Journal of Hydrologic Engineering

The effects of seepage on streambank erosion and failure are less understood compared to fluvial processes, especially the linkage between surface water and groundwater mechanisms. Field data are needed to validate laboratory seepage erosion and instability conclusions and to understand how engineering tools and methods may be applied to field conditions. An innovative trench injection system was utilized to provide a constant head on a near-streambank groundwater system when filled with stream water. This research was performed on a streambank of Dry Creek, a deeply incised stream with near-vertical banks located in Mississippi. Experiments included installing a trench (2.8 m from the bank and 2 m below ground surface) and a network of tensiometers and observation wells to measure soil-water pressures and water table elevations. Bank stratigraphy consisted of a sloping, conductive loamy sand layer between cohesive streambank layers. Groundwater conditions were monitored during a series of induced-seepage experiments. The bank face was outfitted with a seepage collection device to measure seep flow rates and sediment concentrations. Seepage flow rates (as high as $0.4\mathrm{L}/\min$) and corresponding erosion rates (as high as $0.86\mathrm{kg}/\min$) were proportional to estimated hydraulic gradients in the near-streambank region and followed an excess flow rate equation. However, flow paths and hydraulic gradients were largely nonuniform due to local variability in streambank stratigraphy, suggesting difficulty when attempting to apply engineering analyses of bank erosion and stability for seepage processes without accounting for this heterogeneity. Seepage flow and erosion became restricted when small-scale bank failures due to undercutting blocked flow pathways and limited particle mobilization, termed temporary self-healing. Seepage erosion was shown to be an important mechanism of streambank failure, especially when acting in concert with fluvial erosion processes that prevent permanent self-healing of seeps.

The Tarim River is located in the arid areas of northwestern China, where groundwater (GW) and surface water (SW) in different landscape units have undergone regular and duplicate transformation processes, greatly improving the utilization of water resources. Investigation of the interaction between groundwater and surface water is critical to determine proper water resources planning and management in the Tarim River region. A new approach of coupling the soil-vegetation-atmosphere transport model (SVAT) with the groundwater-surface water interaction model (GSI) is presented in this paper. Usually, the surface water recharges to groundwater and groundwater-soil water exchange are not considered in the SVAT model. However, in reality, the soil water content profiles and soil heat profiles are intensely affected by shallow groundwater table, especially in arid riparian zones where groundwater levels fluctuate substantially. A new method linking the SVAT model with the GSI model is proposed in this paper to approach this issue. The groundwater-soil water exchange and groundwater evaporation can be effectively simulated using GW-SW interface in the coupled GSI-SVAT model. The coupled model is validated in the riparian zone of the upper reaches of the Tarim River with the simulation of heat and water transfer between groundwater level and soil surface, evapotranspiration, root water uptake, and groundwater-soil water exchange. The simulation results show good consistency with experimental data, indicating that the use of this coupled model could improve the accuracy of simulation of ecological water consumption in the riparian zone. The coupled GSI-SVAT model could be used for better planning and management of water resources in arid areas.

Irrigation percolation can be an important source of shallow aquifer replenishment in arid regions of the southwestern United States. Aquifer recharge derived from irrigation percolation can be more significant in fluvial valleys overlying shallow aquifers, where highly permeable soils allow rapid water infiltration and aquifer replenishment. This study used data from various irrigation experiments and data at the piezometric level to assess the irrigation percolation effects on the recharge of a shallow aquifer in an agricultural valley of northern New Mexico. The water balance method (WBM) and the water table fluctuation method (WTFM) were used to estimate aquifer recharge at the field scale ($<1\mathrm{ha}$) and the WTFM was used to determine recharge at the entire valley scale ($40{\mathrm{km}}^2$). Also, the temporal and spatial distribution of aquifer response to irrigation percolation and canal seepage inputs was characterized. The results showed that for separate irrigation events at the field scale, aquifer recharge values ranged from 0 to 369 mm when using the WBM and from 0 to 230 mm when using the WTFM. For the cumulative irrigation season at the valley scale, recharge ranged from 1,044 to $\mathrm{1,350}\mathrm{mm}{\mathrm{year}}^{-1}$. A relatively rapid water table response with sharp water table rises and declines was observed in all but dryland location wells in response to canal seepage and irrigation percolation inputs. The results of this study add to the understanding of the mechanisms of shallow aquifer recharge and the interactions between surface water and groundwater in a floodplain agricultural valley of northern New Mexico.

