Journal of Hydrologic Engineering

This study presents a topography-dependent error adjustment technique for observed spatial rainfall estimate (OSRE) with the application of a multivariate t-copula model in a sparsely gauged region in Bangladesh. Multiple satellite precipitation products, including TRMM-3B42v7, TRMM-3B42RT, IMERG-Daily (Final run), PERSIANN-CCS, and PERSIANN-CDR, and reanalysis products, including MERRA-2, ERA-Interim, and JRA55, were tested to remove the topography-dependent errors in OSREs. The Triple Collocation (TC) method was employed as an alternative evaluation method of OSRE in the absence of adequate rain gauges. To address topography-dependent errors, the error adjustment model was separately run over the mountainous and plain zones. After bias correction, the correlation coefficient (CC) values were improved by 17.39%–38.6% for the mountainous region and 23.69%–47.83% for the plain basin. Based on categorical and volumetric performance evaluation matrices, the satellite product IMERG-Daily performed more reasonably than any other OSREs in both the plain and mountainous zones. CC values obtained for IMERG-Daily were high, whereas PERSIANN products exhibited low CC values within the observed datasets. Among the reanalysis products, JRA55 performed poorly, whereas ERA-Interim showed satisfactory results. All reanalysis and precipitation products except for TRMM-3B42v7 provided significant random errors in the coastal area and mountainous basin relative to the plain zone. The TC method provides a similar performance ranking to traditional methods, that is, comparing OSRE with rain gauges. This study works to fill the existing knowledge gap by developing a zone-wise copula uncertainty model for multisource satellite and reanalysis precipitation products. An evaluation technique in a sparsely gauged region using the TC method was adopted to measure OSREs’ performance in the absence of gauge rainfall data.

The importance of discharge measurements is fully acknowledged in many research fields, mostly in hydrology. Numerous measurement devices and various overflow structures have been proposed for discharge measurements; however, their use is based on calibrated simple stage discharge relationships that may cause significant errors when unsteady flow conditions occur. This issue is quite common because of rainfall and runoff temporal variability that inhibits the achievement of the steady state. Although this issue has already been experimentally investigated, it seems that a physically based line of approach has not been attempted before. In this paper, unsteady stage discharge relationships on a theoretical basis are derived for the most common shapes of weirs. According to previous experimental investigations, a transposition of the simple (steady-state) rating curve occurs, with higher input discharges than those under steady state, when the water level rises, and with lower input discharges, when the water level falls. The corresponding time to equilibrium is also studied and two different applications are performed. Finally, when the outflow process is unsteady, an error analysis quantitatively shows that using simple rating curves may determine high errors in discharge measurements, which decrease more and more at the attaining of the steady state, where no transposition of the rating curve occurs. Triangular weirs result more sensitive to the unsteadiness than rectangular weirs, which in turn result more sensitive than the sharp crested orifices. The suggested procedure could be relevant to improve the accuracy in the discharge estimate, as it could be demonstrated by comparing the discharges derived by the proposed procedure and the ones provided by the application of hydrodynamic and rainfall-runoff models.

