The use of Risk Incidence and Diversity Indices to Evaluate Reservoir Water Quality of Semi-arid Reservoirs
Evaldo de Lira Azevêdo, Rômulo Romeu Nóbrega Alves, Thelma Lúcia Pereira Dias,Joseline Molozzi
(Manuscrito publicado na Ecological Indicators, Qualis A2)
34
The use of Risk Incidence and Diversity Indices to Evaluate Reservoir Water Quality of Semi-arid Reservoirs
Evaldo de Lira Azevêdo1, Rômulo Romeu Nóbrega Alves1,2, Thelma Lúcia Pereira Dias1,2,Joseline Molozzi2
1Universidade Federal Rural de Pernambuco – UFRPE, Programa de Pós-Graduação em Etnobiologia e Conservação da Natureza, Rua Dom Manoel de Medeiros, s/n, Dois Irmãos, Recife, PE, CEP:
52171-900
2Universidade Estadual da Paraíba – UEPB, Programa de Pós-graduação em Ecologia e Conservação, CCBS, Rua das Baraúnas, 351, Bairro Universitário, Bodocongó, Campina Grande, PB, CEP: 58429-
500
ABSTRACT
Local ecological knowledge of water quality indicators can be an important element in the management and conservation of aquatic ecosystems from a biocultural conservation perspective. We investigated the ecological indicators of water quality used by local communities that live around reservoirs in the semi-arid region of Brazil to compare their Risk Incidence Indices with their Diversity Indices. This research was undertaken at four reservoirs, and interviewed 126 members of human communities, using a semi-structured format. The macrobenthic fauna was sampled at 332 sites within the reservoirs for the bio- indicator analyses. Among the ecological indicators of water quality cited by the interviewees, the most notable were water color and odor. When the Risk Incidence Index was high, the Biodiversity Index was found to be low. Larger proportions of Oligochaete occurred in reservoirs with low biological diversity, indicating that local ecological knowledge reflected reservoir water quality. Our studies showed that local ecological knowledge of water quality indicators, in conjunction with traditional indicators such as the macrobenthic fauna, can be used for biocultural conservation purposes.
Keywords: Local ecological indicators; Management; Environmental quality; Benthic macroinvertebrates; Risk analysis.
35 1. Introduction
Water supplies in the semi-arid region of Brazil largely depend on surface water accumulated in reservoirs (Araújo, 2003). These artificial ecosystems are essential for the social and economic development of the region and are integral to local historical and cultural contexts (Robock, 2011; Lima et al., 2012). Rural communities living around reservoirs generally use that water directly, without any purification process before consumption (World Health Organization and UNICEF, 2015), but invariably develop some form of local ecological knowledge (LEK) to judge water quality.
Local ecological knowledge corresponds to the accumulated knowledge of a specific group of people about a local ecosystem, and represents a fundamental link between social and ecological systems (Olsson and Folke, 2001; Eufracio-Torres et al., 2016; Silva Junior and Santos, 2017). Understanding this local knowledge is viewed as increasingly important in a new field of conservation – biocultural conservation. Biocultural conservation advocates biological conservation, but emphasizes the insertion of local knowledge, innovations, and practices into bottom-up management systems (Gavin et al., 2015). LEK can help our understanding of the positions taken by local communities in relation to their water resources – a perspective foreseen in the Convention on Biological Diversity (CBD, 1992).
Understanding local ecological knowledge of water quality indicators in hydrographic basins can aid in the management and conservation of those that resource, optimize the work of scientists and managers by furnishing new and valuable information from local communities (Cvitanovic et al., 2016), and encourage the implementation of biocultural conservation approaches (Gavin et al., 2015).
Studies have shown that the main indicators used by human communities to assess water quality are color (David, 1971; Moser, 1984; House and Sangster 1991; Smith et al., 1995;
Cottet et al., 2013; West et al., 2015) and odor (Nare et al., 2006; West et al., 2015). These studies have demonstrated that the organoleptic properties of water are fundamental to the evaluation of their quality by resource users (Espinosa-García et al., 2014). Studies have shown that human communities use biological indicators such as changes in assemblages of aquatic organisms (including the death of fish) to assess water quality (Nicolson and Mace, 1975; Hallwass et al., 2013).
