Mapping and quantifying habitat fragmentation in small coastal areas: a case study of three protected wetlands in Apulia (Italy)

In the Mediterranean Region, habitat loss and fragmentation severely affect coastal wetlands, due to the rapid expansion of anthropogenic activities that has occurred in the last decades. Landscape metrics are commonly used to define landscape patterns and to evaluate fragmentation processes. This investigation focuses on the performance of a set of landscape pattern indices within landscapes characterized by coastal environments and extent below 1,000 ha. The aim is to assess the degree of habitat fragmentation for the monitoring of protected areas and to learn whether values of landscape metrics can characterize fine-resolution landscape patterns. The study areas are three coastal wetlands belonging to the Natura 2000 network and sited on the Adriatic side of Apulia (Southern Italy). The Habitat Maps were derived from the Vegetation Maps generated integrating phytosociological relevés and Earth Observation data. In the three sites, a total of 16 habitat types were detected. A selected set of landscape metrics was applied in order to investigate their performance in assessing fragmentation and spatial patterns of habitats. The final results showed that the most significant landscape patterns are related to highly specialized habitat types closely linked to coastal environments. In interpreting the landscape patterns of these highly specialized habitats, some specific ecological factors were taken into account. The shape indices were the most useful in assessing the degree of fragmentation of habitat types that usually have elongated morphology along the shoreline or the coastal lagoons. In all the cases, to be meaningful, data obtained from the application of the selected indices were jointly assessed, especially at the class level.     

Heavy metal bioaccumulation and metallothionein content in tissues of the sea bream Sparus aurata from three different fish farming systems

The distribution and potential bioaccumulation of dietary and waterborne cadmium and lead in tissues of sea bream (Sparus aurata), a major aquaculture species, was studied in relation to three different fish farming systems. Metallothionein levels in fish tissues were also evaluated. Results demonstrate that metal concentrations in various tissues significantly vary among fish culture systems. Different tissues show different capacity for accumulating heavy metals. The content of both cadmium and lead is not strictly correlated with that of metallothionein. Indeed, the marked accumulation of both metals in liver, as well as the high lead content found in gills and kidney, are not accompanied by a concomitant accumulation of metallothioneins in these tissues. No correlation is present between heavy metals and metallothionein content in muscle tissue. The results also demonstrate that cadmium accumulates mainly via dietary food, whereas lead accumulation is not of food origin. Noteworthy is that the concentration of the two metals found in muscle in all instances is lower than the limits established by European Union legislation for fish destined for human consumption.     

Towards the harmonization between National Forest Inventory and Forest Condition Monitoring. Consistency of plot allocation and effect of tree selection methods on sample statistics in Italy

In the frame of a process aiming at harmonizing National Forest Inventory (NFI) and ICP Forests Level I Forest Condition Monitoring (FCM) in Italy, we investigated (a) the long-term consistency between FCM sample points (a subsample of the first NFI, 1985, NFI_1) and recent forest area estimates (after the second NFI, 2005, NFI_2) and (b) the effect of tree selection method (tree-based or plot-based) on sample composition and defoliation statistics. The two investigations were carried out on 261 and 252 FCM sites, respectively. Results show that some individual forest categories (larch and stone pine, Norway spruce, other coniferous, beech, temperate oaks and cork oak forests) are over-represented and others (hornbeam and hophornbeam, other deciduous broadleaved and holm oak forests) are under-represented in the FCM sample. This is probably due to a change in forest cover, which has increased by 1,559,200 ha from 1985 to 2005. In case of shift from a tree-based to a plot-based selection method, 3,130 (46.7 %) of the original 6,703 sample trees will be abandoned, and 1,473 new trees will be selected. The balance between exclusion of former sample trees and inclusion of new ones will be particularly unfavourable for conifers (with only 16.4 % of excluded trees replaced by new ones) and less for deciduous broadleaves (with 63.5 % of excluded trees replaced). The total number of tree species surveyed will not be impacted, while the number of trees per species will, and the resulting (plot-based) sample composition will have a much larger frequency of deciduous broadleaved trees. The newly selected trees have—in general—smaller diameter at breast height (DBH) and defoliation scores. Given the larger rate of turnover, the deciduous broadleaved part of the sample will be more impacted. Our results suggest that both a revision of FCM network to account for forest area change and a plot-based approach to permit statistical inference and avoid bias in the tree sample composition in terms of DBH (and likely age and structure) are desirable in Italy. As the adoption of a plot-based approach will keep a large share of the trees formerly selected, direct tree-by-tree comparison will remain possible, thus limiting the impact on the time series comparability. In addition, the plot-based design will favour the integration with NFI_2.     

Simulating regeneration and vegetation dynamics in Mediterranean coniferous forests

This study aims to provide a quantitative framework to model the dynamics of Mediterranean coniferous forests by integrating existing ecological data within a generic mathematical simulator. We developed an individual-based vegetation dynamics model, constrained on long-term field regeneration data, analyses of tree-rings and seed germination experiments. The simulator implements an asymmetric competition algorithm which is based on the location and size of each individual. Growth is parameterized through the analysis of tree-rings from more than thirty individuals of each of the three species of interest. A super-individual approach is implemented to simulate regeneration dynamics, constrained with available regeneration data across time-since-disturbance and light-availability gradients. The study concerns an insular population of an endemic to Greece Mediterranean fir (Abies cephalonica Loudon) on the island of Cephalonia (Ionian Sea) and two interacting populations of a Mediterranean pine (Pinus brutia Ten.) and a more temperate-oriented pine (Pinus nigra Arn. ssp. pallasiana) on the island of Lesbos (NE Aegean Sea), Greece. The model was validated against plot-level observations in terms of species standing biomass and regeneration vigour and adequately captured regeneration patterns and overall vegetation dynamics in both study sites. The potential effects of changing climatic patterns on the regeneration dynamics of the three species of interest were subsequently explored. With the assumption that a warmer future would probably cause changes in the duration of cold days, we tested how this change would affect the overall dynamics of the study sites, by focusing on the process of cold stratification upon seed germination. Following scenarios of a warmer future and under the current model parameterization, changes in the overall regeneration vigour controlled by a reduction in the amount of cold days, did not alter the overall dynamics in all plant populations studied. No changes were identified in the relative dominance of the interacting pine populations on Lesbos, while the observed reduction in the amount of emerging seedlings of A. cephalonica on Cephalonia did not affect biomass yield at later stages of stand development.