For several years now, I've worked with a few self-selected faculty members at Greenback U to create and administer undergraduate research projects. Most of those projects have been conducted within the constraints of semester-long courses, and most of them have focused student attention on some aspect either of Greenback's campus or of the surrounding Backboro community. Practicalities (time, money, travel, etc.) mitigate in favor of research that can be conducted locally and, to be honest, emphasizing the local nature of sustainability challenges and solutions is part of my agenda.
From time to time, I've tried to gin up interest in larger umdergrad research projects (more participants than enrollment in a single class, more time than a single semester, more scope than a single campus or community), but without the incentive of course credit, they've proven difficult to sustain. Initial enthusiasm gets overtaken by events, and participation fizzles. I've tried making the subject matter and presentation more attractive, more entertaining, with greater replayability, all to no good end.
But recently I've come across two existing initiatives which, I think, show potential to "have legs". Both of them fall into the general category of "citizen science", but I think they'll still be well received by the professional scientists on campus. (BTW, one of the challenges of planning and designing student-focused initiatives on a twelve-month schedule is that for about four of those months, each year, students and faculty aren't around and often don't even respond to emails.)
The first initiative is one that students can decide to participate in, and yet spend little active time on. Climateprediction.net  offers citizens, including students, the opportunity to donate their unused personal computer capacity to research projects which run very large, very complex simulation models on an ad hoc distributed computing network. It's operated out of Oxford, and some of the projects which define the models they run have a Euro-centric flavo(u)r to them. But if I'm not going to focus students' attention specifically and locally, I guess the next best thing is to have them step back and look at the global picture. Climateprediction.net is kind of like "SETI @ home", the internet-based mechanism by which nerds and geeks and science-fiction fans and I have participated in the search for extra-terrestrial intelligence since 1999. Indeed, the two initiatives run on the same distributed computing platform, called BOINC . One of the features BOINC provides is the ability to form teams, track computing cycles donated, and report team-based statistics. Pretty much all the infrastructure we'll need to administer intramural competition, once we figure out how to market the idea.
The upside of fomenting student participation in Climateprediction.net projects is that the only direct activity required of students is to create and sign in to a participant account. From that point on, students' computers can be chugging away, processing climate models, with no explicit student action and even little student awareness. The downside, of course, is that to maximize the amount of processing power your computer donates to the project --- and, thus, the participation score your team achieves -- you have to leave your computer running all the time. Which burns electricity that otherwise wouldn't be consumed. (There's a lot of variability in the amount of additional electricity various types of computers will consume, but the marginal amount is never zero.) And one of my office's continuing messages to students has been to conserve electricity. Maybe I can come up with some way of estimating the additional consumption and then subsidizing (or, better, getting students collectively to subsidize) the additional cost of purchasing that electricity from renewable sources. A bit tricky on the marketing side, perhaps. We'll just have to see what flies.
The other initiative is one that implies a smaller increase in electrical utilization (perhaps even zero), but more requires conscious time and effort on the part of students. The US Geological Survey has initiated a program called iCoast,  in which citizen scientists visually compare pre- and post-storm aerial photos of sections of US coastline, and then assign pre-defined tags to indicate the degree of coastal change created by the storm event. For climateprediction.net, what's being crowdsourced is pure computing power, because computers can run digital calculations unattended. In the iCoast initiative, what's being crowdsourced is the ability and energy to compare images, detect differences and assess what the image differences mean in terms of real world changes. Compared to running calculations (or even playing chess), that sort of thing is a real bear to program, much less test, much less operate with high levels of accuracy given input photographs which vary in thousands of ways (angle, lighting conditions, time of day, shadowing, focus, centering, scope, etc., etc.) above and beyond the particular differences the project hopes to assess. (In the movies, supercomputer systems can continually analyze billions of high resolution images and automatically identify events/anomalies which pose potential risks to national security. In real life, not so much.)
The upside of the iCoast initiative is that it provides an opportunity for more intense, and thus possibly more meaningful, engagement of students. Additionally, it's US-based (so more conceptually familiar for many of our undergrads) and relevant here in the Northeast where the reality of coastal incursions -- aggravated by coastal erosion and the loss of protective features like dunes -- hits home pretty hard. Downsides (from an administrative management perspective) include the fact that the contribution statistics assembled by the software don't seem to lend themselves as easily to creating team-based competition, and that the actual time/effort requirement seems likely to create significant drop-off from initial student participation levels.
Right now, my overall take is that the existence of citizen science initiatives focused on various sustainability-related challenges creates a range of opportunities to engage undergraduate students' attention. It's not a question of either/or -- not at Greenback U, and not on any other campus. The more, the merrier. Active (or even relatively passive) participation in crowdsourced science initiatives, if it's properly integrated into the existing curricular and co-curricular educational offerings, looks like it has potential to enhance the undergraduate experience and bring awareness that, where sustainabiltiy challenges are concerned, the 21st Century is an experiment and we're all in the test tube together.