Journal of Hydrologic Engineering

A statistical impulse-response model based on multiple-lag correlations was constructed to determine the delayed aquifer response of precipitation events due to the epikarst in a karstic aquifer. The model was used to identify the direction and travel time of slow-diffuse flows and bimodal flows including slow-diffuse and fast-concentrated flows in wet years. Fast-concentrated flows with 0–15-day delays between precipitation events and aquifer response indicated rapid drainage of recharged water from the epikarst. A computed 46–75-day delay in the aquifer response in slow-diffuse flows was the upper bound for the epikarst retention time. The model identified the activation of a new fast-concentrated flow path due to subsurface water storage beneath the precipitation-fed recharge area when the system had high antecedent moisture. For the wettest year following moderately low antecedent moisture conditions, the model identified regional flow pathways extending over a 50–180-km distance. These results suggest that impulse-response models can be used independently to gain insights into the location of preferential pathways and the role of epikarst retention time in aquifer recharge.

A residence time distribution (RTD) provides a complete model of longitudinal mixing effects that can be robustly derived from experimental solute transport data. Maximum entropy deconvolution has been shown to recover RTDs from preprocessed laboratory data. However, data preprocessing is time consuming and may introduce errors. Assuming data were recorded using sensors with a linear response, it should be possible to deconvolve raw data without preprocessing. This paper uses synthetically generated raw data to demonstrate that the quality of the deconvolved RTD remains satisfactory when preprocessing steps involving data cropping or calibration are skipped. Provided noise levels are relatively low, filtering steps may also be omitted. However, a rough subtraction of background concentration is recommended as a minimal preprocessing step. Deconvolved RTDs often include small-scale fluctuations that are inconsistent with a well-mixed fully turbulent system. These are believed to be associated with oversampling and/or unsuitable interpolation functions used in the maximum entropy deconvolution process. This paper describes a new interpolation function—linear interpolation with an automatic moving average (LAMA)—and demonstrates that, in combination with fewer sample points (e.g., 20), it enables smoother RTDs to be generated. The two improvements, to deconvolve raw data and to generate smoother RTDs, have been validated with experimental data. Raw solute transport traces collected from a river were deconvolved after background subtraction. The deconvolved RTDs compare favorably with those generated from the more traditional advection-dispersion equation (ADE) and aggregated dead zone (ADZ) models, but provide more detail of mixing processes. A laboratory manhole solute transport data set was deconvolved with and without preprocessing using 40 sample points and linear interpolation. The raw data were also deconvolved using 20 sample points and LAMA interpolation. The two sets of RTDs deconvolved from the raw data show the same mixing trends as those deconvolved from preprocessed data. However, those deconvolved with LAMA interpolation and 20 sample points are significantly smoother.

The authors present a new set of tools for solving the one-dimensional Saint-Venant equations (1D-SVEs) of flow transport throughout dendritic river networks. The numerical solver is integrated with a set of geoprocessing tools, which include automatic cross-section selection, river bathymetry extraction, and a selection of model parameters, that facilitate the implementation of the 1D-SVE simulation setup. In addition, geographic information systems (GIS)–based preprocessing tools are developed to provide a seamless coupling of the hydraulic model to a hydrological model, which provides estimates of surface and subsurface runoff from hill slopes and performs routing in river networks using simplified ordinary differential equations. The hill slope runoff and streamflow generated by CUENCAS are redistributed as lateral inflows to the channels modeled by the 1D-SVE hydraulic model. The coupling of the hydraulic and hydrologic (H-H) models enables the validation of the hydrological model at internal locations in the basin where stage measurements are made, instead of only at locations where streamflow is estimated. An application of the coupled H-H models is demonstrated in the Squaw Creek watershed, Iowa. Results show that the coupled H-H models serve to validate assumptions in the hydrological model related to the spatial and temporal production of runoff in the watershed and bolster confidence in the estimated discharges at ungauged locations.

