Journal of Hydrologic Engineering

The assessment of sediment transport in rivers is of vital importance in design and management of hydraulic structures such as dams, diversions, hydro-power projects, river training works, bridges, etc. Previously reported studies have shown that data driven techniques such as the artificial neural network (ANN) can give better results in modeling stage-discharge relations than the conventional rating curves. In view of the complexities of rating relationships, compound rating curves are frequently used in place of a single rating curve. Accordingly, this paper investigates the abilities of compound neural networks (CNNs) to model integrated stage-discharge-suspended sediment rating relationship. Using the data of two stations on the Mississippi River and one station on Conococheague Creek, CNNs were trained. A comparison of the results of applying a single ANN and a CNN shows that the estimates of CNN are closer to the observed values than those of single ANN.

The ongoing degradation of the water quality of the Gaza Coastal Aquifer (GCA) is of a great concern for the different authorities and agencies involved in the water sector in the Gaza Strip, Palestine. The GCA is almost the only source of fresh water to over 1.5 million residents where it is utilized extensively to satisfy agricultural, domestic, and industrial water demands. The aquifer is currently being overpumped where pumping largely exceeds total recharge. In addition, manmade sources of pollution endanger the water resources supplies in the major municipalities of the Gaza Strip. Many water quality parameters in the GCA presently exceed the maximum contaminant level (MCL) of the US Environmental Protection Agency drinking water standards, especially for nitrate and chloride. This case study analyzes nitrate concentration distribution for the GCA at different levels such as land use classes and sampling depth. Nitrate concentration data from 1990 and from 2000 to 2004 were compiled and assembled into a single composite database. A geographic information system was used to assess the spatial and temporal variability of nitrate occurrences in the aquifer. Results show that the first quartile of nitrate concentration for the years 1990 and 2000–2004 exceeds the MCL. In addition, the analyses demonstrated a generally increasing trend in groundwater nitrate concentration. The areas with the most elevated nitrate concentrations are areas characterized by heavy agricultural activities and urban areas. Elevated nitrate concentrations in the GCA indicate anthropogenic contamination sources.

The slope-matching method is used for numerical determination of the confined aquifer parameters. In this study, a new slope-matching method is proposed by using a weighted mean of the slope which is implied drawdown with respect to time. The method considers that the slope produces the transmissivity through a simple relationship valid for $u⩽0.01$ . Later the well function, the dimensionless time variable, and the storage coefficient values are calculated from the transmissivity values. As a result, the time-varying aquifer parameter values are obtained for the whole duration of pumping. Average aquifer parameters can be obtained for $u⩽0.01$ similar to the Cooper–Jacob approximation. The method is applied to field data, which are available in the open literature. It is observed that the proposed methodology yielded a sequence of transmissivity and storativity values with the mean error, mean absolute error, and standard error of estimate all less than similar readily available techniques in the literature.

An artificial neural network (ANN) and support vector machines (SVMs) were employed for estimating the removal efficiency of settling basins in canals. The performance of ANN and SVMs was tested using the data from an earlier study carried out by Ranga Raju et al. As compared with the Ranga Raju et al. relationship, the correlation coefficient of ANN as well as SVMs improved from 0.77 to 0.9854 and 0.9853, respectively; whereas the root mean squared error values decreased from 50.66 to 5.712 and 5.7366, respectively. Between SVMs and ANN, SVMs’ performance was found to be better due to its use of the principle of structural risk minimization in formulating cost functions and the use of quadratic programming during model optimization. These advantages led to a unique optimal solution as compared to conventional neural network models.

Hydroclimatological records are increasingly examined for evidence of trends and shifts that may assist in prediction of future climate change scenarios. This study investigates the trend and step changes in U.S. streamflow over a 52-year period (1951–2002) using data from 639 unimpaired streamflow stations categorized according to the hydrologic unit codes. This is particularly relevant since the issue of climate change is of interest to many, and studies have indicated an abrupt change in climate around the year 1976/77. Trends were evaluated using three statistical tests: Spearman’s rho, Mann-Kendall, and linear regression, and step changes were evaluated using the rank sum and student’s $t$ test. The temporal resolution used for the study included water year (Oct–Sept), autumn–winter (Oct–Mar), and spring–summer (Apr–Sept) periods. Additionally, April 1 snow-water equivalent (SWE) data for 121 SNOTEL stations for the period 1941 to 2004 were used to test for the trends in the western U.S. The multiple statistical tests provided robust results for regions with significant changes. Results indicated that the Mississippi and Missouri regions have an increasing trend in streamflow quantity. The Pacific Northwest and South Atlantic-Gulf regions have streamflow decreasing due to a step change in climate. Decreasing trends for the SWE were noted for a number of stations in the states of Oregon and Utah.

