess rainfall release time is one of the most important temporal components of a hydrograph and is often equated with time of concentration; however, this study will demonstrate that these two components are basically different and must be discriminated from each other. A formula for estimation of release time principled on Clark hydrograph is also proposed. This research introduces another concept of partial release time, and a parametrical equation for its estimation is suggested. Furthermore, the fundamental relation between time base and excess rainfall duration for linear systems is clarified, giving rise to a formula for time base of Clark hydrograph. Finally, it will be theoretically shown that release time is the temporal dependency between the rainfall and runoff in each event, and hence it plays a crucial role in event-based rainfall–runoff-forecasting models.

Investigations are carried out to analyze the effect of freshwater pumping on saltwater intrusion in coastal aquifers in the presence of a semipervious subsurface barrier along the coast. A conceptual, unconfined aquifer with a subsurface barrier subjected to simultaneous freshwater pumping at single/multiple locations perpendicular to the coast is considered for the analysis. The location, width, and the conductivity of the subsurface barrier being held constant, the effect of freshwater pumping at seaward and landward locations from the barrier is investigated on the interface motion. The performance of the barrier is compared with a nonbarrier condition. The numerical model developed predicts the behavior of the saltwater-freshwater interface and the piezometric surface due to simultaneous pumping at single/multiple wells across the sea cost. The barrier is effective in checking the progress of saltwater for freshwater pumping on the landward side of the barrier compared to no barrier condition. The freshwater pumping on the seaward side of the barrier should be limited to smaller rates. Additionally, safer zones for simultaneous freshwater withdrawal are identified for single, two, three, and four wells without the advancement of saltwater intrusion. The results obtained are of general nature and would be useful in determining the ideal location and rate of freshwater withdrawal minimizing the saltwater intrusion in coastal aquifers.

Measuring infiltration at the plot scale is difficult, especially for high hydraulic conductivity soils. At the plot scale, the infiltration rate is usually calculated by comparing surface runoff to rainfall. Direct measurement of infiltration beyond the point scale is typically limited to locations where land forming (e.g., infiltration pond) has been performed or fields with basin irrigation systems. The standard method for field measurement of point-scale infiltration is the double ring infiltrometer, which is limited in size (typically 30 cm diameter). In this research, a new method is proposed that uses a temporary berm constructed of a water-filled 15-cm diameter vinyl hose with the edges sealed to the soil using bentonite. The berm is capable of confining infiltration plot areas of various sizes (e.g., $1\times 1$ and $3\times 3\mathrm{m}$ areas in this research). Water tanks with 0.8 and $4.9{\mathrm{m}}^3$ capacity were used to supply water to the plots by gravity flow. A constant head could be maintained within the plot using either an automatic float valve for lower infiltration rates or a manually operated gate valve for higher infiltration rates. Observation wells were installed outside the plots to monitor for water table rise and tracers that leached into the groundwater. Guidelines are provided for tank size and refilling frequency for conducting field experiments. The procedure was tested on soils ranging from silt loam to coarse gravel using 12 $1\times 1$ and $3\times 3\mathrm{m}$ plots at three alluvial floodplain sites. Measured infiltration rates ranged over two orders of magnitude ($0.8–74\mathrm{cm}/\mathrm{h}$) and were typically greater than the estimated permeability of the limiting layer reported in soil surveys, suggesting the need for larger scale field measurements of infiltration rates.

A numerical model is developed for simulating soil moisture movement for a multilayer crop root zone in subtemperate-subhumid agro-climates. Water dynamics is used for a nonuniform root distribution pattern to compute the root extraction rate by plants in a moisture-scarce environment. The root uptake function is incorporated in governing the soil-moisture equation. Sensitivity analysis (SA) has been carried out to evaluate the relative importance of model input parameters in predictions of moisture uptake by plants. The paper describes the results of sensitivity analysis of model parameters for a maize crop and their spatial dimension. An iterative procedure has been used for the sensitivity analysis. The model was run for a range of values ($\pm 10\%$ and $\pm 5\%$ from base value) of different model parameters to obtain plant moisture uptake. Results indicate the model is least sensitive to saturated moisture content (${\theta }_s$), coefficient ($n$), field capacity (FC), and permanent wilting point (PWP). The model is sensitive to residual moisture content (${\theta }_r$), root uptake model parameters ($\alpha$) and ($\beta$), and saturated hydraulic conductivity ($K_s$) since variation in these parameters has significant impact on model output.