A simultaneous survey of several types of droughts, such as meteorological, hydrological, agricultural, economic, and social droughts, is possible by using the multivariate standardized precipitation index (MSPI). In this study, the accuracy of four artificial intelligence (AI) methods, including the generalized regression neural network (GRNN), least-square support vector machine (LSSVM), group method of data handling (GMDH), and adaptive neuro-fuzzy inference systems with fuzzy C-means clustering (ANFIS-FCM), were investigated in forecasting the MSPI of three synoptic stations (Jolfa, Kerman, and Tehran) located in the arid-cold climate of Iran. The data used was monthly precipitation and belongs to a 30-year period (1988–2017). MSPI values were calculated in five time windows, including the following: 3–6 (${\mathrm{MSPI}}_{3–6}$), 6–12 (${\mathrm{MSPI}}_{6–12}$), 3–12 (${\mathrm{MSPI}}_{3–12}$), 12–24 (${\mathrm{MSPI}}_{12–24}$), and 24–48 (${\mathrm{MSPI}}_{24–48}$). The period of 1988–2016 was considered for training (75%) and testing (25%), and 2017 (12 months) was used for long-term forecasting. The methods were evaluated by the root mean square error (RMSE), mean absolute error (MAE), Willmott index (WI), and Taylor diagram. In the short-term forecasting phase, results showed that the methods had their best performances in forecasting multivariate drought types of groundwater hydrology-economic-social (${\mathrm{MSPI}}_{24–48}$), agricultural-groundwater hydrology (${\mathrm{MSPI}}_{12–24}$), surface hydrology-agricultural (${\mathrm{MSPI}}_{6–12}$), soil moisture-surface hydrology-agricultural (${\mathrm{MSPI}}_{3–12}$), and soil moisture-surface hydrology (${\mathrm{MSPI}}_{3–6}$), respectively. Also, among the mentioned methods, the weakest accuracy was reported for GRNN with an $\mathrm{RMSE}=0.673$, $\mathrm{MAE}=0.499$, and $\mathrm{WI}=0.750$ (related to ${\mathrm{MSPI}}_{3–6}$ of the Kerman station); the most accurate performance resulted from the GMDH with $\mathrm{RMSE}=0.097$, $\mathrm{MAE}=0.074$, and $\mathrm{WI}=0.989$ (related to ${\mathrm{MSPI}}_{24–48}$ of the Jolfa station). In spite of the acceptable performance of the models in short-term forecasting, by increasing the forecasting horizons, the models’ errors were increased in the long-term forecasting phase. The models could have acceptable long-term forecasts for just two months (or in some exceptional cases, three months) ahead. Further, according to the investigations, it can be shown that the methods show better performances in mountainous arid-cold regions, compared to desert arid-cold regions. As a theoretical study of multivariate drought forecasting, the AIs have promising results, and this research can be extended for the other regions.

The conventional hydrological modeling framework with a traditional streamflow-alone calibration approach is often challenged with the difficulty of accurately simulating the extreme flow conditions. This study explores the possibility of improving the simulation of extreme hydrological events by incorporating the antecedent soil saturation conditions using the Gravity Recovery and Climate Experiment (GRACE) terrestrial water storage anomaly (TWSA). In this study, a stepwise calibration approach is adopted on the hydrological predictions for the environment (HYPE) model, based on the GRACE-derived TWSA and streamflow observations, and is compared with the conventional calibration approach based only on streamflow observations. The performance of the two modeling frameworks is demonstrated over a tropical basin in Indian subcontinent, Mahanadi Basin, for the study period of 2003–2016. Results showed that the inclusion of GRACE data in addition to the streamflow observations significantly reduced the uncertainty of sensitive parameters and improved the model realizations. The model calibrated using GRACE-derived TWSA helped to improve the model physics and provided more reliable estimates of the terrestrial water storage. Although both calibration approaches are found to be equally good in simulating the mean streamflow conditions, the GRACE-based approach successfully simulates different aspects of the extreme flow conditions. An additional calibration case of remotely sensed top-layer soil moisture instead of the GRACE-derived TWSA illustrate that better model parameterization requires complete information of both surface water and groundwater storage in the modeling framework. Overall, the study highlights the potential of GRACE-derived TWSA estimates in improving the model physics, which further helps in the modeling of extreme flows over basins subjected to frequent extreme events.