Riverbank retreat derives from a complex combination of various processes where a key role is played by the interactions of groundwater and surface water. Recent progress has been made in two main research areas, i.e., the effects of hydrological factors on mass failures, and the role of subsurface flow in seepage erosion and bank stability. This paper aims to review recent progress and identify the main knowledge gaps in modeling riverbank failures, with a particular focus on the interactions between mass failures, bank hydrology and other hydrological factors related to the river hydrograph. Changes in pore water pressures related to transient variations of surface water and groundwater are widely recognized as one of the most important factors controlling the onset and timing of bank instability. Inclusion of negative pore water pressures and groundwater flow modeling has greatly improved the capability to predict mass failures. Fluvial erosion has a fundamental role in deforming the bank profile and, therefore, promoting bank instability. However, fluvial erosion may also affect bank stability indirectly, as deformation of the bank profile alters the pore water pressure field within the bank. Seepage flow determines an important control on bank instability by two mechanisms: hydraulic gradient forces, causing possible mass failure or liquefaction; and seepage erosion and undercutting, eventually inducing a collapse of the upper bank. This review has led to the identification of the main knowledge gaps and needs for future research in two main areas: (1) parameterization of various soil properties that need to be accounted for when modeling hydrologic processes and mass failures; and (2) the need for integrated modeling of seepage and fluvial erosion and how they simultaneously may affect mass failures.

Surface microtopography influences both spatial and temporal distributions of a series of hydrologic processes including infiltration, surface runoff, and unsaturated flow. Importantly, surface roughness and depressions also affect the fate and transport of pollutants throughout surface and subsurface environments. The objective of this study was to experimentally investigate the effects of surface microtopography on overland flow, infiltration, and tracer bromide (Br) runoff, and leaching into the subsurface system under simulated rainfall. Both rough and smooth surfaces were created by using sieved soil and scanned by an instantaneous-profile laser scanner to acquire high-resolution digital elevation models (DEMs) of the soil surfaces. A mold surface, featuring a number of depressions of varied sizes and relationships, was used to create the rough soil surface. The Windows-based puddle delineation (PD) software was applied to characterize surface microtopography, determine flow directions and accumulations, and compute maximum depression storage (MDS) and maximum ponding area (MPA). Two overland flow and tracer transport experiments were conducted using the smooth and rough soil surfaces. Runoff water and soil samples were analyzed for Br. The experimental results indicated that the smooth surface seemed to have earlier and greater water and tracer runoff than the rough surface. Water discharge and tracer loading from the rough surface with varying depressions exhibited a unique, microtopography-controlled threshold behavior. It was found that spatial distributions of the tracer in the subsurface system matched the surface microtopographic features (e.g., depressions, peaks/mounds, and ridges). Enhanced infiltration and tracer leaching under depressions of the rough surface were observed, which resulted in higher Br levels in deeper soil. In contrast, the dominant surface runoff and tracer transport along the horizontal direction for the smooth surface yielded higher Br concentrations in shallow soil. This preliminary study highlighted the important role of surface microtopography in tracer transport in a coupled overland and unsaturated flow system, and demonstrated that surface microtopography influenced the location and timing of ponding, runoff, and infiltration. Increased understanding of the effect of soil microtopography on hydrologic processes is expected to yield improved ability to simulate soil water and solute movement.

