A catchment-scale model of river water quality by Machine Learning.

Water quality is a concern in most river basins worldwide due to the widespread release of pollutants which impacts the freshwater ecosystems. Exploring the relationships between drivers and water quality parameters at the regional scale is key in the identification of appropriate actions for the reduction of water pollution. Regional models are the appropriate tool to achieve this, though their development poses relevant challenges because of the complexity and non-linearity of such relationships. Among the available approaches, Machine Learning (ML) is promising because of its capability to detect complex nonlinear relationships and flexibility in the parameterization, which is learned from data. In this work, we developed regional models of water temperature, dissolved oxygen, arsenic, sulfate and chloride concentrations, as well as electrical conductivity, by using two ML algorithms, Random Forest and Deep feed-forward Neural Network, and compared their performances against the standard Linear Regression model. Our results indicate that the two ML algorithms are much more accurate models for such variables than the classical Linear Regression model, with Deep feed-forward Neural Network being the most effective in identifying the reciprocal importance of the drivers and capturing nonlinear relationships between drivers and water quality variables. Our analysis also revealed that the Julian day and year at which the sample was taken surrogate the air temperature in modeling water temperature and dissolved oxygen, with only a slight performance reduction. Arsenic, sulfate, and chloride show more complex behaviors in which geogenic and anthropogenic sources are intertwined. Dilution exerts a role chiefly for arsenic concentration, which suggests a non-uniform, in space, geogenic origin for this variable.

Priority and emerging organic microcontaminants in three Mediterranean river basins: Occurrence, spatial distribution, and identification of river basin specific pollutants.

There is a worldwide growing use of chemicals by our developed, industrialized, and technological society. More than 100,000 chemical substances are thus commonly used both by industry and households. Depending on the amount produced, physical-chemical properties, and mode of use, many of them may reach the environment and, notably, the aquatic receiving systems. This may result in undesirable and harmful side-effects on both the human and the ecosystem's health. Mediterranean rivers are largely different from Northern and Central European rivers in terms of hydrological regime, climate conditions (e.g. air temperature, solar irradiation, precipitation), and socio-economics (e.g. land use, tourism, crop types, etc.), with all these factors leading to differences in the relative importance of the environmental stressors, in the classes and levels of the pollutants found and their environmental fate. Furthermore, water scarcity might be critical in affecting water pollution because of the lowered dilution capacity of chemicals. This work provides raw chemical data from different families of microcontaminants identified in three selected Mediterranean rivers (the Sava, Evrotas, and Adige) collected during two sampling campaigns conducted in 2014 and 2015 in three different matrices, namely, water, sediments, and biota (fish). More than 200 organic micropollutants were analyzed, including relevant groups like pharmaceuticals, personal care products, perfluorinated compounds, pesticides, pyrethroid insecticides, flame retardants, and persistent organic pollutants. Data obtained were summarized with some basic statistics for all compound families and matrices analyzed. Observed occurrence and spatial patterns were interpreted both in terms of compound physical-chemical properties and local environmental pressures. Finally, their spatial distribution was examined and their ecotoxicological risk in the water phase was assessed. This allowed locating, at each basin, the most polluted sites ("hot spots") and identifying the respective river basin specific pollutants (RBSPs), prioritizing them in terms of the potential ecotoxicological risk posed to the aquatic ecosystems.

Particle bound pollutants in rivers: Results from suspended sediment sampling in Globaqua River Basins.

Transport of hydrophobic pollutants in rivers such as polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyls (PCBs) and heavy metals is often facilitated by suspended sediment particles, which are typically mobilized during high discharge events. Suspended sediments thus represent a means of transport for particle related pollutants within river reaches and may represent a suitable proxy for average pollutant concentrations estimation in a river reach or catchment. In this study, multiple high discharge/turbidity events were sampled at high temporal resolution in the Globaqua River Basins Sava (Slovenia, Serbia), Adige (Italy), and Evrotas (Greece) and analysed for persistent organic pollutants such as PAHs (polycyclic aromatic hydrocarbons) or PCBs (polychlorinated biphenyls) and heavy metals. For comparison, river bed sediment samples were analysed as well. Further, results are compared to previous studies in contrasting catchments in Germany, Iran, Spain, and beyond. Overall results show that loadings of suspended sediments with pollutants are catchment-specific and relatively stable over time at a given location. For PAHs, loadings on suspended particles mainly correlate to urban pressures (potentially diluted by sediment mass fluxes) in the rivers, whereas metal concentrations mainly display a geogenic origin. By cross-comparison with known urban pressure/sediment yield relationships (e.g. for PAHs) or soil background values (for metals) anthropogenic impact - e.g. caused by industrial activities - may be identified. Sampling of suspended sediments gives much more reliable results compared to sediment grab samples which typically show a more heterogeneous contaminant distribution. Based on mean annual suspended sediment concentrations and distribution coefficients of pollutants the fraction of particle facilitated transport versus dissolved fluxes can be calculated.

