Journal of Hydrologic Engineering

Remotely sensed actual evapotranspiration (RS-ET) has been used extensively for multiobjective calibration of rainfall-runoff (RR) models to improve runoff prediction in ungauged basins. This study aims to implement RS-ET during periods when actual evapotranspiration (ET) simulation in RR models contains larger uncertainty and RS-ET is of relatively higher accuracy. Two calibration schemes are proposed and compared over 208 basins in the United States using the Xinanjiang model: (1) calibration against both observed streamflow and RS-ET over the entire period of record and (2) calibration simultaneously against observed streamflow over the entire period of record and RS-ET only during rainless periods. The 208 basins are then regarded as ungauged basins in turn and modeled with parameters transferred from a donor basin (i.e., parameter regionalization) using the spatial proximity method. Results show that compared with conventional calibration solely against observed streamflow, (1) both multiobjective calibration methods degrade streamflow simulation but improve ET simulation in the calibration and validation periods (however, the simulation efficiencies of both variables are improved in regionalization); (2) greater improvements of regionalization results are achieved by using RS-ET only during rainless periods; and (3) the efficacy of using RS-ET is superior for originally poorly simulated basins. Therefore, RS-ET is recommended to be used for multiobjective calibration during rainless periods.

Flow interaction over the interface between main channel and floodplains affects the overall discharge capacity and discharge distribution in compound open channels. Many investigators have attempted to empirically estimate flow interaction in terms of an apparent shear stress acting on the imaginary interface between the main channel and floodplain. However, past models are neither generalized for asymmetric channels nor applied to a wide range of data sets including field data, even though the apparent shear stress for asymmetric channels is found to be higher in comparison to symmetric channels for the same depth of flow. In this paper, using gene expression programming and a back propagation neural network, a generalized dimensionless formula is proposed for predicting percentage shear force and apparent shear stress on the vertical interface between the main channel and floodplain for asymmetric compound channels. The variation of bed characteristics and their dependability on the formula has been tested against a wide range of experimental and river data reported in the previous studies. Statistical analysis shows that the formulas produced in the curve fitting through gene expression and a feedforward back propagation neural network are very satisfactory and better than past models. The exceptionally high accuracy of the proposed models implies that they can be extended to use for a wide range of applications.

A simple hydraulic model is proposed to illustrate the transition mechanism of transient to steady flows in the vadose zone. This model consists of a set of cylinders that are vertically aligned, and each of the cylinders is open at the top and has a narrow outlet at the bottom. An unsaturated soil unit in the vadose zone is considered one of the cylinders, and the entire soil column is modeled as cylinders arranged in series. The cylinder is representative of the soil skeleton of the corresponding soil unit, the water distributed in the cylinder is representative of the water in the soil, and the water flow rate out of the outlet is representative of the soil permeability. Physical tests and theoretical analyses revealed that the water flow in the cylinders can also be obviously divided into transient and steady flow, which resembles the water flow in unsaturated soils. Moreover, the transition process of water flow from the active zone to the steady zone in unsaturated soils can be well captured by the model. The influence of water influx rates, soil permeability, and initial water contents on unsaturated infiltration can also be investigated using this model. The results show that the steady water content of the unsaturated soil depends on the water influx rate and soil permeability and is independent of the initial water contents. As a result, water flow in the vadose zone, especially in the steady zone, which is difficult to detect and monitor in the field, can be better understood with the proposed model.

Forecasting precipitation remains challenging because of its large spatial and temporal variability, and the uncertainty in precipitation forecast leads to an important source of uncertainty in the prediction of other components of a hydrological system. In this study, in order to forecast subseasonal precipitation and better characterize the temporal variability of precipitation, a hybrid precipitation forecast model was developed based on (1) temporal clustering of subseasonal precipitation; and (2) coupling an improved seasonal autoregressive integrated moving average (ISARIMA) model to an artificial neural network (ANN) model to take the advantages of both models and capture precipitation persistence and statistics in each cluster. The performance of the proposed model was compared against different variations of conventional statistical models in the Rasht station with a humid climate and Gorgan station with a Mediterranean climate, both located south of the Caspian Sea in northern Iran. The model evaluation criteria indicated that the hybrid model can remarkably improve forecast accuracy. The root-mean square error score of the forecasted precipitation by the hybrid model against observations decreased 48% and 24% in the Rasht and Gorgan stations, respectively, when compared with the seasonal autoregressive integrated moving average (SARIMA) model and the index of agreement increased 32% and 17%, respectively, when compared with the ANN models. The proposed hybrid model can be a useful tool for forecasting subseasonal precipitation in humid and arid climates with persistent and nonpersistent precipitation patterns.

Hydrologic series in arid and semiarid areas often contain zero values. For such data, the unit impulse response function (irf) of a kinematic diffusion model and a Muskingum flood routing model have been employed as probability distribution functions for frequency analysis. However, the probability weighted moments (PWMs) of these distributions have not been derived for parameter estimation. In this study, the parameter estimation formulas of a two-parameter kinematic diffusion distribution (KD2), a two-parameter Muskingum-based distribution (M-like), and three Muskingum-based three-parameter distributions based on PWMs are derived using mathematical transformation and numerical calculation principles. The fitting effects of PWMs of these distributions are evaluated and compared with the method of moments (MOM) and maximum likelihood method (MLM) using ordinary least square (OLS) criterion, residual square sum criterion (RSS), and Akaike information criterion (AIC). Precipitation data collected from six gauging stations were used as a case study. Results show that for each distribution, PWM yields the smallest OLS, RSS, and AIC values among the three methods and improves the accuracy of estimation.