The three-parameter Burr XII distribution has been seldom used in hydrological applications, although it is particularly appealing because its range covers positive values only, which is convenient when modeling streamflows or rainfall data. Moreover, it has two shape parameters, allowing it to be quite adaptable to different samples because it covers a wide range of skewness and kurtosis values. Parameter estimation methods currently available in the literature require the numerical solution of two joint nonlinear equations to estimate the shape parameters of the distribution. This work proposes a simplified, although accurate, method to analytically compute the two shape parameters starting from the dimensionless L-moments ratios representing the distribution’s variability (L-CV) and skewness (L-skewness). The obtained parameters can be directly used in practical applications or otherwise may be useful to properly initialize the algorithms to obtain a numerical solution for the shape parameters. A detailed analysis of the accuracy of the approximated solution is performed, showing that the errors in the estimation of the distribution quantiles are negligible compared with the sample variability typically affecting hydrological samples. An extensive data set of empirical flow duration curves from stations located in northwestern Italy is considered to demonstrate the suitability of the extended Burr XII distribution to represent flow duration curves in a wide range of situations.

A new drought index termed the “soil moisture drought index (SODI)” is developed to characterize droughts. The premise of the index is based on how much water is required to attain soil moisture at field capacity. SODI captures variations of precipitation, temperature, and soil moisture over time. Three widely used drought indices, including the standardized precipitation index (SPI), the standardized precipitation evapotranspiration index (SPEI), and the self-calibrated palmer drought index (sc-PDSI) are compared with SODI along with local hydrological variables such as streamflow, reservoir storage, and groundwater level for cross-validation. The result indicates that SODI reacts more evidently to relate changes in precipitation and temperature than SPI and SPEI by characterizing soil moisture over time. Results also show that SODI outperforms the existing drought indices in the sense that SODI can detect and quantify the extended severe droughts associated with climate variability and change. SODI will add momentum to build a case toward the use of soil moisture information for drought analysis in a changing environment.

Temporal rainfall disaggregation is an important tool to obtain high-resolution rainfall data, which are needed in many fields of hydrology and water resources management. The multiplicative random cascade model can be used for temporal rainfall disaggregation of daily time series. A resampling algorithm is introduced to implement spatial consistence in disaggregated time series. Spatial consistence is assumed to be represented by four bivariate and distance-dependent rainfall characteristics that complement each other. Relative diurnal cycles of the disaggregated time series are resampled with the aim to reproduce these spatial characteristics while preserving the structure generated by the cascade model. Also, to achieve a final resolution of 1 h the traditional cascade model has been modified. A modification called uniform splitting with a branching number of 3 in the first step is introduced. Results are compared with observations and an approach called diversion. In total 22 recording stations in Northern Germany with hourly resolution were used for the validation of the disaggregation procedure, starting with daily values. Investigation areas are two catchments considering different station densities. The results show that for the disaggregation, errors of time series characteristics between 3 and 12% occur. The nonexceedance curves of rainfall intensities are slightly overestimated. Extreme values are well represented. The uniform splitting method outperforms the diversion method. Spatial rainfall characteristics can be reproduced by the simulating annealing algorithm. However, with an increasing number of stations the reproduction performance declines for some rainfall characteristics. Nonexceedance curves of areal rainfall based on disaggregated and not resampled time series are generally underestimated. By application of the resampling algorithm, a better performance regarding the spatial characteristics can be achieved. The presented resampling algorithm also has the potential to be used for implementing spatial consistence for time series generated by other disaggregation models.

