Towards a 'polycentric' landscape? Reconnecting fragments into an integrated network of coastal forests in Rome

Preserving relict coastal forests from urbanization is a priority target for sustainable land management. Monitoring forests through remote-sensing diachronic approaches is an important tool to identify networks of natural areas and to inform conservation policies for high-quality wood patches. The present study applies the monitoring approach proposed by Hansen et al. (Science 342(6160): 850-853, 2013) to the forest landscape in coastal Rome, central Italy, providing indicators assessing forest cover and short-term wood loss and gain. Forest land (with 1 to 100 % cover) extends through 16.9 % of the total area with prevailing classes at 21-40 % and 71-100 % density. These classes indicate distinct forest types: the largest remnants of mixed coastal forests and the sparse woods embedded in the rural matrix and characterized by structural fragmentation. Dense forest classes (71-100 %) prevail in two urban districts (Castelporziano and Lido di Castelfusano). Percent forest loss (2000-2012) increases more than proportionally with percent forest cover. Forest gain-to-loss ratio decreases for moderately dense classes (31-70 %) and increases for both sparse (>= 30 %) and very dense (>70 %) forest classes. Multivariate analysis identifies a land cover gradient from strictly forest districts to agricultural and peri-urban districts. Spatial planning is increasingly required to preserve and reconnect mesophilous forest patches, pinewoods and isolated woodlots into a green network embedded in the agricultural matrix.