Reconstructing extreme ..ooding events in the Venetian-Friulian Plain (NE Italy) through meta-analysis of a 14C database

1. Introduction The study of the sedimentary record of the past alluvial episodes can support the recognition of extreme events of flooding, back to several hundred or thousands of years (Ely et al., 1993; Knox, 1993; Maas et al., 2001). These studies, among the others, opened the path to the comparison of river systems alteration at various scale, even if at the interhemispheric one (Macklin et al., 2012). The significant increase of dated fluvial units permits the application of meta-analysis techniques which request large datasets. The INQUA group denominated “Hydrological EXtreme Events in Changing Climate” try to explorer this research approach, applying the numerical method developed by Macklin et al. (2006; 2012) for producing probability-based reconstructions of Holocene and Pleistocene flooding periods. As part of this group, we applied the method to the 14C dating database of the Venetian-Friulian plain, which is one of the largest datasets dealing with alluvial stratigraphy. The aim of the whole process is to identify a probabilistic valuation of periods characterized by an increase in the amount of flooding episodes. 2. Study region The Venetian-Fiulian Plain extends for about 10,000 km2 along the South-eastern Alps and was formed by the activity of the main Alpine rivers that are not tributaries of Po River (Isonzo, Tagliamento, Piave, Brenta and Adige rivers). The rainfall regime is bimodal, with a maximum during autumn and a secondary peak in spring; mean annual rainfall values are significantly higher than those of the Mediterranean and central Europe. Those elements, combined, may periodically cause extreme flood events in the plain. 3. Methodology During the last 30 years a large amount of 14C datings has been performed in NE Italy on a widespread type of materials and fluvial units and part of these data have been already collected in a standardized database (e.g. Bondesan et al., 2004; Carton et al., 2009; Fontana et al., 2010). A database of about 400 14C datings was compiled, comprehending published and unpublished realized both with the conventional and the AMS method. The whole dataset was organized following the instructions provided by Johnstone et al. (2006). In particular, for each dating it’s recorded its location, type of sample, stratigraphic position and the geological meaning of the fluvial unit from which the age was obtained. The following step seek to determine if the dated sample represented a “change after” date. This correspond to a variation in the depositional style and it is defined as a mark that lies immediately below a marked sedimentary discontinuity, identified by a reversal in the natural floodplain fining-up sedimentary sequence, or where a peat or soil associated with a period of floodplain stability is overlain by a minerogenic sediment unit (Macklin et al., 2012). It has been demonstrated that the occurrence of these types of stratigraphic breaks can indicate a general change of the climate conditions, expressed by a significant flood event or an increase in flood frequency/magnitude. The various data were subdivided according to their uncalibrated age into Pleistocene and Holocene dates, in order to isolate homogeneous chronological/climate periods. 14C ages were calibrated using the IntCal09 calibration curve and then processed using OxCal 4.1 (Bronk-Ramsey, 2001) to produce cumulative probability function (CPF) plots. Distinct plot were obtained for the “change after” dates, which constitutes a probability distribution that is the best estimate for the chronological distribution of the items dated. 4. Results and perspectives Here are descried some of the preliminary results of the work and possible directions for the future research. Almost half of the database consists of LGM datings carried out on peat or organic silt/clay collected in the distal sector of the Venetian-Friulian Plain with boreholes from the immediate subsoil up to 30 m of depth. These organic layers represent wetlands developed over wide areas comprised between fluvial ridges and periodically inundated by the fluctuating groundwater (Miola et al., 2006). The accumulation of organic material continued until the alluvial minerogenic supply prevailed on the fen organic deposition. When this change occurs, the peat level is incorporated into the stratigraphic record. So, dating the top of these level provide a certain “change after” date. The flooding episodes were identified as those portions of the plot where the relative CPF exceeded the mean value obtained for the whole dataset. Those events were considered rounding the value to the nearest century, in order to avoid over-interpretations. The method has been applied in detail to 2 distinct river systems: Tagliamento and Brenta. Those river, actually, constitute the most studied systems of the area and, in addition, the number of available radiocarbon datings is significantly larger if compared to other Mediterranean systems (see Fig. 1 for a representations of the available datings in the Brenta system). In the Holocene the research evidence at least 2 periods of climate deterioration, during the second part of the Subboreal period (4500-3000 years BP) and the dark ages (ca 500-1000 AD), respectively. These intervals were partly already known as period of significant river systems alteration, testified by frequent avulsions of the Tagliamento, Brenta and Piave rivers. It’s important to compare the results with other independents climate proxies, in order to clarify if the process is consistent with them and if it’s possible to identify some periods of climate deterioration not yet discovered in these part of Italy. It is possible also to compare the results among the various river basins, to point out if some difference occurs, due to geographical/geological conditions. Moreover, the meta-analyses of radiocarbon database allows an objective comparison with the results obtained in other European and extra-European regions.