DINAMICA DI VEGETAZIONE IN AREE DI POST-ABBANDONO DELLA PIANURA PADANA

VEGETATION DYNAMICS IN ABANDONED FIELDS IN THE PO PLAIN (ITALY) Human impact and the mechanization of agricultural activitities have deeply modified the natural landscape in the Po plain (northern Italy), as well as in other countries. Forest vegetation in this area has strongly declined since the Neolithic (CASTELLETTI E ROTTOLI, 1998), with subsequent modifications of forest structure and floristic composition. Nonetheless, the recent decrease of agricultural pressure has enhanced the development of postcultural vegetation dynamics, allowing us to start an examination the vegetation dynamics in the Po plain. In order to identify the dynamic series of the Po plain area, we have analyzed the secondary vegetation dynamics in sample abandoned fields of the upper plain of Milan and Lecco and woody communities of the Gera D\u2019Adda neighbourhood (Bergamo) which, in dynamical terms, represents the final stage of the series. The first case-study concerns the area of Rio Vallone Plis (Milan), which measures 1181 hectares and is located in the Pleistocene and Holocene alluvial sediments of the upper Po plain. The chronosequence or space-for-time substitution approach (PICKETT, 1989; FOSTER & TILMAN 2000) was applied to examine the temporal changes in vegetation. The method included the collocation of relev\ue9s in sites of different \u2013preferably known\u2013 age at given time intervals, assuming that the samples would represent the distinct stages of a temporal series. The historical study was realized by comparing orthocorrected aerial photographs taken in different years (1954, 1980, 1998, 2003, and 2006). Fifty-six relev\ue9s were made by using the phytosociological approach (BRAUN \u2013 BLANQUET, 1932; WESTHOFF & VAN DER MAAREL, 1973). Two matrices (relev\ue9s/species, Landolt ecological index average values/cluster) were constructed from the resulting data set, in order to analyse the relationships between ecological variables and the different clusters of vegetation. Matrices were processed separately through a multivariate analysis approach (PCA) using the SYN-TAX 5.0 software (PODANI, 1995). The third matrice (relev\ue9s/life forms) was processed in order to analyse the relationships between life forms abundance and the age stage of the studied sites, through ANOVA analysis with the SPSS 10.0 software and later analysis by means of Pearson a 2-code correlation. Nomenclature of taxa, corological elements and life forms follow PIGNATTI (1997) and AESCHIMANN & AL (2004). Two main series of secondary succession were distinguished according to their environmental and soil conditions: (a) the acidophylous series of middle Pleistocene terrace (Mindel Auct.) develops from the first herbaceous nitrophylous stage, belonging either to Chenopodion glauci (igrophylous variant) or Dauco-Picridetum hieracioidis (mesoxerophylous variant), to a second herbaceous-shrubland stage, which can be referred to either to Convolvulo-Eupatorietum cannabini (igrophylous variant) or Convolvuletalia sepium (mesoxerophylous variant), to a third forest-edge stage that can be attributed to shrubland communities Prunetalia spinosae, and eventually to the last stage, which is represented by wood referable to Quercetalia robori petraeae. The regressive stage is identified by the communities of Prunus serotina; (b) the mesophylous series of Holocene terrace develops from the first nitrophylous herbaceous stage belonging to Arrhenatheretalia elatioris, to a second forest-edge stage referable to shrubland communities (Sambuco-robinieti) referable to Carpinion betuli, and to the final stage made up mesophylous oak-hornbeam wood attributed to Carpinion betuli. The regressive stage belongs to communities referred to Prunetalia spinosae The second case study regards the phytosociological definition of the mesophylous and mesoigrophylous oak-hornbeam forests set on late Pleistocene and Holocene alluvial sediments of the upper Po plain, in the Gera d\u2019Adda (Bergamo). Thirty-one relev\ue9s were performed by using the phytosociological approach (BRAUN \u2013 BLANQUET, 1932; WESTHOFF & VAN DER MAAREL, 1973). Two matrices were processed separately through a multivariate analysis approach (PCA) with the SYN-TAX 5.0 software (PODANI, 1994). The results indicate that clusters are distributed according to three main ecological gradients (defined by the LANDOLT index): soil moisture value (F); soil pH value (R); soil aeration value (D). The historical study, which was realized by comparing historical maps of different years (1833, 1889) with recent regional topographic maps (CTR, 1991), shows that Orfano Wood has relict characteristic. Orfano Wood is referred to two different associations: (a) Querco-Ulmetum Issler 1926 for the igrophylous communities on Holocene sediments; (b) Polygonato multiflori \u2013Quercetum roboris subass. carpinetosum betuli Sartori 1980, which was included by ANDREIS e SARTORI (2002) in the Alnion incanae Pawlowski et Pawlowski et Wallisch 1928 (Alno-Ulmion minoris Br.-Bl. Et Tx.43), for the mesophylous communities developed on Holocene sediments. Concerning the other wood located on the Pleistocene \u201cmain level of the Po plain\u201d, the mesophylous woods belongs to Erythronio-Carpinion alliance (Hovart 1958) Marin\u10dek in Walln\uf6fer et al. 1993, with igrophylous variant referable to Alnion incanae Pawlowski et Pawlowski et Wallisch 1928 alliance (Alno-Ulmion minoris Br.-Bl. Et Tx.43)