Journal of Irrigation and Drainage Engineering

On‐farm storage facilities can be useful in satisfying the variable diurnal irrigation demand during a day and, at the same time, can prevent waste of water. In this study, an irrigation pond (IP) model is developed for the determination of required minimum storage capacity at a desired reliability level with a given intake operating rule to meet the fluctuating water demands. This IP model is applied to determine the storage capacities of several lateral canals of the Mae‐Taeng Irrigation Project in Chiang Mai Province, Thailand. Simulations are carried out to analyze effects of different cropping patterns and hourly irrigation demand patterns at various reliability levels on the storage capacities of irrigation ponds. The determined storage capacities of irrigation ponds for various lateral canals having different cropping patterns varies from about $1\text{,}870m^3$ to $2\text{,}525m^3$ for a 100 ha command area at 90% reliability. It is observed that introduction of on‐farm irrigation ponds can reduce both water waste and deficit by about 20–30%.

Unsteady flow problems in open channels can be classified as simulation‐ and operation‐type problems, according to the objectives of the study. The simulation problem is for predicting the discharge and water level in the channel during the future time series under given conditions. The operation problem, on the other hand, is for defining inflow at the upstream end of the channel or the operation schedule of controlling structures in order to get the desired outflow at the downstream end of the channel. In this paper, a finite difference computation method is derived for solving operation‐type problems in open channels. The method is explicit and numerically stable. It was applied to solve the St. Venant equations, discretized by the Preissmann scheme. The computation results were compared with the results obtained using the double‐sweep method, and there was good agreement between the two methods. The complete momentum equation was used in both methods. This computation method is called a backward‐operation method since it is for solving operation‐type problems, and the computation is performed backward both in space and time.

A study is made of transients in canal networks using a mathematical model based on St. Venant equations. The transient effects of different combinations of types, location, and operation of control structures are studied. The results of 14 problems on two canal networks are presented. It is found that the wave reflection effects from control structures, especially fixed gates, are significant and, in some situations, may cause embankment erosion and possible overtopping. The time taken for the first wave to reach tail‐end locations, and for the flow to reach steady state, is considerable. It is observed that steady state is reached faster in situations of canal start‐up than in canal shutdown, even though shutdown transient effects manifest earlier. A comparison is made between the complete network model and a simpler model of the main canal system, with the branches treated as bulk lateral outflows. It is found that the depths and discharges in the main canal are significantly underestimated by the bulk lateral outflow model.

Calibration of the rainfall and overland flow erodibility coefficients of a physically based erosion and sediment yield component for the Système Hydrologique Européen (SHE) modeling system is demonstrated using data from $32.54‐m^2$‐field rainfall‐simulator plots with tilled, tilled‐twice, clipped, grazed, and ungrazed land treatments. Good sensitivity to the major classes of treatment is exhibited, but the fine detail of variation for a given land cover is less well represented. This is a result of a lack of quantitative understanding of relevant processes that can be remedied only through detailed process studies. Good simulations of overland flow response were achieved for successive rain applications on wetted soil. Difficulties experienced in simulating overland flow response for dry soil probably resulted from poor definition of the soil hydraulic properties for dry conditions. Tests using other soil erosion data show that laboratory calibration of the erodibility coefficients is unreliable, while sensitivity of the simulated sediment yield to the input flow data indicates the need for accurate flow calibrations.

The presence of vegetation in irrigation canals in Egypt after the Aswan dam was built has become considerable problem in irrigation system. Many irrigation canals were designed by applying the common Manning's equation, but the presence of submerged weeds makes the application of Manning's equation for design of wide channels inadequate, because of the variation of the retardance coefficient over a wide range. Four canals were selected and considerable field data collection was carried out to identify the flow resistance and to modify Manning's equation. The velocity, slope, and discharge were measured in wide channels that had different capacities and that were filled with submerged vegetation (Potamogeton spp.) through the annual life cycle of weed growth. A more general discharge‐depth equation that offers advantages for a modified Manning's equation is developed based on the field data. A relationship was found between Manning's roughness coefficient for earth canals and an equivalent Manning's roughness for submerged weeds. The experimental setup is described and the results of the measurements are given.

The resistance to flow in a grassed channel is linked directly to the relative roughness of the flow through long‐accepted boundary‐layer principles. The difference between flow over grass linings and linings composed of rigid materials such as riprap or concrete lies in the tendency of the grass to deform under an imposed fluid shear. This deformation results in a smoothing of the channel roughness with increasing flow. Mathematically, this can be accommodated with coefficients that define the Darcy‐Weisbach friction factor and are dependent on relative values of boundary shear and vegetative stiffness. The emphasis of the paper is on showing that the relative roughness resistance equation can successfully reproduce the n‐VR retardance curves. Aspects of channel stability are discussed and the paper describes some problems associated with stability design.

Hydraulic conductivity is one of the most important factors controlling water‐table depth in subsurface drained agricultural land. The current design procedure is to calculate lateral drain spacing using some average of hydraulic conductivity measurements within the area to be drained. In this paper, a procedure is presented to take into account uncertainty in hydraulic conductivity and water‐table depth when delineating areas to which a single lateral drain spacing can be assigned. The procedure uses simple block kriging of log‐transform hydraulic conductivity data to estimate the effective hydraulic conductivity of square blocks with sides equal to lateral drain spacing. (Kriging is a statistical method, based on the theory of regionalized variables and developed by D. G. Krige, for local estimation of soil properties.) Kriged estimates and estimation variances are combined to define the cumulative distribution function of the water‐table depth, which is the basis for deciding whether a given area needs to be subdivided to achieve adequate water‐table control. This procedure can be implemented for interactive use into subsurface drainage design software using computer‐aided design (CAD) and geographical information system (GIS) technology. The method is illustrated with the Hooghoudt's equation and applied to a practical design situation in a 80‐ha area in the Nile Delta of Egypt.

