Journal of Hydrologic Engineering

Statistical methods of wavelet, independent component analysis (ICA), and empirical orthogonal function (EOF) analysis were used together with geographical information systems (GIS) to regionalize runoff variability, establish baseline predisturbance hydrologic regimes, and account for runoff heterogeneity across Alberta, Canada as part of an effort to develop future adaptive forest management practices for Alberta. Both ICA and EOF identified three hydrologic clusters from 59 stations of catchment runoff data. However, ICA identified hydrologic clusters that agree better with the five ecoregions of Alberta than that of EOF. These are the Rocky Mountains and foothills, where runoff was characterized by a fairly consistent temporal variability and dominated by a strong annual cycle, southern Alberta/central Alberta, where temporal heterogeneity and a weak annual cycle dominated the runoff variability, and in southwestern Alberta, where the runoff variability was characterized by annual, 4–7, and 11-year cycles. Apparently stagnation moraine dominate hydrologic responses of most catchments in the grasslands and part of the boreal forests of western, central, and eastern Alberta. Empirical equations developed showed that stagnation moraine explained 64% of the runoff variability of selected catchments and that runoff was significantly diminished when stagnation moraines covered at least 74% of the catchment area. The identification of three spatially heterogeneous hydrologic clusters in Alberta, a province that is defined by four spatially homogeneous precipitation regimes indicates the necessity to develop forest management practices that will be suitable to manage Alberta’s forests.

To better handle landscape heterogeneities in distributed hydrological modeling, an earlier work proposed a discretization based on nested levels, which leads to fully unstructured modeling meshes. Upon such a discretization, traditional numerical solutions must be adapted, especially to describe lateral flow between the unstructured mesh elements. In this paper, we illustrated the feasibility of the numeric solution of the diffusion equation, representing groundwater flow, using unstructured meshes. Thus, a two-dimensional (2D) groundwater model (BOUSS2D), adapted to convex unstructured and irregular meshes was developed. It is based on the approximation of the 2D Boussinesq equation using numeric techniques suitable for nonorthogonal grids. The handling of vertical and horizontal aquifer heterogeneities is also addressed. The fluxes through the interfaces among joined mesh elements are estimated by the finite volume method and the gradient approximation method. Comparisons between the BOUSS2D predictions and analytical solutions or predictions from existing codes suggest the acceptable performance of the BOUSS2D model. These results therefore encourage the further development of hydrological models using unstructured meshes that are capable of better representing the landscape heterogeneities.

Typhoon-induced flood and inundation are among the major natural disasters in Taiwan. To understand and mitigate these hydrological hazards, adequate modeling tools on watershed scales are necessary. This paper describes the calibration and validation of a river basin using a physics-based watershed model, WASH123D. The key objective is to demonstrate the enable technology of using physics-based watershed models rather than lump parameter or statistical models. An ad hoc calibration procedure based on our understanding of the target river basin was employed. Systematic parameter optimizations would be addressed in subsequent studies. The Lanyang Creek basin situated in northeastern Taiwan is chosen as the study area. Two years of data (2004 and 2005) were used to examine event-based responses and long-term predictions. Both parametric work and hydrology issues are discussed. The simulated magnitude and time lag of the flow peak are compared with the observation. The results indicate that the proposed model could indeed provide Taiwan Typhoon and Flood Research Institute with a useful protocol for research, prediction, and forecast related to watershed hydrology.

Identifying the optimal drainage network based on digital elevation models (DEMs) is a fundamental task in rainfall-runoff modeling. Rapidly improving geographic information system technology enables hydrologists to use a variety of DEM-based hydrologic models that yield spatially concrete outputs. However, reliable drainage networks are still difficult to represent due to insufficient information about the dynamic behavior of water movement on catchment hillslopes. This study proposes an efficient method for drainage network identification through a comparative analysis of geomorphologic characteristics, such as drainage density, length of hillslope flow path, source area, etc., using area threshold and slope-area threshold criteria that incorporate scaling properties between the local slope and the contributing area. The results demonstrate that both criteria yield different drainage networks from “blue lines” based on topographic map from the Korean National Geographic Information Institute. Although the drainage networks obtained from the two criteria are visually similar, the area threshold yields an incorrect drainage structure due to excessive constraint of the draining source area size. In contrast, use of the slope-area threshold produces a relatively acceptable drainage structure in terms of preserving constant geomorphologic similarity to the study catchment. The proposed drainage network identification procedure may be used to describe landscape evolution for channel initiation in catchment hydrology. In addition, the comparative analysis of geomorphologic characteristics is found to provide important preprocess information for selecting a threshold value to generate reliable drainage networks based on DEMs before the application of hydrologic models.

