Journal of Hydrologic Engineering

This paper and its companion critique the common frequency analysis techniques for hydrological extremes—in particular, the claims that their increasingly refined mathematical structures have increased the accuracy and credibility of the extrapolated upper tails of the fitted distribution models over and above that achieved by the 50-year-old empirical methods. Part 1 compares the common-sense engineering origins of frequency analysis with its present ostensibly “rigorous theory”; some myths advanced under the banner of the latter are analyzed in greater detail in Part 2.

This paper provides a critical examination of some common “theoretically based” approaches to frequency analysis (a general discussion of which appears in Part 1 of the paper) and the myths they have generated about the upper tails of hydrological distributions.

It has been hypothesized that the Solitario Canyon watershed, just west of Yucca Mountain, Nev., could provide a zone of focused ground-water recharge through its ephemeral-stream channels that might be an important factor in the performance of the potential repository for high-level radioactive waste. To examine this possibility, the KINEROS2 rainfall-runoff model was used to estimate the infiltration from surface runoff in each of 79 channel elements using a 100-year simulated rainfall sequence and two sets of saturated hydraulic conductivity KS values. The annual average estimated infiltration ranged from 0.18 to 0.57 mm/year (expressed as depth over the entire 10.7 km² watershed). Evaporation and transpiration by vegetation near the channels were neglected, therefore these quantities of channel infiltration are not assumed to be ground-water recharge. The simulated rainfall intensities compared favorably with depth-duration-frequency curves for nearby stations. The simulated 100-year peak discharge rates for Solitario Canyon and for four subwatersheds exceeded the regionally estimated 100-year rates for the lowest values of KS and were slightly below those rates for the higher values of KS, suggesting the true value of peak discharge (and channel infiltration from surface runoff) is bounded by those obtained for the 100-year simulations.

Physically based spatially distributed (PBSD) river basin models have been available for over 10 years. One of their strengths lies in the way the surface and subsurface are represented as coupled parts of a whole, giving ground-water flows that are controlled by such factors as realistic surface saturation and infiltration, and surface conditions that are controlled by realistic groundwater levels, discharges, and so forth. PBSD sediment and solute transport models can be integrated into PBSD river basin modeling systems, and the integrated systems are powerful tools for studying the environmental impacts associated with land erosion, pollution, and the effects of changes in land use and climate, and also in studying surface-water and ground-water resources and their management. SHETRAN takes PBSD river basin modeling a step further, in that multifraction sediment transport and multiple, reactive solute transport are handled within a single system, fully coupled to water flow, and the subsurface is modeled as a fully 3D variably saturated heterogeneous medium. SHETRAN therefore has a substantial capability for addressing environmental and water resources problems that span the traditional disciplines of river basin and ground-water modeling.

The occurrence probabilities, return periods, and risk of drought events are estimated for dependent hydrologic processes. Traditionally, Markovian models have been used for modeling hydrologic processes having short-term time dependence. However, they are inadequate for processes exhibiting longer time dependence. In this paper, low-order discrete autoregressive moving average (DARMA) models are used for modeling the variability of wet and dry years. Specifically, we center our attention on the occurrence of drought events, particularly their duration, by using the concept of runs. The probability distribution of drought occurrence, expected values and variances of first arrival and interarrival times of drought events, and the associated risks are derived. The derived equations and algorithms are verified by Monte Carlo simulation experiments. The applicability of the proposed methods is demonstrated by using annual streamflow data of the South Platte River in Colorado and the Niger River in Africa. It is concluded that the proposed methods are quite useful for modeling drought events, assuming that low-order DARMA models can describe wet and dry years.

