Journal of Irrigation and Drainage Engineering

Roughness and infiltration parameters in border irrigation are evaluated with reasonable accuracy and simplicity by iterative analysis of advance‐trajectory data and surface‐depth measurements at a station at or near the inlet end of a border. Approximate estimates of the parameters obtained by simple algebraic equations are successively corrected with the aid of an exact solution (zero‐inertia simulation). Numerical experiments are performed to gage the limitations of the method, particularly as regards necessary uniqueness. Results indicate that nonuniqueness is not a problem for the present method. Rate of convergence of the method is significantly greater than that of other methods, based on gradient optimization techniques. Thus, even expensive‐to‐run, but highly accurate, surface irrigation models (i.e. full‐dynamic models) can be used to solve the direct physical problem. The method is also applied to a field border‐irrigation test.

Natural rainfall rate patterns are highly variable. Existing Hortonian infiltration models are not designed specifically to work with such rainfall rate patterns. Thus, for practical applications the models have to be modified to work with the naturally occurring variable‐pattern hyetographs. In this paper, a Hortonian infiltration model, developed specifically to work with the variable rainfall‐rate patterns and constant time intervals of naturally occurring hyetographs, is described. The structure of the model is hypothesized based on a physical appreciation of the Hortonian infiltration process. It is then refined through repeated testing to simulate the overland flow data associated with the process. A numerical infiltration model has been used to simulate the overland‐flow data used in the testing. A three‐parameter conceptual model was able to reasonably simulate the numerically simulated overland‐flow data due to 10 observed hyetographs falling on three different topsoils. The most important parameter in the model is the saturated hydraulic conductivity $\left(\mathrm{KS}\right)$ of a topsoil.

Most models for subsurface drainage in flat lands are based on the Boussinesq equation. This paper compares a semianalytical model called SIDRA, which assumes a constant water‐table shape during the period of water table recession, to a model called SLOP, which is based on a numerical solution to the Boussinesq equation. Solutions were obtained for both steady‐state and transient drainage events on flat lands and for drainage under steady‐state recharge on sloping lands. The two models were in excellent agreement for steady‐state and transient drainage on flat lands. Differences in the shape factors calculated from the two models were always less than 4%. The implicit solutions in SIDRA were also in good agreement with SLOP for sloping lands. The numerical solutions in SLOP were reliable for low slopes but were unstable for conditions in which the streamlines cannot be considered horizontal.

Presented are: (1) A reanalysis of Parshall's original experimental data obtained for the improved Venturi (Parshall) flume; and (2) a reanalysis of data obtained by the writer during model tests of the Parshall flumes at the rate measuring stations installed by the U.S. Department of Agriculture's Soil Conservation Service (SCS) on the Waco, Tex., Blacklands experimental watershed. Parshall's and the Waco data were originally analyzed graphically to develop rating equations. The reanalysis used the least‐squares method. The reanalysis shows that in the head‐discharge equation the Waco coefficient and exponent do not vary with flume size, whereas Parshall's exponent varies with flume size for the 0.30 m (1 ft) to 2.44 m (8 ft) flumes and the coefficient varies with flume size for the 3.05 m (10 ft) and larger flumes. The quality of both sets of data is excellent: no explanation has been found for the albeit small differences. The reanalysis lends confidence that the equations presented by Parshall predict the discharge to within the 5% accuracy claimed by Parshall. Similar accuracy is obtained using the writer's equations. The Waco data show that the entrance wingwalls should be curved and that the head‐discharge relationship is valid for heads as low as 18 mm (0.06 ft).

The Hall technique for computing stream advance down an irrigated border strip is shown to be inconsistent in the selection of numerical parameters for approximating infiltrated volume. This inconsistency leads to violation of mass conservation and to error in computed advance. By viewing it in dimensionless terms, the behavior of this error is demonstrated over the practical range of variation of border‐irrigation parameters with a Kostiakov infiltration formula and Manning roughness formula. It is shown to vary from insignificant magnitudes under 1% to well over 10%. A correction term to neutralize the inconsistency is presented. An alternate formulation free of the inconsistency and suitable for modern high‐speed computers working with a reasonably large number of significant figures is presented. Errors stemming from the basic physical assumption—constant average depth of surface water—are not addressed.