Perception is a phenomenon that involves exposure to stimuli as well as environment familiarity that can be interpreted cognitively (Ingold 2000; Spinosa-Gacía et al., 2014, Albuquerque et al., 2014; West et al., 2015). An individual's exposure to risk situations develop what is called risk perception (Jakus et al., 2009; Albuquerque et al., 2014;
36 Rambonilaza et al., 2016). Perceptions of that sort about water resources can directly affect water use, and can also be associated with socioeconomic factors (Artell et al., 2013). Afroz et al. (2015), for example, evaluated water pollution risk perceptions among Malaysian families, and observed that factors such as gender, age, and educational level influenced each individual's perception of water quality. Although requiring qualitative assessments, risk perceptions can be quantified by calculating the Risk Incidence Index (Ij) (Albuquerque et al., 2014).Although simple, the Risk Incidence Index (Ij) calculation (Albuquerque et al., 2014) may be important for an exploratory analysis and identification of local risks. However, there are other approaches that can be employed, as development of models for assist of decision makers (Li et al., 2016; Li et al., 2017; Chen et al., 2017). In addition to understanding local ecological knowledge in terms of indicators of water quality and applying Risk Incidence Indices, it is also important to investigate how these elements actually reflect water quality in local ecosystems. To that end, a combination of water quality assessment methodologies should be employed. Traditionally water quality assessment is performed through chemical and physical analyzes (e.g. transparency, temperature, oxygen concentration, and nutrient concentrations) (Goulart and Callisto, 2003) or through the analysis of bioindicators such as aquatic macrophytes (Faria et al., 2013; Farnese et al., 2014; Gonçalves et al., 2015), phytoplankton (Vilhena et al., 2014; Schuster et al., 2015), zooplankton (Sousa et al., 2008;
Perbiche-Neves et al., 2013; Azevêdo et al., 2015), fish (Sarpedonti et al., 2013; Osório et al., 2014; Steffens et al., 2015), and the macrobenthic fauna (Pereira et al., 2012; Molozzi et al., 2013; Uherek and Gouveia 2014; Azevêdo et al., 2015; Melo et al., 2015).
The macrobenthic fauna has been widely used in bio-monitoring. Those organisms comprise several taxonomic groups, with Oligochaeta, Mollusca, and Diptera predominating in Brazilian reservoirs (Pamplin et al., 2006; Molozzi et al., 2013; Azevêdo et al., 2017).
Changes in structural aspects and the distributions of their biological communities reflect long-term environmental changes (Morgan et al., 2006; Macedo et al., 2014; Martins et al., 2015) and differ from physical and chemical water assessments (which have greater potential for diagnosing short-term changes) (Goulart and Callisto, 2003).
Within that context, we investigated the indicators of water quality degradation used by local communities living in around reservoirs in the semi-arid region of Brazil, and investigated whether their perceptions of those indicators are related to socioeconomic characteristics. Evaluated whether there are differences in local knowledge between different communities. Moreover, we also applied the Risk Incidence Index and the Biological Diversity Index to relate risk perceptions to macrobenthic fauna. Our work was guided by the hypotheses: 1 - The main local ecological indicators of reservoir water quality are the colors
37 and the odors of the water and the macrobenthic fauna composition; 2 - Local ecological knowledge is correlated with socioeconomic variables (gender, age, activity, income, educational level, and time of residence around the reservoir); 3 - There are differences in local ecological knowledge about water quality indicators among human communities; and, 4 - The highest Risk Incidence Index values occur in communities where local reservoirs have the lowest Biodiversity Index.
2. Materials and methods 2.1. Study area
The study sites were located in the semi-arid region of northeastern Brazil (Pereira- Junior, 2007), which has a marked overall annual water deficit and has recently been experiencing since 2012 one of the most severe droughts in the last 50 years (World Meteorological Organization, WMO, 2015). The region has mean annual temperatures above 20 °C, with annual precipitation varying between 280 and 800 mm. Evapotranspiration is greater than precipitation, with rainfall being extremely variable, but usually concentrated in three or four months of the year (Araújo, 2011; Alvares et al., 2013). These conditions make reservoirs essential for the establishment and persistence of human life in the region.
The present study was developed in two Brazilian states (Rio Grande do Norte and Paraíba), evaluating local communities and reservoirs in two municipalities in each state: the Sumé (7°40'14.86 "S, 36°54'25.57" W, municipality of Sumé) and Poções (7°53'33.20 "S;
37°00'31.54" W, municipality of Monteiro) reservoirs were studied in Paraíba State, focusing on the local communities living around them, both situated in the Paraíba river watershed; the Traíras (6°30'52.99"S, 36°55'58.50" W, municipality of São José do Seridó) and Sabugí (6°39'10.79"S; 37°12'20.55"W, municipality of São João do Sabugí) reservoirs, in Rio Grande do Norte state, as well as the local communities around them, are both located in the Piranhas- Assú river watershed (Fig. 1).
38 Fig. 1. Map of the communities and reservoirs surveyed, in the Piranhas-Assú (A Sabugí, B Traíras) and Paraíba (C Sumé, B Poções) river watershed, Brazil. Where: yellow points = sampling sites of macrobenthic fauna and red points = visited residence.
