Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing stress? A case study in Brazil

STOTEN-17823; No of Pages 11 Science of the Total Environment xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing stress? A case study in Brazil Julio C. López-Doval a,⁎, Sergio Tadeu Meirelles a, Sheila Cardoso-Silva b, Viviane Moschini-Carlos b, Marcelo Pompêo a a b

Institute of Biosciences, Department of Ecology, University of São Paulo, do Matão Str., Travessa 14, 321, Butantã, 05508-090 São Paulo, SP, Brazil São Paulo State University — UNESP “Júlio de Mesquita Filho”, Environmental Sciences Program, 3 de Março Avenue n. 511, PO Box: 18087-180, Sorocaba, SP, Brazil

H I G H L I G H T S • • • • •

Data on water quality of urban reservoirs has been statistically analyzed. Analysis was focused on community and bioassay responses and stressors. Data showed spatial and temporal lacks and non-compliances for some parameters. Bioassays and planktonic community demonstrated impairments in biota. Physical and chemical data compiled do not completely explain biological responses.

a r t i c l e

i n f o

Article history: Received 25 March 2015 Received in revised form 19 May 2015 Accepted 20 May 2015 Available online xxxx Editor: D. Barcelo Keywords: Freshwater monitoring Toxic units Chemometrics Urban reservoirs Tropical reservoirs Brazil

a b s t r a c t The Metropolitan Region of São Paulo (MRSP) is located in the Brazilian State of São Paulo and reservoirs in this region are vital for water supply and energy production. Changes in economic, social, and demographic trends produced pollution of water bodies, decreasing water quality for human uses and affecting freshwater populations. The presence of emerging pollutants, classical priority substances, nutrient excess and the interaction with tropical-climate conditions require periodic reviews of water policies and monitoring programs in order to detect and manage these threats in a global change scenario. The objective of this work is to determine whether the monitoring program of the São Paulo's Environmental Agency, is sufficient to explain the toxicological and biological responses observed in organisms in reservoirs of the MRSP, and whether it can identify the possible agents causing these responses. For that, we used publicly available data on water quality compiled by this agency in their routine monitoring program. A general overview of these data and a chemometric approach to analyze the responses of biotic indexes and toxicological bioassays, as a function of the physical and chemical parameters monitored, were performed. Data compiled showed temporal and geographical information gaps on variables measured. Toxicological responses have been observed in the reservoirs of the MRSP, together with a high incidence of impairments of the zooplankton community. This demonstrates the presence of stressors that affect the viability of organisms and populations. The statistical approach showed that the data compiled by the environmental agency are insufficient to identify and explain the factors causing the observed ecotoxicological responses and impairments in the zooplankton community, and are therefore insufficient to identify clear cause–effect relationships. Stressors different from those analyzed could be responsible for the observed responses. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Preservation of the structure and function in freshwater ecosystems is needed to ensure economic, cultural, and recreational benefits for humans (Turner and Daily, 2008). For this reason, the identification ⁎ Corresponding author. E-mail address: [email protected] (J.C. López-Doval).

and control of potential threats that might impair freshwater ecosystems are required in order to ensure consistent quality of the water supply. Chemical pollution in water bodies is a well-known factor that decreases water quality for human uses and can affect freshwater populations in terms of their abundance, distribution, and interactions with other organisms and the environment, consequently causing effects on the ecosystem (Brack et al., 2005). Other environmental factors

http://dx.doi.org/10.1016/j.scitotenv.2015.05.085 0048-9697/© 2015 Elsevier B.V. All rights reserved.

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

2

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

(nutrient excess, species interactions, introduction of invasive species, climate variations, habitat loss, physical alterations of the river, or overexploitation of water resources) can also be responsible for changes in freshwater ecosystems (Ricciardi et al., 2009). For this reason, the diagnosis, prediction, and forecasting of the impacts of toxic substances requires their discrimination from other stresses in order to obtain reliable cause–effect relationships between chemical pollution and impacts on freshwater populations. The use of statistical methods is a powerful and inexpensive method for the identification of significant relationships and possible explanation of cause–effect linkages (Brack et al., 2005), but requires strong data sets obtained in monitoring programs. Monitoring programs are used to obtain information concerning the quality of water bodies, using measurements of chemical, physical, hydromorphological, biological, or other related parameters. Generally, routine monitoring of water bodies is performed for specific purposes, such as assessing the conformity of measured parameters to quality standards for human uses (recreational, agricultural, industrial, and tap water) and/or quality standards intended to protect the diversity and functionality of aquatic ecosystems. A global change scenario entails the appearance of new stressors, such as the alteration of climatic patterns (including drought and temperature increase) or changes in economic, social, and demographic trends (Stevenson and Sabater, 2010; Carpenter et al., 1992). All these new stressors and changes imposes new challenges for management of the quality and availability of water resources, as well as the conservation of aquatic ecosystems, and requires periodic reviews of water policies and monitoring programs (Araújo et al., 2014; Pahl-Wostl, 2007). In developing countries, biodiversity and water security are likely to be threatened, if no efforts are made to protect water resources (Vorosmarty et al., 2010). Apart from changes in climatic patterns, South America (and especially Brazil as an emerging economy) is facing various other challenges: a) fast growth of populations in urban areas (Varis et al., 2006); b) development of crop areas and high consumption of water, pesticides, and fertilizers (Barletta et al., 2010; Carvalho, 2006; Agostinho et al., 2005; Tundisi, 2003); c) growth of the cattle industry and high levels of water consumption and land use change (Bell et al., 2010); d) development of industry, with weak environmental legislation (Agostinho et al., 2005; Tundisi, 2003); and e) poor sewage treatment infrastructure (Barletta et al., 2010; Agostinho et al., 2005; Tundisi, 2003). All these factors will diminish ecological quality, and increase water consumption and the cost of ensuring a clean and consistent source of water to supply the needs of citizens and economic activities. The economic, social, and demographic changes in this region, and elsewhere around the world, imply the emergence of new chemical threats to freshwater resources, involving new emerging pollutants (Richardson and Ternes, 2014; Petrovic et al., 2011), as well as the classical and well-known priority substances. Therefore, the review of water management policies requires studies aimed at defining and prioritizing the chemical substances that could pose the greatest risks to humans and ecosystems, as well as studies focused on the adaptation of monitoring procedures to enable the detection and quantification of these new substances. Such studies have mainly been performed in the USA and the European Union (Fàbrega et al., 2013), and are scarce in South America. Economic development in South America implies increasing requirements for energy production and water storage. This motivated the construction of a large number of reservoirs in the region during the 20th and 21st centuries, which provide recreational services and play an important role in economic development (Tundisi et al., 1998). Despite of their importance, reservoirs, and in general, water bodies in South America are suffering anthropic pressures which diminish water quality and ecological status (Barletta et al., 2010; Carvalho, 2006; Agostinho et al., 2005; Tundisi, 2003). In addition, some of the

South American reservoirs are under the influence of a tropical climate (Tundisi, 1990), with conditions that enhance primary production (high temperature, high and constant solar irradiation, frequent thermocline instability) but that can cause problems for aquatic organisms, due to rapid oxygen depletion in the bottom, cyanobacteria blooms, and rapid changes in water column stability (Lewis, 1987; Tundisi, 1990). For example, in Brazil, environmental agencies and researchers have reported that the biological and chemical quality of many reservoirs under urban influence are highly impacted by inflows of untreated wastewater containing high levels of pollutants, including nutrients and organic matter, leading to the eutrophication of these water bodies as well as decreased biological diversity (Braga et al., 2006; CETESB, 2013; Fontana et al., 2014). These evidences suggest that the organisms inhabiting these reservoirs are exposed to multiple stressors and policies improving management should be applied (Tundisi and Matsumura-Tundisi, 2003). The objective of this work is to determine whether the data compiled by water agencies, in addition to establishing compliance (or not) with national laws, are sufficient to be able to explain the toxicological and biological responses observed in organisms in reservoirs, and whether these data can be used to identify the possible agents causing these responses. In other words, we want to know if the data compiled in the monitoring programs can identify stressors, or if it only provides indicators of stress. To accomplish this objective, as case study, we used publicly available data on water quality compiled by a Brazilian environmental agency (Environmental Sanitation Technology Company, CETESB) in its routine monitoring program in the Metropolitan Region of São Paulo. Firstly, we performed a general overview of these data, and then we applied a chemometric approach to analyze the responses of biotic indexes and toxicological bioassays, compiled by the environmental agency, as a function of the physical and chemical parameters monitored. Identification of possible stressing agents is a first step for the creation of policies aimed at the elimination or mitigation of these stressors, with the possibility of restoration of the ecosystem, or at least the avoidance of its total degradation. This is the first time that such a study has been performed in South America, and the results and conclusions should be useful to water managers and policymakers involved in the protection of water resources and ecosystems. 2. Materials and methods 2.1. Study site In this study, we analyzed data obtained by CETESB in the monitoring program of urban reservoirs under the influence of the Metropolitan Region of São Paulo (MRSP, see Fig. 1). The MRSP is the largest urban and industrial complex in South America and it is located in the Brazilian State of São Paulo (Braga et al., 2006). The MSRP has experienced dramatic and unplanned growth in the last 50 years, and now includes the city of São Paulo and a further 39 municipalities, with a total population of around 21,000,000 inhabitants and an area of 8000 km2. The main land use in the MRSP is for urban and industrial purposes (Braga et al., 2006; Drucrot et al., 2005; Formiga-Johnsson and Kemper, 2005; SEADE, 2015), and the reservoirs in this region are vital for water supply and energy production (Tundisi et al., 1998; Ducrot et al., 2005). These reservoirs are under the influence of a tropical climate, with conditions that enhance primary production (high temperature, high and constant solar irradiation, frequent thermocline instability) but that can cause problems for aquatic organisms, due to rapid oxygen depletion in the bottom, cyanobacteria blooms, and rapid changes in water column stability (Lewis, 1987; Tundisi, 1990). In addition, environmental agencies and researchers have reported that the biological and chemical quality of many reservoirs under the influence of the MRSP is highly impacted by inflows of nutrients and toxicants, eutrophication