Hydrogeophysical characterization and monitoring of the hyporheic and riparian zones: The Vermigliana Creek case study.

The hyporheic and riparian zones are critical domains in a river ecosystem since they mediate the interactions between surface water and groundwater. These domains are generally strongly heterogeneous and difficult to access; yet their characterization and monitoring still rely mostly on hard-to-perform invasive surveys that provide only point information. These well-known issues, however, can be overcome thanks to the application of minimally invasive methods. In this paper, we present the results of the hydrogeophysical characterization of the Vermigliana Creek's hyporheic and riparian zones, performed at an experimental site in the Adige catchment, northern Italy, by means of electrical resistivity tomography (ERT), distributed temperature sensing (DTS), and hydrological modeling. A major advancement is given by the placement of electrodes and of an optical fiber in horizontal boreholes at some depth below the river bed, put in place via directional drilling. The results of this static and dynamic (time-lapse) geophysical characterization identify the presence of two subdomains (the sub-riverbed and the left and right banks) and define the water flow and solute dynamics. The ERT information is then used, together with other hydrological data, to build a 3D subsurface hydrological model (driven mainly by the watercourse stage variations) that is calibrated against local piezometric information. A solute transport model is then developed to reproduce the variations observed in the dynamic geophysical monitoring. The results show good agreement between ERT data and the model outcome. In addition, the transport model is also consistent with the temperature data derived from DTS, even though some slight discrepancies show that the heat capacity of the solid matrix and heat conduction cannot be totally neglected.

Uncertainty of modelled flow regime for flow-ecological assessment in Southern Europe.

Sustainable water basin management requires characterization of flow regime in river networks impacted by anthropogenic pressures. Flow regime in ungauged catchments under current, future, or natural conditions can be assessed with hydrological models. Developing hydrological models is, however, resource demanding such that decision makers might revert to models that have been developed for other purposes and are made available to them ('off-the-shelf' models). In this study, the impact of epistemic uncertainty of flow regime indicators on flow-ecological assessment was assessed at selected stations with drainage areas ranging from about 400 to almost 90,000km2 in four South European basins (Adige, Ebro, Evrotas and Sava). For each basin, at least two models were employed. Models differed in structure, data input, spatio-temporal resolution, and calibration strategy, reflecting the variety of conditions and purposes for which they were initially developed. The uncertainty of modelled flow regime was assessed by comparing the modelled hydrologic indicators of magnitude, timing, duration, frequency and rate of change to those obtained from observed flow. The results showed that modelled flow magnitude indicators at medium and high flows were generally reliable, whereas indicators for flow timing, duration, and rate of change were affected by large uncertainties, with correlation coefficients mostly below 0.50. These findings mirror uncertainty in flow regime indicators assessed with other methods, including from measured streamflow. The large indicator uncertainty may significantly affect assessment of ecological status in freshwater systems, particularly in ungauged catchments. Finally, flow-ecological assessments proved very sensitive to reference flow regime (i.e., without anthropogenic pressures). Model simulations could not adequately capture flow regime in the reference sites comprised in this study. The lack of reliable reference conditions may seriously hamper flow-ecological assessments. This study shows the pressing need for improving assessment of natural flow regime at pan-European scale.

Driver detection of water quality trends in three large European river basins.

