Plastic debris in lakes and reservoirs.

Plastic debris is thought to be widespread in freshwater ecosystems globally1. However, a lack of comprehensive and comparable data makes rigorous assessment of its distribution challenging2,3. Here we present a standardized cross-national survey that assesses the abundance and type of plastic debris (>250 µm) in freshwater ecosystems. We sample surface waters of 38 lakes and reservoirs, distributed across gradients of geographical position and limnological attributes, with the aim to identify factors associated with an increased observation of plastics. We find plastic debris in all studied lakes and reservoirs, suggesting that these ecosystems play a key role in the plastic-pollution cycle. Our results indicate that two types of lakes are particularly vulnerable to plastic contamination: lakes and reservoirs in densely populated and urbanized areas and large lakes and reservoirs with elevated deposition areas, long water-retention times and high levels of anthropogenic influence. Plastic concentrations vary widely among lakes; in the most polluted, concentrations reach or even exceed those reported in the subtropical oceanic gyres, marine areas collecting large amounts of debris4. Our findings highlight the importance of including lakes and reservoirs when addressing plastic pollution, in the context of pollution management and for the continued provision of lake ecosystem services.

Open and Cost-Effective Digital Ecosystem for Lake Water Quality Monitoring.

In some sectors of the water resources management, the digital revolution process is slowed by some blocking factors such as costs, lack of digital expertise, resistance to change, etc. In addition, in the era of Big Data, many are the sources of information available in this field, but they are often not fully integrated. The adoption of different proprietary solutions to sense, collect and manage data is one of the main problems that hampers the availability of a fully integrated system. In this context, the aim of the project is to verify if a fully open, cost-effective and replicable digital ecosystem for lake monitoring can fill this gap and help the digitalization process using cloud based technology and an Automatic High-Frequency Monitoring System (AHFM) built using open hardware and software components. Once developed, the system is tested and validated in a real case scenario by integrating the historical databases and by checking the performance of the AHFM system. The solution applied the edge computing paradigm in order to move some computational work from server to the edge and fully exploiting the potential offered by low power consuming devices.

Diversity and Cyclical Seasonal Transitions in the Bacterial Community in a Large and Deep Perialpine Lake.

High-throughput sequencing (HTS) was used to analyze the seasonal variations in the bacterioplankton community composition (BCC) in the euphotic layer of a large and deep lake south of the Alps (Lake Garda). The BCC was analyzed throughout two annual cycles by monthly samplings using the amplification and sequencing of the V3-V4 hypervariable region of the 16S rRNA gene by the MiSeq Illumina platform. The dominant and most diverse bacterioplankton phyla were among the more frequently reported in freshwater ecosystems, including the Proteobacteria, Cyanobacteria, Bacteroidetes, Verrucomicrobia, Actinobacteria, and Planctomycetes. As a distinctive feature, the development of the BCC showed a cyclical temporal pattern in the two analyzed years and throughout the euphotic layer. The recurring temporal development was controlled by the strong seasonality in water temperature and thermal stratification, and by cyclical temporal changes in nutrients and, possibly, by the remarkable annual cyclical development of cyanobacteria and eukaryotic phytoplankton hosting bacterioplankton that characterizes Lake Garda. Further downstream analyses of operational taxonomic units associated to cyanobacteria allowed confirming the presence of the most abundant taxa previously identified by microscopy and/or phylogenetic analyses, as well as the presence of other small Synechococcales/Chroococcales and rare Nostocales never identified so far in the deep lakes south of the Alps. The implications of the high diversity and strong seasonality are relevant, opening perspectives for the definition of common and discriminating patterns characterizing the temporal and spatial distribution in the BCC, and for the application of the new sequencing technologies in the monitoring of water quality in large and deep lakes.

Distribution of toxigenic cyanobacteria in Alpine lakes and rivers as revealed by molecular screening.

