CAMBIOS HISTÓRICOS EN EL APORTE TERRÍGENO DE LA CUENCA DEL RÍO DE LA PLATA SOBRE LA PLATAFORMA INTERNA URUGUAYA

The Rio de la Plata Estuary (RdlP) is a fluvio-marine system that drains into the Southwestern Atlantic Ocean with the Parana and Uruguay rivers as main tributaries. The estuary is fed by a 3,100,000km2 catchment area which extends over the territories of Argentina, Paraguay, Brazil, Bolivia and Uruguay (Acha et al ., 2008). The RdlP exhibits significant natural decadal- and annual-scale, hydrodynamic and oceanographic variability associated with the Pacific Decadal Oscillation (PDO), the Atlantic Multidecadal Oscillation (AMO) and the El Nino/La Nina Southern Oscillation (ENSO) (Depetris and Pasquini, 2007b; Chiessi et al ., 2009; Garreaud et al ., 2009). Such variability affects the moisture budget over the surrounding continental areas and leads, thus, to changes in the river discharge. PDO is associated with ENSO as both appear to display similar hydrological responses, though their inherent mechanisms are not yet fully understood (Garreaud et al ., 2009). In this sense, warm and cold PDO phases strongly resemble El Nino and La Nina events, respectively (Garreaud et al ., 2009). During El Nino episodes, an increase in precipitation over the RdlP drainage basin is commonly observed (Boulanger et al ., 2005; Camilloni, 2005; Garreaud et al ., 2009;Garcia-Rodriguez et al ., 2014), and consequently increased Parana and Uruguay river discharges are displayed (Depetris and Pasquini, 2007a). Campo et al . (1999) have recorded a freshwater plume of low salinity and temperature (32, 18 oC respectively) associated with an increase in RdlP discharge during the El Nino event of 1997, expanding northwards up to 23oS. Furthermore, during negative AMO phases it was recorded an increasing trend on the precipitations over the SE South America (SESA) and, as a consequence, a concomitant increase in the Rivers Parana and Uruguay discharge was recorded, while the opposite pattern was observed during positive phases (Chiessi et al ., 2009).The aim of this paper is to infer the link between changes in the delivery of terrigenous sediment to the adjacent Atlantic Ocean with recorded hydrological variability of the RdlP. To achieve this, we used sedimentological and geochemical proxies from two sediment cores, which were retrieved from the inner continental shelf off Uruguay and encompass the past 100 AD. Sediment Core GeoB 13813-4 was taken from the inner-shelf “RdlP paleo-valley mudbelt” (Fig.1;34°44’13’’S, 53°33’16’’W) during research cruiseM76/3a (Krastel et al ., 2012; Lantzsch et al ., 2014).Sediment Core BAR1 was retrieved in the inner shelf “Barra del Indio” zone (Fig. 1; 35°03’00’’S,56°09’00’’W), performed by the Faculty of Sciences, Universidad de la Republica (Uruguay). For CoreGeoB 13813-4, we analyzed the relative abundance of major elements (Ca and Ti) obtained by an X-ray fluorescent sediment core scanner AVAATECH and the Ca/Ti ratio was used to infer continental versus marine influence. This chemical elemental ratio was chosen according to previous successful applications within the Atlantic Ocean (Chiessi et al ., 2009; Mahiques et al ., 2009; Govin et al ., 2012;Bender et al ., 2013; Burone et al ., 2013). Furthermore, the arithmetic mean grain size distribution was studied from both sediment cores, using the GRADISTAT program (Grain Size Distribution and Statistics Package for the Analysis of Unconsolidated Sediments) version 8 (Blott and Pye, 2001) (Fig. 4).For Core GeoB 13813-4 grain size were obtained by Laser Particle Sizer LS200 and for Core BAR1were obtained by Malvern Mastersizer 2000 Laser analyser. The chronology from both cores was assessed by210Pbxs dating (Table 1 and Fig. 2; Appleby, 2001;2008). For core BAR1 we selected the CRS (Constant Rate Supply) model (which is highly used for estuarine systems), while for GeoB 13813-4 the CFCS(Constant Fluxe: Constant Sedimentation Rate)model was applied (Appleby, 2008; Bernal et al .,2010). In the last case, the decision of using the CF-CS model was due to the lack of a complete 210Pb dataset, which would bring very high sedimentation rates uncertainties (Sanchez-Cabeza and Ruiz-Fernandez, 2012). To assess the climatic variability over the pastc entury as inferred from the sediment proxies, we evaluated the climatic indices PDO and Southern Oscillation Index (SOI), of the Joint Institute for the Study of the Atmosphere and Ocean, University of Washington (http://jisao.washington.edu), as well as the AMO from the NOAA (http://www.aoml.noaa.gov/phod/amo_faq.php). We further compared these data