2.2. Experimental design and parameters measures
The study area encompasses two adjacent areas (Fig. 1B) both
deforested 3 years before the experiment started, so collection of
any data on soil structure and tree development before deforestation
was not possible. However, neighbor native forest was used as
reference (Fig. 1B). Unfortunately, no area available for replication
was found at the start of the experiment. Large scale experiments
are sometimes hard to replicate decreasing power of conclusions,
but they may be more realistic than experiments performed at
lower scales, a trade-off imposed by costs and scarcity on experimental
ecosystems (Carpenter, 1989; Hurlbert, 1984). Additionally,
peculiarities of the mangrove ecosystem and features of the studied
areas imposed methodological limitations for dividing areas in
more than one experimental unit without suffering from treatment