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

3

Fig. 1. Metropolitan region of São Paulo and reservoirs included in the study. Image adapted from CETESB.

and decreased biological diversity (Braga et al., 2006; CETESB, 2013; Fontana et al., 2014). In the State of São Paulo, the organ responsible for environmental surveillance is the Environmental Sanitation Technology Company (CETESB), a public organization reporting to the São Paulo State Government, whose remit is to ensure compliance with environmental quality standards in water bodies and other environmental compartments. The quality standards are laid down in Federal and State laws, and aim to ensure the quality of water used for human purposes (CONAMA 357/ 2005). CETESB has monitored chemical and physical water quality parameters in rivers and reservoirs for more than 10 years and could be considered a referential environmental agency in Brazil and South America (Cardoso-Silva et al., 2013). 2.2. Databases and data used for analysis Data used were open access (http://www.cetesb.sp.gov.br/agua/ aguas-superficiais/35-publicacoes-/-relatorios) and were compiled from the CETESB annual water quality reports for the years from 2008 to 2012. The data related to 17 sampling points, distributed over 12 reservoirs surrounding the MRSP. The reservoirs in this area provide water to the metropolitan centers. At each sampling point, CETESB measured a maximum of 45 variables, at least 4 times a year. From the original data, a matrix was constructed with 425 rows and 45 variables. From this primary data set, sampling points were selected based on the availability of data relating to toxicological responses and/or the quality of the biological community (13 sampling points, spread over 10 reservoirs). The bioassay with Ceriodaphnia dubia was chosen as the toxicological response, and the zooplankton community index (ZCI) calculated by CETESB was used as an indicator of the quality of the biological community. Two different matrices were built: one with data recorded for the toxicological responses of

C. dubia tests (TOX matrix) and another with data recorded for the quality of the zooplankton community, based on the ZCI (ZCI matrix). Both variables were used because since the toxicological bioassay is performed with a zooplanktonic organism, possible explicative relationships were expected between the results obtained in the toxicological tests and the values obtained in the ZCI procedure. Toxicological tests with C. dubia were conducted by CETESB, exposing the organism to water sampled at the monitoring point as described by CETESB (1991). In the CETESB reports, toxicological responses in C. dubia are reported as “no effect”, “chronic effect”, and “acute effect”. For the statistical analysis, these results were transformed to an ordinal variable by applying these criteria: “no toxicity” = 0, “chronic toxicity” = 1, and “acute toxicity” = 2, based on the criterion that a higher concentration of toxic substances would cause a greater response in a shorter exposure time. The ZCI is a community index calculated by CETESB based on the trophic state index (TSI) and the presence of Copepoda, Rotifera, and Cladocera, and the taxonomic composition of the copepod community sampled at the monitoring point (Coelho-Botelho et al., 2006). In the ZCI, quality states, or scores, are expressed using an ordinal scale (good, regular, bad, and very bad quality). The results were transformed to an ordinal variable by applying the criteria: good = 1, regular = 2, bad = 3, and very bad = 4. The category “good” was not found for the data used here. From the original number of environmental variables, 11 were finally selected. The criteria used for selection were that: a) the variables should reflect, or related to, toxicological hazards and/or potential environmental impairment, and b) data should be available for different sampling points and during the selected years. The variables included were the following: pH; conductivity; oxygen concentration; concentrations of ammonia, nitrate, nitrite, and phosphorous; chlorophyll a concentration; numbers of cyanobacteria and E. coli; and metal

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

4

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

concentrations expressed as toxic units (TUs). TUs and the concentrations of ammonia, nitrate, nitrite (Alonso and Camargo, 2006; Camargo and Alonso, 2006; Soucek and Dickinson, 2012), and cyanobacteria are considered as variables related to toxic responses; while pH, conductivity, oxygen concentration; concentration of phosphorous, chlorophyll a concentration and colonies of E. coli are related to the environmental impairment. The values of the environmental variables corresponded to those recorded for the same date and sampling point as the ZCI values or toxicological responses. CETESB monitors concentrations of microcystins, but not at all the points included in this study, so to avoid data reduction, the concentration of cyanobacteria was included as a variable related to the toxicological risk due to cyanotoxins. Although not all species produce cyanotoxins and cause the risk of toxic effects, generally freshwater blooms of cyanobacteria produce toxic secondary metabolites (Chorus, 2001). Toxic cyanobacteria, and their respective cyanotoxins, have been detected in São Paulo State, and toxic effects in freshwater organisms have been described (Sotero-Santos et al., 2006; Moschini-Carlos et al., 2009; Piccin-Santos and Bittencourt-Oliveira, 2012). The toxic units (Sprague, 1970) were calculated by the authors and used as a derivative variable originated from the concentrations in water of metals analyzed by CETSB in the monitoring program. We used this derivative variable in order to obtain a single variable more focused on the total potential toxicity of the mixture of metals in the water, rather than considering the concentrations of individual metals and their maximum permissible levels. TUs have been recommended by the European Commission (2011) for the risk assessment of chemical mixtures. Due to the absence of data for other chemical components (such as organic compounds; see Results section), only metals were considered. The TU values were calculated according to Sprague (1970) and Von der Ohe (2009), as the sum of the environmental concentrations (Ci) of n individual metals in the mixture, expressed as a fraction of their corresponding LC50 values for the organism Daphnia magna. Values ≥ 1 imply that a toxicological risk exists. CETESB does not analyze the dissolved fractions of all metals; for metals where dissolved concentrations were available we used that value. When the concentration was given as the total amount of metal, the dissolved fraction was estimated by applying the criterion of Pelletier (1996). TUs were calculated for each sampling period at each sampling point, but only metals that were detected above the limit of detection at least once in the same year were included, and the same criterion was applied to the other variables. When the concentration of an individual metal (or other parameter) was below the limit of detection for a particular sampling period, but compliance was obtained with the previous criterion, the value was considered half of the limit of detection, as stipulated in EU Directive 2009/90. The LC50 values for the individual metals were obtained from the Toxic Unit Calculator (http://www.systemecology. eu/spear/tu-calculator/). For the other variables reflecting toxicological hazards or potentially environment-impairing factors, the environmental quality standards (EQS) applied were those proposed by the Brazilian Federal Government for protection of water bodies of special interest (CONAMA 357, class I water bodies). The most restrictive values were applied to ensure higher environmental standards for the development of biota. 2.3. Chemometric analysis: principal component analysis (PCA) and analysis of variance (ANOVA) The chemometric analysis was performed in two steps, applying PCA followed by ANOVA. Firstly, a PCA based on correlations was run for the environmental data in the ZCI and TOX matrices respectively in order to reduce the high number of environmental variables to a reduced number of components (the principal components, PCs) that maximized the explained variability. After the PCA, two new matrices were obtained for each analysis: the score matrix, whose values represented the position of each sampling point in the new dimensional arrangement, based on

the linear combination of original variables and describing the distribution of variables between samples; and the loading matrix, with information on the weight (or correlation) of each environmental variable in the PCs obtained. Finally, an output was obtained with the percentage of variability explained by each new PC. For the PCA, the environmental data in the TOX and ZCI matrices were transformed (log x + 1) and each variable was scaled using xi − min.var/max.var − min.var, where xi is the value of a variable for a sample, and max.var and min.var are the maximum and minimum values achieved in the variable. After the PCA, two different ANOVA analyses were performed with the values obtained in the two PCA score matrices. The PCs were used as the independent variables, with the ZCI values and the toxicological responses as factors in the corresponding ANOVA tests. Since the score matrix provides information on the arrangement of the sampling points in the new spatial representation, based on the linear combination of the original variables, the objective of these ANOVA tests was to determine whether the PCs provided satisfactory grouping of the sampling points in the new spatial arrangement, considering the pattern of the categories or responses obtained for the ZCI and the bioassays (see Fig. 2). Post-hoc analyses (Tukey's test, p b 0.05) were performed in order to determine the discriminating power of each PC for the different categories of the ZCI and toxicological analysis, by means of significant differences. The loading matrix gives information on the weight of each original variable in the new PC, enabling identification of the variable with the highest explicative value in the significant PCs. Correlation between the ZCI and toxicological responses was studied by means of Spearman's correlation (p b 0.05, with 90 pairs analyzed). This analysis was performed in order to know whether the toxicological responses observed in tests with C. dubia could explain bad quality in the zooplankton community, and therefore whether the toxicological tests could have any explicative or predictive value in the ZCI. The PCA was performed with the free program PAST, while ANOVA and Spearman's correlation were performed with STATISTICA 7.