This study analyses how indicators of water quality (thirteen physico-chemical variables) and drivers of change (i.e., monthly aggregated air temperature and streamflow, population density, and percentage of agricultural land use) coevolve in three large European river basins (i.e., Adige, Ebro, Sava) with different climatic, soil and water use conditions. Spearman rank correlation, Principal Component Analysis, and Mann-Kendall trend tests were applied to long-term time series of water quality data during the period 1990-2015 in order to investigate the relationships between water quality parameters and the main factors controlling them. Results show that air temperature, considered as a proxy of climatic change, has a significant impact, in particular in the Adige and Ebro: positive trends of water temperature and negative of dissolved oxygen are correlated with upward trends of air temperatures. The aquatic ecosystems of these rivers are, therefore, experiencing a reduction in oxygen, which may exacerbate in the future given the projected further increase in temperature. Furthermore, monthly streamflow has been shown to reduce in the Ebro, thereby reducing the beneficial effect of dilution, which appears evident from the observed upward patterns of chloride concentrations and electrical conductivity. Upward trends of chloride and biological oxygen demand in the Adige and Sava, and of phosphate in the Adige appears to be related to increasing human population density, whereas phosphates in the Sava and biological oxygen demand in the Ebro are highly correlated with agricultural land use, considered as a proxy of the impact of agricultural practises. The present study shows the complex relationships between drivers and observed changes in water quality parameters. Such analysis can represent, complementary to a deep knowledge of the investigated systems, a reliable tool for decision makers in river basin planning by providing an overview of the potential impacts on the aquatic ecosystem of the three basins.

River ecosystem processes: A synthesis of approaches, criteria of use and sensitivity to environmental stressors.

River ecosystems are subject to multiple stressors that affect their structure and functioning. Ecosystem structure refers to characteristics such as channel form, water quality or the composition of biological communities, whereas ecosystem functioning refers to processes such as metabolism, organic matter decomposition or secondary production. Structure and functioning respond in contrasting and complementary ways to environmental stressors. Moreover, assessing the response of ecosystem functioning to stressors is critical to understand the effects on the ecosystem services that produce direct benefits to humans. Yet, there is more information on structural than on functional parameters, and despite the many approaches available to measure river ecosystem processes, structural approaches are more widely used, especially in management. One reason for this discrepancy is the lack of synthetic studies analyzing river ecosystem functioning in a way that is useful for both scientists and managers. Here, we present a synthesis of key river ecosystem processes, which provides a description of the main characteristics of each process, including criteria guiding their measurement as well as their respective sensitivity to stressors. We also discuss the current limitations, potential improvements and future steps that the use of functional measures in rivers needs to face.

Contamination sources and distribution patterns of pharmaceuticals and personal care products in Alpine rivers strongly affected by tourism.

Knowledge regarding the impact of tourism on the emergence of pharmaceuticals and personal care products (PPCPs) in Alpine river waters is limited and scarce. Therefore, a study on the occurrence patterns and spatiotemporal variability of 105 PPCPs in an Alpine river basin located in the Trentino-Alto Adige region (North-Eastern Italy) has been conducted. We observed that the total concentration of analyzed PPCPs was generally higher in all sampling sites during winter than in the summer. The analysis of tourist data revealed that during both sampling campaigns the number of tourists was lower in the downstream sites in comparison with the upstream area of the basin (Val di Sole). Particularly, sampling sites located near important tourist resorts have shown the highest abundance of the PPCPs during winter, being analgesics/anti-inflammatories, antihypertensives and antibiotics the most abundant pharmaceutically active compounds (PhACs). Diclofenac showed the highest concentration amongst PhACs, reaching concentrations up to 675ngL-1 in the sampling site situated downstream of the Tonale wastewater treatment plant (WWTP). Antihypertensives were found at concentrations >300ngL-1, while antibiotics were quantified up to 196ngL-1, respectively. Amongst personal care products (PCPs), the most abundant compound was octyl-dimethyl-p-aminobenzoic acid (ODPABA) with concentrations reaching up to 748ngL-1 in the sampling site situated within the Rotaliana district. In general, concentrations and detection frequencies were higher in water than in the sediment samples. The most frequently detected PhACs in sediments from both sampling campaigns were antibiotics, while amongst PCPs in sediments, octocrylene (OC) showed the highest concentration in both sampling campaigns. As a result, this study highlights the potential impact of tourism on the water quality of the Alpine aquatic ecosystems.

Occurrence of halogenated and organophosphate flame retardants in sediment and fish samples from three European river basins.

