Layout 1 INTRODUCTION The application of the multiproxy approach in pale- olimnological studies is becoming increasingly wide- spread as it allows to integrate the responses of lakes and their catchment and to disentangle the effects of combined human and climate impacts (Batterbee, 2000; Bennion et al., 2015). The combination of biological and geochemi- cal proxies allows tracking different aspects of specific impacts (Bennion et al., 2015; Perga et al., 2015), and overtaking the limits of the different studied proxies, such as Cladocera, diatoms, subfossil pigments and lithogenic elements (Rosen et al., 2010). Cladocera represent key players of the lake food web as they act between top-down regulators (fish and inver- tebrate predators) and bottom-up factors (nutrients and phytoplankton; Eggermont and Martens, 2011). In partic- ular, the study of Cladocera remains allows to reconstruct changes in lake trophic status (Jeppesen et al., 2001), water temperature (Nevalainen, 2012; Szeroczyńska, 2006) and pH (Jeziorski et al., 2008), water level (Korhola et al., 2005), macrophyte distribution (Davidson et al., 2007), and food web (Finney et al., 2000; Jeppesen et al., 2001). The studies by Manca et al. (2007) and Korosi et al. (2010) on changes in Cladocera body and appendages length over longer periods highlighted that size varies in relation to predators pressures, water temperature, nutri- ent, pH and Ca. These analyses resulted to be particularly useful when long-term data on fish and invertebrate pred- ators are not available, as well as to support information provided by other proxies. Geochemical analyses of lake sediments have been Advances in Oceanography and Limnology, 2016; 7(2): 220-234 ARTICLE DOI: 10.4081/aiol.2016.6399 This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). Combining sediment Cladocera remains and geochemistry to reveal the role of a large catchment in driving changes in a small subalpine lake (Lake Ledro, N-Italy) Manuela Milan,1* Richard Bindler,1 Monica Tolotti2 1Department of Ecology and Environmental Sciences, Umeå University, Linnaeus väg 6, 90187 Umeå, Sweden 2Sustainable ecosystems and bioresources, Research and Innovation Centre, Fondazione E. Mach, Via E. Mach 1, 38010 San Michele all'Adige (TN), Italy *Corresponding author: milan.manuela@gmail.com ABSTRACT Sediment Cladocera remains and geochemistry were analyzed at Lake Ledro, a small subalpine lake with a large catchment area lo- cated in northern Italy. The aim of the study was to investigate human, climate and hydrological impacts on the Cladocera community and on the geochemical components during the last few centuries. A sediment core was collected from the deepest point of Lake Ledro and radiometrically dated. Cladocera remains were analyzed to track the trophic lake evolution. The core bottom section revealed the dominance of Bosminidae in concomitance with nutrient pulses entering into the lake during major flood events. The abundance of species preferring cold water temperatures confirmed the deposition of this core section during the Little Ice Age. The flood event oc- curred in the first half of the 19th century produced a drastic increase in littoral species, due to the development of new habitats. The de- crease in Cladocera densities during the following lake stage was followed by a rapid increase in planktonic species during the nutrient enrichment after the 1960s. Statistical analyses revealed a clear response of Cladocera community to climate variability during olig- otrophic periods, while no relation to temperature changes was recorded during high nutrient levels. A preliminary study on Bosminidae and Daphnidae body size and appendages length was carried out to reconstruct major changes in the lake food web. Only Bosmina spp. revealed clear body size changes: minor shifts were recorded before the 1930s in relation to the low nutrient concentrations, while the major changes occurred during the 1980s were interpreted as related to the appearance of Cladocera invertebrate predators. Geochemical components were studied using X-ray fluorescence spectroscopy (XRF) analysis in order to recognize the impact of the large catchment area and from the lake-level regulations on the lake hydrology. Moreover the Si:Al ratios profile confirmed the increase in lake produc- tivity after the 1960s. Although both Cladocera and geochemical analysis indicate major changes since the 1960s, they also revealed diverse responses to common external and local forcing, thus confirming the value of