Journal of Hydrologic Engineering

Engineering analyses and designs that involve flood frequency often require an estimate of the log-space coefficient of skewness. The existing procedure may require the use of a flood skew estimate obtained from a skew map, which is known to be inaccurate. The map also lacks a conceptual basis. The goal of this research was to provide a method for estimating flood skew that is accurate and has a sound conceptual basis, then test the method with hydrologic data, and compare the accuracy of the approach to the accuracy of a flood skew map of the same region. The conceptual basis for flood skew estimation involves rainfall skew and watershed and channel storage. Rainfall skew represents an upper bound on the population of runoff skew. Flood skew decreases from the rainfall skew for the same location as storage increases. Using data from the coastal plain of Maryland and Delaware, a rainfall skew of 1.4 was estimated. A regression equation based on storage variables was found to be statistically significant with a correlation coefficient of 0.87. A map based on the same data only provided a correlation coefficient of 0.71, with a standard error that was 40% larger than that of the storage-based model. Greater accuracy was achieved using a storage-based modeling approach than using a skew map for the same region.

The evaluation of crossing probability for a given threshold level of a discrete process is an important modeling issue in stochastic hydrology. Rice’s formula is a tool to measure the instantaneous crossing rate for a given threshold level in a continuous Gaussian process. This paper examines the discrete application of Rice’s formula to estimating the up-crossing probability in typical discrete stationary univariate Gaussian processes. The results based on Rice’s formula are found to follow a biased pattern relative to the exact solutions, and perform poorly in important scenarios, particularly for estimating resilience in these numerical experiments.

A real world case study of water and benefit allocation among stakeholders in the South Saskatchewan River Basin in Alberta, Canada, is presented. To be able to carry out this large-scale water basin system investigation, a useful approach, called the integrated hydrologic-economic river basin model, is developed for computing the optimal benefits of water for various coalitions of stakeholders by allocating available water to the members of a given coalition without reducing specified allocations to those outside of the coalition with respect to both water quantity and quality. Monthly net benefit functions for a rich range of different types of demands are employed in the model. Moreover, a separate irrigation water planning model is designed as a submodel to derive the monthly benefit functions of irrigation water uses. Using these calculated values, cooperative game theory is utilized for fairly reallocating water and benefits among stakeholders in the basin study.

The model CASC2D-SED was applied to the Goodwin Creek experimental watershed in Mississippi to define erosion model response to raster-based grid cell sizes. The model was parameterized at a $30\mathrm{m}$ grid, then calibrated and validated to three representative thunderstorms. The simulated hydrographs replicated the measurements of peak discharge, runoff volume, and time to peak. The model also calculated sediment yields within $\pm 50\%$ of the field measurements. Resampling the watershed digital elevation model at scales from $30\mathrm{m}$ to $330\mathrm{m}$ reduced the land surface slopes and changed the channel topology. In general, very good modeling results are obtained at grid sizes of $30\mathrm{m}$ and $90\mathrm{m}$ , which is comparable to the plot sizes of the universal soil loss equation. At grid sizes coarser than $150\mathrm{m}$ , the sediment source areas became less appropriately depicted and the calculated sediment delivery ratios became unrealistically high. Grid sizes smaller than $150\mathrm{m}$ are recommended for proper watershed simulation of upland erosion and sediment yield.

Automatic time stepping with global error control is proposed for the time integration of the diffusion equation to simulate groundwater flow in confined aquifers. The scheme is based on an a posteriori error estimate for the discontinuous Galerkin finite-element methods. A stability factor is involved in the error estimate and it is used to adapt the time step and control the global temporal error for the backward Euler method. The stability factor can be estimated by solving a dual problem. The stability factor is not sensitive to the accuracy of the dual solution and the overhead computational cost can be minimized by solving the dual problem using large time steps. Numerical experiments are conducted to show the application and the performance of the automatic time stepping scheme. Implementation of the scheme can lead to improvements in accuracy and efficiency of groundwater flow models.