The hyporheic zone has received significant attention in recent years due to its role in regulating the physical, chemical, and biological processes that buffer fluvial systems at a variety of scales. The exchange of water through the hyporheic zone is also important for the regulation of stream and streambed temperatures. Recent research has utilized stream and streambed temperatures to quantify groundwater discharge to streams using a variety of methods, including one-dimensional analytical solutions for vertical flux across the streambed interface. The presence of lateral flows, including hyporheic flow, will cause uncertainty in these analytical solution results. In this study, HydroGeoSphere, a three-dimensional, fully integrated surface/subsurface hydrologic model, is used to simulate the manner in which streambed heterogeneity influences groundwater/surface water interactions along a stream section, and the uncertainty of groundwater/surface water flux estimates based upon streambed temperatures. Groundwater/surface water exchange fluxes from numerical experiments of an idealized linear reach with homogeneous and heterogeneous streambed materials are compared to the results of a one-dimensional analytical solution of groundwater discharge to a stream using the numerically simulated temperature results. Both low- and high-variance distributions of hydraulic conductivity (ln $K$ variances of 0.85 and 17.0, respectively) along the streambed caused hyporheic flow, which caused significant discrepancies between the numerical and analytical groundwater/surface water flux results. A low variance in hydraulic conductivity did result in smaller discrepancies between the numerical and analytical solutions compared to the higher-variance simulations; however, the discrepancies between the low-$K$-variance simulations and the associated analytical solution interpretation still differed by up to an order in magnitude. The quantity of hyporheic flow increased with increased depth of heterogeneous streambed sediments, causing greater discrepancies between the analytical estimate and the numerical solution of groundwater/surface water exchange. In all simulations the analytical solution underestimated the exchange flux, and in the presence of hyporheic flow, the analytical solution was unable to capture the patterns of exchange flux at the streambed. The three-dimensional modeling approach produced an increase in the lateral flow components within the streambed compared to results inferred from one- and two-dimensional models employed in previous research. This three-dimensional approach enabled analysis of the spatial variability of error between the analytical and numerical results and determined that in environments with lateral flows, including hyporheic flow, the analytical solution is indicative of the patterns of the vertical component of fluid flow to the streambed.

Among the key problems in atmospheric and hydrologic sciences are the modeling of the interaction between the atmosphere and land surface hydrology while also quantifying the surface/subsurface hydrologic flow processes both in vertical and lateral directions, and modeling the heterogeneity in surface and subsurface hydrologic processes. Meanwhile, in standard water resources engineering practice, the planning and management of the water resources is performed over the geographical region of a watershed. To address these issues, a model of coupled atmospheric-hydrologic processes at the watershed scale, the Watershed Environmental Hydrology Hydro-Climate Model (WEHY-HCM), has been developed. The atmospheric model PSU/NCAR MM5 (Fifth Generation Mesoscale Model) was coupled to the watershed hydrology model WEHY through the atmospheric boundary layer to form the WEHY-HCM. The WEHY-HCM is especially useful for producing nonexistent atmospheric data as input to the modeling of surface and subsurface hydrologic processes at sparsely gauged or ungauged watersheds. The continuously changing state of the atmospheric boundary layer may be essential information in the computation of evapotranspiration (ET) rates and other land surface fluxes. Because such land surface fluxes are the result of the interaction of land surface hydrologic processes with atmospheric processes, their realistic estimation necessitates the coupled modeling of these processes, as is done in the WEHY-HCM. The model is also useful at watersheds that have heterogeneous topography and land use/cover because the main model components are based on areally averaged, scalable conservation equations and parameters in order to quantify and account for the effect of heterogeneity within watersheds. In this paper, the modeling of an integrated system of atmospheric processes aloft coupled with atmospheric boundary layer processes, land surface processes, and surface and subsurface hydrologic processes is described at the scale of a watershed within the framework of WEHY-HCM.