Drainage reduction through improved irrigation is needed for addressing the problem of drainage water disposal in the San Joaquin Valley in California. Options for improving irrigation include improved management of existing systems (irrigation scheduling, duration of water applications), upgrading traditional surface irrigation systems (reduced field length, increased unit flow rate, surge irrigation, furrow compaction, tailwater recovery), and converting to pressurized irrigation systems (hand‐move and linear‐move sprinklers, low‐energy precision application (LEPA) machines and rip irrigation). The most effective upgrade of surface irrigation systems for reducing subsurface drainage is a reduced field length coupled with reduced irrigation times. Increased furrow flow rates resulted in little change in drainage in some cases. Surge irrigation offers an opportunity of reducing subsurface drainage by only about 1/3. Converting to pressurized irrigation methods can substantially reduce subsurface drainage, but may be uneconomical in some cases. Analyses of large‐scale field comparisons of irrigation methods revealed that generalizing about the best irrigation method is difficult. Economic analyses of these comparisons showed a well‐managed furrow system to be more profitable than a subsurface drip system in one case, but a subsurface drip system to be more profitable compared to a marginally managed furrow system in another case. The analyses also revealed that disposal costs of subsurface drainage water may need to be much higher than projected costs to economically justify converting from furrow irrigation to irrigation systems with high capital costs.

Thousands of free‐flowing wells, drilled in the 1940s and 50s to supply Florida's agricultural needs, have become sources of saline intrusion into freshwater aquifers and sources of salt‐affected soils throughout the state. The ground‐penetrating radar (GPR) with a 120‐MHz antenna was used to assess the salt‐affected area. The study was conducted at an abandoned tomato field in Lee County, Florida. The electrical conductivity (EC) value of 2 dS m⁻¹, previously selected as the lower limit for saline soil, was also found to be a threshold value for GPR application to differentiate the salt‐affected soils. The GPR technique appears to be a useful tool for assessing the salt‐affected area within a field.

Little has been published on the effectiveness of various management practices in small rural lakes and streams at the watershed scale. In this study, statistical techniques were used to test for changes in water‐quality data from watersheds where best management practices (BMPs) were implemented. Reductions in data variability due to climate and seasonality were accomplished through the use of regression methods. This study discusses the merits of using storm‐mass‐transport data as a means of improving the ability to detect BMP effects on stream‐water quality. Statistical techniques were applied to suspended‐sediment records from three rural watersheds in Illinois for the period 1981–84. None of the techniques identified changes in suspended sediment, primarily because of the small degree of BMP implementation and because of potential errors introduced through the estimation of storm‐mass transport. A Monte Carlo sensitivity analysis was used to determine the level of discrete change that could be detected for each watershed. In all cases, the use of regressions improved the ability to detect trends.

An inspection of the original study that developed the rating equations for Parshall flumes indicates that the calibration tests were not conducted for discharges as low as is currently recommended for the flumes' use. A numerical model was developed to predict the effect of fluid viscosity on the depth‐discharge relation. An experimental investigation for several flume sizes indicates that the original rating equations and data overpredict the discharge at flow rates that are less than about 15% of the maximum rated discharge for the flume. The discrepancy can be as much as 25% for the range of flows for which the flume is recommended to be used. The numerical model successfully matches the experimental data for the flume sizes studied. Alternative rating equations are developed for use at low discharges.

The cultivation of eucalyptus trees is being recommended to lower the shallow ground‐water level and to reuse saline‐subsurface drainage waters from tile‐drained cropland in drainage‐impacted areas in California's San Joaquin Valley. To assess the efficacy of trees in using these saline‐problem waters, both field and modeling studies are being carried out. The present paper presents the theory and validation of a dynamic 2D numerical model (D‐HYSAM) to simulate root‐water extraction in a variably saturated soil. The Galerkin‐type finite‐element method is used to solve the governing nonlinear‐water‐flow and salt‐transport equations using three approaches for root‐uptake sink mechanisms. Field data obtained from the 1989 growing season in an irrigated, tile‐drained eucalyptus plantation was used for model validation. The simulated patterns on field‐soil water pressure, volumetric soil moisture, and salinity suggest that the macro root‐extraction approach for the sink term gives the best fit.

Part I of this paper presented a two‐dimensional dynamic finite‐element model of fluid flow and solute transport in root‐soil profiles with different root‐uptake mechanisms. Field data collected in 1989 were used to validate the model. The present paper presents further validation and application of the numerical solution of Richard's equation against the field data collected from an agroforestry demonstration site for sustainable agriculture during the 1990 irrigation season. The site is tile‐drained, isolated from the surroundings by interceptor drains, planted with six‐year‐old Eucalyptus camaldulensis (red gum) over an area of shallow ground water, and irrigated by a furrow system. Computer simulations were made for 220 dayas to evaluate the irrigation management. In comparing the measured with the numerically predicted volumetric moisture content in the soil profile and pressure head, the values were well correlated in space and time $(r=0.82-0.94)$. Therefore, the model can be used to provide management guidelines for subsurface drainage in agroforestry systems.