2.2. Design and ethnobiological data collection
The objectives of the study were explained to the participants before each interview, and their permission was obtained for recording their information by signing a Free and Informed Consent Term, according to Resolution 466/2012 of the Brazilian National Health Council. Ethical approval for the study was obtained from the Ethics Committee of the State University of Paraíba – UEPB (Approval Nº 1.030.872).
Ethnobiological data was collected in September and October/2015. A semi-structured form was developed for the interviews to collect socio-economic data and information about the perceptions of the local community of water quality indicators. A total of 126 people were interviewed, 22 in the community around the Sumé reservoir, 38 in Poções, 31 in Traíras, and 35 in Sabugí. We interviewed one person per residence in houses located up to 150m from the margin of the reservoir. The choice of those interviewees was based on the greater probability of their maintaining contact with the reservoir because they lived quite around it. The residences about 150m from the margin of the reservoirs were previously selected using satellite images. All the residences that had residents were visited. The interviews always were performed with a householder (female or male).
39 Of the individuals interviewed, 64.3% (n = 81) were males and 35.7% (n = 45) females, ranging in age from 18 to 82 years (51.1 ± 14.14); 24.6% (n = 31) of the interviewees were illiterate, 3.2% (n = 4) had completed their elementary education, 60.3% (n = 76) had not completed their elementary education, 3.2% (n = 4) had completed secondary school education, 7.1% (n = 9) had not completed secondary school education, and 1.6% (n = 2) had completed their graduation. Most of the interviewees were farmers (70%, n = 88), 16% (n = 20) fishermen, 6.3% (n = 8) home caregivers, and 7.7% (n = 10) exercised other professions (merchants, civil servants). Residence times around the reservoirs ranged from 1 to 81 years (25.80 ± 18.40). Their average annual income was R$735.71 (± 625.66) (equivalent to approximately US$236.00).
2.3. Design and collection of biological data
A total of 332 samples of macrobenthic fauna were collected during June, September, and December/2014, and March/2015, being 25 sites to Sumé and 25 to Poções, 31 to Sabugí and 10 to Traíras. The samples were collected along the littoral of the reservoirs at average depth of 50 cm. The sampling were performed this zone, since presenting a highest diversity of the macrobenthic community. The Sediment samples were collected with the aid of an Eckman-Birge dredge (225 cm2) and subsequently preserved in 4% formaldehyde. The samples were then removed to the laboratory and washed over a sieve (500 μm) and stored in plastic jars in 70% ethanol. The material was viewed with the aid of a stereomicroscope and identified by consulting the specialized literature (Hawking and Smith, 1997; Mugnai et al., 2010). Most of the organisms were identified to the family level, with only the Chironomidae larva (Diptera, Insecta) being identified to the genus level (Trivinho-Strixino and Strixino, 1995; Trivinho-Strixino, 2011).
2.4. Risk Incidence Index
In order to evaluate the Risk Incidence related to the water quality indicators, a Risk Incidence Index (Ij) was calculated according to the following formula:
I
j= nr/nj
Where:
Ij = Risk Incidence Index
nr = number of times a given risk was cited.
nj = total number of interviewees.
40 The values of Ij can vary between 0 and 1, with the highest values corresponding to higher perceived risks (Albuquerque et al., 2014). The total mean of the Risk Incidence Index as evaluated by the local community was also calculated for an overall assessment of risk.
2.5. Index of Biological Diversity
We used the Shannon-Wiener Diversity Index (Shannon and Weaver, 1949), calculated for each sampling site, to evaluate the diversity of macrobenthic fauna, considering the abundance matrix of the biological community. This index considers the richness and equitability, giving equal weight to rare and abundant species (Magurran, 2013). The Shannon-Wiener Diversity Index was calculated by the following formula:
H’= - ∑ p
i lnp
iWhere:
H’= Shannon-Wiener Diversity Index
pi = number of individuals sampled from the ith species ln = neperian logarithm base
2.6. Data analysis
To analyze the proportions of local ecological indicators for water quality assessment, the percentages of citations provided from interviewees were calculated. Radar graphs were constructed for visualization of these proportions.
The socioeconomic data and local ecological knowledge of water quality indicators were categorized, and the information was replaced by a corresponding number. After categorization, the data were log (x + 1) transformed and the Euclidean Distance was selected as dissimilarity matrix. To verify the existence of differences in local ecological indicators of water quality among local communities, a Permutational Analysis of Variance (PERMANOVA) was performed, using 9999 permutations with p ≤0.05 (Anderson, 2001;
Anderson et al., 2008), considering one factor (community) and four levels (the communities of Sumé, Poções, Traíras, and Sabugí). A "Post-hoc test" was performed to assess differences in local ecological indicators for water quality degradation among pairs of local communities.
