Journal of Hydrologic Engineering

A two-layer soil water balance model is developed to provide an efficient and robust description of land surface dynamics in response to atmospheric evaporative events. Soil, vegetation, and atmosphere are coupled dynamically under the assumption that soil moisture profiles approximately preserve similarity during simultaneous atmospheric drying and gravity drainage. The exfiltration flux at the land surface in response to the atmospheric evaporative demand is limited by relating the surface resistance to water vapor transfer in the Penman-Monteith equation to the near-surface soil status. In addition, the control of deeper soils on both exfiltration and drainage is expressed by performing a time compression approximation water balance over the entire drying profile and by scaling the obtained exfiltration and drainage fluxes to the near-surface soil layers. The reliability and robustness of the proposed formulation is evaluated with rates of evaporation calculated from measurements of the Bowen ratio and soil moisture data obtained from time domain reflectometry measurements for a bare soil field in the Zwalmbeek catchment (Belgium).

This study investigated the use of similarity profiles in space for analyzing surface-runoff hydrographs occurring at the outlet of small watersheds. The spatial behavior of the flow depth over each element (overland region or stream reach) was approximated by a sine function, so that the partial differential equations governing water movement were reduced to ordinary differential equations. Solutions of these reduced equations were obtained at a smaller computational cost, with analytical solutions available in some cases. These solutions then were extended to study surface water movement over small watersheds that could be represented as a sequence of cascading overland flow planes and streams. The performance of this model was studied by comparing with results from another physically based model, KINEROS, and experimental observations. It was concluded that the similarity solutions are robust and almost as accurate as numerical solutions that were obtained from physically based models, when used within their range of applicability explicitly defined in the paper. Some guidelines for spatial discretization of the watershed for appropriate use of the similarity solutions have been provided.

A distributed hydrological model is developed and applied to the Fuji River basin of 3,432 km² in Japan, which is divided into squared meshes of 1-km resolution. Mesh meteorological data of temperature, wind speed, sunshine, precipitation, cloud amount, and humidity are produced from weather stations and geographic information system data by a stepwise regression method that characterizes relationships between meteorology and geography. This method provides spatially allocated daily input data. Then a comparatively simple distributed model is proposed, with a submodel of soil-vegetation-atmosphere-transfer type for any mesh and a conceptual submodel for river routing. Spatial distribution patterns and mesh-specific values of model parameters are calibrated with a geographic information system database and 3 years of meteorologic-hydrologic observations. The daily run model can simulate approximately daily processes of snow accumulation and melting, interception, evapotranspiration, and runoff components, which gives relatively good reproduction of annual water yield. The hourly run model version can simulate well hourly hydrograph in the rainy season, improving the daily model.

Mean areal precipitation values (MAPX) derived from next generation weather radar (NEXRAD) stage III data are compared with mean areal precipitation (MAP) values derived from a precipitation gauge network. The gauge-derived MAPs are computed using Thiessen polygon weighting, whereas the radar-based MAPXs utilize the gridded stage III radar precipitation products that have been conditioned with gauge measurements and have been merged with overlapping radar fields. We compare over 4,000 pairs of MAPX and MAP estimates over a 3-year time period for each of eight basins in the southern plains region of the United States. Over the long term, mean areal estimates derived from NEXRAD generally are 5–10% below gauge-derived estimates. In the smallest basin, the long-term MAPX mean was greater than the MAP. For storm events, a slight tendency for NEXRAD to measure fewer yet more intense intervals of precipitation is identified. Comparison of hydrologic simulations using the two forcings indicates that significant differences in runoff volume can result. This work is aimed at providing insight into the use of a data product that is becoming increasingly available for public use. It also is aimed at investigating the use of radar data in hydrologic models that have been calibrated using gauge-based precipitation estimates.