The Texas water availability modeling (WAM) system consists of the water rights analysis package (WRAP) river/reservoir system simulation model, which is generalized for application any place in the world, and WRAP input data sets for all of the river basins of Texas. The WRAP/WAM system combines historical natural river basin hydrology with specified scenarios of water resources development, management, and use. Multiple-year sequences of monthly river flows for natural and alternative conditions of development are simulated. The paper compares flow-frequency statistics for nine major Texas rivers under natural and current conditions. The WRAP modeling system was recently expanded to provide flexible additional capabilities to develop stream flows representing various other user-specified conditions. An investigation of the Brazos River Basin demonstrates the new methodologies for adjusting river flows to reflect specified water management scenarios of interest.

Statistical downscaling methods describe a statistical relationship between large-scale atmospheric variables such as temperature, humidity, precipitation, etc., and local-scale meteorological variables like precipitation. This study examines the potential predictor variables selected from the National Center for Environmental Prediction and National Center for Atmospheric Research (NCEP/NCAR) reanalysis data set for downscaling monthly precipitation in Tahtali watershed in Turkey. An approach based on the assessment of all possible regression types was used to select the predictors among the NCEP reanalysis data set, and artificial neural network (ANN)–based downscaling models were designed separately for each station in the basin. The results of the study showed that precipitation, surface and sea level pressures, air temperatures at surface, 850-, 500-, and 200-hPa pressure levels, and geopotential heights at 850- and 200-hPa pressure levels are the most explanatory NCEP/NCAR parameters for the study area. It was concluded that ANN-based downscaling models can be implemented to downscale coarse-scale atmospheric parameters to monthly precipitation at station scale by using the above parameters as inputs in the study area.

Flood wave routing is a common problem in water resources engineering, for example, when a hydrograph enters a stream channel and passes downstream to an observation station. In the past, the problem has been approached by making the best possible estimate of the inflow hydrograph. The channel properties such as geometry and roughness are also estimated, along with any lateral inflow. The best estimates are used with a flood wave routing model to predict the hydrograph at the downstream observation station. Making a prediction by this procedure is full of challenges. It is impossible to exactly know the precise form of the hydrograph that will enter the channel. It is also difficult to select the channel properties from the range of values that may be appropriate. Lateral inflow is notoriously difficult to quantify. Making predictions under these circumstances is full of uncertainties. One approach to analyzing uncertainties is to use Monte Carlo modeling to make quantitative estimates of the uncertainty in flood wave routing results. The best estimates used for boundary conditions and model parameters are replaced by probability estimates. The ensemble of results from the Monte Carlo framework can be analyzed to develop probabilistic estimates of the routed hydrograph at the outlet of the channel reach. While this is a powerful approach, it also requires extensive probability data for the boundary conditions and channel properties. Such input data are rare in the published literature and do not appear to exist at all for lateral-inflow-dominated streams. This study examines the flood wave routing problem in a probabilistic framework using the kinematic wave model. It develops a complete data set for a lateral-inflow-dominated stream that includes a probabilistic description of the inflow hydrograph and lateral inflow. It also includes probability density functions for the parameters used in the kinematic wave model. The resulting data set appears to be the first developed for a lateral-inflow-dominated stream, though data sets do exist for streams without significant lateral inflows. Beyond the development of a new data set, this study seeks to evaluate the relative contributions of uncertainty in boundary conditions and channel parameters to the total uncertainty in the routed flood wave. Results for the lateral-inflow-dominated case developed here are compared to a similar example where lateral inflow is not significant. The results found here suggest that in both cases it is the uncertainty in boundary conditions that is most significant and dominates the total uncertainty in the routed flood wave.

A case study was conducted in the Teesta subcatchment in Bangladesh for determining design flood flows and corresponding flood stages for different return periods using frequency analysis and MIKE 11 model. Different distribution functions of frequency analysis were tested for their goodness of fit. The observed discharge data at Kaunia on the river Teesta were used for estimation of design flood. The Pearson type-III distribution was found best fitted by the Kolmogorov-Smirnov, D-index, and L-moment diagram ratio tests, and accordingly 25-, 50-, and 100-year return period design floods were computed. The river network of Teesta River was extracted from SRTM 90-m digital elevation model. The river network of Teesta subcatchment was then simulated by MIKE 11 rainfall-runoff Nedbor-Afstromnings-Model (NAM) and HD model. The resultant time series of river stage was then compared with corresponding observed values. From the model, a stage-discharge relationship $\left(Q\text{-}h\right)$ curve and respective equation were developed for Kaunia station on the river Teesta. The developed equation determines the corresponding flood stage of estimated flood flow of 25-, 50-, and 100-year return periods. The resulting flows and stages will be useful to design hydraulic structures, prepare flood extent maps, assess vulnerability of flood damage for different return periods, and provide flood forecasting for early warnings of floods. The approach presented would be applicable to similar river basin systems where data are limited and scarce.

Using the entropy theory, this note derives the Singh-Yu infiltration equation which is based on two simple constraints: (1) the total probability and (2) the mean logarithmic effective infiltration. The entropy theory permits a probabilistic characterization of the potential infiltration rate and hence the probability density function underlying the Singh-Yu infiltration equation. It also permits a quantitative assessment of the uncertainty of the infiltration yielded by the equation.