Journal of Hydrologic Engineering

Arid coastal regions are more vulnerable to groundwater deterioration problems. Intensive groundwater abstraction from the coastal aquifer of Wadi Ham, United Arab Emirates, caused a severe saltwater intrusion problem. Given the deterioration of groundwater quality, domestic water supply from well fields have been terminated and replaced by desalinated seawater. In addition, many farms in southeast Fujairah city have been abandoned. This paper develops a two-dimensional finite element groundwater flow and solute transport model to simulate the spatial and temporal variations of the salinity distribution in the coastal aquifer of Wadi Ham, taking into account the transition zone between freshwater and seawater bodies. All simulations were conducted in the horizontal view under transient conditions. The available historical records of the water table levels were used to calibrate and validate the developed model. Emphasis was placed on the response of the transition zone to different pumping scenarios in Wadi Ham. The results indicated that the seawater intrusion problem has evolved rapidly during the last two decades. Unlike previous investigations, this study presents the most accurate quantitative and qualitative assessment of available groundwater in the Wadi Ham aquifer under different pumping scenarios.

A shallow seismic method was used in the northwestern part of the Jabal Hafit area, United Arab Emirates (UAE), to delineate the groundwater levels and test the consistency and effectiveness of the seismic refraction technique in detecting groundwater levels. The objectives were achieved through 25 seismic profiles and three electrical resistivity tomography (ERT) sections carried out in the study area. The results were compared with observed groundwater levels. The seismic data analysis indicated that the deepest and shallowest groundwater levels are 15 and 2.5 m, respectively. These results are consistent with the field measurements of the groundwater levels. The maximum value indicates the prevailing groundwater level within the study area, whereas the minimum detected values may refer to the perched subsurface water that may have been infiltrated from the surrounding water bodies, including irrigation water. The applicability of the shallow seismic method for groundwater level detection was confirmed. Periodical monitoring of the groundwater levels in the study area is recommended.

The present study area is primarily underlain by granites, basalts, and a little bit of laterites. Groundwater occurs under unconfined to semiconfined conditions, in weathered and fractured formations, respectively. A three-dimensional groundwater flow model for the Osmansagar and Himayathsagar catchments—a semiarid hard rock area in India with two conceptual layers—is developed under transient conditions using visual MODFLOW software for the period 2005 to 2009. The 15–20 m top layer is a weathered zone, followed by second 20–25 m-layer fractured zone based on hydrogeophysical studies and borehole lithologs. The groundwater recharge estimation is achieved with the help of geographical information system (GIS) and the water table fluctuation method that is well fitted into the flow model with an average recharge value of 21% of the average annual rainfall. The results derived from modeling indicate that the average input to the aquifer system is 321.96 million cubic meters (mcm), and the output is 322.14 mcm. If the same withdrawal is continued up until the year 2020, the water level is believed to decline more than 45 m over the entire study area. To avoid this critical stage, the present draft should be decreased by nearly 40%.

During urbanization, undisturbed land surfaces are altered to create manufactured landscapes. Classifications of these new urban surfaces are utilized in urban planning, environmental monitoring, and other applications such as storm water management and roof runoff harvesting system design. To evaluate runoff volume and design storm water control devices, areas of different urban surfaces need to be identified and defined as pervious (e.g., undisturbed soils and landscaped areas) and impervious surfaces (e.g., roofs, roads, parking lots, sidewalks, driveways). This study presents a means to facilitate urban surface classification and quantification by analyzing high resolution aerial photographs in conjunction with light detection and ranging (LiDAR) data in a custom application for the geographic information system software. This software processes aerial photographs using red/green/blue (RGB) bands to produce a raster with saturation (S) values. In parallel, LiDAR data are used to distinguish the major surface categories of pavement from roofs and pervious surfaces, and topologically integrated geographic encoding and referencing (TIGER) centerlines are used to identify streets. This process was tested for two different land uses: institutional (University of Alabama, Tuscaloosa, Alabama) and residential (also located in Tuscaloosa, Alabama). Compared to manually delineated areas, the urban area classification differences ranged from 0.2 to 5.2% for roofs, streets, parking lots, and pervious areas. The efficiency of the process compared to the manual delineation of surface areas resulted in time and effort savings ranging from 80 to 90% depending on the size of the area processed. Manual verification by field observations of other characteristics, such as the curb-side drainage system type, still needs to be performed regardless of the surface characterization method employed. Determining the basic area measurements greatly accelerates this initial phase into a comprehensive storm water management planning effort.

