Non-consumed water on one place can be consumed in another place. Or in other words, expenses made in one place can become cash in another place. For example seepage brings water to aquifer, which is pumped up in some other place. Conclusion: I think graph is too simple. F.e. it could lead to the misunderstanding that when you have less seepage, you'll have less expenses and more cash. Bottom line: Hydrology requires a basin-wide approach when doing water economics.
I'll add to Johannes comment that the "nonconsumptive" use is far oversimplified, and labeled as such to discourage water from flowing more naturally. Nonconsumptive use doesn't exist at all. Water that gets by crops recharges groundwater, helps clear pollution, moves into riparian, wetland, and marine environments, etc. The closer we get to consumptive volume cutting into nonconsumptive volume, by the flawed logic of the graph, the more likely we are to increase pollution related to what nonconsumptive flow occurs. We need a paradigm shift in understanding the free flow.
Spammers, don't bother. I delete spam.