Journal of Hydrologic Engineering

The distributed large basin runoff model (DLBRM) was designed to simulate the hydrological processes of the Great Lakes watersheds. As part of its development, the DLBRM was recently applied to 18 watersheds in the Lake Erie basin, where it was first calibrated to reproduce the observed discharge in 1950–1964 and then applied to 1999–2006. Four different calibration objective functions: root mean squared error (RMSE) minimization, mean absolute error (MAE) minimization, correlation maximization, and Nash-Sutcliffe index maximization were tested, revealing RMSE minimization as the most successful method and able to achieve results very close to its global minimum. Further, the distribution of the main DLBRM parameters in the 18 watersheds was consistent with regional patterns, although each watershed was calibrated individually, thus adding credibility to the calibration process. Model performances, while generally good, varied across the basin according to a series of environmental factors, including climate, watershed shape, topography, and land cover and observation factors such as gauging station distribution. Gauging station coverage proved to be extremely important in the ability of the model to track flow variability. The DLBRM proved to be able to replicate well the 1999–2006 hydrologies of most watersheds without recalibration. However, its performance declined in heavily urbanized watersheds, where the landscape changed the most. The results described in this paper will lead to improved model performance and increased practical applications of the DLBRM, providing important information to researchers and decision makers for efficient water management programs in Great Lakes watersheds.

The objective of this study was to investigate the presence of trends in extreme precipitation (denoted MAXP and defined as the annual maximum daily precipitation depth) time series for coastal northern New England and to assess changes in the magnitude of the so-called 100-year storm. MAXP depths from 48 stations with long, continuous records in Maine, New Hampshire, and Massachusetts were analyzed. At those same stations, the number of daily precipitation depths $\ge 2\mathrm{in}.$ (51 mm, denoted as GT2in) was also quantified for each year. Although the seasonally averaged MAXP was found to be fairly uniform throughout the year, the frequency of MAXP is highest during August through October, the typical hurricane season in New England. The presence of trends in MAXP and GT2in was evaluated over four time frames (1954–2005, 1954–2008, 1970–2005, and 1970–2008) using two statistical methods (linear regression and the Mann-Kendall trend test) and at two scales (at-site and regional). The trend analysis over the time period 1954–2005 indicated that MAXP was amazingly stationary; however, a trend in GT2in was found at some stations. More trends in both MAXP and GT2in were present in the time period 1954–2008. The majority of stations in southern New Hampshire and eastern Massachusetts showed evidence of trends in MAXP (but not GT2in) for the time period 1970–2008. That the number of trends in MAXP increased despite the shorter record length suggests a strong increase in the magnitude of extreme precipitation in northern coastal New England in the last few decades. The stationarity of the 1954–2005 record was confirmed by the regional trend analysis, as was the presence of stronger trends in coastal stations when the record was extended through 2008. Most stations that had trends in MAXP also had trends in GT2in. The generalized extreme value (GEV) distribution was used to estimate 100-year precipitation depth quantiles for the 1954–2005 record, which were then compared with Technical Paper No. 40 (TP-40) 100-year, 24-h precipitation depths. Estimates for stations along coastal Massachusetts, New Hampshire, and Maine all exceeded 7 in. (180 mm) and exceeded TP-40 by 1 in. (25 mm) or more. Stations in northeastern Massachusetts, southeastern New Hampshire, and southern Maine exceeded 8 in. (200 mm) and also exceeded TP-40 estimates by more than 2 in. (51 mm). These findings indicate that TP-40 underrepresents coastal storm depths. This study, as well as recent record-breaking events in northern New England, strongly suggests the need for updating of design storm estimates. Furthermore, extreme precipitation events of longer than one-day duration have caused large-scale flooding in the region over the last decade. The magnitude of longer duration storms (particularly two-day storms) may also be increasing, calling for engineered infrastructure that can accommodate increases in both storm magnitude and duration.

The Soil Conservation Service–curve number (SCS-CN) method is widely used in hydrologic practice, but its application often goes beyond the purpose of its original use. Common practice makes the method dependent on the antecedent precipitation index, a simple indicator derived from rainfall depth, which can be used to estimate the antecedent moisture condition (AMC) of soil. By taking data from an unspecified location in the United States, the SCS defined the appropriate AMC level based on the total five-day antecedent rainfall for dormant and growing seasons. These values, originally defined as an example practice at the plot scale, were subsequently adopted for general use without full awareness of scale effects and regional differences. We propose a revision to the amount of antecedent rainfall for the definition of AMC of soil, based on the minimization of the errors between an estimated and an observed runoff volume of 347 rainfall-runoff events in large mountain basins in Italy and Switzerland. A significant correlation between the new thresholds and the topographic index is shown, thus permitting extension of the method to basins not included in this study.

