Journal of Hydrologic Engineering

Homogeneous streamflow regions in Turkey were identified using a multivariate statistical tool; namely, principal components analysis (PCA) with a varimax orthogonal rotation. The rotated PCA was applied to monthly streamflow records of 78 gauging stations during the period 1964–1994. The first seven rotated components were found significant at the 0.05 or lower levels. All accounted for 68% of the total variance of the original data assembly. The different groups of highly intercorrelated streams were defined by the patterns of map drawn for the rotated components loadings. Furthermore, an annual cycle analysis was applied to the same monthly streamflow data. Harmonic vectors extracted from a series of 12 monthly means at each station were marked over a map to readily keep track of streamflow seasonal variations across the study domain. The basic statistical characteristics of streamflow were also investigated within the identified regions; thus, some similarities and dissimilarities between the regions were documented. The identified streamflow regions were shown to be consistent with climate zones in Turkey defined by a recent study based on cluster analysis of temperature and precipitation data.

During rainfall events, the runon process allows water from upstream saturated areas to infiltrate into downstream regions where moisture deficit has not been satisfied. To date, several studies have clarified the role of the runon process in describing field-scale infiltration and surface runoff behavior. This study focuses on the role of the runon process in estimating field-scale sediment discharge over hillslopes with spatially-varying infiltration properties. This was accomplished by performing Monte Carlo realizations of saturated conductivity fields with varying properties to reflect different soil conditions. Numerical solutions were sought for flow and sediment transport processes over each realization, and field-scale ensemble averages and variances of sediment discharge hydrographs were obtained both with and without representation of runon in the models. Examination of results indicates that neglecting runon can cause serious discrepancies in estimation of field-scale sediment discharges. Using model differences, the cases when runon is important were identified, and the role of rainfall intensity, duration, slope, and soil type are discussed in this paper. It is observed that only when the rainfall intensity is much larger than the mean conductivity and when variance of the conductivity is very low that runon can be ignored for sediment discharge computations.

In this study we compared applications of mixed and continuous distribution functions to the theory of entropy for the evaluation of rain gauge networks. The use of a mixed distribution function to evaluate rain gauge networks has an important advantage of considering rainfall intermittency in both time and space. Parameters of both mixed and continuous distribution functions were estimated using the daily rainfall data collected in the Choongju Dam Basin, Korea. The optimal number of rain gauge stations is, then, decided by using the entropy theory. The optimal number of rain gauge stations estimated by applying the mixed distribution function was found to be much smaller, but still reasonable, than that estimated by applying the continuous distribution function; mostly due to the small wet probability and the high coincidence of daily rainfall between rain gauge stations.

The specification of the land cover characteristics is essential to have runoff simulated by a rainfall-runoff (RR) model in watershed hydrology. The recent availability of land cover map from remotely sensed data makes the specification of the land cover characteristics easier but it is unlikely that the RR models incorporate all spatial scales of the land cover map in need of application. This paper proposes a simple but physically based aggregation method to upscale natural vegetative heterogeneity, which is represented by Manning’s roughness coefficient, to the grid scale used in typical RR model, and it compares the strength of the new method with the conventional methods and uses a hydrologic model to test the sensitivity of using different methods. The results show the new method provides better performance in terms of both the amount and time to peak flow in relative terms. The proposed method provides a reasonably realistic description of area-averaged vegetation nature and characteristics.

Annual and monthly average temperatures are analyzed with statistical methods to characterize the temperature regime in Italy. Data from 738 weather stations, with homogeneous spatial cover throughout Italy, are used to estimate the annual and monthly air temperature normals. Geographic and morphologic parameters, computed around the points of measure, are considered as explicative variables within a georegression model. Morphometric variables are defined using the USGS GTOPO30 digital elevation model, which has a resolution of $1\mathrm{km}$ . On the basis of a stepwise regression analysis, the significant variables are found to be elevation, latitude, distance from the sea, and a measure of terrain concavity. The relationship between the average annual air temperature and the mentioned variables explains 92% of the variance and produces a standard error of $0.89°\mathrm{C}$ . The temperature regime (normalized mean temperature for each of the $12\text{months}$ ) is reproduced with a two-harmonic Fourier series, with parameters estimated using stepwise regression. Analyses of the reconstruction errors demonstrate that the results are quite satisfactory for many technical purposes, particularly for large-scale climatic characterization.