3. Results 3.1. Data overview As explained in the annual report (CETESB, 2013), the general monitoring program of water bodies includes 9 physical parameters, 52 chemical parameters, 3 microbiological parameters, 4 hydrobiological parameters, 3 toxicological tests, and 3 community indexes. Unfortunately, data for all the parameters were not always available, or were not equally sampled at all sampling points (even in the same reservoir). Such temporal and spatial disparity complicates data comparison. It was therefore necessary to eliminate from the analysis those variables and sampling points for which data were insufficient. During the selected period, pesticides were not analyzed for any reservoir. In year 2007, a one-year intensive analysis was performed for one sampling point, including the 83 chemical compounds that appear on the list of CONAMA 357/2005 (including pesticides, polyaromatic hydrocarbons, and persistent organic pollutants); these compounds were below the limits of detection for all sampling periods. After 2007, only three sampling points were used to monitor the presence of a few organic compounds (toluene, benzene, styrene, ethylbenzene, and xylene), until 2010. Phenolic compounds are currently analyzed, but not at all sampling points. Not all of the 11 metals included in the protocol were measured at all sampling points. The analytical methods were not the same for all the metals, with the total amount or the dissolved fraction being analyzed, depending on the metal. In addition to the bioassays with C. dubia, other toxicological tests were used: the Ames test (performed for 32% of the samples considered in the present study) and Microtox bioassays (performed for 13% of the samples). ZCI was evaluated for 25% of the samples. ZCI and C. dubia tests were only performed simultaneously for five sampling points during the study period.

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

5

Fig. 2. Flowchart of the methodology and data analysis followed in this work.

3.2. Toxic units and other toxicological and environmental risk indicators Out of the 272 cases in the TOX matrix, 28.7% showed chronic responses and 4% showed acute effects. Considering the variables with toxicological relevance, the number of cyanobacteria was the variable with the highest number of non-compliances, followed by the concentration of ammonia. The TU values were ≥ 1 in 3.7% of the cases (Table 1). Out of the 106 cases in the ZCI matrix, 14.1% showed regular, 77.4% bad, and 8.5% very bad scores. Considering the variables with environmental and toxicological relevance the number of cyanobacteria, concentration of total phosphorous and the concentration of dissolved oxygen were the variables with higher number of non-compliances with EQS. The TU values were ≥1 in 3.8% of the cases (Table 2).

3.3. Chemometric analysis and correlation The ZCI and toxicological responses showed a poor but positive and significant correlation (ρ = 0.231; p b 0.05), indicating that the greater the toxicological response, the worse the quality of the ZCI, but maybe other factors are involved in ZCI bad status. Sampling points without toxicological responses reported were associated mainly with “regular” and “bad” ZCI status, while sampling points with “acute effects” were only associated with “bad” ZCI status and “chronic effects” were associated mainly with “bad” and “very bad” status (Table 3). Table 1 Number of non-compliances found in TOX matrix. Thresholds based on CONAMA 357 for class I water bodies, except TU. Nutrients in mg/L.

Threshold % of non-compliances

pH

NH3

NO3

NO2

Number cyano.

TU

6bx≤9 5.1

≤0.5 23.5

≤10 0

≤1 1.1

≤20,000 56.3

b1 3.7

Table 4 summarizes the variance explained by the different PC obtained in PCA with both matrices and Table 5 summarizes the loadings obtained. 3.3.1. Analysis with the zooplankton community In the case of the matrix for the ZCI data, the PCs that significantly discriminated between the quality states after ANOVA were PC1, PC2, and PC3. Only PC2, which explained 21.5% of the variance, was able to discriminate between all the quality states of the ZCI index (Tukey's test; p b 0.05). In this PC, the variables with highest weights were the concentration of Chl. a (0.48), conductivity (0.44), and total P (0.37) (Table 6). 3.3.2. Analysis using toxicological responses in C. dubia In the case of the TOX matrix, the PCs that significantly discriminated between the toxicological responses after ANOVA were PC2, PC3, PC5, and PC9. PC2 explained 22.21% of the variability, while PC3 explained 9.92%. PC5 and PC9 explained minor percentages (Table 7). Tukey's test (p b 0.05) revealed that PC2 discriminated between “toxicity” (states 1 and 2) and “no toxicity” (state 0), while PC3 discriminated between “acute effect” (state 2) and the other states. The variables with highest weights in PC2 were the oxygen concentration and pH, with values of 0.49 and 0.44, respectively. In PC3, TU showed a high weight (0.80). 4. Discussion In this study, a general overview is given of data compiled by CETESB over a period of five years for reservoirs of the MRSP. In addition, chemometric analysis was used to determine whether the environmental data compiled by CETESB could identify the agents that caused the toxicological responses observed in the bioassays and the changes observed in the zooplankton community and, therefore, provide information to water managers to assist in the implementation of policies aimed

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

6

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

Table 2 Number of non-compliances detected in ZCI matrix. Thresholds based on CONAMA 357 for class I bodies, except TU. Nutrients and oxygen concentration in mg/L.

Threshold % of non-compliances

pH

Dis. O2

Total P

NH3

NO3

NO2

Number cyano.

TU

6bx≤9 8.5

≥6 29.2

≤0.025 73.6

≤0.5 34

≤10 0

≤1 0

≤20,000 67

b1 3.8

at reducing or eliminating the observed impacts. Data compiled by CETESB for the studied sampling points showed temporal and geographical information gaps. Potentially toxic substances such as organic compounds were rarely analyzed, and metals were not equally analyzed for all sampling points. It is likely that for these reasons, the chemometric analysis revealed that the data compiled by CETESB were not sufficient to be able to explain the toxicological responses in C. dubia. Likewise, the chemometric analysis revealed that “hidden factors” or methodological artifacts could be responsible for the responses obtained in the zooplankton community at the studied sampling points. Multivariate statistical methods provide powerful tools for the analysis and interpretation of large environmental data sets generated during monitoring programs. They can be useful in obtaining explicative and causative patterns explaining the biological and toxicological responses. Chemometric analysis and multivariate statistics have been applied before in multiple stressor scenarios, in order to describe patterns of pollution, provide ecotoxicological risk assessment, and identify priority stressors, with satisfactory results (García-Reiriz et al., 2014; Ricart et al., 2010; Muñoz, 2009; Peré-Trepat et al., 2007; Terrado et al., 2006; Tauler et al., 2004). This is the first time that such a study has been performed in South America, and the results and conclusions could be useful to water managers and policymakers involved in the protection of water resources and the maintenance or improvement of water quality. Other analytical approaches employing CETESB data have been adopted previously. Bertholdo et al. (2012) used a statistical approach to analyze environmental data provided by CETESB, in order to obtain a tool able to predict the toxicological risk at a given sampling point as a function of the values of the environmental variables. Cunha et al. (2013) performed a spatial and temporal analysis with CETESB data, in order to assess the evolution of water quality in São Paulo water bodies, and described poor improvement over the years, highlighting the absence of investment in sewage treatment and the presence of non-point sources of pollution (41% of untreated wastewater in São Paulo' State, CETESB, 2013). Bioassays are efficient and relatively inexpensive tools used to discern the effects of toxic substances in multiple stressor scenarios. These tests are intended to assess the presence of toxic substances in water and provide information on the causes of overall pollution (Bonada et al., 2006), as well as the presence of effective concentrations of toxic substances. Bioassay responses can be used as early warnings of potential impairment of populations or communities (Schweigert et al., 2002; Vondráček et al., 2004). The use of bioassays is generalized in some South American countries for research purposes or monitoring programs as inexpensive and effective tool (A. Ronco et al., 2000; A.E. Ronco et al., 2000; Ronco et al., 2002). In our study, the toxicological bioassay with C. dubia, out of the 272 water samples taken, approximately 1/3 showed acute or chronic

