Journal of Hydrologic Engineering

Forum papers are thought-provoking opinion pieces or essays founded in fact, sometimes containing speculation, on a civil engineering topic of general interest and relevance to the readership of the journal. The views expressed in this Forum article do not necessarily reflect the views of ASCE or the Editorial Board of the journal.

A comparison is provided between parametric and nonparametric approaches for the calculation of two drought indices: the standardized precipitation index (SPI) and the standardized streamflow index (SSI). The parametric approach has been implemented considering two possibilities: (1) a unique probability distribution for the variable of interest with parameters depending on the month of the year, and (2) the probability distribution depending on the month of the year. The parameters’ estimation is made using the maximum likelihood method, while the distribution selection is operated using the Kolmogorov-Smirnov goodness-of-fit test. For the nonparametric approach, the Weibull plotting position has been used to calculate the cumulative frequency. For both indices, the 1-month, 3-month, and 12-month time scales have been considered. As a case study, two 80-year-long monthly time series in Italy are considered: Roma Collegio Romano for precipitation, and Tevere River Basin at Ripetta for streamflow, both in the period 1921–2000. The analysis of the two time series shows how the differences between parametric and nonparametric approaches are more evident in terms of severity and less in terms of duration, and interarrival time. The mean value of the percentage differences, in absolute value, in terms of severity is $\sim 30–175\%$ for SPI, and $\sim 30–80\%$ for SSI. In terms of duration, it is $\sim 6–13\%$, while in terms of interarrival time it is $\sim 1–3\%$, both for SPI and SSI.

The Green-Ampt equation describing the infiltration of water into soil requires that the rainfall intensity is greater than the infiltration capacity of the soil at all times. The well-known Mein-Larson model modifies the Green-Ampt equation and attempts to avoid this restriction for infiltration of water into a homogeneous soil under rainfall of constant intensity. The rainfall intensity is considered to be less than the infiltration capacity of the soil in the initial phases of a storm. The Mein-Larson model assumes that until the time to ponding of water over the soil surface, the infiltration rate is constant and equals the rainfall intensity. This study shows that the Mein-Larson model underestimates the ponding time and that the cumulative infiltration predicted by the model is always less than the cumulative infiltration potential of the soil. To avoid these drawbacks, an alternative modification of the Green-Ampt equation is presented. In the proposed modification, the ponding time is considered to occur when the cumulative infiltration potential of the soil is fully satisfied. Beyond the ponding time, both the infiltration rate and cumulative infiltration follow the classical Green-Ampt equations. The modification is achieved by using asymptotic matching technique to render the model applicable within a wide range of rainfall rates greater or less than the soil infiltration capacities. This modification will help extend the ability to analyze hydrologic problems related to water infiltration into soil in different environments of high or low rainfall intensities. Case studies are provided to demonstrate the behavior of the proposed model and compare it with the Mein-Larson modification.

Permeable block pavements are designed to treat and delay the water flow in an urban environment. In order to measure the runoff characteristics of the permeable block pavement, a rainfall simulation test was conducted at the accelerated pavement testing facility. This paper focuses on determining the runoff characteristics in terms of the permeability coefficients of the block pavements under different rainfall conditions, such as rainfall intensity, quantity, and time intervals. In this study, four different pavement test sections were constructed and their runoff coefficient values were determined as follows: (1) 0.19–0.45 for the porous block-porous cement treated base (CTB) section, (2) 0.27–0.61 for the porous block-dense CTB section, (3) 0.50–0.78 for the dense block-porous CTB section, and (4) 0.77–0.85 for the dense block-dense CTB section. Particularly, when the rainfall intensity increased from 60 to $200\mathrm{mm}/\mathrm{h}$ on the porous block pavement, 5–45% additional surface runoff occurred. Further, 20% more surface runoff occurred in the saturated surface condition compared to the dry condition.