The increasing frequency of cyanobacteria blooms in waterbodies caused by ecosystem eutrophication could endanger human health. This risk can be mitigated by effective monitoring incorporating molecular methods. To date, most molecular studies on toxigenic cyanobacteria have been limited to microcystins (MCs), disregarding other cyanotoxins, to freshwater planktic habitats while ignoring benthic habitats, and to limited geographic areas (usually one or a few specific waterbodies). In this study, we used PCR-based methods including PCR product sequencing and chemical-analytical methods (LC-MS/MS) to screen many plankton (n = 123) and biofilm samples (n = 113) originating from 29 Alpine lakes and 18 rivers for their cyanotoxin production potential. Both mcyE (indicating MC synthesis) and anaC (indicating anatoxin (ATX) synthesis) gene fragments were able to qualitatively predict MC or ATX occurrence. The abundance of mcyE gene fragments was significantly related to MC concentrations in plankton samples (R2 = 0.61). mcyE gene fragments indicative of MC synthesis were most abundant in planktic samples (65 %) and were assigned to the genera Planktothrix and Microcystis. However, mcyE rarely occurred in biofilms of lakes and rivers, i.e., 4 % and 5 %, respectively, and were assigned to Microcystis, Planktothrix, and Nostoc. In contrast, anaC gene fragments occurred frequently in planktic samples (14 % assigned to Tychonema, Phormidium (Microcoleus), and Oscillatoria), but also in biofilms of lakes (49 %) and rivers (18 %) and were assigned to the genera Phormidium, Oscillatoria, and Nostocales. The cyrJ gene fragment indicating cylindrospermopsin synthesis occurred only once in plankton (assigned to Dolichospermum), while saxitoxin synthesis potential was not detected. For plankton samples, monomictic and less eutrophic conditions were positively related to mcyE/MC occurrence frequency, while oligomictic conditions were related to anaC/ATX frequency. The anaC/ATX frequency in biofilm was related to the lake habitats generally showing higher biodiversity as revealed from metabarcoding in a parallel study.

DNA sequence and taxonomic gap analyses to quantify the coverage of aquatic cyanobacteria and eukaryotic microalgae in reference databases: Results of a survey in the Alpine region.

The taxonomic identification of organisms based on the amplification of specific genetic markers (metabarcoding) implicitly requires adequate discriminatory information and taxonomic coverage of environmental DNA sequences in taxonomic databases. These requirements were quantitatively examined by comparing the determination of cyanobacteria and microalgae obtained by metabarcoding and light microscopy. We used planktic and biofilm samples collected in 37 lakes and 22 rivers across the Alpine region. We focused on two of the most used and best represented genetic markers in the reference databases, namely the 16S rRNA and 18S rRNA genes. A sequence gap analysis using blastn showed that, in the identity range of 99-100%, approximately 30% (plankton) and 60% (biofilm) of the sequences did not find any close counterpart in the reference databases (NCBI GenBank). Similarly, a taxonomic gap analysis showed that approximately 50% of the cyanobacterial and eukaryotic microalgal species identified by light microscopy were not represented in the reference databases. In both cases, the magnitude of the gaps differed between the major taxonomic groups. Even considering the species determined under the microscope and represented in the reference databases, 22% and 26% were still not included in the results obtained by the blastn at percentage levels of identity ≥95% and ≥97%, respectively. The main causes were the absence of matching sequences due to amplification and/or sequencing failure and potential misidentification in the microscopy step. Our results quantitatively demonstrated that in metabarcoding the main obstacles in the classification of 16S rRNA and 18S rRNA sequences and interpretation of high-throughput sequencing biomonitoring data were due to the existence of important gaps in the taxonomic completeness of the reference databases and the short length of reads. The study focused on the Alpine region, but the extent of the gaps could be much greater in other less investigated geographic areas.

Biogeography of bloom-forming microcystin producing and non-toxigenic populations of Dolichospermum lemmermannii (Cyanobacteria).

