5.5 Border irrigation design

With two exceptions, the design of borders involves the same procedure as that for furrow systems. The first difference is that while the depletion and recession phases are generally neglected in furrow design, both phases must be included for borders. The second difference is that the downstream end of a border may be dyked to prevent runoff. One simplification of border analyses is that the geometry of the flow is simpler because it can be treated as wide, plane flow. The values of p₁ and p₂ are always 1.0 and 1.67, respectively.

5.5.1 Design of open-end border systems

The border designs given here assume the advance phase is completed before the inflow is terminated. Many irrigators, in fact nearly all where the downstream end is dyked, actually cut off the inflow before the end of the advance phase. In these cases, the volume of water on the surface will continue to advance along the border until it reaches the lower end where it will run off or pond in front of the dyke. Unless the border system is extremely well designed and operated, the downstream pond often creates a substantial threat to the crop in the submerged areas and although dyked at their lower ends, most farmers provide a surface drain for excess water. Consequently, the border efficiency and uniformity are approximately the same as borders in which excess surface water simply drains off the field after the advance phase is complete. The following procedure is therefore suggested for border systems where the excess surface water is drained from the field either by a completely open-ended border or by a regulated outlet from a blocked-end border.

5.5.2 Design of blocked-end borders

The computations needed to evaluate and design blocked-end borders where the flow is cut off before or shortly after the advance phase is complete are substantially more detailed than the procedures outlined above for furrow and open-end border irrigation systems. In fact, the volume balance methods given previously are relatively weak for this particular case of surface irrigation. Generally, the computations for blocked-end borders are best performed with zero-inertia or full hydrodynamic simulation models which are beyond the scope of this paper.

A number of studies have been made to develop relationships among the most important variables involving border irrigation using a dimensionless approach and the higher level simulation models. The interested reader might want to refer to Strelkoff and Katapodes , Strelkoff and Shatanawi , Shatanawi and Strelkoff , and Yitayew and Fangmeier for some of these reports.

The design procedure outlined below is an extension of the approaches already given and consistent with the level of treatment given herein. The procedure given here is intended to be conservative and will yield designs capable of performing at somewhat lower application efficiencies than is perhaps possible using the more comprehensive methods.

5.5.4 A blocked-end border design example

The problem. Section 8.5.4 illustrated the open-end border design procedure. The option of dyking these borders should be considered as an option for improving application efficiency. From results already available, the required intake opportunity times, r_req, needed to apply a depth of 8 cm (Z_req) were about 389 minutes and 679 minutes for initial and subsequent field conditions, respectively. Assuming the borders will run in the 200 m direction on the 0.1 percent slope as above, Figure 59 indicates the inflows that will complete the advance in the respective r_req times are 0.036 m³/min/m for initial irrigations and 0.0215 m³/min/m for later ones.

The values of r and p need to be generated or extrapolated for these flow rates unless they are already generated as part of the development of Figure 59 or, in this example case, from the previous example problem. For the 0.036 m³/min/m inflow, the values of r and p were determined from the previous example as r =.5635 and p = 6.949. For the 0.0215 m³/min/m inflow, r and p were calculated using the methods outlined in section 5.3.1 rather than extrapolated with the result that r =.6032 and p = 3.916.