Journal of Irrigation and Drainage Engineering

This paper presents an approach for computing economically optimal sustained yield ground‐water extraction strategies and supportable irrigated crop acreages. Computed regional strategies maximize the present value of net economic return. They are useful for long‐term agricultural planning because they are sustainable even beyond the planning period considered within the economic optimization. The hybrid approach uses only steady‐state flow equations and iterative simulation/optimization to reduce optimization memory requirement below that required by conventional models using both steady and unsteady equations. SECTAR, a quadratic optimization planning model, assumes heads will evolve toward the optimal steady state but initially does not know the nonlinear rates of evolution in each cell. Rates of change in head are assumed, and optimization is performed. Subsequently, a simulation model computes actual time‐varying rates of head change that would result from implementing the optimal strategy. Rates of change assumed in the optimization model are then corrected to correspond to simulated values. The process of assuming, optimizing, and then simulating is repeated. Assumed and simulated heads are very similar within three or four cycles. Convergence occurs because of the common tendency to evolve to steady state. The procedure is applied using data from an unconfined aquifer. Optimal water‐use strategies are fairly stable regardless of initial potentiometric surface and tested aquifer and economic parameters.

Field crops such as cotton, wheat, corn, and peas have been irrigated with treated domestic effluent for five years by a subsurface trickle system. For four years the experiments were conducted in the same field without replacing any components of the irrigation system. According to the field results it was found that a subsurface system can be used without failures for years. Emitter clogging by plant roots was prevented by Treflan injection. The uniformity of plants in the experimental plots indicated that the effluent water was well distributed in the soil. Increased yields, as compared to microirrigation systems with emitters on the soil surface, were obtained with subsurface microirrigation.

In urban drainage networks on steep terrain, drop manholes are often employed for energy dissipation and reduction of flow velocities. The hydraulic behavior of such circular manholes located between supercritical pipe segments is studied by means of laboratory experiments. It is found that the local head‐loss coefficient is governed solely by a dimensionless drop parameter, expressed in terms of the drop height and the inflow velocity. An analytical expression based on the experimental results is proposed for practical applications. The water level inside the manhole depends also on the angle between the inflow and outflow pipes, being higher when they are normal as compared to parallel. Optimum ranges of the drop parameter, in order to avoid impediment to the flow by formation of a hydraulic jump or surcharging of the pipe, may be determined and used in engineering practice.

The Al‐Hassa Irrigation and Drainage Authority in The Kingdom of Saudi Arabia covers approximately 8,000 ha of cultivable land. To irrigate this project, a precast concrete open‐channel system has been in operation since 1971. This open‐channel network, which consists of 19 mains, 178 submains, and 1,641 laterals, has more than 1,800 valves and vertical gates to divert water in different reaches of the system. The water is distributed according to a predefined irrigation schedule. In order to decide gate operating policies, a computer package known as channel network model has been developed to simulate depth and discharge variation with time under unsteady‐state condition. This paper introduces dynamic‐node‐numbering concept in channel network model by which the control devices can be completely closed and the original schematic diagram, node numbering, and corresponding data for the modified system are internally updated without altering the original data files. This paper also addresses other features of the model such as gate operation under low differential head, interactive capabilities of the model, and simulation of large‐scale irrigation network.

A new procedure for in situ determination of saturated hydraulic conductivity of an initially unsaturated soil is proposed. The procedure is based on the measurement of the rise of a ground‐water mound in an initially unsaturated zone above an impervious layer receiving uniform recharge from an infiltrometer. In this paper analytical solutions of equations describing the growth of a groundwater mound above an impervious layer are compared with laboratory and field experiments to demonstrate the applicability of the method to the determination of saturated hydraulic conductivity of soils in shallow subsurface.

Numerical investigations of flow over side weirs in rectangular and circular channels are reported. Numerical models are obtained from energy principles and solved by a finite difference method. The results are presented in graphic form. Theoretical and experimental equations are presented for the determination of the length of side weirs for rectangular and circular delivery channels. Comparative analyses of side‐weir lengths computed for several example cases show that rectangular‐channel side‐weir procedures generally should not be used for design purposes or for obtaining approximate water‐surface profiles along side weirs in a circular channel (except for the special case of a side‐weir‐height‐to‐channel‐diameter ratio of $P/D=0.66$). For all other ratios of $P/D\text{,}$ the percent error incurred depends on the basic assumptions involved with the rectangular approximation procedure.

