Planning and Design of Agricultural Drainage Systems Under Uncertainty

Abstract

Drainage systems are major capital investments for irrigated agriculture. Therefore, the goal is to install drainage systems in a manner that will be most beneficial or “optimal” to the agricultural economy as a whole. The investment decisions for drainage projects are structured in a three level hierarchy. The first level is the project evaluation level, that is the level at which the decision is made if and when an area should be installed with drains. This decision is made on the basis of whether all the project benefits exceed all the costs in some multiobjective measurement. Given that an affirmatice decision is made at the first level, the next level in the process is the planning of the network of collector and main drains. This will include the placement, sizing, and design of the collector drains, pumping stations, and main drains. The third level is the design of the field drains. This is referred to as the field level design process. This process will include decisions on the type, material of, depth, spacing, and installation procedures for lateral field drains. These field drains will empty into the collection network designed in level two. The paper will present a description of the decision process at each level, the activities involved in making these decisions, and the interactions that take place between the various levels of planning. The next section will present mathematical tools that have been developed to aid the decision maker and planners to use the data available at each level most efficiently to allow for “optimal” use of limited resources. The third or field level design decisions are the depth and spacing of the lateral drains. A nonlinear optimization model that incorporates the physics, uncertainty of data, and economics of subsurface drainage will be developed to give the “optimal” spacing and depth of lateral drains. In the second level of planning it is important to define homogeneous areas or districts that will allow for field level analysis to be performed most accurately. A heuristic algorithm for efficiently draining these areas by a collection network will be discussed. This algorithm will allow for feedback between level two and level three and provide a multi-level dynamic drainage planning methodology.