Determining the scales, types, and locations of low-impact development (LID) facilities is the basis for LID practice. Spatially available rainwater on site is first calculated in this study to obtain the scales of LID facilities using the geographic information system technique. Spatially available harvested rainwater in different land-use types is then calculated to determine types of LID facilities according to potential facilities’ performance. Next, a $p$-median model and an ant colony optimization (ACO) algorithm are used to obtain spatially optimal locations for LID facilities. Main runoff pollutants [e.g., suspended solids (SS), total phosphorus (TP), dichromate oxidizability (${\mathrm{COD}}_{\mathrm{cr}}$), and ammonia nitrogen (${\mathrm{NH}}_3\text{-}\mathrm{N}$)] are sampled, interpolated, normalized, and spatially overlaid to identify heavily polluted areas and adjust LID facilities for decontamination. Finally, the performance of the ACO for the $p$-median model is compared with that from the genetic algorithm. Determining the scales, types, and locations of LID practices can effectively guide construction of LID for practitioners and stakeholders.

Hydrologic basin models are generalizations of natural hydrologic processes and inevitably contain multiple uncertainties in the simulation of rainfall-runoff events. These uncertainties typically include input uncertainty, model structure uncertainty, and model parameter uncertainty. The input uncertainty is divided into rainfall calculation uncertainty (RCU) and precipitation forecasting uncertainty. For model parameter uncertainty, there are only a few parameters in a hydrologic basin model that make significant contributions to flood forecasting so that only the probability distribution functions of those sensitive parameters need to be evaluated. In this study, a new method of RCU assessment is derived using an inverse sampling gauge (ISG) approach, in which the influencing factors of the RCU are addressed in the calculated area precipitation and standard deviation. Additionally, the coefficient of variation-Nash-Sutcliffe efficiency measure (CV-NS) method is introduced to identify the sensitive parameters of a hydrologic model. The methods of ISG and CV-NS are tested in Huangnizhuang Basin, China, in the evaluation of the RCU and the parameters uncertainty of the Xinanjiang model. It is indicated that the ISG method is useful in RCU estimation by deriving the conditional probabilistic distribution of areal precipitation, and the CV-NS method is effective and simply in the operation of sensitive parameters’ identification. In addition, the continuous ranked probability skill score (CRPSS) is employed as an indicator to show the relative performance of predictions. Three probabilistic predictions considering different sources of uncertainties are conducted and compared with the deterministic prediction. The results of the comparisons suggest that predictions considering one or more uncertainties have higher predictive performance than the deterministic one. Moreover, main source of uncertainty can be identified by the results of CRPSS among different probabilistic predictions. In this study area, the model parameters’ uncertainty is the main uncertainty.

Forecasted changes to climate were used to model variations in the streamflow characteristics of a northern Virginia catchment. Two emission scenarios were applied from international climate projections along with four general circulation models (GCMs) by using two statistical downscaling methods to drive the hydrological simulations in two future time periods (2046–2065 and 2081–2100). Incorporation of these factors yielded 32 runoff simulation models for a $130\text{-}{\mathrm{km}}^2$ watershed located in northern Virginia. These models were compared with historical streamflow data from the late 20th century. Changes in streamflow were compared using median, low, and high flows. Results showed a general increase in median flows in both the mid- and late 21st century. Low flows were projected to decrease, whereas high flows were projected to increase, creating a larger range between low flows and high flows. In addition, statistical tests were conducted to identify the main factors that affected variations in future climate projections. The choice of the downscaling method emerged as the main source of uncertainty. This research quantifies the impacts of climate change as well as uncertainties within climate change projections for regional water resources. Considering the essential role of this watershed for water supply in northern Virginia, the findings of this study illustrate likely impacts of climate change on water supply reliability, supporting climate resiliency in the study area.

The process of river runoff confluence is highly complex and combines both tendency and chaos. Research on the temporal and spatial tendencies of river runoff variability can yield significant information to understand river systems. This paper aims to identify rainfall–runoff variability tendencies using a novel variability metric: the Mei-Wang Fluctuation (MWF) index, which estimates the variability as a compound measurement of quantitative variation (Qv) and contour variation (Cv) and has been successfully applied to quantify the variability of power output. Six rainfall rivers located in North Carolina and Texas with few reservoirs are chosen as the study case. The results show several novel tendencies of the runoff variability, including the following: (1) when the temporal resolution length decreases from 1 day to 1 month, the runoff variability indices, Cv and MWF, conspicuously decrease, while Qv declines slightly; (2) from a spatial perspective, when the drainage area rises, these indices increase, and notably, Qv increases in a linear pattern (sharply in a narrow river basin, slowly in a broad river basin); and (3) these spatial tendencies are further demonstrated by replacing the outflow of a reservoir with its inflow. Additionally, the linear tendency exhibited between Qv and drainage area calls for further exploration that could contribute to hydrological analysis.