An irrigation upgrade can reduce the water applied to one field and still leave a larger question unanswered: what happens to the water no longer used there? The answer depends on the boundary of the analysis. A farm, an irrigation district, and a river basin do not count water in the same way.
Start with the boundary
At the field scale, the comparison may be between water delivered and water needed by a crop. At a wider scale, water that drains or returns to a channel may be available for another user downstream. Those flows are not automatically recoverable, but leaving them out of the accounting can make an intervention appear more conclusive than it is.
Efficiency has more than one meaning
A change in pipes, gates, or application methods can reduce losses at one point in a system. It can also change pumping needs, maintenance work, or the timing of deliveries. The engineering question is how the equipment performs; the economic and public-affairs questions include who pays, who receives the benefit, and how shared supplies are governed.
Keep the claim proportional
When reading a paper about irrigation efficiency, I look for the study boundary before I settle on its headline conclusion. I also note which flows are measured, which are estimated, and whose use is counted. That habit does not dismiss a useful upgrade; it makes the claim match the evidence.