Ice condition forecasts are very important for preventing ice disasters. Because of the complexity of ice conditions, traditional methods could hardly give accurate prediction in the ice condition forecast, especially for the meandering rivers such as the Yellow River, while the artificial neural networks (ANNs) have an obvious advantage over other traditional methods for forecasting ice conditions. An ANN model based on feed-forward back-propagation and improved by the Levenberg-Marquardt algorithm is applied to forecast the ice conditions of the Yellow River in the Inner Mongolia region. The forecast results in the winter of 2004–2005 are in good agreement with the measured ones. Simulation also shows that the ANN model is superior to the multiple linear regression model and the GM (0,1) model.

Estimation of transmissivity $T$ from specific capacity $Q∕s$ is a quick method of acquiring hydraulic aquifer properties by avoiding long-term cost-consuming pumping tests. In a similar way, the hydraulic conductivity $K$ can be estimated from specific capacity index $S_i$ . Studies between $T$ and $Q∕s$ in alluvial aquifers have been known for some time, but are much scarcer in fractured or karst rocks. This paper discusses the correlations between $T$ and $Q∕s$ and between $K$ and $S_i$ , based on a database of 298 wells in dolomite aquifers in Slovenia. According to normality tests, all parameters are log-normally distributed except $K$ by some tests. Results show that transmissivity can be reliably approximated from specific capacity by a simple linear equation, as expected in theory or also by log-transformed correlation. An important new observation is that the correlation coefficient for analyses of $T\text{-}Q∕s$ does not increase with logarithmic transformation, as noted by other authors. Similarly, the hydraulic conductivity can be reliably estimated from specific capacity index with significant results except for some uncertain correlations due to a small quantity of data.

Hydrologic regionalization can be employed for importing data from gauged sites to ungauged sites. A number of techniques have been developed for hydrologic regionalization. This study presents an approach that combines three regionalization techniques in order to obtain homogeneous regions and compute streamflow at desired ungauged sites. First, an agglomerative hierarchical clustering algorithm is used to determine the number of homogeneous regions. Second, homogeneous regions are defined by the $k$ -means partitioning method. Third, the flow duration curve (FDC) method is used to compute streamflow discharge at desired ungauged sites in homogeneous regions. The combined approach was then applied to 26 river basins in Turkey having 1,410 river gauging stations, where homogeneous regions were identified and streamflow discharge was calculated at ungauged sites. At many of the unguaged sites, small hydropower projects are being planned for development. It was found that these 1,410 river gauging stations (26 river basins) can be grouped into six homogeneous regions. Homogeneous regions were found to be better distributed in the nonhierarchical $k$ -means method than in the hierarchical method. FDCs, constructed for each homogeneous region, fit monthly discharge data with correlation coefficients between 95 and 100% for 85.7% of the gauging stations, between 85 and 94% for 12.5% of the stations, and between 72 and 84% for 1.8% of the stations.

Surface-subsurface interactions are an intrinsic component of the hydrologic response within a watershed; therefore, hydrologic modeling tools should consider these interactions to provide reliable predictions, especially during rainfall-runoff processes. This paper presents a fully implicit coupled model designed for hydrologic evaluation in wetlands, agricultural fields, etc. The model uses the depth-averaged two-dimensional (2D) diffusion wave equation for shallow surface water flow, and the three-dimensional (3D) mixed-form Richards equation for variably saturated subsurface flow. The interactions between surface and subsurface flows are considered via infiltration in dynamic equilibrium. A general framework for coupling the surface and subsurface flow equations is adopted, based on the continuity conditions of pressure head and exchange flux rather than the traditional conductance concept. The diffusion wave surface water equation is used as an upper boundary condition for the initial-boundary value problem of variably saturated subsurface flow. The coupled system of equations governing surface and subsurface flows is discretized using the finite-volume method in space and an implicit scheme in time. Component modules and the coupled flow model have been tested by comparing numerical results with published experimental data and analytical solutions. The verified integrated flow model has been applied to simulate the rainfall-runoff processes in a published field-scale experiment and the Deep Hollow Lake watershed, Mississippi. The results have demonstrated that the established numerical model is capable of simulating 3D subsurface flow and 2D surface shallow water flow as well as predicting the interactions between them.