In this paper, a generalized hierarchical multiscale approach for modeling coupled groundwater and surface water systems is demonstrated. Groundwater–lake interactions are simulated by coupling the groundwater equations with the lake’s continuity equation and by providing a two-way iterative feedback between models at multiple scales using specified head/flux boundary conditions. A hierarchical parameter estimation method that allows data and parameters at different scales to communicate between each other is also developed. These methods are applied to simulate a lake augmentation system for the Sister Lakes in southwest Michigan, which involves pumping a large amount of water from an irrigation well into the lakes. This problem requires resolution of time scales ranging from site-scale (hours) to local-scale (months) to watershed-scale (years) and spatial scales ranging from a few meters to a few kilometers. A hierarchical modeling framework consisting of five interlinked models was created, and model calibration was performed using drawdown data from a 72-h pumping test. The calibrated model was then used to simulate the entire lake augmentation system. The results indicate that the proposed modeling and parameter estimation approach can help improve the ability to model real-world complexities.

Regionalization for prediction in ungauged basins at hourly resolution is important for water resources management (e.g., floods and hydropeaking). In the research reported in this paper, calibration of 26 catchments ($39–\mathrm{3,090}{\mathrm{km}}^2$) in mid-Norway was performed using hourly records and three spatially distributed ($1\times 1{\mathrm{km}}^2$) precipitation–runoff models, as follows: (1) first-order nonlinear system model, (2) Hydrologiska Byråns Vattenballansavdelning (HBV) model, and (3) basic grid model. Four regionalization methods for each model [(1) parameter set yielding maximum regional weighted average performance measures (PMs), (2) regional median of optimal parameters, (3) nearest neighbor (NN), and (4) physical similarity (PS)] were evaluated and compared with three benchmarks. Parameter transfer from best regional donor and from an ideal best arbitrary single-donor, and local calibration (LC) were as benchmarks. The PS attributes include hypsometric curves, land use, drainage density, catchment area, terrain slope, bedrock geology, soil type, and combination of all. Comprehensive evaluation of single-donors and multidonors, simple benchmarks, and more advanced regionalization methods using multimodels, two PMs, and their statistical evaluation indicate that the identification of regionalization methods is dependent on the models, the PM, and their statistical evaluation. In general, the hypsometric curves, land use, and best regional donor methods performed better for the Nash–Sutcliffe efficiency based on boxplots and regional median values of both the Nash–Sutcliffe efficiency and relative deterioration or improvement of the Nash–Sutcliffe efficiency from the LC due to the regionalization. The methods also performed better for the individual catchments. The terrain slope, regional median of optimal parameters, maximum regional weighted average, and best regional donor methods performed better for the natural-logarithm-transformed streamflow (i.e., regarding the Nash–Sutcliffe efficiency) based on the same evaluation criteria. Similar performance to the more advanced regionalization methods of transfer of homogeneous parameter sets across the whole region from the best regional donor for both Nash–Sutcliffe efficiency and natural-logarithm-transformed streamflow (i.e., regarding the Nash–Sutcliffe efficiency) indicate the potential of the simple regionalization approach for predicting runoff response in the region.

This study focuses on the anisotropic hydraulic conductivity of a sandstone–mudstone particle mixture. Permeameter tests were performed in the laboratory. The effects of the mixture’s dry bulk density and mudstone particle content on its anisotropic hydraulic conductivity were investigated accordingly. Both the horizontal and vertical hydraulic conductivities were compared based on the calculation results from empirical grain-size formulas and tests. The best predictive formula for the hydraulic conductivities was found to be the Terzaghi empirical formula. It was found that the average horizontal hydraulic conductivity is about 2.0 times more than that of the vertical one. The anisotropy ratio initially increases and then decreases with the increase of the mudstone particle content by weight. A maximum ratio of 2.946 was obtained when the mudstone particle content reached about 60% by weight.