In this paper, artificial neural network (ANN) was applied to model and study the signature of hysteresis phenomena in hydrological processes for the Eel River watershed located in California. Because of the nonlinear and stochastic nature of hysteresis phenomena, it is reasonable to expect ANN to develop a model that efficiently considers hysteretic loops. In this study, hysteretic loops were studied from different aspects such as forms, classification, and effective factors of creation. In rainfall-runoff modeling, counterclockwise loops were mostly observed, whereas in the runoff-sediment process, clockwise loops prevailed. Random or eight-shaped loops were expected in runoff hydrographs with several peaks. A direct relationship was detected between the width of the loops and the area of the subbasin. Larger areas led to wider hysteretic loops. The results showed that ANN efficiently considers hysteresis signs when modeling hydrological processes and can lead to appropriate performance.

The estimation of base flow contributions to streamflow, and in particular the partitioning of base flow into shallow and deep aquifer components, is a challenging problem in water resources management. In this study, the newly developed hillslope-storage Boussinesq analytic-element (hsB/AE) model is applied to the Allen River headwater catchment in southern Quebec, Canada, which is a major recharge area for the transboundary Chateauguay River watershed. This first application to a real catchment serves as an illustration of the basic principles behind the coupled model and an investigation into interaquifer interactions and the origins of base flow contributions to streamflow. The hsB component of the model represents local groundwater flow in sloping, shallow Quaternary sediments, whereas the AE component represents the underlying regional bedrock aquifer. The catchment is partitioned into hillslopes of regular convergent, divergent, or combined planform shapes. Simulations are run for the summer periods of 2008 and 2009, and the resulting base flows show good correspondence with estimates on the basis of standard hydrograph separation techniques, in particular, those for low-flow periods from July to September. Overall, satisfactory matches are obtained for the bedrock aquifer heads, but with some underestimation of heads further away from the river. In terms of exchanges between the aquifers, 95% of total leakage is directed toward the bedrock aquifer, whereas only 5% flows in the reverse direction and feeds the hilllsope aquifer, especially in the steeper portion of the catchment. The simulated base flow is partitioned into roughly 65% originating from the shallow hillslope aquifer and 35% from the deep bedrock aquifer, with small interannual variations on the order of 3–5%. This ratio is relatively insensitive to changes in parameter values. For example, a 50% decrease in aquitard thickness reduces the deep bedrock aquifer contribution by only 12%.

An exact analytical solution for the quantity of recharge from an array of trapezoidal channels underlain by a drainage layer at a shallow depth has been obtained using an inverse hodograph and Schwarz-Christoffel transformation. The symmetry about the vertical axis has been utilized in obtaining the solution for half of the seepage domain only. The solution considers the drainage layer at shallow depth with or without pressure. The solution also includes relations for variation in seepage velocity along the channel perimeter and a set of parametric equations for the location of phreatic line. From this generalized case, particular solutions have also been deduced for an array of triangular and rectangular channels. Also particular solutions for different conditions of the drainage layer for all three shapes of channels are possible like the drainage layer at shallow depth without pressure or absence of drainage layer. The solutions can further be reduced to an isolated channel with drainage layer at shallow depth or at large depth and water table below the drainage layer. The solution is useful in quantifying artificial recharge and/or seepage loss from an array of polygonal channels.

The design of protection/restriction zones around drinking water bodies is not a simple task, as there are many processes that govern transfer of pollutants in corresponding catchments. The width of the protected area is one of the basic factors to be considered when setting up and managing the zones. The presented approach provides a practical methodology for deciding on an optimum protection distance. Its most distinctive outcome is the feasibility of defining protection zones of variable distances for the applications in different catchments, all based on a single scientific reasoning. It basically considers flow travel times within protection zones, subsurface buffer depths, potential diffuse pollution risks from land uses, sediment-bound pollution, and the total utility of land as a resource. Different weighting schemes for merging various factors together and associated sensitivities on the outputs are further examined for providing more realistic assessments than the equiponderant use of the distinctively significant components. By virtue of an unsymmetrical weighting biased more towards the protective parameters, the protection boundary appears essentially to be located at a distance of 5.3 km from the reservoir in the case of Çamlı reservoir catchment under study.