To simulate daily runoff in the middle and lower regions of the Changjiang (Yangtze River) basin, we used an integrated, spatially distributed model consisting of the Hydrological Simulation Program—FORTRAN (HSPF) and two newly developed submodels, the paddy runoff model (PRM), simulating the runoff from paddy fields, and the lake discharge model (LDM), simulating the hydraulic effect of the Changjiang mainstream on the discharge from neighboring lakes. In the PRM, the ratios of the simulated monthly evapotranspiration relative to the observed potential evaporation were close to or exceeded 1.0 in both of the test subcatchments in the full ponding period with plentiful irrigation water, corresponding to many previous observations of evapotranspiraion rates from paddy fields. Whereas the original HSPF could not consider the backwater effect of the water level in the mainstream, the daily discharge from the lakes simulated by the LDM was dependent on the water level in the mainstream and corresponded to the observed data at the hydrologic stations on the channels connecting each lake with the mainstream, as shown by the daily $R^2$ (Nash–Sutcliffe coefficient) values for Dongting Lake (0.89 in the calibration period, 0.90 in the verification period) and for Poyang Lake (0.84 in the calibration period). For the runoff response from the whole basin simulated by the integrated model, the deviations of runoff volume and values of $R^2$ were $-4.2\%$ and 0.95 in the calibration and verification periods and 1.6% and 0.93, respectively, at Datong hydrologic station, which shows the good applicability of this model as a tool for management of water resources in the Changjiang basin.

The saltwater freshwater interface at the intertidal zones of surficial aquifers is continuously influenced by tidal effects. The numerical solution of this continuously changing intertidal saltwater interface profile is difficult. To simulate this dynamic system correctly appropriate definition and use of seaside boundary conditions are necessary. The estimation of submarine groundwater discharge may vary significantly depending on the way these boundary conditions are handled on the intertidal zone. In this study we have investigated this problem numerically to evaluate the effect of the use of different intertidal zone boundary conditions on the magnitude and the direction of submarine groundwater discharge as well as their effect on saltwater interface profiles. Our findings indicate that the numerical results are sensitive to density differences between the freshwater and seawater phases along with the type of boundary condition used at the intertidal zone. Based on these results, we conclude that the adaptable free exit boundary condition introduced by Frind in 1988 is the proper boundary condition to use on the advection dominant intertidal zone boundary. This approach necessitates the use of nontraditional boundary conditions in the transport model to prevent the artificial accumulation of saltwater mass underneath this boundary. The use of an improper boundary condition at this boundary may lead to the development of artificial fingering due to the accumulation of saltwater mass in the intertidal zone.

When we consider a location with a material (e.g., water, pollutant, sediment) passing through it, we can ask: “Where did the material come from and how long did it take to reach the location?” We can quantify the answer by defining the areas contributing to this location during various time periods as “resource sheds.” Various resource sheds and their source material distributions are rigorously defined and properties derived. For watershed hydrology, we compute resource sheds and their source distributions with a spatially distributed hydrology model by tracing water departing from a “cell” (say $1{\mathrm{km}}^2$ ) over one time interval, traveling through intermediate cells soil, groundwater, and surface zones, and arriving at the watershed mouth in another time interval. This requires modeling all cells, but only tracing contributions from one at a time. By then combining these simulations for all cell loadings, we construct a map of the contributions over the entire watershed for specific departure and arrival time intervals. We then combine results of several sets of simulations to determine the source distribution for any time period and infer resource sheds from these mappings. We give examples for the Maumee River watershed in northern Ohio, discuss computation reduction, and suggest future extensions to other materials.