Cost models that provide a preliminary evaluation of the economic feasibility of portable pipe, traveling gun, and center‐pivot irrigation systems on a site‐specific basis are presented. The models require limited input data, with simplifying assumptions about field shape, pipe orientation, and sprinkler layout used to size irrigation system components. Fixed costs are based on market estimates of the cost of irrigation system components, and can also include water‐source development, drainage, and clearing costs. Variable irrigation costs (energy, labor, and maintenance) are calculated as a function of the expected number of agricultural drought days for the site. A sensitivity analysis of the cost models shows that crop price, interest rate, and irrigation system component costs are the most important input parameters affecting model output. The results of applying the cost models to two potential irrigation sites in Virginia are presented.

A subfactor‐based regression technique for estimating hydraulic roughness coefficients for shallow overland flow was developed from simulated rainfall/runoff plots originally collected for erosion studies. The data were collected from 14 different native rangeland areas in the western United States. Rainfall was applied at a constant intensity of 65 mm/hr from a rotating‐boom rainfall simulator. Surfaces evaluated ranged from smooth bare soil to gravelly bare soil and sparsely to densely vegetated rangeland areas. A reference table of “effective roughness” coefficients for shallow overland flow is presented with a description of site characteristics. The derived roughness regression equations predict an “effective Darcy‐Wiesbach roughness coefficient” for native rangeland $(r^2=0.70)$ that incorporates the effect of raindrop impact, soil texture, random roughness, rocks, litter, and canopy and basal plant cover. The sites evaluated in the paper covered a wide range of vegetation types and included short‐, mid‐, and tallgrass prairies; desert shrubs and sagebrush; and oak and pinyon‐juniper woodlands. No trend in effective roughness coefficient associated with type of vegetation (grass or shrub) or soil texture was apparent.

A recently developed two‐dimensional finite element transient water‐flow model was used to simulate infiltration by furrow irrigation with a shallow water table. Suggested applications are primarily those that focus on unsaturated zone water flow. Field soil measurements were used to compare field‐measured with simulated infiltrated water for two‐layered soil systems. Simulated and measured infiltration were fairly close for one irrigation event, but significantly different for another. To further investigate the causes of discrepancies between simulated and measured infiltration, we also applied the water‐flow model to simulate the influence of variable soil and flow characteristics on infiltration in a furrow. We simulated infiltration for varying water levels in the furrow, for water‐filled and dry adjacent furrows, studied the effect of a surface seal on infiltration, and compared homogeneous with heterogeneous soil profiles with preferential flow regions. Lateral water movement was found to be insignificant, and the influence of a variable furrow water level was almost completely described by the changing wetted perimeter. Introduction of a spatially variable hydraulic conductivity in the two‐dimensional soil domain resulted in an immediate response of the water table to ponded infiltration.

The hydrology of two North Carolina muck soils was simulated using four water‐management methods; conventional drainage, improved subsurface drainage, and two types of controlled drainage. One soil was much more responsive to water management than the other. For the more responsive soil, improved subsurface drainage reduced surface runoff by 192 mm/year (66%) but raised outflow by 40 mm/year (10%). Large daily flows $(>25\mathrm{mm}/\mathrm{day})$ accounted for only 11 mm/year of flow from an improved subsurface drainage system versus 31 mm/year from a conventional drainage system. Controlled drainage moderated the effects of improved subsurface drainage. Water management did not alter the likelihood of very large ($>10‐\mathrm{year}$ return period) runoff events on either soil. A three‐year return period event was similated for the Wasda soil on different‐size watersheds. Compared to conventional drainage, improved subsurface drainage reduced peak runoff from 101 to 28 mm/day on a 8.1‐ha field, from 68 to 20 mm/day on a 130‐ha watershed, from 30 to 15 mm/day on a 1,036‐ha watershed, and from 20 to 13 mm/day on a 6,216‐ha watershed.

Grid networks, consisting of nodes and line elements, are often used in porous media flow simulation. An important step in the simulation process is the quantification of the line element conductivity when the pressure or capillary potential at both end nodes is given. Under saturated condition, when the porous medium is layered, the variation of conductivity along the line element is treated by calculating a composite conductivity as the line element conductivity. The composite conductivity under saturated condition can be derived analytically and is the harmonic mean of the saturated conductivity of the layers. For the unsaturated flow in the vadose zone, analytically derived composite conductivity is not available. Heuristic rules, including weighted, geometric, and harmonic means, have been used to estimate line element conductivity in practice. The lack of agreement in estimating composite conductivity indicates that the simulation technique for the vadose zone has not reached its maturity, and that improvements are needed. The purpose of this paper is to derive an analytical method to estimate vadose zone composite conductivity to be used in porous media flow simulation.

The original Horton equation is valid only if the rate of rainfall is always higher than the infiltration capacity of the soil throughout a rainfall-infiltration process. This technical note presents a modified Horton equation that expresses the soil infiltration capacity as a function of the amount of water that has already infiltrated into the soil. The modified equation is derived mathematically by combining the original and the integrated forms of the Horton equation. Therefore, the original Horton parameters are maintained in the modified equation. The modified and the original equations will give the same results when the assumptions of the original equation are valid. The advantage of the modified equation is that it can be used regardless of whether the rate of rainfall is greater than or less than the infiltration capacity. The technical note also includes expressions for the ponding time, and the amount of infiltration until ponding occurs.