A simplified model for estimating the hydrologic response to rainfall events at basin scale is developed in this study. It involves an abstraction of the real basin geometry based on networks of planes and channels and three semianalytical components that give the expected areal-average infiltration rate at field scale, the surface runoff hydrograph as lateral inflow into the channel network, and the water flow routing through the channel network. The model was tested by comparison with Monte Carlo simulations that relied on a detailed representation of spatial variation of excess rainfall and numerical solutions of water routing through planes and channels. The peak flow and the shape of the hydrographs were generally well reproduced. The results were less accurate for cases involving low volumes of effective rainfall, but these are of less importance in hydrologic practice. In spite of the complexity of the processes represented, such as the spatial heterogeneity of soil saturated hydraulic conductivity and rainfall rate as random variables over each subwatershed, the model requires a low computational effort that makes it suitable for application at all levels of spatial discretization.

In its 1988 report, the U.S. National Research Council identified principles to be followed in order to improve estimation of extreme flood quantiles. In spite of the many advances attained by more recent research on extreme flood estimation, the NRC principles remain useful recommendations to be considered, which are used in the methodology to be described herein. According to the proposed method, flood peaks that have exceeded an arbitrary threshold and their associated flood volumes are modeled as a marked point process. The estimation method consists of separately estimating the marginal density function of flood volumes and the density function of flood peaks conditioned on volumes. The annual probability distribution of flood peaks can then be estimated by double integrating the product of both densities. Regionalized quantiles of rainfall depths have been used as auxiliary variables for guiding the estimation of the upper tail of the marginal density function of flood volumes by postulating simple assumptions regarding the rainfall-runoff transformation under extreme conditions. The inclusion of the regional frequency analysis of rainfall depths, for the same duration as that of flood volumes, is performed by fitting a regional two-component extreme value distribution to rainfall data observed at a number of gauging stations within a homogeneous region. This paper summarizes the basic ideas of the method and the details of its application to a $\mathrm{1,620}{\mathrm{km}}^2$ watershed in southeastern Brazil.

Impacts of climate change on water resource in the Bosten Lake basin in the south slope of the Tianshan Mountains in Xinjiang, China, were evaluated using an artificial neural network model. The model was trained using the error backpropagation algorithm and validated for a major catchment that covers 82% of the Bosten Lake basin and has the only available weather and streamflow data. After validating the model it was used to examine the surface hydrology responses to changes of regional temperature and precipitation. Major results showed that because of an additional effect on glacier melt in the upper reach of the basin temperature increase can cause large increases of streamflow. Model results also showed that if the current climate trend continues, the annual streamflow would increase by 38% of its current volume, and the summer and winter streamflow would increase by 71.8 and 11.4% of their respective current volume in the next $50–70\text{years}$ , highlighting challenges for the basin’s water resources management and flood protection.

An at-site based rainfall frequency analysis was performed using hourly precipitation data for 53 selected stations in Indiana. The combined use of generalized extreme value and extreme value type I distributions was found to provide the best statistics, and was applied to construct the at-site estimates. This at-site result was then compared to other commonly used regional rainfall estimates, including Indiana Department of Natural Resources rainfall, Huff and Angel rainfall, and National Weather Service Atlas 14. Non-negligible differences between at-site and regional estimates were found even though Indiana was previously reported as a homogeneous region for rainfall. Results suggest that discretion is necessary when using regional rainfall estimates for site-specific problems. The at-site estimate was also applied to compute the depth ratios between different durations. This ratio exhibited significant variability with increasing return period implying that the use of a constant depth ratio to estimate rainfall might not be appropriate for larger return periods.

For a catchment comprising a network of open channels and overland planes, the time of travel and the equilibrium detention storage in the channels have profound effects on the runoff characteristics of the catchment. For channels with long time of travel and large detention storage, they produce smaller catchment outflow, and the converse is true. Intuitively, the shape of the channels must have an effect on both the time of travel and the detention storage. Hence, the channel shape can be used as a means to manage the runoff from a catchment. In this technical note, the effect of channel shape on the time of travel and the equilibrium detention storage has been compared for seven channel shapes. The channels are subject to a uniform lateral inflow and a constant upstream inflow. The seven channel shapes are (1) square; (2) wide rectangular; (3) deep rectangular; (4) triangular; (5) vertical curb; (6) parabolic; and (7) circular. The comparison shows that channels with long time of travel also have large equilibrium detention storage, and vice versa. It also shows that channel shape can cause a sixfold increase in the time of travel, and a fourfold increase in the detention storage. Of the seven channel shapes, the one that produces the longest time of travel and largest detention storage is the deep rectangular channel. Hence, the use of this channel produces smaller catchment outflow. The channel that produces the shortest time of travel and smallest detention storage is the parabolic channel with large flow depth. Hence, the use of this channel produces larger catchment outflow.