Despite advances in data collection and modeling, there still remains a significant gap in predicting flows in ungauged locations. This study presents an approach for using gridded data to regionalize flow indices (long-term average streamflows and flows that are equaled or exceeded for 95% of the time) along an entire streamflow network grid. The methodology is based on using the Terrain Analysis Using Digital Elevation Model (TauDEM) toolset to obtain input variables for the regionalization model as averaged to the catchment area of each pixel in the streamflow network grid. The variables used as input for the regionalization regression were the catchment area of each pixel in the streamflow network grid, average slope, annual rainfall, and annual evapotranspiration of the corresponding catchment area. These variables were downscaled to the resolution of the Multi-Error-Removed Improved-Terrain (MERIT) digital elevation model (DEM) ($90\times 90\mathrm{m}$). The result was a $90\times 90\text{-}\mathrm{m}$ stream network grid with corresponding values for the modeled flow indices. Thus, the use of satellite-derived or gridded products improved the capacity of the grid to capture spatial patterns that fail to be captured in poorly gauged watersheds. The methodology can be adapted to other cases, for example, to model the parameters of hydrological models along the streamflow network grid. The prediction of the streamflow/flow indices over the entire basin can provide water managers with greater confidence in granting water rights and in sustainably managing the resources in ungauged catchments.

Due to its simplicity, the kinematic wave is commonly applied to open-channel routing in watershed hydrologic modeling. However, these applications must fulfill certain conditions, such as relatively large riverbed slopes and long time of rise in the flow. This study attempts to determine the most appropriate kinematic wave scheme for open-channel routing in the highly regulated Peace River, Canada. Five schemes were used to simulate a 5-day hydrograph with sudden plunges and hikes. The outputs from these five schemes were compared visually and statistically with the observed hydrograph. It was found that all five schemes are applicable to open-channel routing for the highly regulated Peace River if the temporal and spatial increments are set properly. However, one scheme, which is a total variation diminishing (TVD) high-resolution scheme, is the most appropriate scheme for this purpose. This scheme allows large temporal and spatial increments while relatively high accuracy can be achieved.

This study employs two heuristic algorithms, including the genetic algorithm (GA) and ant colony optimization for continuous domains (${\mathrm{ACO}}_{\mathrm{R}}$), for optimizing the parameters of two soft computing models, namely adaptive neuro-fuzzy inference system (ANFIS) and least-squares support vector machine (LSSVM), which were used for modeling monthly precipitation for all 12 months of the year. Data from 40 meteorological stations situated in different parts of Iran were used. The effectiveness of input data was determined by internal correlation-coefficient and nonlinear sensitivity analysis. Selected input data were further evaluated by another sensitivity analysis method, cosine amplitude (CA). Considering different evaluation months, LSSVM was more accurate and reliable than ANFIS. It was also found that both algorithms improved the performance of models for most months of the year. ${\mathrm{ACO}}_{\mathrm{R}}$ was better and more reliable than was GA in optimizing the models. ${\mathrm{ACO}}_{\mathrm{R}}$ produced the best results in autumn that led to the improvement of performance of ANFIS in terms of correlation coefficient ($R$) and root-mean square error (RMSE) by 35% and 0.40 mm for October; 42% and 0.99 mm for November; and 31% and 0.74 mm for December. The performance of LSSVM was enhanced by 6% and 0.28 mm for October; 22% and 0.20 mm for November; and 4% and 0.10 mm for December, respectively. For July and August, the suggested algorithms could not improve the performance of ANFIS. The algorithms did optimize LSSVM in all months, so the RMSE and mean absolute error were improved by 0.15 and 0.28 mm for July and 0.28 and 0.56 mm for August, respectively.

