Journal of Irrigation and Drainage Engineering

Of the 15,000,000 ha (37,000,000 acres) of irrigated land in the U.S., 21% is affected by soil erosion to some extent. Irrigation‐induced soil erosion has been studied, primarily in the Northwestern United States, since 1940. A number of studies have measured annual sediment yields from furrow‐irrigated fields exceeding 20 t/ha (9 tons/acre) with some fields exceeding 100 t/ha (45 tons/acre). Under the center‐pivot sprinkler method, sediment yields as high as 33 t/ha (15 tons/acre) have been measured. Annual sediment yields as high as 4.5 t/ha (2 tons/ acre) were measured from irrigation tracts. Erosion is seldom excessive on slopes less than 1% and is often excessive on slopes greater than 2%. Erosion reduces the agricultural productivity of the fields and causes off‐farm damages. In southern Idaho, crop yield potential has been reduced by 25% due to 80 years of irrigation‐induced erosion. Some irrigation districts spend more than $50,000 annually to remove sediment from drains. Sediment in irrigation return flows causes major water‐quality degredation problems in several rivers in the Western United States.

Soil erosion is sometimes excessive during furrow irrigation and under center pivot sprinkler systems. An understanding of erosion processes is required to predict and develop management practices to reduce irrigation induced erosion. Little erosion process research has been carried out under irrigation, but much of the extensive channel sediment transport and rainfall‐induced erosion process research can be adapted to irrigated conditions. Soil erosion occurs when fluid in motion detaches and transports soil particles. Sedimentation occurs when the fluid transport capacity decreases to less than the sediment load. Hydraulic forces of moving water and soil factors such as aggregate stability and particle size determine erosion and sedimentation. Under furrow irrigation, the shear of the overland flow against the soil provides the detachment force and is a primary factor determining channel transport capacity. With sprinkler irrigation, water drop energy detaches particles, some of which may be transported downslope by shallow interrill flow if the water application rate exceeds the soil infiltration rate.

Recent research efforts have shown that soil erosion decreases soil productivity. Erosion‐caused crop production decreases of 15–40% are commonly reported with some values over 50%. Furrow erosion on irrigated land in Idaho decreases topsoil depth on the upslope approximately 33% of the field area and may increase topsoil depth on the downslope 50–55%. Crop yields are generally decreased where topsoil depths are decreased, but yields are not generally increased where topsoil depths are increased beyond a critical depth. Crops vary in their sensitivity to decreases in topsoil depth, but all crops studied exhibited lower yields on the eroded areas. Soil productivity potential of one area representing several million ha of furrow irrigated land was reduced at least 25% by furrow erosion over 80 irrigation seasons. Technology is not available to restore soil productivity potential to the level that would exist had there been no erosion except for returning topsoil to eroded areas. Research and technology applications are needed to reduce or eliminate topsoil loss and redistribution by irrigation erosion.

Irrigation‐induced erosion and subsequent sediment loss is a serious agricultural and environmental problem. Recent recognition of this problem has stimulated the development and evaluation of erosion and sediment‐loss‐control technology. Research results indicate that the application of the technology available today can reduce sediment loss by 70–100%. Important practices include irrigation‐water management, sediment‐retention basins, buried‐pipe tailwater‐control systems, vegetative filter strips, tailwater‐recovery systems, keeping crop residues on the soil surface and in furrows, and implementing conservation tillage practices.

A new computational procedure is presented that addresses the complex phenomena found in declining‐flow regimes and does not require that the boundary conditions of the declining‐flow segment be defined. These conditions are determined as a part of the solution of the larger problem of determining the flows in the entire conveyance system. The procedure utilizes the principle of conservation of momentum in the declining‐flow segment, coupled with standard water‐profile procedures in the constant‐flow segments. The algorithm determines whether the flow is supercritical, subcritical, or a combination of both. If a hydraulic jump occurs, its location is found, even if it lies within the declining‐flow segment. The procedure is applied specifically to the side‐flow weir, and theoretical computations are compared to experimental values that have been published by others. Excellent correspondence exists between the theoretical and experimental results. Examples of the results for a variety of conditions are presented, including the circumstance in which a hydraulic jump occurs in the weir reach.