Classic (polybromodiphenyl ethers, PBDEs) and emerging halogenated flame retardants (HFRs) such as decabromodiphenyl ethane (DBDPE) and halogenated norbornenes, as well as organophosphate flame retardants (OPFRs) were analysed in 52 sediments and 27 fish samples from three European river basins, namely the Evrotas (Greece), the Adige (Italy) and the Sava (Slovenia, Croatia, Bosnia and Herzegovina and Serbia). This is the first time that FR levels have been reported in these three European river basins. The highest contamination was found in the Adige and Sava rivers, whereas lower values were obtained for the Evrotas. The levels in sediment samples ranged between 0.25 and 34.0ng/g dw, and between 0.31 and 549ng/g dw, for HFRs and OPFRs respectively. As regards levels in fish, concentrations ranged between 9.32 and 461ng/g lw and between 14.4 and 650ng/g lw, for HFRs and OPFRs, respectively. Thus, whereas OPFR values were higher in sediments, similar concentrations (in the Evrotas) and even lower concentrations than HFRs (Sava) were found for OPFRs in the fish samples, indicating the lower bioaccumulation potential of OPFRs. Biota to sediment accumulation factors (BSAFs) were calculated and higher values were obtained for HFRs compared to those assessed for OPFRs.

Hydroclimatic and water quality trends across three Mediterranean river basins.

Water resources are under pressure from multiple anthropogenic stressors such as changing climate, agriculture and water abstraction. This holds, in particular, for the Mediterranean region, where substantial changes in climate are expected throughout the 21st century. Nonetheless, little attention has been paid to linkages between long-term trends in climate, streamflow and water quality in Mediterranean river basins. In the present study, we perform a comparative analysis of recent trends in hydroclimatic parameters and nitrate pollution in three climatologically different Mediterranean watersheds (i.e., the Adige, Ebro and Sava River Basins). Mann-Kendall trend analyses of annual mean temperature, precipitation and streamflow (period 1971 to 2010) and monthly nitrate concentrations, mass fluxes and flow-adjusted concentrations (period 1996 to 2012) were performed in these river basins. Temperature is shown to have increased the most in the Ebro followed by the Sava, whereas minor increases are observed in the Adige. Precipitation presents, overall, a negative trend in the Ebro and a positive trend in both the Adige and Sava. These climatic trends thus suggest the highest risk of increasing water scarcity for the Ebro and the lowest risk for the Adige. This is confirmed by trend analyses of streamflow time series, which indicate a severe decline in streamflow for the Ebro and a substantial decline in the Sava, as opposed to the Adige showing no prevailing trend. Concerning surface water quality, nitrate pollution appears to have decreased in all study basins. Overall, these findings emphasize progressive reduction of water resources availability in river basins characterized by continental climate (i.e., Ebro and Sava). This study thus underlines the need for adapted river management in the Mediterranean region, particularly considering strong feedbacks between hydroclimatic trends, freshwater ecosystem services and water resources availability for agriculture, water supply and hydropower generation.

Evaluation of genotoxic potential throughout the upper and middle stretches of Adige river basin.

In this study a comprehensive genotoxicological survey throughout the upper and middle stretches of Adige river basin is presented. The study was carried out at 7 sites located along the Adige main course and one the most significant tributaries, the Noce creek, both presenting different levels of pollution pressure. To give an insight into the nature of the genotoxic activity we employed the battery of prokaryotic and eukaryotic assays. Mutagenicity in water samples was evaluated by SOS/umuC test in Salmonella typhimurium TA1535/pSK1002. The level of DNA damage as a biomarker of exposure (comet assay) and biomarker of effect (micronucleus assay) and the level of oxidative stress as well (Fpg - modified comet assay) were studied in blood cells of Salmo cenerinus Nardo, 1847 and Salmo marmoratus Cuvier, 1829. Within the applied bioassays, comet assay showed the highest potential for discriminating the sampling sites which are under lesser extent of pressure (sampling sites 1-Barnes at Bresimo and 4-Noce downstream S. Giustina) from the sites under high pressure (sampling sites 5-Noce at Mezzolombardo and 6/7-Adige upstream and downstream municipality of Trento). Significant correlation between the standard and Fpg - modified comet assay indicated that oxidative stress could be a major contributor to observed DNA damage in collected specimens.

A review of hydrological and chemical stressors in the Adige catchment and its ecological status.