The purpose of this paper was to investigate the effect of the uncertainty of historical flood information on flood frequency analyses using a Bayesian approach. The uncertainty of the magnitude was described using variations in lower or upper bound values, but the return period was not changed. Two types of hydrologic data—synthetic data generated from the Pearson type three distribution population and observed peak discharge data from the Yichang hydrologic station located below the Three Gorges Dam on the Yangtze River, China—were used. For the measured and the synthetic data, when the magnitudes were underestimated and the errors increased, the accuracy of the historical flood samples decreased, the estimated values of the parameters and design floods decreased using the Bayesian and probability weighted methods, and the values using the Bayesian method were more sensitive to errors and more quickly decreased relative to the probability weighted method. In contrast, overestimated magnitudes did not improve the representativeness of the historical flood samples but merely appended their values with an increase in errors, and the estimated values were hardly affected by changes in the errors using the Bayesian approach. At the same time, the estimated values continued to increase using the probability weighted method and did not accord with the variations in the errors for the flood frequency analysis. In practice, when the return period could be set with a specific value using several historical or paleoflood events, accurately quantifying the peak values of historical floods with known information was difficult, and we overestimated their magnitudes within a specified interval using the existing approaches and available information to estimate their bounds. The estimated values of the parameters and design floods and their uncertainty were less affected when using the Bayesian approach.

Hydrology in low-relief coastal plains is especially challenging to simulate in flood modeling applications. Two-dimensional (2D) hydrodynamic models are often necessary, but creating such models for regional-scale systems at a high spatial resolution presents significant data challenges. The objective of this research is to explore these challenges using a 2D hydrodynamic model built for a $\mathrm{5,800}\text{-}{\mathrm{km}}^2$ region in the coastal plain of Virginia as a case study. Systematic enhancements to the hydrodynamic model’s topographic, bathymetric, streamline, surface roughness, and rainfall representations are tested to assess their impact on the model’s predictive skill. Results showed that incorporating high-resolution terrain and land use data sets alone only produced minor improvements to model accuracy. However, the addition of river cross-section data collected through site visits and careful, detailed quality control (QC) of observed rainfall data produced much more substantial improvements to accuracy. Based on these findings, increased focus should be placed on integrating topographic and river bathymetric data sets for low-relief coastal plain regions along with improved methods for QC of observed rainfall data, especially for extreme weather events.

This work reports climate change signals and long-term trend analysis of climate variables, meteorological drought, and extreme climate indexes over the tropical state of Kerala in India. The trend analysis reveals statistically significant decrease of annual and southwest monsoon rainfall (as much as 63 mm and 55 mm per decade, respectively). A decrease in number of annual rainy days (up to $2.8\mathrm{days}/\mathrm{decade}$) is also reported. Temperature trend analysis indicates an increasing trend with as high as $1.3°\mathrm{C}/\mathrm{decade}$. The spatio-temporal variation of extreme climate indexes across Kerala shows a decreasing trend of extreme precipitation indexes and an increasing trend of extreme temperature indexes. R95 and R95p decreased in northern and southern Kerala whereas R5 index increased in central and southern regions. Warm days have significantly increased whereas cold days exhibit a decreasing trend across the state. The increase in warmer nights is statistically significant whereas colder nights are decreasing in central and southern regions. Meteorological drought using Standardized Precipitation Index (SPI) reveals increasing occurrence of droughts in Kerala with higher frequencies over southern and central Kerala.

We compared precipitation intensity-duration-frequency (PIDF) curves developed for four small forested watersheds to spatially interpolated estimates from the National Oceanic and Atmospheric Administration’s (NOAA) Atlas-14. We also evaluated the Rational Method (RM) using on-site PIDFs and USGS Regional Regression Equations by comparing their estimated design discharges with a given exceedance probability $p$ ($Q_p$) to values computed from on-site data fitted to the Log-Pearson (LPIII) distribution. Overall, NOAA’s PIDF estimates were not substantially different from the on-site PIDFs. The 25-year and larger $Q_p$ by the RM were in closer alignment with LPIII estimates in the smaller watersheds, whereas $Q_p$ by the USGS were a better fit for the larger ones in most cases. Adapting return period-dependent runoff coefficient improved estimates by the RM in the large lowland watershed, but not in the other smaller high-relief watersheds. We recommend RM with 1-h duration NOAA-PIDF for designing road drainage structures in small and possibly the USGS method for large forested watersheds. However, future studies should focus on validation in watersheds of different sizes and topography.