The objective of this study is to evaluate the potential of the Watershed Environmental Hydrology Hydro-Climate Model (WEHY-HCM) for modeling runoff at ungauged or sparsely gauged watersheds. The WEHY-HCM employs an atmospheric module (fifth generation mesoscale model, MM5) that is coupled with its process-based watershed environmental hydrology (WEHY) module. In this study the atmospheric component of the WEHY-HCM was utilized for the dynamical downscaling of the coarse U.S. National Center for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) historical global reanalysis atmospheric data over a foothills region in Northern California in order to reconstruct the historical hydro-climate data over four watersheds in the foothills region at 3 km grid scale at hourly intervals. The WEHY-HCM’s atmospheric module performance was evaluated by the comparison of model-reconstructed precipitation and air temperature against ground observations in time and space with satisfactory results. These results lead to the conclusion that WEHY-HCM may be useful at sparsely-gauged or ungauged watersheds for producing nonexistent atmospheric data as input to the modeling of surface and subsurface hydrologic processes at such watersheds. By means of the reconstructed atmospheric data as its input, the WEHY module of WEHY-HCM was then applied to the Sierra foothills region, encompassing Big Chico Creek ($192{\mathrm{km}}^2$), Little Chico Creek ($78{\mathrm{km}}^2$), Upper Butte Creek ($407{\mathrm{km}}^2$) and Deer Creek ($508{\mathrm{km}}^2$) watersheds in northern California. The model simulations of daily and monthly runoff at these watersheds, when compared against historical observations by means of visual inspections and statistical tests during a validation period yielded satisfactory results. Therefore, it is concluded that the WEHY-HCM may be useful in producing both atmospheric data and runoff simulations over ungauged and sparsely gauged watersheds.

Return flows from irrigation form an important component in the overall management of water in a basin. The Nagarjunasagar irrigation project is one of the biggest irrigation projects in India. At the Nagarjunasagar project formulation stage, the return flows were assumed to be 7.5% of the total water drawn into the command area (area irrigated under the canal), which could be approximately $538.08\times {10}^6{\mathrm{m}}^3$. This paper aims to calculate the return flows from the Nagarjunasagar canal command areas. At present, the calculations are made during the period 1982–2004 by using multiple linear regression analysis. Return flows form 15% of the total releases from the canals for the whole study period. During wet, normal, and dry periods, return flows are 29, 20, and 10%, respectively. Similarly, they are 29 and 18% during the kharif (June–October) and rabi (October–February) seasons, respectively.

Many applications require estimation of reference evapotranspiration (${\mathrm{ET}}_o$) in areas where meteorological measurements are limited. Previous studies have shown that simple evapotranspiration models based on radiation and temperature perform relatively well in humid climates but underpredict ${\mathrm{ET}}_o$ in drier and windier climates. In this paper, estimates of ${\mathrm{ET}}_o$ based on existing simple models were compared with ${\mathrm{ET}}_o$ calculated with the more comprehensive Penman-Monteith equation using meteorological measurements at 106 locations in the contiguous United States for a range of climates. Results showed that the simpler models were closest to the more comprehensive model at sites where the annual mean relative humidity (RH) was approximately 70% and annual 2-m wind speed (U) was less than $2\mathrm{m}·{\mathrm{s}}^{-1}$. Equations for adjusting the model coefficients were developed based on annual averages of RH [or vapor pressure deficit (VPD)] and $U$ to improve the performance of these models for drier and windier sites. Publicly available data sets of spatial distributions of annual RH and $U$ were used to estimate local coefficients for the contiguous United States. The new coefficients were tested with additional data from 22 sites, not used for coefficient development. At the test sites, the performance of both tested models improved with the revised coefficients. Depending on the model, 63–90% of the stations had ${\mathrm{ET}}_o$ within 10% of the Penman-Monteith ${\mathrm{ET}}_o$ for the growing season. The revised coefficients can be used to improve estimation of reference ${\mathrm{ET}}_o$ in data-limited applications such as remote sensing and distributed hydrologic modeling.

Changes in climate may alter extreme rainfall intensity in most regions. In order to incorporate potential future changes, tools for planning and design should be capable of considering nonstationary climate conditions. In this study, potential changes in intensity-duration-frequency (IDF) curves, which are often used for assessment of extreme rainfall events, were explored using climate model historic runs, reanalysis runs, and future climate projections. The concept of the adjustment factor, which represents the general ability of climate models in representing the reanalysis data that is a reasonable estimate of actual gridded historical climate at the model grid scale, was applied to investigate potential impacts of climate change on IDF curves. For most of the study regions, future climate projections suggested an increase in the intensity of extreme storms for a given return period and duration with strong regional variations. The results also revealed the dependence of the adjustment factor values on return period and storm durations as related to the characteristics of extreme rainfalls, as it varied significantly with region, storm duration, and return period.