a multi-proxy approach for disentangling the lake responses to different environmental stressors. Moreover, it outlined the importance of larger catchment areas on small lakes as they are to a larger extent influenced by the modifications occurring in the drainage basin. Key words: Cladocera; geochemistry; wavelenght-dispersive X-ray fluorescence spectroscopy; paleolimnology; hydrological impact; Bosmina morphology. Received: November 2016. Accepted: December 2016. No n c om me rci al us e o nly Sediment Cladocera and geochemistry of Lake Ledro 221 successfully used to reveal responses of lake ecosystems to catchment change (Battarbee, 2000). In fact, soil stabi- lization and erosion are strongly influenced by land-use (García-Ruiz et al. 2010), while physical and chemical weathering of the catchment area are affected by climatic variability (Vannière et al., 2013). The analysis of differ- ent geochemical components can provide information on lake productivity (e.g., Si:Al as a proxy of biogenic silica), redox conditions (i.e., Fe:Mn ratios), atmospheric pollu- tion (e.g., Pb concentration), land use (lithogenic ele- ments), climate change and weathering rates (K:Al ratios; Martín-Puertas et al., 2011). Several paleolimnological studies were conducted on subalpine lakes located on the northern slope of the Alps (Alefs and Müller, 1999, Berthon et al., 2013; Wessels et al., 1999), while only a few sediment records were studied on the southern side (Marchetto et al., 2004; Guilizzoni et al., 2006; Milan et al., 2015). As already anticipated by Battarbee (2000), the recent work on the largest Italian lake, Lake Garda (Milan et al., 2015), highlighted the ne- cessity to expand the study to other lakes of the Italian subalpine region in order to understand both the individ- ualistic response to common external forcing, and the im- portance of local factors in driving the lake dynamics at secular scale. Lake Ledro appeared to be the ideal site to begin this expansion as it is located very close to Lake Garda and its large drainage basin has a great influence on the lake hydrology. Several physical analyses were re- cently conducted on the sediments of this lake in order to understand the impact of flood events on the lake (Magny et al., 2012; Vannière et al., 2013; Simonneau et al., 2013b). However, the response of the biological communi- ties to long term changes is still unknown as sediments of Lake Ledro were never analyzed for biological proxies. The aim of this work was to reconstruct the long-term influence of human, climate and hydrological impacts on biological and geochemical components of Lake Ledro during the last few centuries. The multiproxy approach was applied in order to identify and disentangle the effects of the different impacts on the lake-catchment systems. This appears to be of particularly importance for this lake as its large drainage basin is related to major source of disturbance. Cladocera remains were identified in the sed- iment samples and related with independent environmen- tal and climatic variables in order to reconstruct the influence on the lake biological community exerted by major climate changes as occurred for example during the Little Ice Age, by the post-war nutrient enrichment and by major flood events. A preliminary study on the body size and appendages length was carried out in order to overtake the lack of long-term predator information. Finally, sediment samples were analyzed by X-ray fluorescence spectroscopy (XRF) for the reconstruction of the geochemical changes. The different elements and ratios have been compared in order to understand magnitude and impacts on the lake dynam- ics of major flood events and water-level regulations. METHODS Study site Lake Ledro is a small glacial lake (area: 3.7 km2, Vol: 0.08 km3, Zmax: 49 m) with a catchment area of ca. 111 km2, ranging from 2254 m asl down to 652 m asl (Fig. 1). Several stages of high flood frequency during the Holocene were outlined by previous studies on Lake Ledro sediments and interpreted in relation to combined effects of the torrential regime of the two temporary trib- utaries, Massangla and Pur rivers, the steepness of their valleys and the high ratio between catchment and lake area (30:1 ratio, Vannière et al., 2013; Simonneau et al., 2013b). Lake Ledro is situated very close to the northern extrem- ity of Lake Garda (Fig. 1), the largest Italian lake, to which it is connected through an underwater pipe located at 25 m depth in Lake Ledro. Water is forced through a pump-stor- age