effects on the organism, demonstrating the presence of substances capable of causing chronic or acute effects. However, despite the presence of toxic responses in the C. dubia tests, it was not possible to determine the factors causing these responses, using the chemometric analysis. The total variance explained by the two PCs that showed significance was not so high (32%), and (with the exception of metals, calculated as TUs) the other variables routinely included in the CETESB analyses and that could cause toxicological responses (ammonia, cyanobacteria, nitrates, and nitrites) showed little or no correlation with the PCs. The number of cyanobacteria was the variable with the highest number of noncompliances, in both the TOX and the ZCI matrices. Due to eutrophy, cyanobacterial blooms are common in South America reservoirs and are becoming a problem in many countries in the world (Dörr et al., 2010; Sukenik et al., 2012). Cyanobacterial blooms constitute a potential risk to human and environmental health (Soares et al., 2013; Piccin-Santos and Bittencourt-Oliveira, 2012; Moschini-Carlos et al., 2009). However, despite the high number of noncompliances, and the relationships observed in other studies between cyanobacterial blooms and toxic responses (Chorus, 2001; Sotero-Santos et al., 2006), this variable did not show any explicative value in the chemometric analysis for responses on C. dubia. Ammonia, another toxic substance, exceeded the EQS in 23% of the samples, but was not significant in the chemometric analysis. The percentage of points showing toxicological risk following application of the TU criterion was very low. However, in the chemometric analysis with the TOX matrix, TU was the only variable with expected toxicological effects that showed strong correlation with PC3. As is showed by the ANOVA, this PC was able to discriminate between "acute effects" and "chronic or no effects", suggesting a high toxicological relevance of the TU variable and, therefore, the concentrations of metals. Nonetheless, the variability explained by PC3 was very low. In toxicological risk assessment, the dissolved fraction is normally considered the main cause of potential toxicity in aquatic organisms (Schmidt et al., 2010). Dissolved concentrations were only analyzed for Cu and Al. In our opinion, dissolved or bioavailable fractions should be included in monitoring programs to improve risk assessment. Due to the lack of data for the dissolved fractions of many metals in the database, they were estimated using the total metal concentrations in the water. It is possible that the model chosen underestimated the dissolved fraction. Other models can also be used to calculate the bioavailable fraction from the total concentration of metal in water (Ohio EPA, 1996), and more complex models are available for assessment of the toxicity of

Table 4 Eigenvalues and variance explained by the different PC in PCA performed with both matrices. TOX matrix

Table 3 Scores of ZCI index in the different sampling points in function of toxicological responses. In percentage. ZCI scores Toxicological responses

Good

Regular

Bad

Very bad

Total toxicological responses

No effects Chronic Acute Total ZCI cases

0 0 0 0

18.57 5.56 0 14

75.71 72.22 100 67

5.71 22.22 0 8

70 18 1 89

ZCI matrix

PC

Eigenvalue

% variance

Eigenvalue

% variance

1 2 3 4 5 6 7 8 9 10 11

3.374 2.443 1.091 0.918 0.766 0.598 0.530 0.465 0.408 0.219 0.182

30.67 22.21 9.92 8.35 6.97 5.44 4.82 4.23 3.71 1.99 1.65

3.01 2.37 1.46 1.10 0.79 0.61 0.52 0.45 0.29 0.23 0.15

27.34 21.55 13.31 10.03 7.24 5.52 4.75 4.11 2.70 2.07 1.33

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

7

Table 5 loading obtained in both PCA analysis for each variable. Only significant PC are showed.

Loadings TOX matrix

Loadings ZCI matrix

PC 2 PC 3 PC 5 PC 9 PC 1 PC 2 PC 3

pH

Cond

O2

Total P

NH3

NO3

NO2

Chl. a

E. coli UFC

TU

Num. cyano.

0.44 −0.21 0.04 0.57 0.44 0.21 0.18

0.04 0.14 0.12 0.11 −0.10 0.44 −0.38

0.49 −0.20 0.12 0.08 0.50 0.02 0.07

−0.22 −0.31 0.35 −0.34 −0.18 0.38 0.42

−0.36 −0.09 0.13 −0.15 −0.40 0.26 0.17

−0.02 0.19 −0.04 0.09 −0.03 0.25 −0.67

−0.12 0.28 −0.75 0.06 −0.12 0.37 −0.14

0.20 −0.10 0.03 −0.22 0.28 0.48 0.07

−0.51 −0.13 0.19 0.67 −0.40 0.11 0.25

0.10 0.80 0.48 0.03 0.17 −0.08 −0.15

0.26 −0.02 −0.03 −0.13 0.27 0.33 0.23

metal mixtures (Farley et al., 2014; Schmidt et al., 2010). However, the environmental variables required as inputs to the algorithms (dissolved organic carbon and alkalinity) were not consistently measured at the sampling locations. TUs offer an easy methodology for assessment of the toxicological risks of mixtures of toxic substances. TUs have been demonstrated to show good correlation with negative responses in communities and individuals when levels of pollutants are correct and widely analyzed (Beketov et al., 2013; Höss et al., 2011; Ginebreda et al., 2010, 2014; Barata et al., 2007; Liess and Ohe, 2005). For some of the studied locations, CETESB also performed Microtox and Ames tests (data not shown), in addition to the bioassay with C. dubia. The results obtained suggested the presence of chemical substances with mutagenic, carcinogenic, or lethal activities, such as pesticides, antibiotics, chlorinated compounds, aromatic hydrocarbons, and aromatic amines, amongst others (Coleman and Qureshi, 1985; Brack et al., 2007; Umbuzeiro et al., 2001). Unfortunately, although these bioassays can provide information on the presence of effective concentrations of toxic substances, they are unable to identify the substances. The general overview of the data compiled by CETESB between 2008 and 2012 revealed a lack of analyses of organic substances such as pesticides and other organic compounds that might cause toxicological responses. Although these compounds were not detected during an intensive sampling campaign conducted at one sampling point in 2007, this does not imply the absence of these substances at other locations or in the future, because inputs could occur due to point source spills, as well as due to economic or demographic shifts (as indicated by the bioassay responses in the next years). Due to inadequate wastewater treatment and sewage disposal, high concentrations of nutrients, pesticides and emerging and unregulated pollutants have been detected in freshwater ecosystems of South America (Magdaleno et al., 2014; Valdés et al., 2014; Barletta et al., 2010; Barra et al., 2006) and, specifically, in the State of São Paulo (De Sousa et al., 2014; Montagner and Jardim, 2011; Locatelli et al., 2011; Sodré et al., 2010; de Almeida and Weber, 2009; Silva et al., 2008). Inadequate wastewater treatment and disposal is still a main problem in South America and source of pollutants in freshwater ecosystems [from 2.4 to 83.3% of population connected to wastewater treatment depending on the Country in South America (UN Stats, 2011)]. Bioassays have the disadvantage that they are not able to indicate other environmental stressors (Kimball and Levin, 1985), for which community based indexes are useful tools. Such indexes have more ecological relevance, because they are based on changes in the biological

community (including zooplankton, phytoplankton, and benthic macroinvertebrates) that occur at a given location. Due to the ecological roles of the species in a community, changes in the community structure provide integrated information on long-term exposure to a wide variety of environmental conditions, including toxic substances. For all the sampling points, the ZCI revealed different degrees of community impairment, according to the criterion of Coelho-Botelho et al. (2006), demonstrating the existence of environmental factors that cause alterations in the abundances of Copepoda, Rotifera, and Cladocera. The ZCI and ecotoxicological bioassays were only performed simultaneously for five sampling points. Since the toxicological bioassays were performed with the planktonic species C. dubia, toxicological responses observed in this organism could predict possible toxicological disturbances in the planktonic community. A slight but significant correlation was observed between the ZCI and the toxicological responses, suggesting a small influence of toxic substances on the structure of the zooplankton community, as well as the possible effects of environmental factors other than toxic substances on the quality states in the ZCI. Chemometric analysis showed that in the case of the ZCI matrix, only PC2 was able to discriminate between quality states, but only explained 21.5% of the variance. The most correlated variables in this PC were conductivity, Chl. a, and total P. Increases in conductivity has been closely related to pollution in reservoirs (Marcé et al., 2008). The contribution of Chl. a and total P seems to indicate that eutrophy had an important effect on the zooplanktonic community composition. However, this relationship could be an artifact. The ZCI is based in the composition of the zooplankton community, but the final scoring includes the trophic state index (Toledo, 1990). The TSI includes the concentrations of Chl. a and total P in the calculation, which explains the correlation of these variables in PC2. Since the zooplankton community structure is depending on multiple factors, in our opinion, the inclusion of the TSI in the final scoring of the ZCI used by CETESB could introduce a bias and give an artificial weight to eutrophy, which could decrease the capacity of this index to indicate the presence of other environmental pressures. This could prevent a good interpretation of the stressors affecting the zooplankton community, and consequently the detection of current threats to the biota in the reservoirs. Despite its drawbacks, due to the high spatial and temporal variability of zooplankton communities and the complexity of factors regulating community dynamics (Beck and Hatch, 2009; Gannon and Stemberger, 1978), the zooplankton community structure could be

Table 6 Tukey's post hoc test for ANOVA with ZCI scores and PCs. Asterisk means significant differences p b 0.05. Multiple comparisons Tukey Dependent variable

ZCI quality states

PC1

2

PC2

2

PC3

3

3 4⁎ 3⁎ 4⁎ 2⁎ 4⁎

Mean difference (I-J)

−0.65 −2.05 −2.18 −4.16 −0.95 −1.26

Std. error

0.472 0.709 0.332 0.498 0.318 0.397

Sig.