A simple cost‐efficient model is developed to predict advance phase in border irrigation. Assuming a power function for the water‐depth profile, a volume‐balance equation is applied using an adjusted surface shape factor. In addition, the zero‐inertia motion equation is evaluated at the upstream end of the border. A system of two equations containing two unknowns (the advance position and the water depth at the head of the border) is thus constructed and solved at each time step. Comparison of dimensionless advance trajectory and surface‐profile solutions indicate that predictions of the present method are in agreement with zero‐inertia solutions. The proposed model is also tested for four well‐documented field irrigations. It is shown that even for level or small slope borders for which the traditional volume‐balance models give poor results, the proposed method is accurate. Programming requirements and computation cost of the present method are significantly less than for sophisticated models.

Single sink depths providing maximum ground‐water table decrease during a fixed time interval within a selected area are found for the 2‐D and 3‐D cases. The curve of the maximal phreatic surface position (underflooding curve) in the aquifer from flood induced variation in water level of the ground‐water reservoir is calculated. Well‐known analytical solutions based on nonlinear and linear potential theories and the Dupuit‐Forchheimer approximation are applied to calculate the objective function, decision variables, and boundary of the fully saturated zone. In the linear case, an explicit analytic solution gives the unique maximum of the water table decrease at the compliance point for a given pumping duration. For small values of sink depth, the linear approach is invalid. In the nonlinear case, complex analysis and series expansions are used. For small values of drain depth, the series technique becomes untenable. For the reservoir‐aquifer problem the spreading phreatic surface is a rotating straight line and the underflooding curve is a parabola.

A side sluice gate is a flow‐regulation device widely used in irrigation works for diverting the flow from a main channel to a secondary channel. The discharge of a side sluice gate may be obtained through the concept of an elementary discharge coefficient for an elementary strip along the gate length. Similar to the case of a normal sluice gate, the elementary discharge coefficient for a side sluice gate has been found to be a function of channel flow depth to the gate opening ratio for free‐flow conditions. It depends on an additional parameter, namely, the ratio of the crest width to the gate opening for submerged‐flow conditions. For a broad‐crested side sluice gate, the discharge coefficient involves still one more parameter: the ratio of the crest width to the gate opening.

For design problems where annual‐event analyses are inappropriate, such as in the case of channel scour due to floods above a threshold, multiple‐event analyses with partial duration series should be used. A procedure for calibrating a partial‐duration, balanced‐hydrograph model for regional design work is presented. The model can be used either as a series of synthetic storm events or as part of a simulation model. Data from Los Angeles were analyzed to develop a model for ungaged urban watersheds in southern California.

The benefits of water‐table management for increasing crop yields, and reducing nitrate pollution were assessed for the St. Lawrence lowlands region of eastern Canada. Soybeans [Glycine max (L.) Merr. cvs. Apache and KG 30] were grown in 0.4‐m I.D. field lysimeters in relatively wet (1990) and dry (1991) years, and subjected to controlled water tables of 0.4, 0.6, 0.8, and 1.0 m. The latter was regarded as conventional, free‐outlet drainage. Plant parameters, including yield, as well as nitrate, ammonium, and moisture distributions in the soil profile (0.3 m and 0.7 m) were measured. Seed and pod number per plant, and percent nitrogen in the seed were approximately 20% higher for the 0.4‐, 0.6‐, and 0.8‐m water‐table treatments, compared to the conventional drainage treatment. No significant differences were observed between the three shallowest water‐table depths. Soil nitrate levels were higher in the conventional drainage treatment, compared to the controlled water tables, particularly at the 0.7‐m sampling depth. There were no significant differences in nitrate between the three shallower water‐table depths. Soil ammonium levels were lowest under conventional drainage and increased throughout the growing season, particularly at the 0.7‐m sampling depth. Due to lack of significant differences in yield and soil nitrate for the 0.4‐, 0.6‐, and 0.8‐m water tables, it is recommended that water tables be maintained at 0.6 to 0.8 m. This practice will result in reduced nitrate contamination of water resources and maximum soybean yield.