power plant built on the River Ponale down to Lake Garda, and then the water is pumped back up. The lake-level of Lake Ledro has been regulated for hydroelectricity pro- duction since AD 1929 (Vannière et al., 2013). Sediment coring and chronology A gravity corer (UWITEC, Austria) was used to col- lect a short core (83 cm) from the deepest point of Lake Ledro (45°52’44”N, 10°45’10”E) in December 2011. The core was vertically extruded and sliced in the laboratory at 0.5 cm intervals from 0 to 30 cm, and at 1 cm intervals from 31 cm down to the core bottom. Sediment aspect and texture were annotated during the slicing. The core chronology was established applying the CRS dating model (Appleby, 2001) to direct gamma assay radiometric analyses of210Pb, 226Ra, 137Cs and 241Am, which were con- ducted at ENSIS Ltd-University College London, UK. Geochemistry and subfossil pigments Wet density (WD), water content (H2O, measured from dry weight) and total organic matter (OM, measured as loss on ignition at 550°C) were determined for all the subsamples as described in Milan et al. (2015). About 0.2 g of dried and homogenized sediment were used to meas- ure major and trace elements using a Bruker S8 Tiger WD-XRF analyzer equipped with an Rh anticathode X- ray tube (further detail in Rydberg, 2014). The analysis showed accuracy within 10% and analytical precision within 5% for the majority of the elements. Concentra- tions of key groups (i.e., major elements, redox, litholog- No n c om me rci al us e o nly 222 M. Milan et al. ical and trace elements) and ratios between specific ele- ments are discussed in this study. In particular, the Mn:Fe ratio was used to infer changes in the sediment redox con- ditions, while the Si:Al ratio was used to infer changes in biogenic silica. The Pb enrichment factor (PbEF) was cal- culated in order to outline the atmospheric pollution in this region by using titanium (Ti) as the reference element. The value of Pb:Ti ratio in the deepest sediment layers was used as reference. The Zr:Ti ratio indicated the changes in grain size patterns and was used to understand the possible impacts of secular changes of hydrological dynamics the on the sediment quality (Boyle, 2000). The ratio K:Al was considered to identify mineral matter changes, since lower values generally reflect more-weath- ered mineral matter and viceversa (Kauppila and Salonen, 1997). All the stratigraphic plots were drawn with the soft- ware C2 version 1.7.2 (Juggins, 2007). Around 0.5 g of wet sediment of each subsample were extracted in 90% acetone overnight in the dark and under nitrogen atmosphere for photosynthetic pigments analy- sis, which were carried out at CNR-ISE (Verbania, Italy). After centrifugation (3000 rpm, 10 min) the sediment was removed and the obtained extract was used to quantify total carotenoids (TCar) by double beam spectrophotome- ter (SAFAS, UVmc2), and astaxanthin (Asta) by Reversed Phase High-Performance Liquid Chromatography using a Thermo Separation HPLC (Ultimate 3000). TCar was used to infer past total phosphorus concentration (Car-TP) according to Guilizzoni et al. (2011). Subfossil Cladocera Cladocera remains were analyzed every fourth sample along the core following the methods described by Sze- roczyńska and Sarmaja-Korjonen (2007). About 2 cm3 of wet sediment were treated with KOH (10%) and HCl (10%). Asafranin-glicerol mixture was added to the cleaned subsamples of 0.1 ml in order to facilitate the identification of the remains under an optical microscope (LEICA DM2500) at 100-400x magnification. All Clado- cera remains (headshield, shell, postabdomen, postab- dominal claws, mandible, caudal furca) were counted, and converted to number of individuals following Frey (1986). Taxonomical identification was based on Flössner (2000), Margaritora (1983) and Szeroczyńska and Sar- maja-Korjonen (2007). Three to six slides were counted for each sample in order to obtain a minimum of 100 Cladocera individuals (Kurek et al., 2010). This minimum was not achieved in a few samples with extremely scarce Cladocera remains. In order to preliminary explore the in- fluence of environmental variables on Cladocera body size changes, ca. 30 carapace, mucro and antennules of different Bosminidae species and 30 postabdominal claws of Daphnidae were measured in each sample (Korosi et al., 2008). Environmental variables and data analysis Homogenized monthly average air temperature data for the period 1870-2008 were obtained from the HISTALP webpage (2013) for the station Torbole-Riva del