Due to a number of hydrologic and meteorologic factors, Houston, Texas, is subject to major flooding problems. The National Weather Service's (NWS) WSR-88D radar (NEXRAD) was used to estimate the areal and spatial distribution of rainfall for three storms over the Brays Bayou watershed in Houston for hydrologic modeling purposes. Rainfall rates estimated from the NEXRAD radar compared favorably to the point rainfall measurements at the operating rain gauges in the watershed for the large October 1994 storm, but tended to overpredict the point rainfall measurements at the rain gauges for two smaller events. Using HEC-1, two model runs were made using the unadjusted NEXRAD-estimated rainfall rates and also using the rain gauge network, and the modeled outflow hydrographs were compared to the measured hydrograph for all three storms. The results from the radar proved to be as accurate, and in some cases more accurate, than the rain gauge model when predicting runoff volume, peak flow, and time of peak. Results from this research are being used to create a local flood warning system on the Brays Bayou watershed using real-time NEXRAD radar.

Discharge estimates obtained from gauged data are generally considered to be more accurate than model-estimated discharges. While most designs do not occur at the location of a gauging station, many designs are required at sites near but not at the gauged location. Transposition methods transfer discharge estimates from a gauged location to a nearby location on the same river. Despite their frequency of use, little is known about their accuracy and sensitivity. Because the state of Maryland is considering using discharge transposition, an assessment of the accuracy and sensitivity of two methods was undertaken: the area-ratio method and Sauer's weighting function method. Gauged data from nine states were used to evaluate the two methods for recurrence intervals of 2, 10, and 100 years. The criterion used to reject a data pair on the basis of time-sampling variation was when the T-year flood for the larger drainage area was less than the T-year flood for the smaller drainage area. Approximately 50% of the station pairs of gauged data had to be discarded because the data were collected during different periods, which is known as the time-sampling-variation problem. Sauer's method provided slightly better accuracy than the area-ratio method. Sensitivity analyses of the two methods are used to assess their rationality. Overall, both methods provide improved accuracy when the ungauged site is near the gauged site. The accuracy results suggest that Sauer's method can be reasonably applied if the drainage area of the ungauged site is within ±25% of the area of the gauged station, but the sensitivity analysis suggests that the method should be applied with caution because of its potential irrationality.

Hydrological homogeneity is often assumed whenever there is lack of data in a basin and hydrologic analysis is carried out utilizing the data from neighboring basins. However, the neighboring basins may not be hydrologically homogeneous with the basin under consideration. The identification of hydrological homogeneity of basins or clustering of basins based on homogeneity demands a high degree of subjective judgment. Although an expert integrates multivariate, nonlinear, and unquantifiable factors quite well based on his subjective judgment, a different expert may not reproduce the same results. Hence, there is a need to have a rational procedure for clustering or grouping of basins based on hydrometeorological homogeneity. In this study, the application of artificial neural network for clustering the basins on the basis of hydrological homogeneity is investigated. First, an attempt is carried out to check whether the classifications in the data hyperspace have any physical meaning or not. Subsequently, it is attempted to check whether the clustering with factors that affect runoff has any effect in runoff values of each cluster. The statistics presented indicates that there is congregation about the cluster center. Finally, use of clustering of basins based on homogeneity in data hyperspace is investigated.

A precipitation time series is often a necessary input for the analysis and design of hydrologic and hydraulic systems. The precipitation records employed for these purposes can be either measured observations or generated by stochastic simulation. One common problem with recorded and generated precipitation data is that often it is not in small enough time increments for use in engineering applications (e.g., continuous hydrologic simulation). To solve this problem, rainfall amounts can be disaggregated into shorter time increments. This paper evaluates the use of artificial neural networks (ANNs) for the disaggregation of hourly rainfall data into subhourly time increments. Two different ANN models are introduced and evaluated in this paper. A back-propagation/steepest-descent algorithm trains one model and the other model uses the idea of self-organization in a competitive learning ANN. The results indicate that the performance of both ANN models are: (1) comparable to other disaggregation schemes in terms of predicting the overall disaggregated rainfall hyetograph; and (2) improvements over other disaggregation models in the prediction of the maximum incremental rainfall intensity (depth) within a storm hyetograph. Based on these results, the use of ANN models can be recommended as a viable alternative for hourly rainfall disaggregation.

