Journal of Irrigation and Drainage Engineering

As water is becoming a scarcer commodity, savings in the irrigation sector could enhance water development in areas currently not being irrigated, and arrest the rapid environmental degradation due to waterlogging in arid areas. The agro-hydrological model SWAP is used to investigate possible water reductions for wheat and cotton crops under shallow water table conditions prevailing in the Fourth Drainage Project in Punjab, Pakistan. The simulations are performed for both drained and undrained conditions considering three different irrigation water qualities. The overall objective is to save good-quality irrigation water. The results indicate that when good-quality canal water is available, a reduced application to wheat (195 mm) and cotton (260 mm) will keep the soil healthier under both drained and undrained conditions. For poor-quality irrigation waters (mixed canal and tubewell or tubewell alone), this water conservation strategy will be insufficient. Therefore, more water (325 mm for wheat and 325 mm for cotton) should be applied to keep crop production and soil salinity within desirable limits. However, this will only be applicable to the areas where proper subsurface drainage systems are present. For undrained areas, this strategy will not be feasible due to rising water tables; other options like growing more salt-tolerant crops should be considered. Drainage cannot solve salinity buildup problems under all circumstances because relatively dry monsoons provide insufficient leaching water, and salts added by tubewell irrigation can only be evacuated from the soil profile if the drainage system is very intense. Reduced irrigation inputs is a proper short-term solution, although wheat production tends to decline in all areas without drainage, even when irrigated with canal water. Large-scale drainage investments associated with adjusted irrigation planning seem unavoidable in the long run.

Accurate estimates of crop evapotranspiration ET_c, that quantify the total water used by a crop, are needed to optimize irrigation scheduling for horticultural crops and to minimize water degradation. During early growth, accurate assessments of ET_c are difficult in vegetable crops because of high soil evaporation due to frequent irrigation. A model to estimate ET_c for vegetable crops, using only daily reference evapotranspiration data as an input parameter, was developed. It calculates crop transpiration and soil evaporation based on ground cover and daily radiation intercepted by the canopy. The model uses a two-stage soil evaporation method adapted to conditions of variable reference evapotranspiration. The model was evaluated against data using measurements from two seasons of lettuce crop, two tomato fields in the same season, and one season of broccoli crop production. Using all of the crop data, the root-mean-square error for measured versus modeled daily ET_c was 0.72 mm day⁻¹, indicating that the model works well.

The flow behavior along an open-channel irrigation network can be assessed by determining the sensitivity of the irrigation structures. The importance of governing factors of the sensitivity and proportionality indicators is analyzed using a theoretical approach as well as practical results from historical data on a gated system in Sri Lanka and a gated and structured system in Pakistan. The influence of the discharge on the sensitivity indicator varies with the type of flow (overshot-undershot). For a common offtake (e.g., an orifice followed by a crest), sensitivity is at its peak at maximum discharge. The available head through the structure appears to be the most influential factor for sensitivity; the greater the head, the lower the sensitivity is. The submergence of the flow downstream of the structure is a reducing factor of the sensitivity that is important to consider for sensitive structures. Different methods for assessing the governing factors of sensitivity are discussed and illustrated using case studies. It is recommended that the sensitivity indicator be determined for the full supply depth in the parent canal and the maximum discharge through the offtake.

Numerical flow models usually represent drains as a system dependent boundary condition. If soil is saturated, drains act as the Dirichlet boundary condition with pressure head set equal to zero, and if soil is unsaturated, drains act as the Neumann boundary condition with flow set equal to zero. The underlying assumption is that drains exhibit ideal behavior. In reality, however, this is generally not so, and the flow encounters additional resistances due to pipe slotting and clogging of the envelope material around the drains. To account for the resulting resistance, a Hooghoudt-type boundary condition was developed that prescribes drain flow in relation to the groundwater level at a reference position. The measured drain discharge in an old drainage system was compared with calculated discharge assuming an ideal drain. It was found that the ideal drain assumption led to large errors in simulated discharge. With a correctly formulated and calibrated Hooghoudt boundary condition, however, more accurate drain discharges were obtained.

The WAVE model describes the transport and transformations of matter and energy in the soil, crop, and vadose environment. A lateral field drainage subprogram was added to the WAVE model to simulate lateral subsurface drainage flow. The subsurface drainage is considered as the drainage provided by evenly spaced parallel drains with a free outlet: drain tubing or ditch. The rate of subsurface water movement into drain tubes or ditches depends on the hydraulic conductivity of the soil, drain or ditch spacing, hydraulic head in the drains, profile depth, and water table elevation. Hooghoudt's steady-state equation was selected for incorporation in the WAVE model. The subsurface drainage subprogram was calibrated and validated by comparison with the SWAP model (The Netherlands) and DRAINMOD (the United States) and partially by using 7 years of drain outflow data from an experimental field under fallow and cropped conditions. The comparative study revealed that the three models performed equally well and that the models were reliable and accurate tools for predicting the drainage flux as a function of rainfall-evapotranspiration and local conditions. The WAVE model, in comparison to the SWAP and DRAINMOD model, provided as good a prediction of the lateral subsurface drainage flow to drains. The statistical analysis between each model and observed data revealed that the three models were able to predict with sufficient accuracy the observed drainage discharge. The DRAINMOD model, however, has the advantage of giving a more accurate estimate of the discharge, resulting in a more precise modeling. The models were consistent in predicting water table levels, but they could not be verified against field data because of a lack of suitable measurements.

A new solution to the multidimensional linearized Richards equation was derived using a Fourier integral transform. Exponential functional forms k = k_se^αψ and θ = θ_r + (θ_s − θ_r)e^αψ were used to represent the hydraulic conductivity and pressure relation and the soil water release curve. The analytical solutions consider the conditions of time dependence and nonuniform distribution of rainfall intensity and arbitrary initial water content distribution with a water table. The analytical solutions can be used to predict the ponding time and to obtain the volumetric water content distribution over time and space.

To describe falling water tables between two drains lying on a horizontal/sloping impermeable barrier, analytical solutions of the Boussinesq equation linearized by Baumann's and Werner's methods and numerical solutions of the nonlinear form of the Boussinesq equation using finite-difference and finite-element methods were obtained. A hybrid finite analytic method, in which the nonlinear Boussinesq equation was locally linearized and solved analytically after approximating the unsteady term by a simple finite-difference formula to approximately preserve the overall nonlinear effect by the assembly of locally analytic solutions, was also used to obtain a solution of the Boussinesq equation. Midpoints of falling water tables between two drains in a horizontal/sloping aquifer as obtained from various solutions were compared with already existing experimental values. Euclidean L2 and Tchebycheff L_∞ norms were used to rank the performance of various solutions with respect to experimental data. It was observed that the performance of the hybrid finite analytic solution is the best, followed by finite element, finite difference, analytical with Werner's linearization method, and analytical with Baumann's linearization method, respectively.

Normal depth is an important parameter for the design of irrigation canals. Direct analytical solution of normal depth is not possible because of the implicit nature of governing equations. The solution usually requires tedious techniques of trial and error. Tabular and graphical methods are only available for more common cross sections. This problem has been previously addressed for some cross sections (i.e., rectangular, triangular, trapezoidal, and circular) and the calculation of normal depth in these cross sections has been reported in the literature. Reported herein are graphical solutions for normal depth in other cross sections (i.e., round-bottomed triangular, parabolic, and round-corner rectangular).

A mathematical model is developed to determine the optimal spacing of shallow tube wells and the optimal length of field channel lining to minimize the total cost. The model takes into account the monthly varying crop water requirements. The optimization model is also applied to a set of basic data pertaining to North Bihar, India.