Rainwater harvesting (RWH) is a decentralized practice that provides both water supply and runoff reduction benefits that are often difficult to assess. To assist in this evaluation, a model was developed that simulates a single RWH system in Richmond, Virginia, using storage volume, roof area, irrigated area, an indoor nonpotable demand, and a storage dewatering goal as independent design variables. Water supply and runoff capture reliability are assessed for a wide variety of cases. Tradeoff curves were developed to evaluate the design variable substitution when reliability was held constant. A reliability function was fit to the simulation results, and a solution method was developed to solve for an unknown variable as a function of the others. This method evaluates different design cases that provide the same water supply and/or runoff reliability, demonstrating that the design variables can be substituted for each other, using care to restrict substitutions between functional inputs or (separately) functional outputs. This method can provide guidance for designers in selecting equivalent RWH systems and regulators in assessing runoff reduction goals. Results indicate that the dewatering goal enhances runoff capture reliability but reduces water supply reliability moderately. Increases in storage volume increased both water supply and runoff capture reliability. Irrigated area has a much larger, negative effect on water supply reliability, and roof area has a similar negative effect on runoff capture reliability. As irrigated area increases for the same population, runoff capture reliability increases but eventually remains constant, reflecting the dominance of indoor demand, which in turn reflects the simulation’s assumption of seasonal irrigation. Applications indicate that land uses that provide larger demands, such as offices, commercial sites, and high-density residential sites, may be better suited than lower-density residential lots where RWH is more commonly employed.

Hydrologic alteration can be attributed to climate and human effects with some confidence. However, when quantifying the hydrologic alteration induced by dams using historical flow records, there is a lack of consideration concerning the impacts of climate variability on hydrologic alteration. Evaluating dam-induced and climate-influence-induced hydrologic alteration has emerged as a key problem, which is the process of quantifying the extent that they have altered the river’s hydrologic attributes. Central to this process is restoring the hydrologic attributes without dam influence. This study simulates hydrologic regimes without dam influence and evaluates the dam and climate influences on hydrologic processes in the Lancang-Mekong River. The results showed that the meteorological variables were closely related with runoff. Therefore, a stepwise-regression model was established and validated using meteorological variables with their time lags, and it showed good performance. The simulative evaluation revealed that runoff from the Lancang River was simultaneously and significantly affected by dam construction and climate variability, and the damming effects were mainly concentrated in the primary stage of reservoir operation. Meanwhile, the dam operation increased the minimum and maximum monthly runoff over a year. During the study period, the degree of hydrologic alteration (DHA) indicated that the Manwan Dam compensated for the impacts of climate variability on the Lancang River.

A database containing more than 17,700 discharge values and ancillary hydraulic properties was assembled from summaries of discharge measurement records for 424 U.S. Geological Survey streamflow-gauging stations (stream gauges) in Texas. Each discharge exceeds the 90th-percentile daily mean streamflow as determined by period-of-record, stream-gauge-specific, flow-duration curves. Each discharge therefore is assumed to represent discharge measurement made during direct-runoff conditions. The hydraulic properties of each discharge measurement included concomitant cross-sectional flow area, water-surface top width, and reported mean velocity. Systematic and statewide investigation of these data in pursuit of regional models for the estimation of discharge and mean velocity has not been previously attempted. Generalized additive regression modeling is used to develop readily implemented procedures by end-users for estimation of discharge and mean velocity from select predictor variables at ungauged stream locations. The discharge model uses predictor variables of cross-sectional flow area, top width, stream location, mean annual precipitation, and a generalized terrain and climate index (OmegaEM) derived for a previous flood-frequency regionalization study. The mean velocity model uses predictor variables of discharge, top width, stream location, mean annual precipitation, and OmegaEM. The discharge model has an adjusted R-squared value of about 0.95 and a residual standard error (RSE) of about 0.22 base-10 logarithm (cubic meters per second); the mean velocity model has an adjusted R-squared value of about 0.67 and an RSE of about 0.063 fifth root (meters per second). Example applications and computations using both regression models are provided.