A generalized error analysis is discussed in this paper to investigate solution accuracy for watershed simulation results using overland flow equations. The total or compounded error for watershed simulations may increase with increasing spatial resolution, scale, and number of iterations to convergence if inaccurate time-step is chosen for the numerical simulations. A relationship is derived between watershed scale error, overland flow plane error, finite element, and node errors. The scale versus compounded error relationship for watershed scale simulations is a function of element spatial discretization level, dynamic time-step criteria, and the relationship between element and nodal errors. The Courant number and wave number space solution is simulation optimal and would require separate solutions of Fourier-transform (and its inversion) of finite element nodal equations to obtain solution accuracy within 5%. Dynamic FEM square grid discretization criteria is derived from the results of eigenvalue analysis and dynamic time-step results for various finite element schemes—consistent, lumped, and upwind. Courant number and amplification factor evolutions have been determined for a reference case for consistent, lumped, and upwind finite element schemes. The number of amplification factors that are greater than 1.0 essentially determine the reason why oscillations could be noticed in numerical simulations even though the Courant number is less than 1.0 for few of the finite element numerical schemes.

The formulation of Kinoshita curves takes into account a number of physiographic characteristics and the configuration of the river; however, it is the curve amplitude, the one that represents the main characteristic of this formulation. This main feature is known as the angular sinuosity coefficient ($\theta$). In this paper an alternative expression for meander prediction formulation, based on a stochastic multivariate analysis of the geomorphologic and physiographic characteristics of a river is proposed. Stochastic models are used to simulate 3,480 occurrences of the sixteen characteristics proposed to characterize the meandering of Cahuacan River in the Mexican state of Chiapas. The prioritization of the variables obtained through an empirical orthogonal functions (EOF) analysis clearly showed the existence of three groups of parameters, which altogether explain the behavior of the meandering of Cahuacan River. The first group is formed by the morphologic characteristics of the river. The second group corresponds to the hydrologic features of the basin, and the third one to the morphologic and geometric characteristics of the river. The computation of the confidence limits, although from a statistical approach, constitutes a good tool to consolidate the arguments that define the zones at potential risk. The stochastic simulation of the future conditions of the river allows the precise definition of the zones directly in the field.

A formal Bayesian methodology is presented for integrated model calibration and risk-based water quality management using Bayesian Monte Carlo simulation and maximum likelihood estimation (BMCML). The primary focus is on lucid integration of model calibration with risk-based water quality management and total maximum daily load (TMDL) estimation under conditions of uncertainty. The sources of uncertainty considered in the analysis are modeling errors, observational data errors and fuzziness of the water quality standard. The difference between observed data or transformation thereof and corresponding model response is assumed to follow first-order Markov process, a specific case of which is statistically independent Gaussian errors. The BMCML method starts with sampling parameter sets from prior probability distributions of the model parameters and uses Bayes theorem and the maximum likelihood technique to estimate the triplicate (variance of residual errors, bias and autocorrelation coefficient of total errors) for each parameter set and the corresponding likelihood value. By approximating integration over the entire parameter space discretely, analytical expressions are derived for the cumulative probability distributions of model outputs and probability of violating water quality standards. The solution of the TMDL problem and related margin of safety (MOS) is then framed in the context of the developed Bayesian framework. Three example applications of varying complexities are utilized to demonstrate the versatility of the Bayesian methodology for water quality management. The BMCML methodology is validated using a hypothetical lake-phosphorus model and familiar statistical benchmarks. It is shown that the risk-based framework can estimate the reliability of an arbitrarily selected MOS as demonstrated in the Fork Creek bacteria and Shunganunga Creek dissolved oxygen TMDL case-studies. It is also shown that neglecting covariation among model parameters (i.e., by sampling parameter values from their posterior marginal distributions) influences the estimation of probability of exceedance and could potentially lead to the overestimation of the MOS at low risk levels.

