There is a well-known misconception that improving farm irrigation efficiency can save water in an entire water system. Often the argument associated with irrigation efficiency is analyzed with a large and complex model of a real system—I have done some of these models. With these large models it becomes difficult to see the clearly the relevant hydrologic principles—one gets lost in the trees and loses sight of the forest. Often it is useful in both science and engineering to use a zero-order analysis, a back of the envelop calculation, to illustrate a basic principle. A simple explanation will illustrate the problem associated with improving farm irrigation efficiency:
Nevada has several rivers that flow east out of the Sierra Mountains into the western part of the state where irrigation uses all of the water in the river, except in the wettest years. The Walker River is one of these; Walker Lake is drying up because of irrigation upstream along the Walker River. The irrigators are very efficient in capturing the streamflow.
Let’s use the Walker River as a rough prototype for a very simple irrigation system. Let’s assume that the River flows during the average irrigation season at 100cfs, and that there are 10 farmers along the river, with more or less equal water rights of 20cfs each. Typically the irrigation system including, the canals and laterals and the farms, are 50% efficient; we need not complicate our analysis with where the inefficiencies occur. So we have:
- 100cfs of streamflow
- 10 farms
- 20cfs water rights/each
- 50% efficiency
If we now examine the entire system, 90% of the water is used productively, even though the farm efficiency is only 50%. We need to distinguish a system efficiency, as well as a farm efficiency.
Now let’s improve the farmer’s efficiency. For the sake of argument, to make our point, let’s assume the efficiency is improved to 100%: meaning we line canal and laterals, and go to drip irrigation on the farms: Now we have a decision -- What is meant by a water right?
- Is the farmer entitled to use all the water his right allows him to divert, or
- does the water right apply to water that he previously consumed?
Let’s first look at the situation where the farmer is entitled to use his entire 20cfe right. With 100% efficiency, the first 5 farmers use their 20cfs, drying up the stream, and the other 5 farmers get no water. The system is 100% efficient, all the water goes to crops. Of course, only five of our ten farmers get water.
In the second scenario each farmers gets 10cfs—his previous consumptive use. Now all the farmers get water. Again, the stream is dried up, and all the water goes to crops—the system is 100% efficient.
COMPARISON OF THE TWO SYSTEMS—50% versus 100%
When we compare our two systems; the 50% farm efficient valley uses at least 90% of all the water available. Of course, the 100% efficient system uses all the water. In terms of the entire valley, we gained, at best, 10% by going to 100% farm efficiency. This is hardly a policy one wants to subsidize. There is the issue of how to pay for the increase in farm efficiency. An individual farmer may increase his output by using his water more efficiently, but usually his increased water use comes at the expense of less water in the system for his neighbors—he beggars his neighbors.
Bottom Line: In focusing on increasing farm efficiency, one often loses sight of looking at the efficiency of the entire system. From a policy perspective, it is usually the performance of the system that is of ultimate concern. Increasing farm efficiency usually does not save water in the system.
* In response to my posts (here and here) criticizing the Pacific Institute's claims about "saving" water from efficient irrigation.
For academic studies on this issue, see:
- Ward, F. A. & Pulido-Velazquez, M. "Water conservation in irrigation can increase water use" in Proceedings of the National Academy Of Sciences, 2008, 105, 18215-18220.
- Pfiffer, L. & Lin, C.-Y. C. "Incentive-Based Groundwater Conservation Programs: Perverse Consequences?" in ARE Update [PDF], 2009, 12, 1-4.