Analysing methodological issues in short-term monitoring of rare European beech forests restoration

Monitoring restoration effectiveness implies that restoration measures should be evaluated using multi-scale analysis and modelling approaches. In this way, the scale-dependent nature of ecosystem processes and dynamics can be adequately described. Here we evaluated how grain size (i.e. the dimension of sampling units) and type of data (abundance vs. presence/absence) impact the analysis of the response to restoration management activities aiming at halting the loss of European beech forests with Abies alba in the cCentral Apennines (Italy). The proposed analysis relies on forests sampled in three Sites of Community Importance (SCIs) located in Ccentral Italy. For each described treatment (reference, control and impact), randomly located multi-scale plots were selected. The total number of plots was 35, grouped as follows: 18 plots in Pigelleto SCI (9 reference and 9 impact, Before/After-Reference/Impact design); 9 plots in Badia Prataglia SCI (5 reference and 4 impact, Before/After - Reference/Impact design), and 8 plots in Valle della Corte SCI (4 control and 4 impact, Before/After Control-Impact design). Vascular plant occurrences were recorded twice in every plot during summer 2010 or 2011 (before restoration) and in 2013 or 2014 (after restoration). Analysis of species abundances , recorded at the largest grain size available (20x20m plot), showed a significant change in plant species composition, in restored sites just after two years from the restoration management. These changes referred oninvolved tree species composition, object of thinning or removal, but also toon herbaceous species that are typical of the understory vegetation of deciduous woodlands as well as on synanthropic species, these latter showing an increase in their presence in managed plots. When presence/absence data were used, however, previous results were less visible for all the examined grains in two out of the three sites where restoration measure were carried out, also suggesting a site-dependent response. In conclusion our results suggest that the type of data determines the ability to discriminate control/reference and impact areas better than the grain size alone, since most of the observed differences are related to the variation in species abundances .