In this paper, an overlapping control volume method is presented for the numerical solution of transient 2D solute transport problems in ground water. The method is applicable for nonorthogonal grids and uses an isoparametric formulation for computing the dispersion and for second-order upwinding. Time integration is performed using an implicit approach. Three test cases are considered for comparing the numerical with analytical solutions. The scheme is second order in space and, when combined with Crank-Nicholson, is also second order in time. The results using Crank-Nicholson and full implicit (first-order) time integration methods are compared for problems with a variety of boundary conditions. For diffusion-dominated flows (P_Δ ≤ 2) Crank-Nicholson works well, but for convection-dominated flows it produces spurious oscillations due to numerical dispersion errors. These oscillations are controlled by a flux limiter but only for Courant numbers below unity. It is shown that for high Courant and Péclet numbers a slight weighting of the time stepping toward fully implicit is effective against spurious oscillations and offers an optimum compromise between numerical dissipation and dispersion errors for a wide range of Courant and Péclet numbers. The scheme is shown to work on mildly nonorthogonal grids.

The linear correlation of air temperatures and stream temperatures was investigated in this paper. Time lag, time scale, and mesoscale air temperature variability influence the magnitude of the regression coefficients of the linear model, as well as the strength of correlation. Monthly and weekly time scales were found to produce the best correlations. Streams affected by impoundments, reservoirs, and/or artificial heat inputs generally limit the air/stream interaction and produce poor correlations. Ground-water inflows, stream shading, and wind sheltering limit the influence of air temperature on stream temperature thereby reducing the slope of the relationship, but do not typically affect the strength of correlation. Air/stream temperature correlations from Minnesota remained linear throughout the extent of air temperatures experienced, whereas air/stream temperature correlations from Oklahoma departed from linearity as air temperatures exceeded approximately 25°C due to evaporative cooling. Linear air/stream temperature correlations will overestimate stream temperatures at extreme air temperatures.

The propagation of flood waves down very long prismatic rectangular channels was numerically simulated using the full St. Venant equations in the full equations (FEQ) model. The importance of nonkinematic effects on the behavior of each flood wave was quantified by calculating the dimensionless width of the simulated looped rating curve, which is directly related to the sum of the nonkinematic terms in the equation of motion. Dimensionless loop widths, W_L, were found to be closely related to the dimensionless wave period, τ, of Ponce et al. for a wide range of simulated bed slopes, channel widths, Manning n-values, and inflow hydrograph rise times. Since the variables defining τ can be estimated prior to detailed modeling, ranking of streams in order of decreasing values of estimated τ should roughly correspond to ranking in order of increasing nonkinematic effects. However, backwater and expansion/contraction effects are not captured by τ.

In the past, flash floods were characterized by rainfall and limited effort was made to examine them from a runoff perspective. For example, the question of how flash floods differ from other floods remains to be investigated. In order to satisfy this need, runoff hydrographs of 30 flood events from four watersheds in eastern Kentucky are selected and analyzed. In flash flood situations, certain portions of the hydrograph, such as the rising limb, time to peak, and magnitude of the peak, are important. To recognize this, runoff hydrograph characteristics such as the rising curve gradient, K, flood magnitude ratio, M, and flood response time, TP, are evaluated directly from observed floods. Using these characteristics, relative severity factors RK, RM, and RT are determined and summed to obtain a flash flood index, RF. Results from this study show that a reported flash flood event of May 18, 1995, has values RK = 3, RM = 16, and RT = 6 and a flash flood index, RF = 25. Overall, half of the flood events have flash flood indices greater than or equal to 15.

New York City transports drinking water from distant surface reservoirs. Accurate understanding of the city's water demand pattern is needed to identify possible conservation measures and plan future water supply projects. Previous studies of other U.S. cities have suggested that weather conditions affect daily water use. This hypothesis is studied through the daily water use data for New York City from 1982 to 1991. It is found that, during the winter months, weather has little effect on water use, while during summer periods, precipitation causes a decrease in daily water use. Graphs of heat functions (daily water demand deviations from its annual average value versus daily average temperature) show that above 78°F, water demand increases linearly with average daily temperature.