Field studies show that the variance of travel distance often increases nonlinearly with time elapsed after release of solute tracers. The nonlinear relationship between variance of travel distance and time is attributed to the heterogeneity of the porous media. To describe the transport in such a heterogeneous system, a time-dependent dispersivity is necessary. Though more attention has been devoted toward the study of non-Fickian dispersion at early time, there are no known studies that explicitly describe the dispersivity behavior in a fracture–matrix-coupled system. The observation from numerical results suggests that dispersivity has a time-dependent behavior and it reaches asymptotic values after a long time. The preasymptotic behavior of a solute front in fracture is characterized by increasing effective dispersivity with time. The role of fracture and matrix transport parameters on this behavior is analyzed for linearly sorbing solutes. Approximate expression is provided for the time-dependent dispersivity of the solute front in a single fracture with matrix diffusion and the expression for the time required to attain the asymptotic behavior is also obtained. A comparison of the front dispersivity behavior between parallel multiple fractures with a constant aperture width model and smooth parallel multiple fractures with a varying aperture width model is done.

The Soil and Water Assessment Tool (SWAT) was implemented in Northwest Arkansas to investigate flow and sediment predictive ability at multiple subbasin and hydrologic response units (HRU) distributions. The objectives of this study were to use SWAT and identify differences in annual predicted flow and sediment response of a watershed considering two subbasin delineations with six different HRU distributions each; quantify the uncertainty in SWAT output when sensitive model parameters are considered to have stochastic distribution; and evaluate the ability of the model to describe flow and sediment predictions of an ungauged watershed by stochastically validating the model. Flow results from the Single-Factor Between-Subjects Analysis of Variance test $(\alpha =0.05)$ indicated that predicted flow means were not significantly different from each other for all SWAT subbasin/HRU combinations; however, predicted flow means and measured flow mean were significantly different. Sediment simulation results suggested significant differences were present amongst the different model subbasin/HRU delineations and measured values. The Monte Carlo simulation of the model, using curve number (CN), soil evaporation compensation factor (ESCO), groundwater revap coefficient (GW_REVAP), and peak rate adjustment factor for sediment routing in the subbasin (AMP) as uncertain parameters, indicated that generally ESCO induced most uncertainty in predicted flow. However, sediment prediction uncertainty was affected most by uncertainty in AMP. Results indicated that SWAT applications on ungauged watersheds should include small subbasin sizing ( $\sim 2\%$ of watershed area), HRUs that reflect actual land cover composition, a check to evaluate surface runoff and ground water contributions and modification of parameters as needed to reflect site conditions, sensitivity analysis, and uncertainty analysis that includes several sensitive parameters.

Two of the most widely used approaches for flood estimation are the rational method (RM) and the Soil Conservation Service (SCS) method. Although the rainfall to runoff conversion process takes place during certain time durations, these models do not take time variability directly into their structure. Attempts to model the physical hydrology of a drainage basin are not considered; instead their preference and strength lie in their potential for easy use in converting the design rainfall into runoff discharge estimation. Some researchers have already noted that the parameters used in these models such as the runoff coefficient, $C$ , are not constant but have temporal and spatial variations. A digression is adopted in this note where $C$ is defined as the ratio of runoff to rainfall for infinitesimally small time duration. This is equivalent to the instantaneous runoff coefficient concept where $C$ varies with time. A set of plausible variation conditions are preset and then the constancy of the $C$ is assumed only for the infinitesimally small time intervals. An exponential model is suggested which subsumes all the properties of RM and SCS method. The mathematical structure of the suggested model reduces to the rational and SCS methods under a set of convenient assumptions. The possible range of this parameter is presented and simple application procedures are explained.

The relationship between the depression area $\left(A\right)$ , volume $\left(V\right)$ , and depth of water $\left(h\right)$ for wetlands, ponds, and lakes is often required for model studies. Area-depth $\left(A\text{-}h\right)$ and volume-depth $\left(V\text{-}h\right)$ relationships for wetlands, ponds, and lakes, which are important inputs to hydrologic models can be developed from detailed surveys and/or bathymetric maps that are costly and labor intensive to produce, or the use of simple geometric models. An analytical approach is introduced to describe wetland stage-storage $\left(V\text{-}h\right)$ relationships using a single shape parameter, simple geometric data available from geographic information system coverages or field studies, and a power-function model. The dimensionless shape parameter defines the wetland prismatic conic shape, and together with the reference wetland pool area and associated maximum pool depth defines the $V\text{-}h$ relationship to a high degree of accuracy. Shape parameters were developed for 42 wetlands to generate $V\text{-}h$ relationships for each wetland. The study findings were: (1) there is little variability in the dimensionless shape parameter; (2) the mathematical model accurately represents the $V\text{-}h$ relations for all wetlands analyzed; and (3) only the upper 80% of the wetland storage is needed to develop accurate stage-storage predictions using this approach. The developed wetland storage equations should prove useful in field studies, theoretical investigations, and surface/groundwater modeling.

It is pointed out that the hydrological models introduced in recent papers by Swamee are contained by at least two families of distributions known since the 1940s. Some elementary statistical properties, history, and related issues of these families are discussed.