0.3566 0.0128 0.000 0.000 0.009 0.005

95% confidence interval Low. bound

Up. bound

−1.775 −3.741 −2.976 −5.3545 −1.709 −2.206

0.473 −0.366 −1.397 −2.983 −0.196 −0.315

⁎ p b 0.05

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

8

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

Table 7 Tukey's post hoc test for ANOVA with toxicological responses and PCs. Asterisks mean significant differences p b 0.05. Multiple comparisons Tukey Dependent variable

Toxic responses

PC2

0

PC3

2

PC5

2

PC9

0 1 2

Mean difference (I-J)

1⁎ 2⁎ 0⁎ 1⁎ 0⁎ 1⁎ 1 2 0 2 0 1

−0.87 −1.54 2.099 2.219 1.058 1.294 0.169 0.428 −0.169 0.259 −0.428 −0.259

Std. error

0.202 0.465 0.297 0.308 0.262 0.271 0.085 0.196 0.085 0.203 0.196 0.203

Sig.

0.0001 0.0031 0.0000 0.0000 0.0002 0.0000 0.1193 0.0762 0.1193 0.4125 0.0762 0.4125

95% confidence interval Low. bound

Up. bound

−1.348 −2.632 1.397 1.491 0.441 0.654 −0.032 −0.034 −0.370 −0.221 −0.891 −0.739

−0.393 −0.440 2.799 2.946 1.675 1.934 0.370 0.891 0.032 0.739 0.034 0.221

⁎ p b 0.05

used for assessment of the ecological status in reservoirs, and to provide information on long-term environmental disturbances (Ferdous and Muktadir, 2009; Gannon and Stemberger, 1978). Relationships between the community composition (Copepoda and Rotifera) and the concentrations of chlorophyll a, total phytoplankton, and total phosphorus in Brazilian reservoirs were described by Perbiche-Neves et al. (2013). However, Brito et al. (2011) found different trophic states in Brazilian reservoirs, depending on whether the criterion was based on the zooplankton community or separate application of the TSI. The response of the zooplankton community to environmental factors other than eutrophy has been described in several studies. Pesticides (Hanazato, 2001), pharmaceuticals (Richards et al., 2004), metals (Jak et al., 1996), ammonium (Branco et al., 2002), the hydrological regime (Naselli-Flores and Barone, 1994; Nogueira et al., 2008), or other ecological factors such as predation could produce effects on the community structure (Beck and Hatch, 2009). Therefore, alterations in community structure could be due to multiple factors and their interactions (Holmstrup et al., 2010). Such a multistressor scenario is more likely in the case of polymictic tropical reservoirs with high nutrient concentrations, mixtures of pollutants, and hydrological instability, which cause continuous resets in planktonic communities (Lewis, 1987; Tundisi, 1990). ZCI should be able to detect all these different pressures but the inclusion of the TSI could impair this capability. Bioassays and community-based indexes provide information on impairments in biota and, in our opinion, their implementation in monitoring programs can help to improve water management in situations involving multiple stressors (as in the case of inland tropical waters). The observation of community impairment, together with good chemical quality of the water (according to the established regulatory standards), suggests that monitoring strategies may be inadequate, with incomplete chemical screening that is unable to detect the presence of other possible threatening substances (Brack et al., 2009). López-Doval et al. (2012) suggested the inclusion of bioassays in combination with good chemical screening in European monitoring programs, in order to confirm the presence, bioavailability, and toxic capacity of chemical substances in water samples. A tiered approach, involving ecological indexes, bioassays, and improved chemical screening, has been suggested as a good option when there is evidence of ecosystem impairment and it is necessary to discriminate between the effects of toxic substances or other environmental stressors (Hein et al., 2010). Several studies have tested this tiered approach for different rivers, with successful identification of the main stressors impairing the biological community (Damásio et al., 2007, 2011; Wolfram et al., 2012; De Castro-Català et al., 2015). Monitoring programs should be able to provide data for correction of noncompliances and environmental impairments (Cairns and McCormick, 1992), and for prediction of future threats to water quality and freshwater ecosystems (Holt et al., 2000). The monitoring program

of CETESB is in accordance with Brazilian federal legislation, but has a poor ability to provide information on the stressors disturbing freshwater communities and organisms. For this reason, it is not currently possible to ensure satisfactory protection of these ecosystems. In our opinion, this is due to the small number of chemical substances monitored, and the temporal and spatial gaps in the data for the parameters that are analyzed. As has been demonstrated by means of the chemometric approach, data obtained from the CETESB database are not sufficient to explain the observed toxicological and community responses. Therefore, without identification of the substances or environmental factors responsible, it is difficult to implement specific policies aimed at controlling them. For this reason, it would be desirable to conduct more frequent intensive monitoring campaigns, in order to update information on the chemical substances present in the reservoirs and other water bodies (at least in those where toxic responses have been observed), and to prioritize those substances or environmental factors known to present human or environmental risks. This type of approach has been widely used in Europe (Guillén et al., 2012), but is very rare in Brazil or other countries of South America (Montagner et al., 2014). Models provide useful tools for predictive studies of the dynamics, trends and anticipation of management problems in reservoirs, and can assist managers in decision-making processes (Tundisi, 1990; Straškraba, 1994). However, a strong data background, with adequate temporal resolution, is needed in order to enhance the predictive power of a model. With the current data available for the study region, the development of predictive models is difficult. As reported by Bertholdo et al. (2012), this lack of chemical resolution results in a poor capacity of models to predict toxicity responses, using the primary data from the CETESB database. Reservoirs are key water resources for economic development in South America, and this region is facing present and future challenges in the management of water quality and supply. This is amply demonstrated during the current episode of water scarcity in São Paulo's State (New Scientist, 2015). It is essential to not only reduce pollution, improving wastewater treatment and collecting systems, and protect the services provided by ecosystems, but also invest in research in order to acquire the information needed to predict present and future threats. The investment in this kind of fundamental research would be of strategic benefit, helping managers to improve water quality and protect the ecosystems in reservoirs of South America, while providing new data on the biology, ecology, and chemistry of reservoirs (Tundisi and Matsumura-Tundisi, 2003). The effective management of reservoirs and other water resources is critical for the sustained economic development of the region, and actions developed in the MRSP could provide an example of management for other developing countries.