Garda, which is located at 5 km away from Lake Ledro. In this work only annual and seasonal average air temperature were considered, as previous statistical analy- ses (Milan, 2016) showed no relation between precipita- tion data and Cladocera. Information on lake nutrients and plankton were collected from the literature (Casellato, 1990; Boscaini et al., 2012) and from the Environmental Agency of the Autonomous Province of Trento (unpub- lished data). Past lake total phosphorus concentrations were in- ferred from both concentrations of total carotenoids (TCar, see above) and subfossil diatoms. Diatom-inferred total phosphorus concentrations (DITP) were recon- structed basing on a weighted-average regressions (WA) with inverse deshrinking, calibrated against the NW-Eu- ropean datasets (NW-Eu, Bennion et al., 1996). Further details on diatoms and pigment composition as well as TP reconstruction are available in Milan (2016). The optimal partitioning method based on the sum of squares criterion and implemented in the ZONE software (Lotter and Juggins, 1991) was used to identify the ho- mogenous Cladocera zones along the master core, while the number of significant zones was established through comparison with the broken stick model (Bennett, 1996). The binary logarithm-based Shannon Index (Shannon and Weaver, 1949) was applied to determine the diversity of Fig. 1. Study site including catchment area for Lake Ledro. Dot indicates the coring point. No n c om me rci al us e o nly Sediment Cladocera and geochemistry of Lake Ledro 223 Cladocera assemblages. The ratio of planktonic to littoral taxa was calculated for every analysed subsample, since planktonic species are expected to be dominant under warmer conditions and/or water level increase (Sarmaja- Korjonen, 2001). Bosmina longirostris (O.F. Müller) was excluded from this classification due to its capacity to live in both the pelagic and littoral zones (Szeroczyńska, 1998). Specific ecological preferences were defined as in Korhola (1990), Frey (1986) and Margaritora (1983). In order to identify patterns in the temporal evolution of Cladocera assemblages, a non-metric multidimensional scaling (NMDS, Kruskal and Wish, 1978) was applied to a Bray & Curtis dissimilarity matrix computed on the square root of Cladocera abundances. The NMDS analy- sis was performed with R 3.3.1 (R Core Team, 2016), vegan package version 2.4-1 (Oksanen et al., 2016), and after 20 trials the NMDS solution providing the lowest ‘stress’ (i.e. 0.16), which is measure of the configuration stability (Legendre and Legendre, 1998), was selected. A scree plot analysis was performed in order to determine the final number of NMDS dimensions to be considered (Legendre and Legendre, loc. cit.). For the identification of the major environmental, cli- matic and geochemical drivers for the studied proxy, vec- tor and surface fitting analyses were applied respectively to the sample scores of the NMDS configurations. Only variables significantly (P<0.05) correlated to sample score were considered in this work and plotted in the graph. The two analyses were computed using R 3.3.1 (R Core Team, 2016), vegan package version 2.4-1 (Oksanen et al., 2016). A LOWESS interpolation of the non-contiguous ra- dioisotopic ages was performed (R 3.3.1; R Core Team, loc. cit.) in order to attribute ages and sedimentation rates to each subsample within the 210Pb dated core section. A Kruskal-Wallis test was applied as non-parametric test for non-normal distributed variables to sediments sections characterized by homogeneous Clacodera body size, in order to verify the significance of different average sizes (Kruskal and Wallis, 1952). PAST software was used for the statistical analysis on the Clacodera body measure- ments (Hammer et al., 2010). RESULTS Core chronology 210Pb equilibrium was reached in the core of Lake Ledro at 28 cm depth, while 137Cs showed a well resolved peak at 21.25 (coincident with a peak in 241Am) and a minor one at 10.25 cm, corresponding to the fallout from atmospheric testing of nuclear weapons in 1963 and from the Chernobyl accident in 1986, respectively. The chronologies and the sediment accumulation rates were calculated by the CRS model using the 137Cs peak at 21.25 cm as reference level for 1963. The sediment accu- mulation rates (Fig. 2) showed a gradual increase in the last hundred years and a peak in the middle 1960s, which could be the result of either a sediment slumping or of the “century” flood event, which interested the north and cen- tral Italy in November 