In hydrocarbon-contaminated sites, it is common practice to estimate the total hydrocarbon volume based on the hydrocarbon thickness measurements made in monitoring wells. In most cases, the hydrocarbon thickness measurement data obtained from a well field and the volume estimate analysis based on these data show significant variations over time. In this paper we consider an important cause of this variation, which is neglected in most field applications, i.e., the unsteady ground-water table fluctuations and its effect on monitoring well measurements. The analytical solutions we present indicate that, under fluctuating ground-water table conditions, if the hydrocarbon volume estimates in the formation are based on monitoring well measurements made at one point in time, or if these estimates are based on a sequence of measurements made over a period of time, which then are evaluated independently, then the volume estimate results will be in significant error. The analysis discussed in this paper is based on the analytical solution of volumetric equilibrium equations for constant residual saturation levels. Thus, results discussed in this paper only may represent the best estimates of hydraulic equilibrium conditions at a contaminated site. For a complete analysis, in addition to the effect of ground-water table movement, the effect of entrapment on residual saturation, capillarity, hystereses, threshold pressures, and non-Darcian behavior should be included in the analysis. However, because of inherent nonlinearities introduced, such cases cannot be solved using analytical methods and the correlation between light nonaqueous phase liquid (LNAPL) thickness in the aquifer and the monitoring well cannot be determined easily. Based on our findings, we conclude that present-day volume estimation techniques, which are based solely on physical equilibrium models, are not reliable tools to estimate light nonaqueous phase liquid volume at contaminated sites under unsteady conditions.

The usefulness of a hydrologic model depends on how well the model is calibrated. Therefore, the calibration procedure must be conducted carefully to maximize the reliability of the model. In general, manual procedures for calibration can be extremely time-consuming and frustrating, and this has been a major factor inhibiting the widespread use of the more sophisticated and complex hydrologic models. A global optimization algorithm entitled shuffled complex evolution recently was developed that has proved to be consistent, effective, and efficient in locating the globally optimal model parameters of a hydrologic model. In this paper, the capability of the shuffled complex evolution automatic procedure is compared with the interactive multilevel calibration multistage semiautomated method developed for calibration of the Sacramento soil moisture accounting streamflow forecasting model of the U.S. National Weather Service. The results suggest that the state of the art in automatic calibration now can be expected to perform with a level of skill approaching that of a well-trained hydrologist. This enables the hydrologist to take advantage of the power of automated methods to obtain good parameter estimates that are consistent with the historical data and to then use personal judgment to refine these estimates and account for other factors and knowledge not incorporated easily into the automated procedure. The analysis also suggests that simple split-sample testing of model performance is not capable of reliably indicating the existence of model divergence and that more robust performance evaluation criteria are needed.

Hydrologists and water resource planners in the Pacific Northwest region of the United States are under pressure from competing water users to make long-range forecasts of streamflow. At present, long-range streamflow forecasts are not offered. Yet in the Pacific Northwest, a significant lag relationship exists between (El $\mathrm{Ni}ño-\mathrm{Southern}\mathrm{Oscillation})$ and streamflow. Using this lag relationship, this study proposes to extend the prediction of spring-summer runoff in the Pacific Northwest from the current 1- to 3-month lead time to a 3- to 7-month lead time. This study looks at the Columbia River basin, where testing is being conducted on a long-range seasonal streamflow forecasting model that uses, as predictors, persistence in streamflow along with two El $\mathrm{Ni}ño-\mathrm{Southern}\mathrm{Oscillation}$ indicators: the Southern Oscillation Index and the Wright sea surface temperatures. A probabilistic streamflow forecast is made from an optimal linear combination of persistence, Southern Oscillation Index linear discriminant analysis and Wright sea surface temperature linear discriminant analysis forecasts. This final combination forecast is referred to as a “consensus forecast.” This consensus forecast model is tested on data covering the period 1933–1992 from seven of the Columbia River basin streamflow stations. Data on an additional station, Columbia River at The Dalles, covers a period from 1911 to 1992. It was found that, at some stations, a 3- to 7-month lead time forecast of spring-summer runoff had better skill than climatology.