This paper presents a cubic-spline multiscale finite element method (CMSFEM) for solving groundwater flow problems in porous media. The main idea of this method is using the multiscale finite element method (MSFEM) to efficiently solve the hydraulic heads and nodal Darcian velocities. CMSFEM employs the cubic-spline technique to obtain continuous nodal head derivatives, which ensures the continuity of velocities. Furthermore, CMSFEM can not only solve velocities at coarse-scale grid nodes but also solve those at the fine-scale nodes in each coarse element grid. Instead of solving the full study region problem, the computation of the fine-scale velocities is decoupled from coarse element to coarse element, which can be implemented in parallel. Therefore, CMSFEM saves much computational costs in solving heads and velocities, which is important for high computation problems. The applications in this paper demonstrate that the CMSFEM has high accuracy and efficiency in solving velocities and heads.

Large-scale networks of sewer pipes, control gates, and pump stations are the primary facilities of urban drainage systems for mitigating flood risks. During flooding periods, cooperation of pumps and control gates plays a major role in efficiently reducing flood inundation. In this study, a real-time optimization approach is developed to find optimal policies for collaborative operation of drainage facilities, including detention storage and control gates, and multiple pump stations. The online approach is a predictive control scheme based on a harmony search algorithm that provides actively optimal operational policies according to real-time rainfall information. Operational models are applied to a prototype network in a portion of the Seoul urban drainage system, and results are compared with those of traditional operating rules. A high efficiency for the operational model of both the river diversion system and pump stations in terms of flood mitigation is demonstrated. For the river diversion system, considering several rainfall events, active control of the gates provides in average 48% peak discharge reduction in target points against 26% reduction obtained by a passive one. The active control can provide mostly 47% improvement compared with the passive policy just by delaying and controlling the release of storm water into detention storages. In parallel, the optimal online pump station operation decreases the collected storm water peaks by an average of 33 and 21% in two studied pump stations compared with the conventional approach even when a smaller number of pump switches (15% in average) is used, an advantage related to maintenance costs. More specifically, for the six separate rainfall events considered, the optimal policies mostly eliminate the storm water depth violations in pump stations, whereas there is a considerable violation for all events when the traditional operating rules are used. These results show that optimizing pumping operations is a practical and highly effective way to reduce flood volumes and urban inundation without making changes to the existing system infrastructure.

Soil aggregates as the basic units of soil structure greatly affect soil properties and fertilizer. They have an important theoretical and practical significance in studying the change in solute transport with different aggregate-size soil on the Loess Plateau. Based on a soil column experiment in the laboratory, breakthrough curves (BTCs) of chloride displaced through columns of different aggregate sizes of a clay soil were measured under water-saturated steady-flow conditions. The data were fitted with a convection and dispersion equation (CDE), a mobile-immobile model, and a two-flow-region model. In the CDE, all soil water was assumed to be mobile with a physical equilibrium existing in the system. For the mobile-immobile model, soil water was partitioned into mobile and immobile regions and convective diffusive solute transport was limited to the mobile water region; however, chloride transfer between the two regions was assumed to occur at a rate proportional to the difference in tracer concentration between them. In the two-flow-region model, soil water was divided into two regions based on flow velocities, but neither region had a nonzero flow rate. All three models could fit the solute transport curves in columns packed with all aggregate sizes at a small pore water velocity ($0.68\mathrm{cm}/\mathrm{h}$), but the values of the fitted parameters varied greatly. The Péclet number derived from the mobile-immobile model and the two-flow-region model showed the same changing tendency with aggregate size, increasing as aggregate size increased. The Péclet values derived from the CDE were approximately two orders of magnitude smaller than those obtained by the mobile-immobile model and the two-flow-region model. The mobile water fraction obtained with the two-flow-region model decreased as aggregate sizes increased whereas the mass transfer coefficient $\omega$ decreased as pore water velocity increased because of the short resident time of chloride transport through the soil columns.