It is a familiar exercise in hydrology to characterize river basins into hydrologically homogeneous regions by using parameters suited to explain typical hydrological variables such as extreme and annual flows. This paper discusses regionalization of the Blue Nile River Basin (BNRB) by using statistical techniques and describes the selection of best-fit distribution models to estimate the flood frequency of the basin; such undertakings have not been made before. The BNRB is delineated into five homogeneous regions based on statistical parameters of station data. Approximately 14 different distributions are analyzed. The generalized logistic model is the best fit for Regions I and IV. Log-Pearson Type III distribution is appropriate for Region II. Lognormal and generalized extreme value distributions are selected for Regions III and V, respectively. For all distributions, probability weighted moment parameter estimation method is most efficient, but for log-Pearson Type III, ordinary moments are chosen. For each region, a unique regional flood frequency curve is developed. These curves are important to estimate the flood quantiles of ungauged catchments in the data scarce area of the basin; hence, they meet the needs of water engineers who currently face tremendous challenges in designing small and medium hydraulic structures in the basin.

The significance of uncertainty analysis (UA) to quantify reliability of model simulations is being recognized. Consequently, literature on parameter and predictive uncertainty assessment of water resources models has been rising. Applications dealing with urban drainage systems are, however, very limited. This study applies formal Bayesian approach for uncertainty analysis of a widely used storm water management model and illustrates the methodology using a highly urbanized watershed in the Los Angeles Basin, California. A flexible likelihood function that accommodates heteroscedasticity, non-normality, and temporal correlation of model residuals has been used for the study along with a Markov-chain Monte Carlo-based sampling scheme. The solution of the UA model has been compared with the solution of the conventional calibration methodology widely practiced in water resources modeling. Results indicate that the maximum likelihood solution determined using the UA model produced runoff simulations that are of comparable accuracy with the solution of the traditional calibration method while also accurately characterizing structure of the model residuals. The UA model also successfully determined both parameter uncertainty and total predictive uncertainty for the watershed. Contribution of parameter uncertainty to total predictive uncertainty was found insignificant for the study watershed, underlying the importance of other sources of uncertainty, including data and model structure. Overall, the UA methodology proved promising for sensitivity analysis, calibration, parameter uncertainty, and total predictive uncertainty analysis of urban storm water management models.

Vegetation in a constructed storm water wetland (CSW) is essential for nutrient removal and thermal pollution reduction. However, a portion of the storage volume designed to retain the water quality event is occupied by vegetation, which may result in the CSW being undersized from a hydrologic perspective. The fraction of the volume associated with vegetation mass was unknown and designers have been unsure as to (1) how they should account for volume occupied by vegetation and (2) whether this was a significant design issue. Twelve storm water wetlands and one hybrid wet pond in North Carolina were sampled to assess their percent vegetative occupancy by volume. Vegetation was harvested from randomly selected $0.1{\mathrm{m}}^2$ grids at the normal pool elevation, and displacement tests were used to calculate the volume added by vegetation at elevations of 30 and 45 cm. Combining these data with observed vegetative cover, a water quality volume factor of safety accounting for vegetative occupancy was calculated for use in CSW design. Of the thirteen practices sampled, the highest factor of safety calculated for temporary pool elevations of 30 and 45 cm are 1.019 and 1.018, respectively. Therefore, a conservative recommended factor of safety is 1.025. Species-specific factors of safety (FOS) were also calculated; at the 30-cm temporary pool elevation, the Iris (Iris spp.) FOS was highest at 1.020, and the mixed vegetation FOS was lowest at 1.005. When designing a wetland, a more specific FOS may be calculated using a weighted average of the FOS for each species in the planting plan. The outcomes of this research allow designers to account for the volume taken up by emergent wetland vegetation; however, the uncertainty in other factors such as discrepancies between designed and as-built bathymetry probably have a greater impact on the design volume loss when compared with vegetation occupancy ($<2\%$).