Journal of Hydrologic Engineering

Time-series analysis in hydrology plays an important role in the efficient use of water resources, prediction of flood risks, and crop production. However, considering many parameters in the solution of hydrological problems complicates the time-series analysis. For this reason, the use of artificial intelligence methods in hydrology studies has become interesting. The support vector regression (SVR) model and random forest (RF) model are frequently used in hydrological time-series forecasting. The SVR model reduces the experimental measurement errors as well as giving the statistical properties of the data used. The RF model is often used in the solution of regression. However, in both models, realistic values can be reached with different parameters in predicting total monthly rainfall (TMR). The reason for this is that the parameters used in the SVR model and RF model in time-series analysis should be used in different values for each period. In this study, the SVR model is used to forecast monthly rainfall amounts. The artificial bee colony (ABC) algorithm is used to obtain optimum values of SVR model parameters. In the proposed approach, TMR values in the years between January 1938 and December 2016 of Elazig, Erzincan, Erzurum, and Van stations in the eastern part of Turkey are used. The performance of the proposed model is compared with search-based SVR and RF models and hybrid SVR-genetic algorithm and RF-genetic algorithm. For the four stations, the results showed that the hybrid model performed significantly better. The proposed SVR-ABC model provided more realistic values than the search-based SVR and RF models.

To reveal the temporal variability of hydrological elements in the Daqing River Basin (TDRB), an index system with temporal variability indices of hydrological elements, namely jump variability, trend variability, and period variability, is constructed. These indices are calculated with the help of variability diagnosis methods. Additionally, the entropy weight method and technique for order preference by similarity to an ideal solution (TOPSIS) evaluation model are combined to grade and evaluate the temporal variability of hydrological elements. The results show that (1) there is no temporal variability in rainfall, but runoff shows jump variability in 1979 with a sharp decrease. Meanwhile, groundwater depth presents jump variability in 1998 with a significant increase. (2) The period variabilities of runoff and groundwater depth are mainly exhibited in the medium-long periods. After the variability, the decomposed component periods become shortened in runoff while longer in groundwater depth. Furthermore, there is a significant change in the entropy of the medium-long periodic components, indicating higher variability in these components. (3) The importance of temporal variability indices of runoff and groundwater depth is that jump variability is the largest, followed by trend variability, and period variability is the least. The variability significance plays a paramount role in jump variability and trend variability, whereas in period variability, the entropy after variability should be focused. The comprehensive evaluation scores of the runoff and groundwater depth showing medium-strong temporal variability are 0.645 and 0.678, respectively. This study provides a new approach for studying the temporal variability of hydrometeorological elements in other meteorologically similar regions.

The present paper focuses on the numerical model development of the colloid-facilitated contaminant transport using an equilibrium approach in the vadose zone of subsurface porous media. Flow in the vadose zone of subsurface porous media plays a prominent role in predicting the transport of contaminants from the ground surface to the water table. Ignoring the fact that standard flow equations and colloids are ubiquitous in subsurface environments can lead to a severe misjudgment of the distances traveled by the contaminants. Although contaminant transport with colloids in saturated porous media has been studied, the equilibrium-based colloid-associated contaminant transport in the unsaturated zone has received inadequate attention. The present study encompasses the flow equations for a vadose zone with a colloid-facilitated contaminant transport model. The present study encompasses the first numerical model of colloid-associated contaminant transport with an equilibrium approach under unsaturated flow conditions. The mixed form of the Richards equation is solved using a fully implicit finite-difference method with Picard’s iteration and coupled with the solution of the transport equation. The breakthrough profiles and sensitivity analyses culminate in indicating that colloids enhance the mobility of contaminants by reducing the retardation factor. However, an engrossing finding is that the mobility of contaminants also relies upon the degree of interaction of pollutants with stationary porous matrix and suspended colloids. As the degree of interaction of the contaminants with the stationary solid matrix increases, retardation is noticed in the contaminant movement, even in the presence of colloids. In contrast, contaminants move faster as the degree of interaction of the contaminants with the suspended colloids increases.