Eucalyptus trees are being recommended as a management option for lowering the shallow ground water and reusing saline subsurface drainage water in drainage‐impacted areas of the San Joaquin Valley in California. Both field and modeling studies are being carried out. To assess the efficacy of eucalyptus trees for these purposes, a finite‐element method based on the Galerkin technique was used to solve the nonconservative transport governing equation. The nonconservative nature of salinity was addressed by the dissolution and precipitation of gypsum in the vadose zone. Water fluxes and volumetric moisture content needed to solve the solute model were obtained from outputs of the water‐flow submodel. The present paper shows the mathematical formulation of the 2D salinity submodel as well as an example of an application during the 1990 irrigation season to an agroforestry system at the Mendota site. The measured and predicted salinity distribution was found to have good linear correlation coefficients (an average of 0.856). The sensitivity analysis of different irrigation practices on salinity distribution and shallow ground water quality are also presented.

A three‐dimensional Green‐Ampt analysis is developed, and the infiltration capacity curve for a three‐dimensional infiltration model is presented. The infiltration‐capacity curve represents the time distribution of wetting‐pattern volume of a water source with unlimited inflow supply. A significant feature of the infiltration‐capacity curve is that the infiltration‐capacity rate increases with time during extended infiltration periods. This feature is different from the one‐dimensional infiltration model where the infiltration‐capacity rate decreases with time. The infiltration‐capacity curve is applied to describe the wetting pattern of a surface emitter with a constant discharge by matching the emitter discharge with the average infiltration‐capacity rate. Algebraic solutions of the wetted radius and the maximum wetting‐pattern depth are derived. Graphical solutions of the infiltration‐capacity curve are presented. A numerical example is prepared to illustrate the procedure for estimating time of application, wetting pattern volume, emitter discharge, and irrigation period of a surface emitter in a drip‐irrigation system.

Knowledge of seasonal distributions of peak runoff rates is needed for use with watershed models, and for the development of new techniques of estimating peak flow rates and their frequencies of occurrence from small agricultural watersheds. The peak flows in 312 station years of data from seven small agricultural watersheds in Ohio were analyzed to determine the seasonal and land‐use distributions of peak flows and their return periods. The assumptions often made when using design storms that peak flows occur during the seasons of highest rainfall intensities was largely confirmed, but there were exceptions. Between 92 and 94% of the peak flows, regardless of land‐use and soil type, with return periods greater than two years, occurred during the seasons of highest rainfall intensities. A similar seasonal distribution of peak flows was found from an analysis of specific agricultural land uses. More peak flows occurred during the winter season for flows with return periods smaller than about five years for some land uses. Seasonal distributions of peak flows greater than five years are largely independent of soil type and land use, allowing extrapolation of results beyond the study area in Ohio.

Small ponds created by crop residue serve to store water on upland areas. The present study is conducted to identify the maximum surface storage provided by crop residue. Equations for estimating surface storage are derived assuming that residue elements were oriented end to end, at uniform intervals, in a direction perpendicular to flow. Separate equations are developed for conditions where average slope was greater than or equal to residue cover, or less than residue cover. Both equations relate average surface‐storage depth to residue cover, residue diameter, and average slope. Data to test the reliability of the equations are obtained in a laboratory investigation. Artificial residue elements are glued onto a 1‐m² test section positioned at slopes of 1%, 10%, or 20%. Four sizes of residue elements and three surface‐cover conditions are examined. Surface‐storage depth for each experimental condition is measured. Close agreement is found between predicted and measured surface‐storage values. Surface storage occurring under field conditions may be substantially less than the estimates obtained using the predictive equations.

The present study estimated evaporation $E$ and transpiration $T$ losses separately during various growth stages of a tomato crop (Lycopersicon escallen mill. UC 82 variety), in Davis, California. Micro‐Bowen‐ratio systems were used during the summer of 1984 to determine net radiation and temperature, and humidity gradients close to the soil surface. Transpiration was determined by subtracting $E$ from crop evapotranspiration $\mathrm{ET}$ as measured by a large weighing lysimeter. Results show considerable effects of soil moisture on evaporation and above‐soil net radiation during early growth stages, but the effects diminished greatly as crop cover approached a maximum. With crop cover less than 50%, soil‐surface evaporation one day after, irrigation was much higher than for two or three days later. Increased transpiration occurred as the soil surface dried, apparently in response to decreased evaporation. Irrigation timing and crop cover substantially affected partitioning of $\mathrm{ET}$. At 34% crop cover, $\mathrm{ET}$ was evenly split into $E$ and $T$. Later, at 43% and 85% cover, calculated $E/\mathrm{ET}$ values were 0.28 and 0.05, respectively.