Intensity-duration-frequency (IDF) curves are commonly used in engineering planning and design. Considering the possible effects of climate change on extreme precipitation, it is crucial to analyze potential variations in IDF curves. This paper presents a quantile-based downscaling framework to update IDF curves using the projections of future precipitation obtained from general circulation models (GCMs). Genetic programming is applied to extract duration-variant and duration-invariant mathematical equations to map from daily extreme rainfall quantiles at the GCM scale to corresponding daily and subdaily extreme rainfall quantiles at the local scale. The proposed approach is applied to extract downscaling relationships and to investigate possible changes in the IDF curves for the City of Saskatoon, Canada. The results show that genetic programming is a promising tool for extracting mathematical mappings between extreme rainfall quantiles at the GCM and local scales. The duration-variant mappings were found to be more accurate than the duration-invariant relationships. Using the extracted relationships, future changes in IDF curves in the City of Saskatoon are estimated using projections obtained from the CGCM3.1 based on A1B, A2, and B1 emission scenarios. The results show that future IDF curves in the City of Saskatoon are subject to change, but the sign, magnitude, and uncertainty in the estimates of possible changes depend on the emission scenario, storm duration, return period, and mapping equations. Regardless of the emission scenario and/or the mapping relationships, the results of this study show increases in short-duration extreme rainfall with short return periods in Saskatoon. This study shows that the downscaling of extreme precipitation quantiles directly from the corresponding large-scale estimates can be an efficient approach when estimating the design precipitation values under climate change are sought.

Unsteady shallow-layer flows may be described through full dynamic models or using simplified momentum equations, based on kinematic, diffusion, and quasi-steady approximations, which guarantee a reduction of the computational effort. This paper aims to investigate through linear analysis the applicability range of simplified shallow-wave models with special concern to unsteady flows of mud. Considering a three-equation depth-integrated Herschel-Bulkley model, the applicability of the approximated wave models is discussed comparing the propagation characteristics of a small perturbation of an initial steady uniform flow as predicted by the simplified models with those of the full dynamic model. Based on this comparison, applicability criteria for the different wave approximations for mud flows of Herschel-Bulkley fluids, which account for the effects of the rheological parameters, are derived. The results show that accounting for the fluid rheology is mandatory for the choice of an appropriate simplified model.

Concentration of suspended sediment is of fundamental importance in environmental management, assessment of best management practices, water quality evaluation, reservoir ecosystem integrity, and fluvial hydraulics. Assuming time-averaged sediment concentration along a vertical as a random variable, a probability distribution of suspended sediment concentration is derived by maximizing the Tsallis entropy subject to the constraint given by the mean concentration and under the assumption that the sediment concentration is zero at the water surface. For deriving the sediment concentration profile along the vertical, a nonlinear cumulative distribution function is hypothesized and verified with observed data. The derived sediment concentration profile is tested using experimental and field data; however, the clear water surface assumption does not seem to be valid for field data. The Tsallis entropy-based concentration profile method is compared with three sediment concentration profile methods. Comparison shows that the Tsallis entropy-based method is, in general, more accurate than these three methods when judged by the root mean square deviation (RMSD) from observations.

The uncertainty in the spatial rainfall distribution and basin mean rainfall for flood discharge estimation has not been considered in hydrologic practice. The geostochastic simulation has potential for the assessment of uncertainty of the spatial rainfall distribution and basin mean rainfall. This study compares three geostochastic simulation methods, including the circulant embedding method (CEM), the sequential Gaussian simulation (SGS), and the turning bands method (TBM), for assessing the uncertainty in the spatial distribution of rainfall. These methods are found to be comparable in terms of spatial standard deviation, coefficient of variation, interquartile range, overall range, and the distribution of basin mean rainfall. Further, the estimates of basin mean rainfall by these methods are almost similar to those by conventional spatial interpolation methods. However, CEM and TBM are found to be superior to SGS in the simulation performance based on mean error (ME), mean percentage error (MPE), mean absolute error (MAE), and root-mean-square error (RMSE). The geostochastic simulation methods have potential for flood risk assessment.