Overland flow is often generated by moving rainstorms and is modeled using the kinematic wave theory. In overland flow modeling, it is usually assumed that rainstorms are stationary and occur over the entire watershed. Consequently, studies on overland flow modeling considering moving storms have been limited. Storms may move from upstream to downstream, downstream to upstream, or across stream. Likewise, storms can occur over the entire watershed or a portion thereof, which can be upstream, downstream, in the center, or at different portions. Studies that have considered moving rainstorms have assumed that storm velocity is the same as flow velocity. However, it is not uncommon that storms move at a velocity slower than flow velocity, and such storms have not been considered in the studies. For such storms, the structure of the solution domain and, in turn, of the kinematic wave solution becomes quite different and has not yet been reported in the hydrologic literature. The objective of this paper therefore is to derive an analytical solution of the kinematic wave equation under the condition that a rainstorm is moving at a velocity slower than flow velocity. Field or laboratory observations on storms moving at a velocity slower than flow velocity are not available. Therefore, validation of the derived solution is not the objective here, because without data, the analytical solution cannot be verified.

Water quality control bioretention cells are usually sized so that they are large enough to provide storage of runoff generated from its catchment resulting from the water quality control design storm of the location of interest. A bioretention cell can have three parts where runoff may be temporarily stored: (1) the depressed surface ponding area, (2) the growing media layer, and (3) the storage layer. In this study, the detailed sizing calculation methods adopted by 21 jurisdictions in the US and Canada were reviewed and compared. It was found that for satisfying the required storage volume to store runoff generated from the water quality control design storm, some jurisdictions only allow the storage provided by the surface ponding area to be counted, whereas other jurisdictions allow the sum of the storages provided by two of the three or all of the three parts to be counted. These differences can result in significant differences in the sizes and performances of bioretention cells. By examining and analyzing the detailed hydrologic processes involved in the operation of bioretention cells, it was demonstrated in this study that many of the adopted sizing calculation methods are flawed or inappropriate. Presented here are a summary of the findings and the recommended sizing calculation method. The possible degrees of over- or underdesigns resulting from some of the incorrect sizing calculation methods are also estimated.

The use of a natural and low-impact way of managing surface water and groundwater in a sunken plaza is meaningful for urban ecological development. To achieve water storage, purification, and recycling, this study investigated a water circulation system in a sunken plaza based on the groundwater and surface water transformation and combination. A joint water-stop curtain system was constructed with underground continuous walls, RC wall, building main structure, adjacent support structure, and waterproof stratum at the bottom of the main structure. On the basis of the stratum data and parameters of the study area, a three-dimensional (3D) groundwater model was built using FEFLOW 7.5 comparing the actual groundwater level monitoring data from the field pumping test with the simulation data, the accuracy of the model was verified, and the water volume change of the sunken plaza was calculated. The low permeability of the curtain system calculated by a field test and the numerical model proved that the pumping activities in the sunken plaza had extremely low impact on the groundwater level outside the plaza. On the basis of this result, the pumping and seepage wells and water quality and quantity observation wells were designed scientifically and rationally. The purification and circulation control of water were realized in combination with rainwater, drainage, landscape, paving, and other facilities. According to the circulation system designed in this study, the annual water recycling capacity of the sunken plaza can reach more than $\mathrm{60,000}{\mathrm{m}}^3$, and the total suspended solids (TSS), chemical oxygen demand (COD), total nitrogen (TN), and total phosphorus (TP) indicators in the rainwater can be reduced by more than 15% on average. The sunken plaza area forms a small spongy body that realizes water circulation, economic benefits, and ecological functions for the first time using groundwater storage. The developed method can be used for the water circulation and ecological construction of sunken plazas with similar geological environments.

