Sustainability, to the extent that it's achieved, will be achieved regionally. This is fairly easy to understand, if we look at what are called "GHG emissions offsets".
Offsets are at the heart of a wide range of schemes for achieving ecological sustainability. If the term "cap and trade" enters the discussion, there are offsets in play. An offset is an agreement by which I get to emit (or continue to emit) a certain quantity of greenhouse gases because you agree (for a price) to reduce your emissions by like amount, in my stead. Offsets have been satirized by comparing them to an agreement that I get to cheat on my wife because (for a price) you've agreed to remain faithful (to my wife) on my behalf. But the truth is that some form of offsets will eventually be required by each and every signatory to the President's Climate Commitment (all 677 of them, to date). The PCC commits its signatories to achieve zero net GHG emissions, but there's no practicable way that any college or university can ever reduce actual emissions to zero. Whatever emissions can't be eliminated will have to be offset -- that's the difference between "zero" and "net zero".
There is already a small market in GHG offsets in place. Signatories to the Kyoto Protocol buy offsets, and countries in the European Union buy offsets, and some other parties (including some colleges and universities) buy offsets voluntarily on the open market. Most of the offsets purchased to date are sold by developers/national entities in the third world (or, less frequently, in underdeveloped portions of first-world countries). The basic idea is that third-world countries agree not to take some environmentally destructive action (like clearing tropical rain-forests or building new coal-fired power plants), thereby decreasing their GHG emissions in current and future years. The problem, of course, is that if the forest wouldn't have gotten cleared anyway, or the coal-fired plant wasn't really going to be funded, then no actual reduction in emissions ever occurs. Similarly, if one area of rain-forest gets protected but an equivalent area somewhere else (perhaps even in a different country) gets cleared as a result, no real reduction ever occurs. The first problem with globally-distributed offset agreements is that they're unenforceable and un-auditable. The second problem is that, given the first problem and a continuing stream of willing buyers, a perverse incentive develops to create totally bogus offsets for fun and profit.
As a signatory to the PCC, Greenback U knows that it will require GHG emissions offsets. We don't know just how much emission we'll have to pay to offset, but we know it will be some positive amount. What we've decided to do, however, is to offset our irreducible emissions locally. We're going to buy offsets by paying to weatherize local low-income housing, to improve local mass transit, to generate more electricity (utilized off campus) renewably, to make local infrastructure and residence patterns more energy efficient. By keeping things relatively local (within our region), we can assure that the reductions we claim credit for really do occur, and really wouldn't have happened except for our funding, and aren't just shifting emissions to some other location.
We expect to pay more for the offsets we eventually use than is currently charged on the open market. Indeed, even if market prices rise in future, we'll probably still end up paying a premium. But the truth is that the whole concept of a global GHG offset market -- the idea that it's cheaper to reduce GHG emissions in poor countries than in rich ones -- is dependent on the premise that poor countries remain poor and rich countries remain rich. In a very real sense, a global GHG offset scheme attempts to achieve global ecological sustainability by trading away any hope of global economic and social sustainability. Students of international trade are familiar with the concept of "the tragedy of resources" -- the pattern by which the majority populations in countries which base their economies on export of abundant natural resources lose both economic and political power to outside forces. Global GHG offset schemes simply create another category of natural resources which can be "exported"; there's no reason to believe that the tragedy-of-resources pattern is likely to change as a result.
An organization's (university's) GHG emissions can be reduced using offsets, but the offsets have to be fairly local to that organization. Emissions need to be offset in more or less the same location, more or less the same economy, more or less the same local culture as they're generated. Not only can this allow ecological sustainability to be achieved (at least in theory) without creating social and economic problems, it can also allow the offsetting organization to assure the quality (including what's referred to as "additionality") of the offsets it's claiming.
Scale this thought process up from a university to a city, or a county, or a state. Sustainability, if it's to be achieved, must be achieved more or less regionally. Sustenance economy may be only a soft model, and autarky may be too hard a term, but the idea of balance achieved, audited, and (when necessary) enforced regionally is the only one that makes any real-world sense to me. Unfortunately, it doesn't particularly resemble the model of the world ("business as usual") we're implicitly preparing Greenback students to participate in. Or expect. Or demand.