This study analyzed the variability of the hydrological regime in a Chinese watershed and in the available water resources and used this data to develop an optimal environmental flow allocation model that lets decision-makers work towards sustainable management of the watershed’s water resources. First, historical hydrological data was analyzed to determine its statistical parameters, flow duration characteristics, and trends based on data from two gauging stations (at the West Dayang and Wangkuai reservoirs) from 1956 to 2000 and data for Baiyangdian Lake from 1950 to 2005. The monthly and annual inflows and lake water levels were found to be highly variable. The optimization model that was developed for environmental flow allocation to the lake seeks a rational balance of providing appropriate environmental water to the lake, maximizing economic returns for humans, and minimizing agricultural yield losses. The appropriate quantity of environmental water was defined by minimizing the difference between the regulated and natural flow regimes. Scenario analysis then provided a range of potential management strategies for the socioeconomic, agricultural, and environmental water uses in the watershed surrounding the lake and revealed the effects of assigning different weights to these three objectives. The optimal results will help the lake’s managers to develop aquatic ecosystem restoration strategies, establish sustainable water regimes, and perform environmentally and economically sound management.

In this paper the process to implement the Soil and Water Assessment Tool (SWAT) in Catamayo-Chira binational basin ($\mathrm{11,910.74}{\mathrm{km}}^2$) in the border area between Ecuador and Peru (South America) is detailed. The basin under study has a low density of meteorological and hydrological stations as well as a notable variation in its biophysical characteristics, which further highlights the lack of information. The SWAT model requires a great amount of parameters for its implementation. Its performance was studied, even though the writers had scarce information. The SWAT model made a relatively satisfactory reproduction of the historical record of flows, i.e., Nash-Sutcliffe model efficiency coefficient of 0.73, with certain limitations in calculation of sediment production. Two land-use scenarios were analyzed, observing that an increase of natural forest and shrub areas could significantly reduce sediment production.

Machine learning (ML) techniques have been popular data-driven approaches for hydrological studies during the last few decades owing to their capability to identify complex nonlinear relationships between input and output data without the requirement for physical understanding of the system. This paper aims to predict river flows using various ML methods [feed forward neural network (FFNN), adaptive neuro fuzzy inference system (ANFIS), and genetic programming (GP)] and also a non-ML method (multiple linear regression) in the Euphrates Basin in Turkey. Infilling the missing data in the runoff record of the selected stations in Euphrates Basin is also an objective of this study. The ML methods were applied to the three main sub-basins of the Euphrates Basin, namely the Upper, Middle, and Lower Euphrates Basins. ANFIS and FFNN methods were the most successful ML methods for runoff estimation in the Upper and Lower Euphrates Basins, whereas GP and ANFIS models were the best ones in the Middle Euphrates Basin. Missing flow data were constructed successfully in the selected stations.

The Zoige alpine wetland on the eastern edge of the Tibetan Plateau is an important headwater area for the Yellow River Basin. The White and Black Rivers are two major tributaries in the Zoige Basin. However, the alpine wetland has experienced a rapid degradation due to human and other recent environmental changes in the region. The effects of climate variations and human activities on runoff in this region are still unclear. In this study, those changes in runoff were quantified and categorized using the hydrologic sensitivity analysis method and a monthly water balance model. The temperature index-based snow melting submodel was integrated into a monthly water balance model to account for the considerable snow meltwater from the wetland in the summer season. The nonparametric Mann-Kendall test was used to analyze the annual and seasonal climatic trends in the Zoige Basin. Results suggest that during the past 55 years (1957–2011), annual precipitation was significantly decreasing at a rate of $-0.978\mathrm{mm}/\mathrm{year}$ while the air temperature and potential evapotranspiration (PET) have increasing trends at the rates of $0.029°\mathrm{C}/\mathrm{year}$ and $0.755\mathrm{mm}/\mathrm{year}$, respectively. Using three change-point detection methods, one change point was detected at 1989 in the annual streamflow series (1960–2011) and was adopted to divide the data set into two study periods—the baseline period (1960–1989) and the human-induced period (1990–2011). The mean annual runoff depth from the White (Black) River in the baseline period was 448.9 mm (151.5 mm), while the mean of the human-induced period decreased by 28% (35%). It was determined that decreases in runoff between the two periods can be attributed to 55% (64%) from climate variations and 45% (36%) from human activities in the White River (Black River). Thus, climatic impact exerted more important influence on runoff decline in the Zoige wetland basin as compared to human activities. This study enhances our understanding of the changes caused by climate variations and runoff (and their contributing factors) in the alpine wetland.