The writers have formulated quasi-analytical nonlinear solutions to investigate the uncertainty of one-dimensional ground-water flow in heterogeneous formations in which the standard deviation of the log hydraulic conductivity σ_ln(K) > 1. The writers have studied three cases for fixed boundary conditions: a semibounded domain, a free domain, and a both-side-bounded domain. The writers have also studied the random boundary conditions for the bounded domain. The writers have obtained variances and covariances of head for semibounded and free domains, and have calculated variances of head for the both-side-bounded domain. Monte Carlo simulations are generated to obtain the head variances and are used to verify the quasi-analytical solutions. The ergodicity assumption is proven to be rigorously correct in a semibounded or a free domain but not correct in a both-side-bounded domain by using a linear autocorrelation function.

This paper describes the development of simplified momentum equations, in stage as well as in discharge formulations, governing the transition between the diffusion and the kinematic waves (including the latter). It also describes the application of these equations to arrive at the approximate convection-diffusion equations. The appropriateness of these approximate convection-diffusion equations to model flood waves in the transition range is established, and the characteristics of these equations are studied from the point of view of description of the loop-rating curve. The development of these simplified equations provides a theoretical justification for their use in the well-known “Jones formula” expressed as Q/Q₀ = {1 + [1/(S_0c)]∂y/∂t}^1/2 for converting the stage to discharge of a diffusive flood wave—an approach that has hitherto been considered to be logically incorrect.

A hydrologic model, capable of approximating surface and subsurface runoff from a forested shallow-soil watershed without the time of concentration parameter, is developed. A watershed is modeled as a series of vertically and horizontally linked nonlinear storages, each represented by a lumped mass-balance equation. The resulting system of ordinary differential equations is evaluated by the fourth-order Runge-Kutta finite-difference method. Comparison of this simpler modeling approach with method of characteristics and Petrov-Galerkin finite-element solutions shows good agreement. However, because the performance of the preliminary formulation on a forested watershed was unsatisfactory, an extended subsurface model is developed that explicitly accounts for the presence of both the soil and the small channels between the soil layer and bedrock face. Application of the extended model to the same forested watershed shows significantly improved agreement between the observed and computed runoff hydrographs.

Gridded precipitation products created as part of the U.S. National Weather Service's Next Generation Weather Radar (NEXRAD) program are referenced to a national grid called the Hydrologic Rainfall Analysis Project (HRAP) grid, which is used to mosaic precipitation estimates from different radars into a single national precipitation map. The HRAP grid is defined in a polar stereographic map projection that is formed on a plane intersecting a spherical earth datum at 60° N. This paper describes a method for transforming HRAP grid cells into a coordinate system commonly used for mapping geographic information system data sets, quantifies mapping errors associated with using the HRAP coordinate system, and outlines an approach to reduce these mapping errors. The NEXRAD radar rainfall processing software assumes that the earth is spherical rather than using a more accurate ellipsoidal representation. This assumption causes east-west distances to be distorted relative to north-south distances. The magnitude of this shape distortion and the magnitude of the scale factor distortion associated with the polar stereographic projection it self are quantified for different latitudes.

In level pool routing, which is the simplest hydrological routing method, the downstream discharge may be expressed explicitly in terms of the inflow and the channel or reservoir characteristics. The level pool routing equation can also be used to estimate the inflow hydrograph given the outflow hydrograph and the water level in the reservoir. Unfortunately, use of the traditional level pool routing method, which is based on the implicit finite difference scheme, for reverse routing has been unsuccessful, despite the simplicity of the problem. If a simple explicit centered differencing scheme is used instead for simulating the inflow hydrograph, the problems associated with traditional schemes, which require the application of filtering techniques, are bypassed. This is demonstrated using a realistic hypothetical example and a case study. The explicit scheme results are comparable in accuracy with results from the implicit scheme without resorting to the use of filtering techniques. The simple explicit scheme produces a direct solution to the problem and should be the preferred method for both level pool and reverse level pool routing.