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

5. Conclusions Toxicological responses have been observed in the reservoirs of the MRSP, together with a high incidence of regular or bad quality of the zooplankton community. This demonstrates the presence of stressors that affect the viability of organisms and populations. In this study, by means of a statistical approach, we show that the data compiled by CETESB are insufficient to identify the factors causing the observed ecotoxicological responses, and are therefore insufficient to identify clear cause–effect relationships. In addition, impairments in the zooplankton community are not well explained using the data currently available. Other stressors, different from those analyzed, could be responsible for the observed responses. The identification and quantification of stressors (emerging pollutants, metals, pesticides, other organic pollutants, environmental stressors) is a prerequisite for the implementation of policies focused on the reduction or elimination of stress factors in freshwater ecosystems. In addition biological indices and bioassays should be improved. Current monitoring programs should be improved in order to provide predictive capacity and enable corrective policies to be adopted, in order to ensure the quality and availability of water resources, and the protection of ecosystems. Acknowledgment Authors acknowledge the financial support of the Fundação de Amparo à Pesquisa do Estado de São Paulo (proc. 2012/11890-4, 2012/16420-6 and 2013/08272-0). Cardoso-Silva holds a PNPD fellowship from the Coordenação de Aperfeiçoamento de Pessoal de Nível. References Agostinho, A.A., Thomaz, S.M., Gomes, L.C., 2005. Conservation of the biodiversity of Brazil's inland waters. Conserv. Ecol. 19, 646–652. Alonso, A., Camargo, J.A., 2006. Toxicity of nitrite to three species of freshwater invertebrates. Environ. Toxicol. 21, 90–94. Araújo, R.S., Alves, M.G., Condesso de Melo, M.T., Chrispim, Z.M.P., Mendes, M.P., Silva Júnior, G.C., 2014. Water resource management: a comparative evaluation of Brazil, Rio de Janeiro, the European Union, and Portugal. Sci. Total Environ. 511, 815–828. Barata, C., Damasio, J., Lopez, M.A., Kuster, M., De Alda, M.L., Barceló, D., Riva, M.C., Raldua, D., 2007. Combined use of biomarkers and in situ bioassays in Daphnia magna to monitor environmental hazards of pesticides in the field. Environ. Toxicol. Chem. 26, 370–379. Barletta, M., Jaureguizar, A.J., Baigun, C., Fontoura, N.F., Agostinho, A.A., Almeida-Val, V.M.F., Val, A.L., Torres, R.A., Jimenes-Segura, L.F., Giarrizzo, T., Fabre, N.N., Batista, V.S., Lasso, C., Taphorn, D.C., Costa, M.F., Chaves, P.T., Vieira, J.P., Correa, M.F.M., 2010. Fish and aquatic habitat conservation in South America: a continental overview with emphasis on neotropical systems. J. Fish Biol. 76, 2118–2176. Barra, R., Colombo, J.C., Eguren, G., Gamboa, N., Jardim, W.F., Mendoza, G., 2006. Persistent organic pollutants (POPs) in eastern and western South American countries. Reviews of Environmental Contamination and Toxicology. Springer, New York, pp. 1–33. Beck, M.W., Hatch, L.K., 2009. A review of research on the development of lake indices of biotic integrity. Environ. Rev. 17, 21–44. Beketov, M.A., Kefford, B.J., Schäfer, R.B., Liess, M., 2013. Pesticides reduce regional biodiversity of stream invertebrates. Proc. Natl. Acad. Sci. 110, 11039–11043. Bell, A.R., Lemos, M.C., Scavia, D., 2010. Cattle, clean water, and climate change: policy choices for the Brazilian agricultural frontier. Environ. Sci. Technol. 44, 8377–8384. Bertholdo, L., da Silva, C.G., de Aragão Umbuzeiro, G., Júnior, L.C., 2012. Técnicas De Mineração De Dados Na Classificação De Ecotoxicidade De Água Para Aplicação Na Gestão De Corpos Hídricos. VIII Congresso Nacional de Excelência em Gestão. Bonada, N., Prat, N., Resh, V.H., Statzner, B., 2006. Developments in aquatic insect biomonitoring: a comparative analysis of recent approaches. Annu. Rev. Entomol. 51, 495–523. Brack, W., Bakker, J., de Deckere, E., Deerenberg, C., van Gils, J., Hein, M., Jurajda, P., Kooijman, B., Lamoree, M., Lek, S., 2005. MODELKEY — models for assessing and forecasting the impact of environmental key pollutants on freshwater and marine ecosystems and biodiversity. Environ. Sci. Pollut. Res. 12, 252–256. Brack, W., Klamer, H.J.C., de Ada, M.L., Barcelo, D., 2007. Effect-directed analysis of key toxicants in European River Basins. A review (9 pp). Environ. Sci. Pollut. Res. Int. 14, 30–38. Brack, W., Apitz, S.E., Borchardt, D., Brils, J., Cardoso, A.C., Foekema, E.M., van Gils, J., Jansen, S., Harris, B., Hein, M., Heise, S., Hellsten, S., de Maagd, P.G.J., Müller, D., Panov, V.E., Posthuma, L., Quevauviller, P., Verdonscho, P.F.M., von der Ohe, P.C., 2009. Toward a holistic and risk‐based management of European river basins. Integr. Environ. Assess. Manag. 5, 5–10. Braga, B.P.F., Porto, M.F.A., Silva, R.T., 2006. Water management in metropolitan São Paulo. Water Resour. Dev. 22, 337–352.

9

Branco, C.W., Rocha, M.I.A., Pinto, G.F., Gômara, G.A., Filippo, R.D., 2002. Limnological features of Funil Reservoir (RJ, Brazil) and indicator properties of rotifers and cladocerans of the zooplankton community. Lakes Reserv. Res. Manag. 7, 87–92. Brito, S.L., Maia‐Barbosa, P.M., Pinto‐Coelho, R.M., 2011. Zooplankton as an indicator of trophic conditions in two large reservoirs in Brazil. Lakes Reserv. Res. Manag. 16, 253–264. Cairns Jr., J., McCormick, P.V., 1992. Developing an ecosystem-based capability for ecological risk assessments. Environ. Prof. 14, 186–196. Camargo, J.A., Alonso, A., 2006. Ecological and toxicological effects of inorganic nitrogen pollution in aquatic ecosystems: a global assessment. Environ. Int. 32, 831–849. Cardoso-Silva, S., Ferreira, T., Pompêo, M.L.M., 2013. Diretiva Quadro D'Água: uma revisão crítica e a possibilidade de aplicação ao Brasil. Ambient. Soc. 16, 39–58. Carpenter, S.R., Fisher, S.G., Grimm, N.B., Kitchell, J.F., 1992. Global Change and freshwater ecosystems. Annu. Rev. Ecol. Syst. 23, 119–139. Carvalho, F.P., 2006. Agriculture, pesticides, food security and food safety. Environ. Sci. Pol. 9, 685–692. CETESB, Companhia Ambiental do Estado de São Paulo, 1991. Norma Técnica CETESB SQ PR/LB 088: Teste de toxicidade de águas superficiais utilizando Ceriodaphnia dubia. Companhia Ambiental do Estado de São Paulo, São Paulo. CETESB, Companhia Ambiental do Estado de São Paulo, 2013. 2012 Qualidade das águas superficiais no estado de São Paulo. Companhia Ambiental do Estado de São Paulo, São Paulo. Chorus, I., 2001. Cyanotoxin occurrence in freshwaters — a summary of survey results from different countries. Cyanotoxins. Springer, Berlin, pp. 75–82. Coelho-Botelho, M.J., Carvalho, M.C., Kuhlmann, M.L., Salvador, M.E.P., Ribeiro de Souza, R.C., Watanabe, H.M., Araújo, R.P.A., Bressan Jr., H., Brandimarte, A.L., Anaya, M., Guereschi, R.M., 2006. Desenvolvimento de índices biológicos para o biomonitoramento em reservatórios do Estado de São Paulo. Companhia Ambiental do Estado de São Paulo, São Paulo. Coleman, R.N., Qureshi, A.A., 1985. Microtox® and Spirillum volutans tests for assessing toxicity of environmental samples. Bull. Environ. Contam. Toxicol. 35, 443–451. Cunha, D.G.F., Calijuri, M.C., Lamparelli, M.C., Menegon Jr., N., 2013. Resolução CONAMA 357/2005: análise espacial e temporal de não conformidades em rios e reservatórios do estado de São Paulo de acordo com seus enquadramentos (2005–2009). Eng. Sanit. Ambient. 18, 159–168. Damásio, J.B., Barata, C., Munné, A., Ginebreda, A., Guasch, H., Sabater, S., Caixach, J., Porte, C., 2007. Comparing the response of biochemical indicators (biomarkers) and biological indices to diagnose the ecological impact of an oil spillage in a Mediterranean river (NE Catalunya, Spain). Chemosphere 66, 1206–1216. Damásio, J., Fernández-Sanjuan, M., Sánchez-Avila, J., Lacorte, S., Prat, N., Rieradevall, M., Soares, A.M.V.M., Barata, C., 2011. Multi-biochemical responses of benthic macroinvertebrate species as a complementary tool to diagnose the cause of community impairment in polluted rivers. Water Res. 45, 3599–3613. de Almeida, G.A., Weber, R.R., 2009. Fármacos na represa Billings. Rev. Saúde Ambient. 6, 7–13. De Castro-Català, N., Muñoz, I., Armendáriz, L., Campos, B., Barceló, D., López-Doval, J., Pérez, S., Petrovic, M., Picó, Y., Riera, J.L., 2015. Invertebrate community responses to emerging water pollutants in Iberian river basins. Sci. Total Environ. 503, 142–150. de Sousa, D.N.R., Mozeto, A.A., Carneiro, R.L., Fadini, P.S., 2014. Electrical conductivity and emerging contaminant as markers of surface freshwater contamination by wastewater. Sci. Total Environ. 484, 19–26. Dörr, F.A., Pinto, E., Soares, R.M., 2010. Microcystins in South American aquatic ecosystems: occurrence, toxicity and toxicological assays. Toxicon 56, 1247–1256. Ducrot, R., Kruitwagen, M., Wagemans, F., Van Essen, P., Jacobi, P., 2005. Land and Water Management in the Metropolitan Region of Sao Paulo: Presentation of the Geographic and Institutional Context. PROCAM-USP, São Paulo (18 pp.). European Commission, 2011. Guidance document no. 27 Technical guidance for deriving Environmental Quality Standards Common implementation strategy for the Water Framework Directive (2000/60/EC). Fàbrega, F., Marquès, M., Ginebreda, A., Kuzmanovic, M., Barceló, D., Schuhmacher, M., Domingo, J.L., Nadal, M., 2013. Integrated risk index of chemical aquatic pollution (IRICAP): case studies in Iberian rivers. J. Hazard. Mater. 263, 187–196. Farley, K.J., Meyer, J.S., Balistrieri, L.S., De Schamphelaere, K.A., Iwasaki, Y., Janssen, C.R., Kamo, M., Lofts, S., Mebane, C.A., Naito, W., Ryan, A.C., Santore, R.C., Tipping, E., 2014. Metal Mixture Modeling Evaluation project: 2. Comparison of four modeling approaches. Environ. Toxicol. Chem. http://dx.doi.org/10.1002/etc.2820. Ferdous, Z., Muktadir, A.K.M., 2009. A review: potentiality of zooplankton as bioindicator. Am. J. Appl. Sci. 6, 1815–1819. Fontana, L., Albuquerque, A.L.S., Brenner, M., Bonotto, D.M., Sabaris, T.P., Pires, M.A., Cotrim, M.E.B., Bicudo, D.C., 2014. The eutrophication history of a tropical water supply reservoir in Brazil. J. Paleolimnol. 51, 29–43. Formiga-Johnsson, R.M., Kemper, K., 2005. Institutional and policy analysis of river basin management: the Alto-Tietê River Basin, São Paulo, Brazil. World Bank Policy Research Working Paper, (3650). Gannon, J.E., Stemberger, R.S., 1978. Zooplankton (especially crustaceans and rotifers) as indicators of water quality. Transactions of the American Microscopical Society 16–35. García-Reiriz, A.G., Olivieri, A.C., Teixidó, E., Ginebreda, A., Tauler, R., 2014. Chemometric modeling of organic contaminant sources in surface waters of a mediterranean river basin. Environ. Sci.: Processes Impacts 16, 124–134. Ginebreda, A., Muñoz, I., de Alda, M.L., Brix, R., López-Doval, J., Barceló, D., 2010. Environmental risk assessment of pharmaceuticals in rivers: relationships between hazard indexes and aquatic macroinvertebrate diversity indexes in the Llobregat River (NE Spain). Environ. Int. 36 (2), 153–162. Ginebreda, A., Kuzmanovic, M., Guasch, H., de Alda, M.L., López-Doval, J.C., Muñoz, I., Ricart, M., Romaní, A.M., Sabater, S., Barceló, D., 2014. Assessment of multi-chemical pollution in aquatic ecosystems using toxic units: Compound prioritization, mixture