The hydrologic performance of three partial-infiltration permeable pavement (PP) systems was evaluated at the Kortright Centre for Conservation in Vaughan, Ontario, Canada over 22 months. Partial-infiltration systems allow some stormwater to infiltrate into native soils and drain excess water by way of underdrains. The native soils at Kortright are composed of clayey silt and silty clay till, with clay content ranging from 7 to 30%. Flow restrictors on the underdrains were adjusted to the smallest orifice possible to assess the potential for stormwater outflow volume reductions. The hydraulic behavior of the PP systems was compared with runoff from an asphalt parking lot control. Peak outflow rates from PP were 91% smaller than peak flowrates of asphalt runoff on average, and attenuation of stormwater was observed during all seasons. Stormwater was found to infiltrate at the surface of the PP systems throughout two winters. Although increases in outflow were observed during periods of seasonal thawing due to the delayed release of infiltrating stormwater, the PP systems (with restricted flows from the underdrains) reduced overall stormwater outflow volume by 43% and completely captured (i.e., infiltrated and evaporated) most rainfall events that were less than 7 mm in depth. The study demonstrated that in cold climates and over low permeability soils, partial-infiltration permeable pavements reduce the volume, peak flow, and frequency of storm flows. These changes to the hydrology of stormwater are important for achieving water quality benefits as well as sustaining a more natural water balance and flow regime.

A model that couples the rock matrix and fracture flow is employed to simulate the flow and solute transport in fractured porous media. The porous matrix is represented using an equivalent continuum model, and discrete fractures are described using a discrete-fracture network model. Data on the fracture properties were collected from the study area to randomly generate the fractures. The model was calibrated and validated using the measured groundwater levels and tracer test data. The results indicate that the groundwater levels and concentrations simulated using the coupled model agree well with the observed data. In contrast, those obtained using the continuum model for the entire domain do not. In the fracture areas in which abnormal groundwater levels were observed (i.e., where the level was lower than the Yalong River stage), the coupled model effectively captures and replicates the primary geological and hydrogeological characteristics. Once again, the continuum model results do not. A sensitivity analysis of the parameters showed that an order of magnitude variation in the hydraulic conductivity corresponds to a 0.05–0.13% variation in groundwater levels, implying that changes in hydraulic conductivity have little effect on groundwater level. Additionally, a half order of magnitude increase or decrease in the dispersion parameters leads to a 0.6–4.26% variation in concentration. Thus, the response of the concentration to changes in dispersion parameters is not significant.

The optimal reservoir operation in dry seasons is an important topic in water resources management due to conflict of interest. This paper tends to address this issue by providing possible reservoir operation for individual months, together with multiobjective optimization and other constraints. A multiobjectives genetic algorithm optimization model is presented to determine the optimum releases of the Bigge reservoir, Germany, by assuming two inflow scenarios for dry seasons. The first one represents the minimum recorded monthly inflow during the period between 1995 and 1996; whereas the second scenario was the minimum monthly inflow during a five consecutive years generated using Monte Carlo model. The objectives of this study are to maximize energy production, the benefits of recreation, as well as the benefits of the energy produced, and to minimize the total penalty due to deviation from the targets. Several trade-off Pareto optimal solutions were obtained. A compromise solution is presented from a set of Pareto optimal solutions to help the decision maker. Details of the model formulations and implementation are described. The results demonstrate the efficiency of the developed model to determine the optimum releases effectively in the two inflow scenarios of dry seasons achieving all constrains.

Data of 14 standard duration annual maximum rainfall series with durations of 5 min to 24 h and lengths between 30 and 73 years up to 2010 that were recorded in Turkey are used. Mann-Kendall and linear regression trend, von Neumann independence, Wald-Wolfowitz stationarity, and Mann-Whitney homogeneity tests are applied on 155 complete series with a 24-h duration and 23 complete series with 14 standard durations. Next, only the linear regression test is applied on 174 incomplete series of 14 standard durations. The results of these tests indicate that almost 90% of all annual maximum rainfall series are trend free, independent, stationary, and homogeneous at a critical probability of 5%. Finally, the newly proposed Sen’s $1:1$ trend line method is applied to some series and its results are observed to be in agreement with both the Mann-Kendall and linear regression tests. It is concluded that standard duration annual maximum rainfall series in Turkey can be generally treated as independent and identically distributed random variables, which allow conventional intensity-duration-frequency calculations to be performed by statistical frequency analysis.

