Expermiental Design: How to avoid pseudoreplication

My research group struggled with experimental design to avoid pseudoreplication in all of our data as we will all be sampling data in the same stream, Slaughter Creek. We found ourselves asking if there was a great enough spatial scale to avoid treatments not being replicated even though our samples would be. We also wondered if our replicates would not be statistically independent due to our limited spatial scale within one creek. To avoid pseudoreplication, I planned the location of my impacted sample sites throughout the whole watershed of Slaughter Creek, both upstream and downstream of the WQPLs. In this way, the treatment I am measuring (protected land) will be compared to impact sites both upstream and downstream of the WQPLs. Because I will not be able to measure multiple treatments (sites at multiple WQPL tracts in multiple creeks), I will not be able to isolate all random factors from fixed factors, but I will remain keenly aware of this fact.

The map below shows the WQPL tracts along Slaughter Creek (Baker, Hafif, Hielscher).