Quantifying the effects of multiple stressors on Alpine freshwater ecosystems is challenging, due to the lack of tailored field campaigns for the contemporaneous measurement of hydrological, chemical and ecological parameters. Conducting exhaustive field campaigns is costly and hence most of the activities so far have been performed addressing specific environmental issues. An accurate analysis of existing information is therefore useful and necessary, to identify stressors that may act in synergy and to design new field campaigns. We present an extended review of available studies and datasets concerning the hydrological, chemical and ecological status of the Adige, which is the second longest river and the third largest river basin in Italy. The most relevant stressors are discussed in the light of the information extracted from a large number of studies. The detailed analysis of these studies identified that hydrological alterations caused by hydropower production are the main source of stress for the freshwater ecosystems in the Adige catchment. However, concurrent effects with other stressors, such as the release of pollutants from waste water treatment plants or from agricultural and industrial activities, have not been explored at depth, so far. A wealth of available studies address a single stressor separately without exploring their concurrent effect. It is concluded that a combination of extended experimental field campaigns, focusing on the coupled effects of multiple stressors, and modeling activities is highly needed in order to quantify the impact of the multifaceted human pressures on freshwater ecosystems in the Adige river.

Soil-plant interaction monitoring: Small scale example of an apple orchard in Trentino, North-Eastern Italy.

Accurate monitoring and modeling of soil-plant systems are a key unresolved issue that currently limits the development of a comprehensive view of the interactions between soil and atmosphere, with a number of practical consequences including the difficulties in predicting climatic change patterns. This paper presents a case study where time-lapse minimal-invasive 3D micro-electrical tomography (ERT) is used to monitor rhizosphere eco-hydrological processes in an apple orchard in the Trentino region, Northern Italy. In particular we aimed at gaining a better understanding of the soil-vegetation water exchanges in the shallow critical zone, as part of a coordinated effort towards predicting climate-induced changes on the hydrology of Mediterranean basins (EU FP7 CLIMB project). The adopted strategy relied upon the installation of a 3D electrical tomography apparatus consisting of four mini-boreholes carrying 12 electrodes each plus 24 mini-electrodes on the ground surface, arranged in order to image roughly a cubic meter of soil surrounding a single apple tree. The monitoring program was initially tested with repeated measurements over about one year. Subsequently, we performed three controlled irrigation tests under different conditions, in order to evaluate the water redistribution under variable root activities and climatic conditions. Laboratory calibration on soil samples allowed us to translate electrical resistivity variations into moisture content changes, supported also by in-situ TDR measurements. Richards equation modeling was used also to explain the monitoring evidence. The results clearly identified the effect of root water uptake and the corresponding subsoil region where active roots are present, but also marked the need to consider the effects of different water salinity in the water infiltration process. We also gained significant insight about the need to measure quantitatively the plant evapotranspiration in order to close the water balance and separate soil structure effects (primarily, hydraulic conductivity) from water dynamics induced by living plants.

Impact of climate change and water use policies on hydropower potential in the south-eastern Alpine region.

Climate change is expected to cause alterations of streamflow regimes in the Alpine region, with possible relevant consequences for several socio-economic sectors including hydropower production. The impact of climate change on water resources and hydropower production is evaluated with reference to the Noce catchment, which is located in the Southeastern Alps, Italy. Projected changes of precipitation and temperature, derived from an ensemble of 4 climate model (CM) runs for the period 2040-2070 under the SRES A1B emission scenario, have been downscaled and bias corrected before using them as climatic forcing in a hydrological model. Projections indicate an increase of the mean temperature of the catchment in the range 2-4K, depending on the climate model used. Projections of precipitation indicate an increase of annual precipitation in the range between 2% and 6% with larger changes in winter and autumn. Hydrological simulations show an increase of water yield during the period 2040-2070 with respect to 1970-2000. Furthermore, a transition from glacio-nival to nival regime is projected for the catchment. Hydrological regime is expected to change as a consequence of less winter precipitation falling as snow and anticipated melting in spring, with the runoff peak decreasing in intensity and anticipating from July to June. Changes in water availability reflect in the Technical Hydropower Potential (THP) of the catchment, with larger changes projected for the hydropower plants located at the highest altitudes. Finally, the impacts on THP of water use policies such as the introduction of prescriptions for minimum ecological flow (MEF) have been analyzed. Simulations indicate that in the lower part of the catchment reduction of the hydropower production due to MEF releases from the storage reservoirs counterbalances the benefits associated to the projected increases of inflows as foreseen by simulations driven only by climate change.