1966 (Malguzzi et al., 2006). The core showed major discontinuities at 40, 51, 59, 68 and 78 cm depth, which were interpreted as flood events by comparison with the structure of the core collected at 46 m depth in 2008 by Simonneau et al. (2013b). Other minor discontinuities detected in the core top 30 cm were related to minor floods (e.g., 1966) or to particularly rainy years (e.g., 1996, 2002, 2007). Geochemistry and subfossil pigments Wet density values oscillated in the core studied be- tween 1.3 and 1.6 g cm-3. The highest values recorded be- tween the core bottom and 28 cm (i.e., AD 1860±36) indicated high proportion of mineral matter during the Little Ice Age (Fig. 2). The upper 28 cm were character- ized by gradually decreasing sediment densities with two peaks, one in the 1990s and one in the 2000s. Water con- tent oscillated around ~50% of fresh weight (FW) from the core bottom to 55 cm depth, and increased thereafter up to >80% in the upper 5 cm (Fig. 2). Organic content slightly decreased from core bottom to 28 cm depth (from 14% to 8% of DW), while it irregularly increased in the upper core section and reached maximum values in the early 1970s (at ~18 cm) and in the second half of the 2000s (at ~3 cm; Fig. 2). The depth profile of astaxanthin (a marker for N-lim- ited cyanobacteria and aquatic invertebrates, including zooplankton) oscillated between ~2 and 30 nmol g-1 LOI throughout the core (Fig. 2). Total carotenoid con- centration were low in the deeper core sections, while they increased up to 3.4 U g-1 LOI during the 1980s and to ~4 U g-1 LOI in the early 2000s (Fig. 2). The past lake TP inferred on total carotenoid concentration (Car-TP, Fig. 2) oscillated between 5 and 8 µg L-1 from the core bottom up to ~20 cm depth (late 1960s), while the upper section showed two rapid increases up to mesotrophic level (>25 µg L-1), the first one in the 1970-1980s and the second one between the late 1990s to the beginning of the 2000s. These two maxima were separated by a sharp drop down to base line TP values in the 1990s (Fig. 2). Sedi- ment surface Car-TP concentrations (10-15 µg L-1) agreed with present mesotrophic status of Lake Ledro. The Car-TP pattern at Lake Ledro was highly compa- rable to the TP profile inferred from subfossil diatoms (DI-TP, Fig. 2). However, DI-TP concentrations reached in general higher values than Car-TP and showed en- hanced values also around 30 cm depth (i.e., during the middle 19th century). DI-TP values oscillated around 10 No n c om me rci al us e o nly 224 M. Milan et al. µg L-1from the bottom till the late 1960s (~20 cm depth), when they started to increase up to maximum values around 70 µg L-1 in the second half of the 1970s and the end of the 1980s, and to values of ~78 µg L-1 in the second half of the 2000s (Fig. 2). Subfossil Cladocera The 33 Cladocera taxa (6 planktonic and 27 littoral), which were identified in Lake Ledro belonged to the families Leptodoridae, Daphnidae, Bosminidae, Chydoridae, Cerco- pagidae. Three major zones were identified based on the Cladocera assemblages and abundances (Fig. 3). The first zone (LC1, 82.5-36.5 cm depth), was mainly characterized by the presence of B. longirostris, which was also responsible for the high values of total Cladocera abundance (Total) in the bottom core section. Bosmina (E.) longispina Leydig and a few individuals of Chydoridae were also found in this sec- tion. The sediment layer around 38.5 cm depth revealed the lowest Cladocera abundance value of the entire core, and in this core section only a few individuals of B. (E.) longispina, B. longirostris, Acroperus harpae (Baird) and Alona affinis (Leydig) were identified. In agreement with the species scarcity of this zone, the Shannon Index showed low values around 1. A rapid increase in littoral species, especially those connected with high water turbidity, such as Alona rectan- gula Sars and Chydorus sphaericus (O.F. Müller), or those associated with abundant detritus, like Disparalona rostrata Fig. 2. Depth profiles of geochemical and biological proxies in the Lake Ledro. Sed rate, sedimentation rate; WD, wet density (x-axis from 1 to 1.6); Water, water content; OM, organic content; Asta, astaxanthin (LOI, loss on ignition); TCar, Total carotenoids concen- trations; Car-TP, total phosphorus concentration reconstructed from total carotenoids levels; DI-TP, diatoms inferred total phosphorus concentrations. No n c om me rci al us e o nly Sediment Cladocera and geochemistry of Lake Ledro 225 (Koch), characterized the zone LC2 (36.5-29.25 cm, Fig. 3). High abundances of A. harpae and A. affinis were also found in this zone, while B. longirostris and B. (E.) longispina showed very low density in comparison to LC1. Cladocera diversity reached the highest values of the entire in this sec- tion (Shannon Index = 3). The zone LC3 (29.25-0.0 cm) was characterized by an increase in density and diversity of planktonic species (Fig. 3), which were used to define the two subzones LC3a and LC3b. LC3a (29.25-19.25 cm) showed a rapid decrease in total Cladocera abundance, while only spo- radic individuals of B. longirostris were identified. Daph- nia longispina O.F. Müller increased gradually, and Bosmina (E.) coregoni Baird appeared for the first time in this section (i.e., after the 1930s). Compared to the pre- vious zone, the Shannon Index decreased in association to the abundance of planktonic species. The dominance of planktonic and the sporadic presence of littoral species marked zone LC3b (19.25-0.0 cm). Planktonic species further increased since the early 1960s (19.25 cm) and culminated in the late 1980s, when they determined the peak in total Cladocera abundance. A few individuals of B. longirostris were identified at the beginning of LC3b subzone, while it completely disappeared in the surface layers. The cladoceran predators Bythotrephes longi- manus Leydig and Leptodora kindtii (Focke) appeared for the first time in this zone and reached their maximum val- ues at 10.25 and 12.25 cm, respectively. Cladocera diver- sity showed a further regular decrease since the beginning of the subzone LC3b (Fig. 3). The highest Cladocera abundance along the core largely corresponded to stages of higher astaxanthin concentrations (Fig. 2). Fig. 4 summarizes the results of the Cladocera meas- urements. No trend was observed for D. longispina (not represented in Fig. 4), while Bosminidae exhibited a major shift during the 1980s. In particular, antennulae of B. (E.) coregoni clearly increased after the 1960s (~20 cm depth), and in particularly during the 1980s, while cara- pace and mucro sizes increased only since the 1980s (~10 cm depth) and following a decreasing stage from the 1960s to the 1980s (Fig. 4). Antennules and mucro of B. (E.) longispina showed a relatively comparable trend, while carapace size clearly increased only after the 1980s. Despite the high abundance of B. longirostris especially Fig. 3. Depth profiles of: subfossil Cladocera total abundance (Total); ecological classification: Planktonic (white), B. longirostris (grey), Littoral (black); key species and Shannon Index. LC1-LC3b represented the homogeneous Cladocera zones. Species codes and full names of taxa are presented in Supplementary Tab. 2. No n c om me rci al us e o nly 226 M. Milan et al. in the deeper core sections, the remains were often broken and the measurement of the body parts was hard. As major changes in Bosmina body size occurred around ~10 cm depth, average sizes above and below this level were compared through the Kruskal-Wallys test (Fig. 4, Supplementary Tab. 1). Only the mucro of B. longispina were analyzed for three groups, i.e. between 0-10 cm, 10-26 cm and 26-82 cm. Size differences be- tween the two core sections were significant (P<0.05) for antennules and mucro of B.(E.) coregoni and for anten- nules and carapace of B.(E.) longispina. While differences between the three groups were significant (P<0.001) for mucro of B.(E.) longispina. Environmental variables and data analyses The scree plot analysis indicated two dimensional NMDS as sufficient to describe the diversity of subfossil Cladocera data. The analysis revealed an evident grouping of the planktonic species in the NMDS lower left quad- rant, while littoral species were concentrated in the two upper quadrants. B. longirostris, resulted to be separated from all the other species in the lower right quadrant (Fig. 5a), in agreement with its ecological characteristics (i.e., its capacity to live in both the pelagic and the littoral lake habitat). The separation of planktonic from littoral species along the first NMDS dimension (DIM1) corresponded to the species change which identified the separation of Cladocera subzones LC3a and LC3b along the core (Fig. 3). Moreover, the sample score on the second NMDS di- mension (DIM2, not shown) highlighted a drastic change at 38.5 cm depth, as already outlined by the Cladocera as- semblages (Fig. 3). The vector fitting outlined strong pos- itive relations of the Cladocera assemblages with water content (Water), Tcar, Car-TP and DI-TP (0.30