According to the hydrodynamic dispersion theories, the dispersion parameter is proportional to a power $n$ of the velocity; the power ranges between 1 and 2. Based on the value $n=1$, analytical solutions of the dispersion problems along temporally dependent flow domains were obtained in previous works. In the present work, two dispersion problems are addressed for $n=2$. Using the Laplace transform technique, analytical solutions are obtained for two-dimensional advection-diffusion equations describing the dispersion of pulse-type point source along temporally and spatially dependent flow domains, respectively, through a semi-infinite horizontal isotropic medium. Point sources of a uniform and varying nature are considered. The inhomogeneity of the medium is demonstrated by the linearly interpolated velocity in the space variable. Introduction of new space variables enable one to reduce the advection-diffusion equation in both problems to a one-dimensional equation with constant coefficients. The solutions are graphically shown.

In this study, the spatial and temporal patterns of trends for reference evapotranspiration (RET) at 34 meteorological stations (between 1957 and 2007) in the Haihe River basin, China, were analyzed using the Mann-Kendall (MK) test and the Sen’s method. To reveal the possible causes and main driving forces of the changing patterns of RET, the spatial distribution and temporal patterns of trends for four meteorological variables (i.e., temperature, wind speed, relative humidity, and sunshine duration) were examined for each station. In addition, partial relative analysis between RET and meteorological variables and a sensitivity analysis of RET to meteorological variables were conducted. The results show the following: First, the Haihe River basin is dominated by a significant decreasing MK trend in annual RET at $>95\%$ confidence level, which is observed at most stations in the eastern and southern areas of the basin. There are no observed trends or significant increasing MK trends in annual RET in the western plateau. For the intra-annual variability of RET, similar spatial-changing patterns are only observed in summer. Second, there are no obvious trends in RET from the late 1950s to the early 1970s for the whole basin. However, opposite trends of RET in the plateau area are identified after the 1970s in comparison with those in the plain and mountain area of the Haihe River basin. These two trends become significant in the late 1990s. Third, wind speed and sunshine duration are recognized as the two major driving forces for the decreasing trends in RET. Relative humidity is mainly responsible for the increasing trends in the western Haihe River basin during recent decades.

Identification of critical source areas (CSAs) of pollution in a watershed is important for effective implementation of best management practices (BMPs). Process-based watershed models are often used for this purpose. One of the main inputs to these models is the spatially explicit soils data. The objective of this study was to evaluate whether the use of two commonly used soil data sets, the State Soil Geographic (STATSGO) and the Soil Survey Geographic (SSURGO) data, can lead to differences in location of CSAs of sediment. A watershed model, Soil and Water Assessment Tool (SWAT), in combination with the Tukey-Kramer test was used for locating CSAs in the Fish River watershed located in coastal Alabama. The model was calibrated and validated using flow data from a U.S. Geological Survey (USGS) gauging station located within the watershed. The locations of the CSAs of sediment were analyzed at subwatershed and hydrologic response unit (HRU) levels. Results show that the locations of the CSAs were different for the two soil data sets. The locations of the CSAs varied at both subwatershed and HRU levels. The use of STATSGO soil data resulted in higher soil erodibility factor and surface runoff. As a result, higher sediment yield was obtained from the use of the STATSGO data as compared with the sediment yield obtained from the use of the SSURGO data. Therefore, for accurate identification of CSAs of sediment (and potentially other pollutants) and for effective implementation of economically feasible BMPs, it is important to use the most detailed spatial data set available.

Historical floods on the Delaware River were analyzed using 15-, 25-, and 50-year-long records to determine whether floods in recent years are worse than those in previous periods. The 111-year-long record of annual maximum floods at Trenton, New Jersey, was used. Predictions of the 10-, 50-, 100-, and 1,000-year floods were seen to vary significantly with the record length used to predict them.

The Green-Ampt model is an approximate analytical solution to Richards’ equation that is commonly used to simulate infiltration processes in hydrological models and land surface schemes. The Green-Ampt model assumes that neither a water table nor an impermeable layer (e.g., bedrock or a frost table) exist near the soil surface. In regional-scale applications these idealized conditions will often not be met, and it is presently unclear what implications this has for regional water resource models. This paper investigates the limiting conditions under which the Green-Ampt model is appropriate and how individual assumptions about lower boundary conditions affect the validity of the model. Guided by the comparison between the Green-Ampt model and numerical solutions to Richards’ equation, various simple revisions to the Green-Ampt model are suggested. Results demonstrate that even when the traditional assumptions are relaxed, the Green-Ampt model often still provides reasonable results for regional-scale analysis and can be amended to account for conditions for which it was not intended.