In the last decades, the cyanobacterium Dolichospermum lemmermannii showed an increasing spread to Southern Europe, raising serious concerns due to its ability to produce cyanotoxins. The widening of its geographic distribution and the observation of strains showing high optimum temperature underline its ecological heterogeneity, suggesting the existence of different ecotypes. To investigate its biogeography, new isolates from different European water bodies, together with strains maintained by the Norwegian Institute for Water Research Culture Collection of Algae, were genetically characterised for the 16S rRNA gene and compared with strains obtained from public repositories. Geographic distance highly influenced the differentiation of genotypes, further suggesting the concurrent role of geographic isolation, physical barriers and environmental factors in promoting the establishment of phylogenetic lineages adapted to specific habitats. Differences among populations were also examined by morphological analysis and evaluating the toxic potential of single strains, which revealed the exclusive ability of North European strains to produce microcystins, whereas the populations in Southern Europe tested negative for a wide range of cyanotoxins. The high dispersion ability and the existence of toxic genotypes indicate the possible spread of harmful blooms in other temperate regions.

Planktic Tychonema (Cyanobacteria) in the large lakes south of the Alps: phylogenetic assessment and toxigenic potential.

This work allowed assessing a widespread occurrence of Tychonema bourrellyi in the largest lakes south of the Alps (Garda, Iseo, Como and Maggiore). The taxonomy of the species was confirmed adopting a polyphasic approach, which included microscopic examinations, molecular (16S rRNA and rbcLX sequences) and (Lake Garda) ecological characterisations. Over 70% of the 36 isolates of Tychonema sampled from the four lakes tested positive for the presence of genes implicated in the biosynthesis of anatoxins (anaF and/or anaC) and for the production of anatoxin-a (ATX) and homoanatoxin-a (HTX). A detailed analysis carried out in Lake Garda showed strong ongoing changes in the cyanobacterial community, with populations of Tychonema developing with higher biovolumes compared to the microcystins (MCs) producer Planktothrix rubescens Moreover, the time × depth distribution of Tychonema was paralleled by a comparable distribution of ATX and HTX. The increasing importance of Tychonema in Lake Garda was also suggested by the opposite trends of ATX and MCs observed since 2009. These results suggest that radical changes are occurring in the largest lakes south of the Alps. Their verification and implications will require to be assessed by extending a complete experimental work to the other large perialpine lakes.

Variability of microcystin cell quota in metapopulations of Planktothrix rubescens: causes and implications for water management.

In this study, we investigated the relationships between microcystin (MCs) concentrations and the biovolumes of Planktothrix rubescens (BPr) in 2 natural lakes (Pusiano and Garda) and 2 artificially dammed reservoirs (Occhito and Ledro) in Italy. In all the considered water bodies, P. rubescens was the dominant cyanobacterium. All the lakes were characterized by significant relationships between MCs and BPr, with limited variability in the MC quota (the content of MCs per unit of biovolume) within each water body compared with the variability between sites. The results were consistent with the development of specific MC-genotypes, with moderate seasonal and spatial changes in the proportion between toxic and non-toxic strains. The MC cell quota obtained in our work (ECQ, Environmental Cell Quota) were in the same range of values computed on the basis of analyses made on environmental samples dominated by P. rubescens or Planktothrix agardhii, and on isolates of the same two species (<1 to over 10 µg mm(-3)). Besides the usual ordinary least square regressions, models have been evaluated by using quantile regression, a method that allows estimating the conditional median or other quantiles of the response variable. We showed that the use of quantile regressions has different advantages, which included the computation of MC quota based on the whole range of available data, the robustness against outliers, and the ability to estimate models also in cases where there is no or only weak relationships. The highest ECQ values estimated from 95% quantile regressions in specific water bodies might be used to estimate the worst-case MC concentrations from algal abundances. Nevertheless, it was stressed that a realistic assessment of toxicity and potential adverse health effects necessarily should take into account the toxicity potential of the more abundant MC-congeners produced by specific cyanobacteria populations.