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

10

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx

characterization and relationships with biological descriptors. Sci. Total Environ. 468, 715–723. Guillén, D., Ginebreda, A., Farré, M., Darbra, R.M., Petrovic, M., Gros, M., Barceló, D., 2012. Prioritization of chemicals in the aquatic environment based on risk assessment: analytical, modeling and regulatory perspective. Sci. Total Environ. 440, 236–252. Hanazato, T., 2001. Pesticide effects on freshwater zooplankton: an ecological perspective. Environ. Pollut. 112, 1–10. Hein, M., Rotter, S., Schmitt-Jansen, M., von der Ohe, P.C., Brack, W., de Deckere, E., Schmitt, C., de Zwart, D., Posthuma, L., Foekema, E.M., Muñoz, I., Sabater, S., Segner, H., Semenzin, E., Marcomini, A., van Gils, J., van Hattum, B., van Vliet, L.A., 2010. MODELKEY. Umweltwiss. Schadst. Forsch. 22, 217–228. Holmstrup, M., Bindesbøl, A.M., Oostingh, G.J., Duschl, A., Scheil, V., Köhler, H.R., Loureiro, S., Soares, A.M.V.M., Ferreira, A.L.G., Kienle, C., Gerhard, A., Laskowski, R., Kramarz, P.E., Bayley, M., Svendsen, C., Spurgeon, D.J., 2010. Interactions between effects of environmental chemicals and natural stressors: a review. Sci. Total Environ. 408, 3746–3762. Holt, M.S., Fox, K., Griessbach, E., Johnsen, S., Kinnunen, J., Lecloux, A., Murray-Smith, R., Peterson, D.R., Schröder, R., Silvani, M., ten Berge, W.F.J., Toy, R.J., Feijtel, T.C.M., 2000. Monitoring, modelling and environmental exposure assessment of industrial chemicals in the aquatic environment. Chemosphere 41, 1799–1808. Höss, S., Claus, E., Von der Ohe, P.C., Brinke, M., Güde, H., Heininger, P., Traunspurger, W., 2011. Nematode species at risk — a metric to assess pollution in soft sediments of freshwaters. Environ. Int. 37, 940–949. Jak, R.G., Maas, J.L., Scholten, M.T., 1996. Evaluation of laboratory derived toxic effect concentrations of a mixture of metals by testing fresh water plankton communities in enclosures. Water Res. 30, 1215–1227. Kimball, K.D., Levin, S.A., 1985. Limitations of laboratory bioassays: the need for ecosystem-level testing. Bioscience 35, 165–171. Lewis, W.M., 1987. Tropical limnology. Annu. Rev. Ecol. Syst. 159–184. Liess, M., Ohe, P.C.V.D., 2005. Analyzing effects of pesticides on invertebrate communities in streams. Environ. Toxicol. Chem. 24, 954–965. Locatelli, M.A.F., Sodré, F.F., Jardim, W.F., 2011. Determination of antibiotics in brazilian surface waters using liquid chromatography–electrospray tandem mass spectrometry. Arch. Environ. Contam. Toxicol. 60, 385–393. López-Doval, J.C., De Castro-Català, N., Andrés-Doménech, I., Blasco, J., Ginebreda, A., Muñoz, I., 2012. Analysis of monitoring programmes and their suitability for ecotoxicological risk assessment in four Spanish basins. Sci. Total Environ. 440, 194–203. Magdaleno, A., Juárez, Á.B., Dragani, V., Saenz, M.E., Paz, M., Moretton, J., 2014. Ecotoxicological and genotoxic evaluation of Buenos Aires city (Argentina) hospital wastewater. J. Toxicol. 1–10. Marcé, R., Moreno-Ostos, E., López, P., Armengol, J., 2008. The role of allochthonous inputs of dissolved organic carbon on the hypolimnetic oxygen content of reservoirs. Ecosystems 11, 1035–1053. Montagner, C.C., Jardim, W.F., 2011. Spatial and seasonal variations of pharmaceuticals and endocrine disruptors in the Atibaia River, São Paulo State (Brazil). J. Braz. Chem. Soc. 22, 1452–1462. Montagner, C.C., Jardim, W.F., Von der Ohe, P.C., Umbuzeiro, G.A., 2014. Occurrence and potential risk of triclosan in freshwaters of São Paulo, Brazil — the need for regulatory actions. Environ. Sci. Pollut. Res. 21, 1850–1858. Moschini-Carlos, V.M., Bortoli, S., Pinto, E., Nishimura, P.Y., de Feitas, L.G., Pompêo, M., Dörr, F., 2009. Cyanobacteria and cyanotoxin in the Billings reservoir (São Paulo, SP, Brazil). Limnetica 28, 273–282. Muñoz, I., López‐Doval, J.C., Ricart, M., Villagrasa, M., Brix, R., Geiszinger, A., Ginebreda, A., Guasch, H., López de Alda, M.J., Romaní, A.M., Sabater, S., Barceló, D., 2009. Bridging levels of pharmaceuticals in river water with biological community structure in the Llobregat river basin (northeast Spain). Environ. Toxicol. Chem. 28, 2706–2714. Naselli-Flores, L., Barone, R., 1994. Relationship between trophic state and plankton community structure in 21 Sicilian dam reservoirs. Nutrient Dynamics and Biological Structure in Shallow Freshwater and Brackish Lakes. Springer, Netherlands, pp. 197–205. New Scientist, 2015. Brazil Hit by Worst Drought in 80 Years. New Scientist (January 28th, 3006). Nogueira, M.G., Oliveira, P.C.R., de Britto, Y.T., 2008. Zooplankton assemblages (Copepoda and Cladocera) in a cascade of reservoirs of a large tropical river (SE Brazil). Limnetica 27, 151–169. Ohio EPA (Environmental Protection Agency), 1996. Dissolved metals criteria. Ohio EPA Great Lakes Initiative Issue Paper, July 1996. Ohio Environmental Pollution Agency, Columbus, OH. Pahl-Wostl, C., 2007. Transitions towards adaptive management of water facing climate and global change. Water Resour. Manag. 21, 49–62. Pelletier, G., 1996. Applying metals criteria to water quality-based discharge limits. Empirical Models of the Dissolved Fraction of Cadmium, Copper, Lead, and Zinc. Department of Ecology Publications, P. O. BOX 47600, Olympia, WA 98504-7600 (USA). Perbiche-Neves, G., Fileto, C., Laço-Portinho, J., Troguer, A., Serafim-Júnior, M., 2013. Relations among planktonic rotifers, cyclopoid copepods, and water quality in two Brazilian reservoirs. Lat. Am. J. Aquat. Res. 41, 138–149. Peré-Trepat, E., Ginebreda, A., Tauler, R., 2007. Comparison of different multiway methods for the analysis of geographical metal distributions in fish, sediments and river waters in Catalonia. Chemom. Intell. Lab. Syst. 88, 69–83. Petrovic, M., Ginebreda, A., Acuna, V., Batalla, R.J., Elosegi, A., Guasch, H., de Alda, M.L., Marce, R., Munoz, I., Navarro-Ortega, A., Navarro, E., Vericat, D., Sabater, S., Barcelo, D., 2011. Combined scenarios of chemical and ecological quality under water scarcity in Mediterranean rivers. TrAC Trends Anal. Chem. 30, 1269–1278. Piccin-Santos, V., Bittencourt-Oliveira, M.C., 2012. Toxic cyanobacteria in four Brazilian water supply reservoirs. J. Environ. Prot. 3, 68–73.