Four two‐stage pipe expansions were designed, constructed, and tested to estimate their annual operating cost in irrigation service. The hydraulic performances of these fittings were compared with past two‐stage expansion research. The fittings were compared to other commonly used pipe expansions to determine their economic competitiveness. Measured head‐loss coefficients for the two‐stage pipe expansions averaged 0.32 and 0.28 for the 50–100 mm and 75–150 mm expansions, respectively. Loss coefficients for the 100–200 mm and 125–250 mm averaged 0.17 and 0.16, respectively. The smaller two fittings performed worse than past research on two‐stage expansions would predict. The two larger fittings (100–200 mm and 125–250 mm) worked substantially better than predicted. The two‐stage fittings presented in this paper are practical alternatives to other pipe expansions under the pumping conditions typically seen in irrigation service. The two‐stage pipe expansions designed with commonly available parts were more efficient than comparable abrupt and single‐stage pipe expansions, and cost about the same.

Evapotranspiration (ET) estimation has not been made for certain parts of the world, partially because some basic climatological data are not available. One alternative is to adopt the available ET data directly from other areas where crop and climatic conditions are similar. Another is to use synthetic data to estimate the ET using the conventional methods of Penman (PEMM), the original Blaney‐Criddle (BCOF), Blaney‐Criddle modified by Shih (BCMS), Blaney‐Criddle modified by the United Nations' Food and Agriculture Organization (FAO) (BCMF), or Thornthwaite (THNM). The Everglades Agricultural Area and Puerto Rico were used to exemplify the application. The results showed that both alternatives are feasible to estimate the ET for an area that is short on data. If the generated climatic data are still insufficient to implement the PEMM in Puerto Rico, the preferred conventional methods are, in descending order, BCMS, BCMF, BCOF, and THNM.

The effects of controlled‐drainage‐subirrigation (CD‐SI) systems on subsurface and outflow water quality must be evaluated before these systems can be promoted for use in the Georgia flatwoods region of the eastern Coastal Plain. In this paper, Nitrate‐nitrogen (N), ammonium‐N, and orthophosphorous (P) concentrations are measured from shallow subsurface wells and outlets of a CD‐SI system in the Georgia flatwoods for comparison to adjacent forested and grassland areas (the most common current land uses). Nitrate‐N concentrations did not exceed 10 mg/L from any of the outflow samples although almost 6% of the well samples taken from the CD‐SI area had concentrations exceeding 10 mg/L, Nitrate‐N concentrations in most of the wells in the CD‐SI area were significantly higher than in the forested and cleared areas. Ammonium‐N and ortho‐P concentrations were relatively low and do not appear to be a problem based on this limited data. No significant differences in ammonium‐N and ortho‐P concentrations were shown between any of the sites. Ammonium‐N concentrations did exceed 2 mg/L from at least one sample of the scrub forest, one of the CD‐SI blueberry sites, an open ditch (1.8 m depth), and one of the cleared unirrigated sites. Ortho‐P concentrations ranged from 0 to 1.94 mg/L, with average concentrations from all sites near 0.1 mg/L.

A kinematic-wave model was developed to solve for time to advance to specified locations along the furrow where infiltration was measured as a function of ponding time and wetted perimeter. Time step increased during advance to take advantage of the more slowly changing flow profile. In an earlier model developed by Rayej and Wallender, the system of nonlinear flow equations was solved implicitly for flow area and time step using the double-sweep technique. Coding requirements were extensive and computation time was protracted, however. Performance of the time-iteration model was compared with the double-sweep model. Advance trajectories for both uniform and nonuniform soils are essentially the same for both time-iteration and double-sweep methods. For both models, advance time decreases as space step decreases for both soil types. Computation time for the time-iteration model is one-third and one-fourth of the computation time for the double-sweep model for uniform and nonuniform soil, respectively. As expected, the number of nodes decreases as the space step increases and computation time decreases.