The peaks over threshold (POT) is a widely used technique to describe the exceedances of hydrological data above a threshold. It is well known that, under some conditions, the exceedances distribution can be approximated by a generalized Pareto distribution (GPD). The lack of a generally accepted methodology for selecting the optimal threshold is a major issue of the POT technique. In this paper an integrated approach is proposed that combines some graphical approaches with some analytical approaches to identify the optimal threshold and estimate the shape parameter of the exceedances distribution. Such a combination intends to reduce the subjectivity in graphical methods, and to refine their finding by using rigorous mathematical tools of analytical methods. First, a statistical test is used to select the appropriate GPD fitting the exceedances. Then, three numerical approaches, namely the likelihood ratio test, square error method, and multiple threshold method, are applied to detect the optimal threshold above which exceedances can be approximated by a GPD. These techniques are illustrated in a case study of Loukkos basin, a water resource of great importance in Morocco.

A scenario of particular importance in water resources management occurs when reservoir release decisions must be made well in advance of accurate hydrologic forecasts because of long travel times between reservoir releases and demand. This type of situation is evaluated using the Washington metropolitan area (WMA) water supply as a case study. Several classes of operating rules are evaluated using a state-of-the-art multiobjective evolutionary algorithm linked to a hydrologic simulation/decision model. Operating rules were evaluated using historical Potomac River streamflows (1929–2007) and synthetically generated time series. The proposed optimization framework is effective for a wide range of water resources vulnerability studies and was successful in improving the efficiency of the WMA system with respect to competing objectives ranging from reservoir storage to recreation and environmental flow requirements.

This study was conducted to investigate the recharge of surface water to groundwater in the Chinese loess area. A monitoring site was set up in Zhengning County, Gansu Province, China. The results show that the shallow layer (within 2 m) is greatly affected by rainfall, evaporation, and the moisture content changes in the annual cycle; the trend is consistent with the change of evaporation. When the precipitation exceeds a certain threshold, the moisture content changes at a deep position. The increment of the soil moisture content rises with the increasing precipitation, decreases with the depth, and lags with time. The results appear to confirm that water conducts in the unsaturated soil below the soaked zone through unsaturated migration. The motion of the water vapor is very small but can not be ignored. The water would accumulate when encountering the paleosol bed of low permeability; as a result, loess landslides might eventually be induced.

Controlling runoff from impervious surfaces such as roadways is an important part of storm water management in both urban and rural societies. Infiltration basins are one type of structure used to control storm water runoff. Design of these structures often involves using a single design storm and the assumption that the basin is initially dry. However, in tropical and subtropical climates where rainfall occurs frequently, the assumption of a dry basin is invalid. The objective of this study was to develop a new infiltration basin model that includes the effects of closely spaced storms on basin size. The model was used to design an infiltration basin for a section of unpaved road in João Pinheiro, Minas Gerais, Brazil. The performance of the model was evaluated with respect to infiltration rate, storm return period, and basin dimension parameters. Results showed that the infiltration rate was the most important input variable affecting basin size and that small but closely spaced storm events more strongly influenced basin size than isolated large events. Thus, while infiltration basins are often sized to capture small storm events (e.g., 25 mm or 90th percentile) to capture the most pollutants, this model indicates that such a sizing guideline may be inadequate if the basin is not empty. For unpaved roads, it is recommended that basins have an infiltration rate $5–10\mathrm{mm}{\mathrm{h}}^{-1}$ to balance hydrologic performance and size.