Ricart, M., Guasch, H., Barceló, D., Brix, R., Conceição, M.H., Geiszinger, A., López de Alda, M.J., López-Doval, J.C., Muñoz, I., Postigo, C., Romaní, A.M., Villagrasa, M., Sabater, S., 2010. Primary and complex stressors in polluted Mediterranean rivers: pesticide effects on biological communities. J. Hydrol. 383, 52–61. Ricciardi, F., Bonnineau, C., Faggiano, L., Geiszinger, A., Guasch, H., Lopez-Doval, J., Muñoz, I., Proia, L., Ricart, M., Romani, A., Sabater, S., 2009. Is chemical contamination linked to the diversity of biological communities in rivers? TrAC Trends Anal. Chem. 28, 592–602. Richards, S.M., Wilson, C.J., Johnson, D.J., Castle, D.M., Lam, M., Mabury, S.A., Solomon, K.R., 2004. Effects of pharmaceutical mixtures in aquatic microcosms. Environ. Toxicol. Chem. 23, 1035–1042. Richardson, S.D., Ternes, T.A., 2014. Water analysis: emerging contaminants and current issues. Anal. Chem. 86, 2813–2848. Ronco, A., Sobrero, C., Grassi, V., Kaminski, L., Massolo, L., Mina, L., 2000a. WaterTox bioassay intercalibration network: results from Argentina. Environ. Toxicol. 15, 287–296. Ronco, A.E., Castillo, G., Díaz-Baez, M.C., 2000b. Development and application of microbioassays for routine testing and biomonitoring in Argentina, Chile and Colombia. New Microbiotests for Routine Toxicity Screening and Biomonitoring. Springer, US, pp. 49–61. Ronco, A., Gagnon, P., Diaz‐Baez, M.C., Arkhipchuk, V., Castillo, G., Castillo, L.E., Sánchez, A., 2002. Overview of results from the WaterTox intercalibration and environmental testing phase II program: part 1, statistical analysis of blind sample testing. Environ. Toxicol. 17, 232–240. Schmidt, T.S., Clements, W.H., Mitchell, K.A., Church, S.E., Wanty, R.B., Fey, D.L., Verplanck, P.L., San Juan, C.A., 2010. Development of a new toxic‐unit model for the bioassessment of metals in streams. Environ. Toxicol. Chem. 29, 2432–2442. Schweigert, N., Eggen, R.I., Escher, B.I., Burkhardt-Holm, P., Behra, R., 2002. Ecotoxicological assessment of surface waters: a modular approach integrating in vitro methods. ALTEX 19, 30–37. SEADE, 2015. Portal de Estadísticas do Estado de São Paulo, Fundação Sistema Estadual de Análise de Dados. Available from, http://produtos.seade.gov.br/produtos/perfil_ regional/ (23 November 2015). Silva, D.M.L.D., Camargo, P.B.D., Martinelli, L.A., Lanças, F.M., Pinto, J.S., Avelar, W.E.P., 2008. Organochlorine pesticides in Piracicaba river basin (São Paulo/Brazil): a survey of sediment, bivalve and fish. Quim. Nova 31, 214–219. Soares, M.C., Huszar, V.L., Miranda, M.N., Mello, M.M., Roland, F., Lürling, M., 2013. Cyanobacterial dominance in Brazil: distribution and environmental preferences. Hydrobiologia 717, 1–12. Sodré, F.F., Pescara, I.C., Montagner, C.C., Jardim, W.F., 2010. Assessing selected estrogens and xenoestrogens in Brazilian surface waters by liquid chromatography–tandem mass spectrometry. Microchem. J. 96, 92–98. Sotero-Santos, R.B., Silva, C.R.D.S.E., Verani, N.F., Nonaka, K.O., Rocha, O., 2006. Toxicity of a cyanobacteria bloom in Barra Bonita Reservoir (Middle Tietê River, São Paulo, Brazil). Ecotoxicol. Environ. Saf. 64, 163–170. Soucek, D.J., Dickinson, A., 2012. Acute toxicity of nitrate and nitrite to sensitive freshwater insects, mollusks, and a crustacean. Arch. Environ. Contam. Toxicol. 62, 233–242. Sprague, J.B., 1970. Measurement of pollutant toxicity to fish. II. Utilizing and applying bioassay results. Water Res. 4, 3–32. Stevenson, R.J., Sabater, S., 2010. Understanding effects of global change on river ecosystems: science to support policy in a changing world. Hydrobiologia 657, 3–18. Straškraba, M., 1994. Ecotechnological models for reservoir water quality management. Ecol. Model. 74, 1–38. Sukenik, A., Hadas, O., Kaplan, A., Quesada, A., 2012. Invasion of Nostocales (cyanobacteria) to subtropical and temperate freshwater lakes — physiological, regional, and global driving forces. Front. Microbiol. 3. Tauler, R., Lacorte, S., Guillamón, M., Cespedes, R., Viana, P., Barceló, D., 2004. Chemometric modeling of main contamination sources in surface waters of Portugal. Environ. Toxicol. Chem. 23, 565–575. Terrado, M., Barceló, D., Tauler, R., 2006. Identification and distribution of contamination sources in the Ebro river basin by chemometrics modelling coupled to geographical information systems. Talanta 70, 691–704. Toledo, A.P., 1990. Informe preliminar sobre os estudos para a obtenção de um índice para a avaliação simplificada do estado trófico de reservatórios de regiões quentes tropicais — outubro de 1990. Companhia Ambiental do Estado de São Paulo, São Paulo. Tundisi, J.G., 1990. Perspectives for ecological modelling of tropical and subtropical reservoirs in South America. Ecol. Model. 52, 7–20. Tundisi, J.G., 2003. Agua no seculo XXI: enfrentando a escassez. RiMa/IIE, São Carlos, Brasil (in Portuguese). Tundisi, J.G., Matsumura-Tundisi, T., 2003. Integration of research and management in optimizing multiple uses of reservoirs: the experience in South America and Brazilian case studies. Hydrobiologia 500, 231–242. Tundisi, J.G., Rocha, O., Matsumura-Tundisi, T., Braga, B., 1998. Reservoir management in South America. Int. J. Water Resour. Dev. 14, 141–155. Turner, R.K., Daily, G.C., 2008. The ecosystem services framework and natural capital conservation environmental and resource economics. Environ. Resour. Econ. 39, 25–35. Umbuzeiro, G.D.A., Roubicek, D.A., Sanchez, P.S., Sato, M.I.Z., 2001. The Salmonella mutagenicity assay in a surface water quality monitoring program based on a 20-year survey. Mutat. Res. Genet. Toxicol. Environ. Mutagen. 491, 119–126. UN Stats, 2011. United Nations Statistics Division. Inland Water Resources. http://unstats. un.org/unsd/environment/wastewater.htm. Valdés, M.E., Amé, M.V., de los Angeles Bistoni, M., Wunderlin, D.A., 2014. Occurrence and bioaccumulation of pharmaceuticals in a fish species inhabiting the Suquía River basin (Córdoba, Argentina). Sci. Total Environ. 472, 389–396.

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085

J.C. López-Doval et al. / Science of the Total Environment xxx (2015) xxx–xxx Varis, O., Biswas, A.K., Tortajada, C., Lundqvist, J., 2006. Megacities and water management. Int. J. Water Resour. Dev. 22, 377–394. Vondráček, J., Chramostová, K., Plíšková, M., Bláha, L., Brack, W., Kozubík, A., Machala, M., 2004. Induction of aryl hydrocarbon receptor-mediated and estrogen receptor-mediated activities, and modulation of cell proliferation by dinaphthofurans. Environ. Toxicol. Chem. 23, 2214–2220. von der Ohe, P.C., De Deckere, E., Prüß, A., Muñoz, I., Wolfram, G., Villagrasa, M., Ginebreda, A., Hein, M., Brack, W., 2009. Toward an integrated assessment of the ecological and chemical status of European river basins. Integrated environmental assessment and management 5 (1), 50–61.

11

Vorosmarty, C.J., McIntyre, P.B., Gessner, M.O., Dudgeon, D., Prusevich, A., Green, P., Glidden, S., Bunn, S.E., Sullivan, C.A., Liermann, C.R., Davies, P.M., 2010. Global threats to human water security and river biodiversity. Nature 467, 555–561. Wolfram, G., Höss, S., Orendt, C., Schmitt, C., Adamek, Z., Bandow, N., Großschartnera, M., Kukkonen, J.V.K., Leloup, V., López Doval, J.C., Muñoz, I., Traunspurger, W., Tuikka, A., Van Liefferinge, C., von der Ohe De, P.C., Deckere, E., 2012. Assessing the impact of chemical pollution on benthic invertebrates from three different European rivers using a weight-of-evidence approach. Sci. Total Environ. 438, 498–509.

Please cite this article as: López-Doval, J.C., et al., Ecological and toxicological responses in a multistressor scenario: Are monitoring programs showing the stressors or just showing ..., Sci Total Environ (2015), http://dx.doi.org/10.1016/j.scitotenv.2015.05.085