Numerical weather prediction (NWP) can provide vital information for pan evaporation (${\mathrm{E}}_{\mathrm{p}}$) forecasts for the 16 days ahead, which is of great help to water resources management. However, the information for forecasting ${\mathrm{E}}_{\mathrm{p}}$ usually requires bias corrections. This study was based on three bias correction methods [the equidistant cumulative distribution function method (EDCDFm; M1), XGBoost (XGB) with a single meteorological factor input (M2), and XGB with multiple meteorological factor input (M3)] and the meteorological data from 18 weather stations in southern China, the bias correction of meteorological factors forecasted by the second-generation Global Ensemble Forecast System (GEFSv2) was carried out. The results indicated the bias correction ability of the M3 method for GEFSv2 outputs was better than that of the M1 and M2 methods. It was a model-data error between GEFSv2 outputs and the corresponding observation data. Solar radiation exhibited the lowest error, whereas minimum temperature exhibited the highest. However, the M3 method decreased the forecast model-data error. In addition, this study compared the ability of three tree-based models to forecast ${\mathrm{E}}_{\mathrm{p}}$, namely, M5Tree (M5T), random forest (RF), and XGB. The XGB model had the highest forecasting accuracy for ${\mathrm{E}}_{\mathrm{p}}$. When the NWP outputs corrected by M1, M2, and M3 methods were used as the input of the XGB model, the averages of mean absolute errors (MAEs) at the 18 stations during the 1–16 day period ranged at 0.99–1.69, 0.78–1.14, and $0.78–1.07\mathrm{mm}/\mathrm{day}$, respectively. ${\mathrm{E}}_{\mathrm{P}}$ forecast showed the most significant error in the summer. Further, the relative humidity contributed the most to the ${\mathrm{E}}_{\mathrm{p}}$ forecasting error. By addressing the issue of NWP outputs applied to ${\mathrm{E}}_{\mathrm{p}}$ forecast, this study improves understanding of the bias correction method of NWP outputs and tree-based models to forecast ${\mathrm{E}}_{\mathrm{p}}$. It also improves understanding of the seasonal performance of ${\mathrm{E}}_{\mathrm{p}}$ forecast and the impact of meteorological factors on forecast error that can inform future studies and models.

Abstract

This study attempts to build nonstationary indices for assessing meteorological drought in two different climate zones in India: the arid Saurashtra and Kutch and humid-tropical Coastal Karnataka. Time and climate indices are considered as covariates to develop nonstationary models using the generalized additive model in location, scale, and shape (GAMLSS) for the period, 1951–2004. A comparative study has been conducted to assess the statistical performance of stationary and nonstationary models on various time scales (3, 6, 12, and 24 months). The best model is selected to conduct copula-based bivariate drought analysis. For this purpose, drought properties such as drought severity, duration, and peak are calculated. The annual and seasonal rainfall departures are also analyzed, and more rainfall-deficient years are detected in Saurashtra and Kutch regions than in Coastal Karnataka. The nonstationary index performed better in capturing drought properties in statistical analysis over both the study areas at all time scales. The nonstationary drought index shows better consistency with historical drought and flood events than the stationary index. Cooccurrence and joint return periods are calculated and compared with univariate return periods. A significant difference is observed between bivariate and univariate return periods, and more risk is detected in Saurashtra and Kutch than in Coastal Karnataka. The impacts of rainfall and drought on the yield of major crops in study areas are also analyzed. The yield loss rate of bajra significantly correlates with the nonstationary standardized precipitation index (NSPI) in Saurashtra and Kutch, whereas rice yield has no significant correlation with the index in Coastal Karnataka. This new aspect of drought analysis provides feasible results in both arid and humid regions in a changing environment.

Practical Applications

El Niño-Southern Oscillation (ENSO) refers to changes in ocean temperatures and atmospheric pressure in the Pacific Ocean, while Indian Ocean Dipole (IOD) refers to differences in sea surface temperatures in the Indian Ocean. Both ENSO and IOD can affect rainfall and drought conditions in India, making them essential factors to consider in understanding and predicting drought events. The nonstationary drought index developed using climate indices could accurately assess the important aspects of drought, like how severe it was, how long it lasted, and when it reached its peak. By evaluating the return period of a specific drought severity and duration, decision makers can assess the likelihood of experiencing such an event in a given year or over a specific period. It helps to prioritize resources, plan adaptation measures, and design drought management strategies to mitigate potential impacts. If specific ENSO or IOD conditions are associated with prolonged drought periods, water resource managers can implement proactive measures like water conservation initiatives, irrigation scheduling, and crop diversification. This research examines the impact of rainfall and drought on major crop yields (bajra in arid regions and rice in humid-tropical regions). This knowledge can aid farmers and policymakers in predicting crop losses, optimizing irrigation strategies, and implementing timely interventions to minimize agricultural productivity losses.