We posit that a more complete understanding of runoff response from urban catchments requires an assessment of the hydrologic behavior of composite impervious-pervious patches. We therefore examined how the factors of impervious extent, connectivity, and antecedent moisture content of pervious areas might affect mechanisms of runoff production at small spatial scales in a laboratory setting. We used rainfall simulation (with a storm comprised of $20,30,40\mathrm{mm}{\mathrm{hr}}^{-1}$ rainfall rates for 48, 24, and $24\min$ , respectively) to observe surface runoff from $0.6{\mathrm{m}}^2$ boxes (impervious or pervious-soil) $0.2\mathrm{m}$ deep that were connected together in series to produce different arrangements of impervious and pervious surfaces (0, 25, 50% impervious) with different connectivity to the outlet (disconnected, connected), and under two different antecedent moisture conditions for pervious areas (drier, wetter). In general, an increase in percent impervious area led to fewer opportunities for infiltration, and a quicker onset of runoff, which was intensified by wetter antecedent moisture conditions and connectivity to the outlet. Runoff rate ratios were strongly affected by antecedent moisture condition and somewhat less significantly by an interaction between impervious area extent and its connectivity status. In each impervious treatment, we observed a decreased time to runoff initiation and higher final runoff rate ratio for wetter than drier treatments. Interestingly, we found that the connectivity of 25% impervious area accounted for differences in runoff rate ratio only early in the simulation. The patterns in runoff from connected and disconnected 25% treatments eventually converged, leaving antecedent moisture conditions the only relevant factor. As impervious area was increased to 50%, we noted a precipitous decline in infiltration rates due to a reduction in infiltration opportunities and infiltration behavior of the pervious surfaces in these treatments. Evidence of return flow in the 50% disconnected treatment is presented and discussed in the context of saturation-excess runoff mechanisms. These experimental results are then discussed in terms of their potential extension and application to better understand aspects of urban hydrology and models thereof.

One of the challenges in groundwater modeling is the prediction of hydraulic head in close proximity to a pumping well using a regional-scale model. Typical applications of numerical models to field-scale problems generally require large grids that can seldom accommodate cells as small as the actual well diameter. In this paper, we apply a dynamically integrated “hierarchical patch dynamics paradigm (HPDP)” to model detailed near-well dynamics and interactions. The HPDP enables converting a large, complex problem into a network of hierarchically nested and dynamically coupled patch models that can be easily solved. The performance of the HPDP is verified against the analytical solution for a single well, against a superposition of analytical solutions for a wellfield, and against a numerical solution in a three-dimensional heterogeneous system. The results show that the HPDP is capable of providing an accurate and efficient representation of head in a wellfield in large-scale hydrogeologic systems.

Gaps in measured discharge series are common in Iceland, especially during winter, but reconstructed series are needed for a number of applications. Modeled series are therefore sometimes used to fill the gaps to create a continuous series, but the modeled inserts may have different statistical characteristics from the observed series, whilst their initial and final values rarely coincide with the corresponding measured values. Where gaps are numerous, individual treatment of each gap is time-consuming and often impractical. A method for adjusting modeled sequences to be inserted into gaps was therefore designed with the objective of reducing the difference between model-generated inserts and recorded series. This paper reports on the performance of the method and shows how an improved fit can be attained by adjusting the model-generated inserts so as to smoothly match the discharges observed before and after a gap.

This note investigates different procedures to improve radar-based estimates of rainfall accumulation, using the City of Winnipeg, Canada, as an example. Specific contributions include the implementation assessment of an advection scheme to track the movement of storms, a scheme to account for wind-induced drift, and an attenuation correction scheme that includes a simple hail identification method. It was found that a combination of these schemes improved the correlation between gauge radar $10\text{-}\min$ accumulations from 0.48 to 0.86 for six rainfall events that occurred in the summer of 2000 in Winnipeg.