Bioretention, a key structural practice of low impact development (LID), has been proved to decrease peak flow rates and volumes, promote infiltration and evapotranspiration, and improve water quality. Exactly how well bioretention mimics predevelopment (or “natural”) hydrology is an important research question. Do bioretention outflow rates mirror shallow groundwater interevent stream recharge flow associated with natural or nonurban watersheds? Streamflow from three small, nonurban watersheds, located in Piedmont, part of central North Carolina, was compared with bioretention outflow from four cells also in North Carolina’s Piedmont region. Each benchmark watershed drained to a small stream, where flow rate was monitored for an extended period of time. After normalizing the flow rates and volumes by watershed size, data were combined to form two data sets: bioretention outflow and stream interevent flow. Results indicate that there is no statistical difference between flow rates in streams draining undeveloped watersheds and bioretention outflow rates for the first 24 h following the commencement of flow. Similarly, there is no statistical difference between the cumulative volumes released by the two systems during the 48 h following the start of flow. These results indicate that bioretention cells behave comparably to watersheds in natural or nonurban conditions, with respect to both flow rates and flow volumes, and suggest that bioretention outflows may mirror post–storm event shallow groundwater interevent stream recharge flow. Solely considering bioretention outflow as a conjugate to runoff may be a misinterpretation of a flowrate that actually resembles shallow interflow.

This study presents a hydrogeological analysis of a humid tropical, seasonal river in the context of climate change, increasing demand for water, and uneven distribution of rainfall. We investigate the Netravathi basin, a tropical river basin of south India. The climate change effect on the basin was evident in terms of increasing trend in temperature by about $0.7°\mathrm{C}/100\mathrm{years}$ and decreasing trend in the river flow during the monsoon by about 0.8% of average annual flow per year using the Mann-Kendall trend test. Even though rainfall was found to be decreasing, no significant trend could be established. From the trend analysis of the river flow, it was found that there is an overall declining trend with longer scarcity periods. In addition, the trends of magnitude and frequency of high flows are declining. Even though the region receives an average annual rainfall of about 3,930 mm, it has nonuniform distribution with most of the rainfall confining to a few months of a year. In view of this, the region suffers from a prolonged dry period during February to May. The projected domestic water demand of the region for the next 25 years is estimated to be increasing from the present $0.09{\mathrm{mm}}^3$ to $0.25{\mathrm{mm}}^3$ per day because of rapid urbanization and industrialization. The purpose of this investigation is to highlight the effects of climate change and uneven distribution of rainfall in the river basin. This may assist in proper planning of the basin through strategies such as river water harvesting, which is investigated in the companion paper. Because the Netravathi River is a seasonal and tidal river, and saltwater intrusion along the river during the summer months is affecting the development of the basin. It was found that the river water is affected up to distance of about 22,000 m from the Arabian sea and the wells on the banks of the river are found to be highly vulnerable to saltwater intrusion during the summer period (March to May).

Tropical rivers are predominantly seasonal in nature, and managing water resources during the deficit period is becoming more difficult because of the rapidly increasing demand for water. The present investigation focuses on harvesting Netravathi River water in the southern Indian peninsula through a series of vented dams with an estimated storage capacity of $102{\mathrm{Mm}}^3$ for use during the deficit period. A brief hydraulic design of a vented dam at a specific location is presented. The spacing and capacity of these reservoirs were worked out on the basis of the dam height and the river characteristics. The proposed vented dams are seasonal dams, and the closure of the vents will be decided on the flow available (i.e., 95% dependable flow), the storage capacity, and the minimum water release required for the downstream ecosystem. The appropriate time to start storing water in the vented dams was estimated to be in the month of November, and the entire process of storing water in the vented dams may last for about 41 days. An operational protocol for the storing process is presented. The investigations of aquifer parameters were performed by using electrical resistivity, pumping, and soil tests. The results indicated that the aquifer is shallow, unconfined in nature, and had a depth ranging from 18 to 30 m and hydraulic conductivity ranging from 62.6 to $406\mathrm{m}/\mathrm{\text{day}}$ . A multiple regression model developed to assess the groundwater recharge in the adjoining well fields indicated that water table fluctuations may be 30% of reservoir level fluctuations. Because the river is also tidal in nature, a saltwater exclusion dam is proposed at the lower reaches of the river to prevent the entry of saltwater along the river during the summer period.