A correct estimation of evapotranspiration (ET) is required to determine the amount of available water in any watershed. Potential evapotranspiration (PET) is used to estimate actual evapotranspiration (AET). The Soil and Water Assessment Tool (SWAT) has different methods to compute PET; in data-scarce watersheds hydrological modeling is typically challenging. Although there are a lot of remotely sensed PET (RS-PET) data, validation of hydrological models using these data has rarely been studied in such watersheds. Thus, the purpose of this paper is to assess the impacts of direct assimilation of RS-PET on water balance components using SWAT for a basin in Iran with limited data. To this end, we changed the SWAT source code to automatically integrate RS-PET data. The performance of the model was then evaluated using streamflow data and AET for the period 2001–2005. The results reveal that the Default PET model overestimates soil moisture, underestimates AET, and ultimately fails to appropriately capture streamflow at the watershed’s outlet. However, by incorporating RS-PET, the model improves the accuracy of PET, AET, soil moisture, and streamflow. For example, in the Default PET scenario, the Nash–Sutcliffe efficiency streamflow was 0.66, whereas it was 15% higher in the RS-PET scheme. The findings also demonstrate the SWAT parameters’ sensitivity to PET values and the efficiency of RS-PET in enhancing hydrologic modeling in regions with limited data.

Sediment trap efficiency ($TE$), defined as the sediment trapped in a reservoir divided by the sediment flowing into the reservoir, or the sediment delivery ratio ($\mathrm{SDR}=100\%\text{-}TE$), has been studied for small or ponded reservoirs, but for large reservoirs it is still not well understood even though it is important for predicting and managing reservoir sedimentation. In this case study, we analyzed the changes in water and sediment regime caused by the operation of two cascade dams (the Sanmenxia and Xiaolangdi Dams just upstream of the lower Yellow River) during 2000–2020. We found that more than 50% of the water discharge ($Q$) entering and flowing out of the reservoirs was smaller than $\mathrm{1,000}{\mathrm{m}}^3/\mathrm{s}$, while over 85% of the sediment was transported by larger discharges. Among different ranges of $Q$, the range $\mathrm{2,000}–\mathrm{3,000}{\mathrm{m}}^3/\mathrm{s}$ transported the most sediment (30%–50%). Due to the Water and Sediment Regulation Scheme of the dams, the maximum sediment concentration of floods tended to lag behind the maximum water discharge. The longer the lag time, the greater the water volume released to the downstream, resulting in a higher SDR of floods. A calculation method for the SDR at the Xiaolangdi Reservoir during floods and flood seasons was proposed by considering the water retention time in the reservoir. The sedimentation delta in the Xiaolangdi Reservoir changed rapidly after the dam closure and gradually became stable, and its geometric changes also impacted sediment transport through the reservoir. This study may shed light on sediment transport through large reservoirs under the operation of water and sediment regulation.

This study develops a four-module-based multimodel framework, coupled with the projections of general circulation models (GCMs), a hydrological model (SWAT), and two sets of evaluation indicators [Indicators of Hydrologic Alteration (IHA) and River Regime Index (RRI)], to systematically quantify the response of ecohydrological regimes to climatic change. The Jinsha River Basin (JRB) was selected as a target area for both framework verification and application analysis because of its precious natural conditions, privileged abundant hydropower, and extremely rich fish resources. Results revealed that: (1) temperature and precipitation in the JRB would show increasing trends to varying degrees in the future. The increase in temperature would be much higher than that in precipitation, and the increment would be more significant during the end of the century (2068–2097). (2) The future runoff in the JRB may face an overall decreasing trend, leading to more and more frequent drought disasters. (3) The disturbances in the ecohydrological regime would result in more concentrated runoff and smoother hydrological pulse fluctuations in the JRB, meaning human-assisted participation would be required for the conservation of reproduction and development of ecological resources like Coreius guichenoti. Our findings suggest that effective management of future watershed resources can only be accomplished if an in-depth and comprehensive hydrological evaluation of changing climate is made, and the protection and sustainable development of ecological resources of the JRB in the future will require the participation of stakeholders. Figuring out the alteration in ecohydrological regimes under natural variation and its potential hazards may guide the degree of human participation in futural river protection.

