Journal of Irrigation and Drainage Engineering

Open drainage channel design involves variables that are uncertain. Because the performance of the channel system is also uncertain, reliability analysis is used to measure the reliability of the system performance. In this paper, the reliability of open drainage channels under three possible failure modes is examined. The first failure mode occurs when the runoff exceeds channel capacity. The runoff and channel capacity are random variables that are estimated using the rational method and the Manning equation, respectively. The second failure mode occurs when the actual flow velocity exceeds the maximum allowable velocity for erosion control. The third failure mode occurs when the actual flow velocity is less than the minimum allowable velocity for deposition control. The minimum and maximum allowable velocities are considered random variables. The failure probability of each mode is estimated using the advanced first‐order second‐moment (AFOSM) method. The overall failure probability of the system that accounts for the correlations between the failure modes is presented. The method was applied to an example and was verified using Monte Carlo simulation. In practice, the method can be used to find the reliability of an existing channel under multiple failures, to evaluate the effects of alternative improvements, and to design a new channel for a specified reliability level.

Simulation of aggregate water demands by areas commanded by canal and lateral turnouts is an important input to the operation and management of the irrigation delivery system. A computer simulation model to predict these command area water‐demand functions was developed and reported by Keller in 1987. To test the validity of the model, data from an existing irrigation project in the Delta, Utah, area were used to simulate the aggregate irrigation demand of five turnouts operated on a demand basis. During the model calibration process, the influence of three important variables in the operation stage on the prediction of demand was identified using the cumulative water‐demand curve. The key variables were: (1) Management‐allowable depletion (MAD); (2) irrigation application uniformity (IAU); and (3) initial soil moisture (ISM).

The cumulative water demand (CWD) curve was used for validating and calibrating the unit command area (UCA) model, which predicts the aggregate water demands of individual command areas for the entire crop season. During the model calibration process, as presented in the companion paper, three important variables in the operation stage on the prediction of demand were found to be dominant in determining the CWD curve profile. The logistic function with three parameters, which is a particular form of the exponential‐type function, was adopted for fitting the CWD curve. High correlations were observed between the individual curve parameters and the field variables. Three parameters are management‐allowable depletion (MAD), irrigation application uniformity (IAU), and initial soil moisture (ISM). Effectiveness of the equation developed for a particular UCA using the regression analysis was successfully demonstrated, and applicability for some other UCAs in the same water system was also recognized.

This paper presents a procedure for estimating the total pollutant mass loss resulting from the interrill component of agricultural runoff. The procedure involves successive applications of a flow model and a mass transport model. The two‐dimensional flow model developed by Limback et al. in 1985 provides depth and velocity distributions in the interrill flow domain, where the flow is longitudinally uniform. These distributions are used as input for a diffusional mass transport model to estimate pollutant losses from soil interstices into the overland flow. Empirical relationships are used in the mass transport model to express dispersion coefficients as functions of overland flow and soil parameters. The procedure, based on the interrill flow model, can be used with other mass transport models instead of the diffusion model presented here. Such models might include infiltration, adsorption, convection, or other pertinent physical processes.

The operation of a new irrigation canal system in central Arizona is described. Water is relatively expensive, and thus farmers pressure the irrigation district to improve delivery service and keep water losses to a minimum. The system was designed and constructed to function with supervisory (remote‐manual) control, with the potential for automatic (remote‐computer) control. Experiences with hardware and software difficulties are described. While water control and delivery service are currently very good, further improvements through automation are possible. Real‐world operating conditions are described that are somewhat different from the assumptions made for the development of canal‐control algorithms. Field data collection and canal‐flow computer simulation were used to develop a test case for developers of automatic control algorithms. A set of canal‐control performance measures is suggested. For the conditions studied, upstream manual control performed better than a simple local downstream controller.

Chemical irrigation is being used increasingly as a means of applying agricultural chemicals, but the effects of different irrigation practices on the movement and distribution of these chemicals in the soil are not well understood. This paper reports the results of a field study where bromide is applied as a tracer under both sprinkler and ponded irrigation. The travel‐depth bromide concentration curves under sprinkler irrigation were smooth and unimodal, with very little variability between soil cores. For ponded irrigation, however, the curves were bimodal and showed much greater variability between cores. The writers attribute the differences in shape and variability to the presence of preferential flow in the ponded case. Results of this study have important implications for ground‐water pollution potential and monitoring of the vadose zone.