Urbanized watersheds produce instantaneous response to rainfall, resulting in storm-water runoff exceeding the carrying capacity of drainage systems—the management of which is important to prevent flooding and erosion of streams. Management can be achieved using structural storm-water best management practices (BMPs). The city of Austin, Texas, envisioned a plan to mitigate future scenarios for flooding and erosion, resulting in the development and integration of storm-water BMP algorithms into the subhourly version of the soil and water assessment tool (SWAT) model. Algorithms were developed to simulate the physical processes governing the flow transport mechanisms through these BMPs. They were tested using a previously flow-calibrated watershed in the Austin, Texas, area. The retention irrigation (RI) system is one such BMP commonly found in Austin, Texas. This paper describes the development of a modeling tool for the RI system and its potential applications. From the test results obtained, it appears that the RI algorithm functions rationally. The algorithm developed could be used to evaluate the functionality of an individual BMP structure to analyze the benefits of such structures at a watershed scale and as a design tool.

In the interest of assessing an alternative source of precipitation input for the Chesapeake Bay Watershed model, the goal of this study is to compare the next generation weather radar (NEXRAD)-based multisensor precipitation estimator (MPE) with currently available gauge-based precipitation inputs for the model at both point-pixel and spatially aggregated scales. The comparison of gauge-based data with MPE at a point-pixel scale offers insight into how commonly known characteristics of gauge-point data are reflected in the mean areal gauge-based precipitation product. The time and effort for recalibration using MPE can be significantly reduced by strategizing the procedure according to the results of this study. Techniques and findings described here should be of general value for model users to anticipate changes in model parameters and output when moving from gauge-based to MPE precipitation in continuous simulation models.

In high-density urban areas, flooding affects a large number of people. A rapidly implementable nonstructural measure is the development of an early flood warning mechanism based on observations from ground-based rainfall stations, especially where radars are not yet installed. To increase the lead time for issuing warnings, a reliable short-term rainfall forecasting model is required, specifically for fast-responding urban catchments where the time of concentration is less than 45 min. With this objective, a rainfall forecasting methodology has been developed using the least-squares support vector machine (LS-SVM) and the probabilistic global search–Lausanne (PGSL) technique. The study’s focus was Mumbai, which receives all of its annual rainfall of 2,430 mm during June to September. The Mumbai storm drainage system had been designed to drain $25\mathrm{mm}/\mathrm{h}$ rainfall and is being upgraded for $50\mathrm{mm}/\mathrm{h}$. Storms less than $25\mathrm{mm}/\mathrm{h}$ do not cause flooding; hence, the proposed methodology was developed using rainfall events greater than $25\mathrm{mm}/\mathrm{h}$. The model developed in this study has been evaluated using statistical performance criteria for four observed high-intensity rainfall events ($50–100\mathrm{mm}/\mathrm{h}$) during 2011. The results indicate that the proposed methodology using LS-SVM and PGSL can effectively forecast high-intensity rainfall with lead time from 5 to 20 min. This study improves upon the 1-h forecast limitation of earlier studies and has the potential to forecast rainfall in real time, especially where radar data are not available.

In this paper, the authors analytically derive the solution for the spatial moments of groundwater solute concentration distributions simulated by a one-dimensional model that assumes advective-dispersive transport with first-order degradation and rate-limited sorption. Sorption kinetics are assumed to be governed by a first-order rate model, and degradation is assumed to occur in the aqueous (i.e., mobile) phase only. As an extension of a previously published moment analysis of a model with similar assumptions, an explicit analytical solution is presented for the zeroth, first, and second spatial moments, as well as equations that describe the behavior of the first and second moments at long times. Moment simulations show that when the degradation rate is relatively large compared to the sorption rate, there are time periods in which the first and second spatial moments of the solute concentration distribution (the mean and variance, respectively) decrease over time. At long times, it is seen that the first and second moments increase linearly with time, indicating that velocity and dispersion are constant.