Abstract

Remote sensing is widely used for monitoring reservoir capacities. The relationship between two water levels (WLs) and the corresponding satellite-derived water spread areas provides the reservoir volume between the two WLs. However, the accuracy of this method depends on the ability to capture the water spread areas at fine increments of WLs so that the cross section of a reservoir is represented in adequate detail. In a monsoon climate, persistent cloud cover during the rainy season limits the number of usable optical satellite images. Hence, the time interval between two successive WLs for which cloud-free images are available is generally large. Using only a few WLs increases the likelihood of missing a significant break of slope in the reservoirs that may lead to gross error in the storage measurement. We showed that the combination of freely available optical (e.g., Landsat 8) and cloud-penetrating microwave [e.g., European Space Agency (ESA) Sentinel-1] images may improve storage estimation and mitigate the inherent uncertainty of using remote sensing data for monitoring reservoir sedimentation. Three reservoirs in the Upper Godavari Basin in India constituted the study area. Findings show that the average annual sedimentation rate was overestimated by $\sim 10\%$ when using only Landsat-8 images compared to combined radar and optical remote sensing databases (2016 and 2017). The suggested method made significant improvements in the estimation of reservoir capacities for narrow-bottomed and convex-edged reservoirs.

Practical Applications

The capacity of a reservoir is typically monitored on a periodic basis as a means of determining the reservoir’s overall health. This is a common application of satellite imagery. Satellite images are used to determine the volume of water stored in a reservoir by analyzing the relationship between two water levels and the area covered by the water. The greater the difference between the two water levels, the less precise this measurement will be. Under monsoon climate, cloud cover prevents the use of optical images during the rainy season when reservoir levels fluctuate greatly. Because there are fewer water levels to choose from due to the lack of images, the accuracy of the estimate suffers. Using three reservoirs in the monsoon region of peninsular India, this study shows that combining optical data from satellites like Landsat-8 with publicly available cloud-penetrating Sentinel-1 radar images greatly increases the number of water spread area–water level data points and decreases the error and uncertainty in the storage estimation. Based on our findings, it appears that the reported rate of capacity loss, derived from limited optical images, is likely to be inaccurate. The authors also found that the storage estimates of the narrow-bottomed conical-shaped reservoirs will be greatly improved by increasing the number of water levels in volume calculation.

Abstract

Hydropower dams can induce flash floods, leading to a severe cataclysm in flood-prone areas at downstream regions. On the catchment scale, flooding is not contributed solely by the reservoir releases, and there can be significant flow contributions from tributaries downstream of the dam. The major challenge in estimating the lateral flow contribution is that most tributaries are ungauged and situated in inaccessible areas. To overcome this inconsistency and to increase the precision of downstream flood warnings, a modeling framework was developed to quantify the flow contribution by ungauged tributaries to the mainstream using the drainage area ratio (DAR) method. The model parameters were estimated using optimization algorithms, and the best parameters were selected based on the error metrics. The modeling framework constitutes a reservoir operation model and hydrodynamic model developed in MATLAB version 2020b environment with the ease of coupling the two models. The estimated flow from the lateral tributaries based on the optimal model parameters of DAR and hourly inflow hydrographs were incorporated into the model. Two scenarios were analysed with and without lateral flow from ungauged tributaries. Results impart that the flood peaks have increased by more than 75% with the incorporation of the lateral flow. The model was validated with downstream stage and discharge data. The results indicated that the magnitude of the model generated and actual flow data were in the same range.

Practical Applications

Flooding downstream due to sudden release from a hydropower dam is a matter of serious concern worldwide. To evaluate the potential flooding situation downstream, a dam release is generally routed by a hydrodynamic model. However, because hydropower dams are mostly located in remote areas, the tributaries located at inaccessible downstream areas remain ungauged and, therefore, obtaining precipitation/streamflow data of such tributaries become difficult. In absence of downstream flow contribution, the water level obtained by routing the reservoir release underestimates flood magnitude. The dam release flood falls in the high-hazard category because of its suddenness characteristics and, therefore, adverse consequences of underestimation cannot be overemphasized. This paper presents a framework that couples a reservoir operation model, a hydrodynamic model, and a simplified area–proportionate model to estimate downstream tributary contribution, so that a more reliable estimation of the downstream flood situation can be made. The modeling framework has been tested in the Ranganadi Hydropower Project situated in northeastern part of India. The coupled model can be applied to any reservoir with proper calibration of model parameters. By applying this model, a disaster manager would be in a position to disseminate in advance a more reliable downstream flood warning.