Maximum annual 24‐hour rainfall maps for return periods of 2, 5, 10, 25, 50, and 100 years were developed by using hourly precipitation data. Data from 92 rain gauges for the period of 1948 to 1987 were compiled and combined into 26 synthesized stations. Log‐Pearson Type 3 distribution (LPEAR3) with parameter estimation by the method of moments (MOM) was used to compute the 24‐hour rainfall quantiles. This combination was found to be the best for the Louisiana rainfall data. The new isohyetal maps were compared to the U.S. Weather Bureau Technical Paper 40 (TP‐40) maps using the performance indices of the standardized mean square error (MSE) and the standardized bias. On the average, for all of the 26 synthesized stations corresponding to six return periods, the new maps reduced the MSE by 58% and the bias by 80%, as compared to the TP‐40 maps. Application of the predicted quantiles to the Soil Conservation Service (SCS) runoff model showed that the computed runoff was equally sensitive to both parameters, rainfall input, and the curve number, of the SCS rainfall‐runoff model.

A study was conducted in the Hydraulics Laboratory of Colorado State University to evaluate how settlement affects the accuracy of 30.5‐ and 61.0‐cm (1‐ and 2‐ft) prefabricated metal Parshall flumes. The flumes were installed and tested for lateral, longitudinal, and combined lateral‐longitudinal slopes settlements. Flume slopes varied ±7%, and were conducted for free outfall conditions. The results indicated that the Parshall flume accuracy is slope‐dependent. The flume rating was in error approximately 28% for longitudinal slopes of ±5%, and approximately 10% for lateral settlement of ±5%. Flow rates used in this study ranged from 0.01 to 0.20 m³/s for the 30.5‐cm flume and 0.03–0.28 m³/s for the 61.0‐cm flume. A procedure was presented to adjust the flow rating for lateral, longitudinal, and combined lateral‐longitudinal settlement as a function of the flume slope, flow depth, and flume throat width. The results of the 30.5‐ and 61.0‐cm flumes were compared to the results of similar tests with 2.54‐, 5.08‐, and 7.62‐cm Parshall flumes.

Alfalfa yields decrease when water supply is inadequate. In drought years, starting and continuing with full irrigation until the water supply has been exhausted will limit full production to earlier harvests. The objective of this study was to examine the effect of discontinued irrigation on alfalfa yield during the irrigation season and when water supply is restored in the following year. Three irrigation levels, water for two harvests (i1), water for 3 harvests (i2), and water for four harvests (i3), were imposed in 1981–83 and one irrigation level (i3) in 1984. Yield response was compared among il–i3 by harvest for 1981–83 and separately for 1984, between flood and sprinkler irrigation, and between drainage and no drainage. Generally, yields were significantly less for nonirrigation conditions, they fully recovered in 1984 after three years of deficit irrigation, yield of harvests 1 did not differ between 1984 and 1982–83, drainage condition did not affect yield, and irrigation method had a significant influence on yield for harvests 3 and 4 in 1981–83.

The optimal management of a cutback‐furrow‐irrigation system with spatially variable infiltration based on an average intake function was analyzed. The problem was formulated as a cost‐minimization function subject to meeting a specified fraction of the irrigation requirement. Optimal solutions were examined in the context of developing a real‐time control system for furrow irrigation. Although total infiltration was adequately predicted with the average function, final water distribution was not. Consequently, the optimal policies resulted in actual requirement efficiencies less than the target value. Nonetheless, relative changes in performance as a function of the constraint were well predicted. The performance index was relatively insensitive near the optimum, and cutback time had the least impact on application efficiency and uniformity. Satisfactory performance was therefore still obtained by reducing the inflow after the final advance time. Similar values of application efficiency were generally computed with decreasing application depths, but smaller efficiency resulted when the optimized cutoff time was less than the final advance time. There were small performance differences between discrete and continuous‐time cutback functions.