Styrofoam is already replacable by fungus(http://www.youtube.com/watch?v=zmDENxTPn8Q) - and can be returned to our farming industries as input. Much packaging may be possible to replaced by similar.
Clothing is already somewhat recyclable and if natural materials, biodegradable. Broken peices to equipment are also likely recyclable - it's just how much effort it takes.
Water should be pretty much almost entirely re-harvestable. Between centrafuges, filters, and evapourators and freezers(different materials exhibit differing melting/boiling/freezing points, and can assist seperations) and catalysts 99.9% or thereabouts should be recoverable.
Of more concern, personally, was the byproducts of the various manufacturing industries. Located orbitally they are naturally suitably contained but they will require dealing with. Some things may be reacted against others to provide for something useful, but I still sense a lot bound for Sol, which is the easiest way to deal with anything that would otherwise pile up with no practical purpose in sight. Centrafrugal launchers could make it quite "cheap" to regluarly throw waste matter into the nearest star for inceneration. Being a large ball of nuclear plasma the "waste" should incinerate a long time before it actually gets there somewhat permenantly solving the problem, whilst causing little other problems in the process.
The problem with thermal dissipation in space is commonly the only method afforded is via radiation - the infra-red variety typically. Unfortuntely, this is possibly the least effective method of transferring heat. I personally can't see - from a mechanical aspect - what would seperate a "terrarium" and a station. They're both(for purpose of argument) sealed containers, bleeding via IR. It's likely the combination of organic thermal output, the thermal output of the equipement keeping them entertained and the equipements keeping them alive combined will eventually lead to not bleeding fast enough and require something like radiator panelling to provide more surface area to dissipate through. However, if this was placed on or in the surface of an asteroid or a celestial body it would potentially allow for "geothermal" conduction - and conduction is a much better method than radiation(ofc, the object you're conducting through will essentially become a radiator and if it doesn't bleed fast enough you'll end up using it to cook you).
IMHO one of the largest problems to solve is thermal dissipation. Scaling existing systems results in a lot more potential for things to go wrong - each panel in the array requires valves in case the panel is physically damaged to prevent hemorrhaging coolant, as well as many, many more pumps required to maintain pressure/flow. I'm to understand each curve in the system reduces flow/pressure as does distance and I suspect there to be a lot of corners and distance involved pushing fluid through a 286 square mile+ radiator array. I suspect Tesla's "valvular conduit" should be suitable for assuring unidirectional flow, regardless of gravity which should be one good thing in the stack.
Clearly what is required is either a all-round better method of thermal dissipation (unlikley) or severe improvements on existing methods. The likes of stirling engines to convert heat»electricity will generate heat themselves in the process (basic thermodynamics) and still require to bleed heat themselves for the "compression cycle" to operate. The seerback effect may be able to be employed to dissipate some energy, but it's likely to be too low on the scale of output to significantly consume the input. The same with things like metamaterial panelling that uses themal energy to excite into the visible light spectrum. Even combined, I cannot see a sensible way to use that energy (dissipating is such a waste).