Wow, sounds like you really thought this out. Ok, this sounds pretty solid in theory but I have some concerns (pretty minor).
Watersheds can 'spill over' or 'divide' in wet and dry years. Multiple watersheds often combine during extremely wet years (though this is likely to help the causes) and in dry years, watershed can splinter. Say 2 watersheds get cut off due to lower waters, how does this impact removal rates? The differences in water volume could be huge. What about other factors that can influence removal rates. Temperature, water clarity (particulate matter), conc. of bio-organisms, water velocity, etc. These can fluctuate quite drastically from year to year, and can cause some pretty big changes in removal rates. Ideally, the limits would be buffered for such changes ut how do you accommodate changes of this magnitude?