One aspect of the riparian‐stream ecosystem, which is imperative for the restoration or reestablishment of a streambank and streambed that has been disrupted, is the ability of the stream system to entrap and retain sediment. Sediment deposition in a degraded stream system is essential to the natural rebuilding of channel morphology. Although it is recognized that herbaceous vegetation enhances sediment deposition, a quantitative measure of entrapment has not been developed. Furthermore, once sediment has been deposited or entrapped, a measure of sediment retention during flushing has not been identified. A meandering stream was physically simulated in the laboratory. A series of 29 tests was conducted to evaluate the sediment deposition and entrapment potential of Kentucky bluegrass when placed in the streambed. Sediment was injected into the stream for 6 h and then subjected to sediment‐free flushing flows. The maximum amount of sediment deposition and the amount of sediment retained after flushing was documented. The results indicate that the presence of vegetation significantly enhances not only the deposition, but also the entrainment/entrapment of sediment on the vegetative bases. The amount of sediment entrapped is a function of the flow and vegetation blade length. It was determined that the vegetation could retain from 30–70% of the deposited sediments.

This work presents analytical and numerical models for the analysis of steady‐state bed‐level variation in gradually expanded open channels with movable beds. The analytical model employs the one‐dimensional flow assumption, and equations for flow depth and bed‐level variation have been derived. The numerical model uses a false transient principle with the quasi‐steady uncoupled approach to solve the two‐dimensional governing equations in a transformed coordinate system, which allows an orthogonal grid and an easier inclusion of sidewall boundaries. A second‐order accurate, explicit finite‐difference scheme (Mac‐Cormack scheme) has been used for water flow computation. A comparison between the two shows that the analytical results are comparable to those obtained by the more complicated two‐dimensional numerical model for small expansion angles. Though not discussed in this technical note, the numerical model can be applied to any type of expansion and is capable of predicting the temporal variations of flow depth and bed level.

The prediction of sediment yield from catchments is necessary for studies of reservoir sedimentation, morphologic modeling, and soil‐conservation planning. Though records on sediment yield are generally not available, methods that can be used to predict sediment yield from other commonly available hydrologic data, are available in the literature. In this study some of these methods are verified for their accuracy, using carefully collected data from experimental catchments. It is found that existing methods do not adequately account for the process of sediment delivery; hence these produce a less accurate prediction of sediment yield. A new method, based on the routing of surface erosion through time‐area segments, is proposed. This method is found to produce the sediment yield more accurately than those it has been compared with. The method being proposed for the prediction of sediment yield has a more sound basis for further use in distributed models.

Values of reference evapotranspiration $\left({\mathrm{ET}}_0\right)$ are used with crop coefficients $\left(\mathrm{KC}\right)$ for many aspects of irrigation and water resources planning and management. More than a score of methods are used for estimating ${\mathrm{ET}}_0$. Estimated values vary widely due to the lack of standardization of the reference. Evaporation (ET) measured by lysimeters of various grasses and/or alfalfas has been used as the standard for developing estimating equations. Due to variation in the references used, some international organizations now wish to promote the use of a single equation or method to avoid the confusion caused by the current diversity. Different versions of the Penman combination equation have been proposed. The Research Center for the European Community and the ASCE Committee on Irrigation Requirements have evaluated various equations for estimating ${\mathrm{ET}}_0$. Due to its simplicity and the accuracy of estimates, the 1985 Hargreaves et al. equation is recommended for general use. ${\mathrm{ET}}_0$ is used in irrigation planning, design, and scheduling and for other water adequacy studies.

Although living alfalfa plants show no elongation during dormancy, evapotranspiration $\left(\mathrm{ET}\right)$ can occur. The yield of the first harvest and dormancy $\mathrm{ET}$ could be statistically correlated as it is correlated during the growing season. The objective of this study was to evaluate the statistical correlations of yield and $\mathrm{ET}$ and of water use efficiency (WUE) and $\mathrm{ET}$ during dormancy and nondormancy. The correlations involving data from several years, with four harvests per growing season, were evaluated based on slopes and intercepts. Six data points for which dormancy $\mathrm{ET}$ was known were added to evaluate the statistical fit. Correlations between yield and $\mathrm{ET}$ for various harvests had one common slope but different intercepts; the same was true for the correlation between WUE and $\mathrm{ET}$, except the latter correlation was weak in terms of $R^2$. Average values of $\mathrm{ET}$ and the corresponding estimated yield showed that between one‐third and half of the observed average water loss during dormancy fit the estimated yield and $\mathrm{ET}$ correlation. The remaining amount appeared to be an extra water loss during dormancy.