with temporal series (encompassing the last century) of the Parana and Uruguay fluvial discharges (http://www.hidricosargentina.gov.ar/acceso_bd.php), river-flow anomalies were calculated following the approach of Piovano et al . (2004). The generated proxy data were analyzed by running cluster analyses using the stratigraphically constrained Moristia similarity index, in PAST program version 3 (http://folk.uio.no/ohammer/past/) . The generated groups are represented with red lines in figure 3. The sedimentation rate of coreGeoB 13813-4 was assumed to be constant with a mean value of 1.3 cm yr-1 (Table 1; Perez Becona, 2014), while for the sedimentation rate of core BAR1three groups were observed: 1911-1973; 1976-1984and 1986-2010. The mean sedimentation rate for the above groups showed an increasing trend from 0.24± 0.13 cm yr-1 to 0.31 ± 0.14 cm yr-1 and 0.37 ± 0.10cm yr-1, respectively. The most positive and stable values of SOI (LaNina events) were recorded during 1910-1970. After1970, a higher variability and a trend towards more negative values was observed (El Nino events). After the year 2005, very negative SOI values occurred (Fig. 4). PDO showed either negative or close to zero values during the early period 1910–1970 (cold phase). During the subsequent interval, i.e., 1970–2005, positive values (warm phase) were observed. Regarding with AMO, a positive phase was observed from 1925-1960, followed by a negative phase (1960-2000), but then a shift to a positive phase until the present was observed. The Parana river discharge anomalies for the years 1910–1970 were mostly associated with negative values (Fig. 4), while between 1970 and 2010 positive anomalies were documented. Between the years 2000 and 2010, we mostly registered values close to zero. Furthermore, the trends in AMO and SOI indices were negatively associated with the anomalies of both Parana and Uruguay rivers flows, while PDO index were positive associated with such anomalies. A change in mode polarity observed for PDO and AMO took place by the middle 1970s, in addition to more frequent and intense El Nino events that led to the increased rainfall over SESA (Garreaud et al ., 2009). Thus, the increase in rainfall over SESA was concomitant with positive anomalies inthe Parana and Uruguay river discharge rates after1970 (Camilloni, 2005). In this sense, the Parana river discharge was 20% higher during the past 30years than the historical average of the 20th century(Mauas et al ., 2008).The results of the cluster analyses groups (Fig. 3)showed a differentiation in both sediment cores that corresponds to the beginning of the 1970s, which could be associated with the increasing discharge trend recorded for the Parana and Uruguay rivers over the ast three decades. The increase in RdlP discharge led to a higher accumulation rate of terrigenous sediments, as inferred from the high sedimentation rate and mean grain size (BAR1), and the lowest Ca/Ti ratio (GeoB 13813-4), and explains both the spatial and temporal sedimentological and geochemical variability. Ca/Ti ratio in the RdlP was successfully used to infer marine vs. continental influence, as Ti is associated with a continental RdlP discharge, while Ca is associated with autochthonous marine productivity (e.g. foraminifera, Burone et al ., 2013).Thus, the highest continental sediment supply to the inner continental shelf is observed in GeoB 13813-4after 1970, associated with a decrease in the Ca/Ti ratio (Fig. 4). Regarding with Core BAR1 the grain-size distribution and the sedimentation rate were both associated with the estuarine hydrodynamic changes. After 1970, the highest and most variable sedimentation rate and mean grain size was found, probably associated to an increase in both the Parana and Uruguay river discharges during the past three decades, while the lowest and more stable sedimentation rates and mean sediment grain size recorded before 1970, is indicating a reduced RdlP freshwater supply to the study area. This study shows that both sediment cores contain a distinct continental runoff record as the result of climatic changes (PDO, AMO and ENOS), which have influenced the precipitation patterns over SESA. Both sites reflect similar responses to these environmental changes for the last 100 yr in continental terrigenous sediment supply from the RdlP watershed towards the inner continental shelf. We conclude that it is possible to assess the temporal RdlP discharge patterns variability within the estuarine and adjacent oceanic area through the study of terrigenous proxies from sediment cores retrieved within the continental shelf.