The spatiotemporal variations of the drought scenario over the Bundelkhand region in central India have been studied using the standardized precipitation index (SPI) for the time scales of 3, 6, and 12 months. The space-time variability of the drought characteristics has been analyzed using the SPI, which is based only on long-term monthly rainfall data. From this analysis, it is determined that the frequency and severity of droughts have increased over the last decade in the Bundelkhand region and varied between once in 3 years and once in 5 years. The nonparametric Mann-Kendall test for trend analysis of climatic variables has been applied to investigate the regular occurrences of drought during the last few decades. The sign test indicates that there is a falling trend in the seasonal rainfall and number of rainy days in a few blocks. However, statistically significant falling trends at a 95% confidence level have been observed with the 3-month SPI for many blocks in the region, indicative of the increasing severity of drought during the principal rainy months (July, August, and September). The spatiotemporal analysis, which included progression and withdrawal of drought over the region, indicates that every drought event has its own unique characteristics, including its onset, severity, frequency, duration, area of influence and the magnitude of the losses caused by it. The severity of recent drought events shows a more widespread aerial extent in the region. It is observed that the area under the influence of drought in the Bundelkhand during the year 2007 varied between 79.5% (July) and 100% (August). The comprehensive analysis does not indicate any climate change signal and study recommends that effective use of water resources may be the key to address the drought in Bundelkhand region.

Digital elevation models (DEMs), because of their simple structure and high performance, play an important role in rainfall-runoff modeling. In this paper, six different methods were used to create DEM. The TOPMODEL, which is a semidistributed and topographic-based rainfall-runoff model, was employed to investigate the effects of the DEM creation methods on the results of such models. The results showed that the topographic index is highly sensitive to the interpolation method used in DEM creation. The consequences of this sensitivity were illustrated by the extent and distribution of the saturated zone across the study watershed. While the effect of the DEM creation method on the amount of surface runoff is significant, the effects on the mean daily discharge could be ignored. The shape and peak flow of simulated hydrograph showed a meaningful sensitivity to the type of DEM. The maximum difference of the Nash-Sutcliffe values for hydrograph simulation among all interpolation methods were obtained as approximately 2.5 and 23.3% for daily and hourly time steps, respectively.

Hydrological changes of the Ili River were analyzed concerning the changes of annual runoff and its distribution features within a year measured by coefficient of variation and concentration degree. Abrupt changes were detected by the heuristic segmentation method. Possible causes of the hydrological changes were investigated considering climate changes and human activities (especially the reservoir operation and irrigation extension). The Mann-Kendall method was applied to estimate whether the temperature and precipitation were changed. The increased temperature and the decreased precipitation in the flood season may decrease the runoff. At the middle reaches, the impact of the reservoir is significant and may be the main factor leading to the abrupt decreases in annual runoff and its intra-annual variability and concentration. The increased water surface area of the reservoir aggravates the evaporation and leads to annual runoff reduction. The reservoir regulates runoff by storing water in the flood season and releasing water in the dry season.

Two-peak hydrographs after a single rain event are observed in watersheds and storms with distinct volumes contributing as fast and slow runoff. The authors developed a hydrograph model able to quantify these separate runoff volumes to help in estimation of runoff processes and residence times used by watershed managers. The model uses parallel application of two advection-diffusion equations and calibrates the model’s fast and slow time parameters as well as a coefficient representing the relative size of the smaller hydrograph peak. The model provides an accurate representation of hydrograph timing, volume, peak, points of inflection, and recession rate, and its parameters represent physical processes of advection and diffusion and relate to watershed scale. The authors calibrated the model to match observed two-peak hydrographs with high efficiency on a watershed with distinct urban and rural land cover, and another watershed with distinct fast runoff from saturated areas. The Nash–Sutcliffe efficiency (NSE) of the simulated discharge was 0.93 for the urban watershed and 0.92 for the rural watershed. For the urban watershed, the simulated slow runoff volume was 89.6% of total runoff, and the fast runoff volume was 10.4% of total runoff; and for the rural watershed, the simulated slow runoff volume was 93.1% of total runoff, and the fast runoff volume was 6.9% of total runoff. This parsimonious two-peak hydrograph model can help researchers investigate how different storms and land cover types partition fast and slow flow and impact rainfall-runoff dynamics.