Principal Component Analysis (PCA) was performed to visualize the spatial patterns and local ecological indicators of water quality most important for explaining the data in each
41 community (Anderson et al., 2008). Spearman correlation analysis with a correlation factor of 0.3 was performed to evaluate the existence of a correlation between local ecological indicators of water quality and socioeconomic variables (sex, age, activity, schooling level, income, and residence time around the reservoir).
A similarity percentage analysis (SIMPER) was used to analyze which taxa of the macrobenthic fauna contributed up to 95% to the composition of the fauna in each reservoir.
Biological data were square root transformed and the Bray-Curtis coefficient was considered (Clarke and Warwick, 2001).
The value of the Shannon-Wiener Diversity Index was used to evaluate the existence of differences in diversity between the reservoirs, using an analysis of significance with 9999 permutations, p ≤ 0.05 (Permutational Analysis of Variation Univariate) (Anderson, 2001;
Anderson et al., 2008). The Euclidean Distance coefficient was used as the dissimilarity matrix for this analysis; the mean diversity for each reservoir was also calculated.
A "Box plot" was constructed to visualize the relationship between the values of the Shannon-Wiener Diversity Index and the values of the Risk Incidence Index, as calculated from the local ecological indicators of water quality in each reservoir (Ayres, 2007).
Except for the Spearman correlation analysis, which was performed in the Bio Estat 5.0 statistical package (Ayres, 2007), all of the other analyses were performed using PRIMER-6 + PERMANOVA software (Systat Software, Cranes Software International Ltd., Anderson et al., 2008).
3. Results
3.1. Local ecological knowledge: indicators of water quality
The majority of the interviewees (96.82%) stated that they could perceive changes in water quality in the reservoirs. The main indicators for evaluating water quality in terms of local ecological knowledge were color and odor; there were also citations related to the taste of the water and the presence of organisms of the macrobenthic fauna, although in smaller proportions.
The green color of degraded water was frequently cited, mainly by community members around the Poções and Traíras reservoirs (82% and 61.3% of the respondents respectively) (Fig. 2, A and B). Bad smells were mentioned by most of the interviewees, who reported odors as indicators of poor water quality (68% in Traíras, 37% in Poções, 36% in
42 Sumé, and 20% in Sabugí). A brackish taste was reported as an indicator of poor water quality in the Traíras (39%), Sumé (32%), and Poções (8%) communities (Fig. 2, A and B).
Approximately 8% of the interviewees in the Poções community noted increases in the abundance of the exotic mollusk Melanoides tuberculata (Müller, 1774) (known locally as Aruá) as an indicator of water quality degradation; interviewees from the Poções and Traíras communities cited insect larvae as an indicator of water quality degradation (5% and 3%
respectively); interviewees from the Poções (3%) and Sumé (5%) reservoir communities cited beetles as indicators of poor water quality; and interviewees from the Poções (3%) and Sabugí (3%) reservoir communities cited red insect larvae as poor water quality indicators (Fig. 2, A and B).
43 Fig. 2. A - Local ecological indicators for communities living around the Poções and Sumé reservoirs, Paraíba river watershed. B - Local ecological indicators for communities living around the Traíras and Sabugí reservoirs, the Piranhas-Assú river watershed.
44 Some interviewees from the communities around the Poções and Sabugí reservoirs cited dead fish as an indicator of bad water quality (8% and 3% respectively). Additional local ecological indicators of water quality degradation were cited in smaller proportions, including the disappearance of aquatic animals, the presence of aquatic plants, the disappearance of aquatic plants, and other diverse observations (e.g., the lack of growth of fish, green substances in the water, bad water for washing clothes, the presence of garbage, water staining aluminum utensils and foaming when boiled) (Fig. 2). There were significant differences in local ecological indicators of water quality between communities, except for the communities living around the Sabugí and Traíras reservoirs (Table 1).
Table 1 Results of the PERMANOVA and Post-hoc tests analysis to assess the differences between local indicators for the degradation of water quality among the communities around the researched reservoirs.
DF MS F P-perm Permutations
Local ecological indicators
Communities 4 3.1449 4.6575 0.0001 9916
Resídual 133 0.67524
Total 126
Post-hoc tests
Communities T P-perm
Poções x Sumé 1.8951 0.0042
Poções x Traíras 2.1791 0.0003
Poções x Sabugí 2.6952 0.0002
Traíras x Sumé 1.637 0.0336
Traíras x Sabugí 1.4679 0.0716
Sabugí x Sumé 2.5777 0.0006
45 The local indicators of water quality that showed the highest correlation for data segregation in the Poções community were odor (-0.490-Axis1, -0.759-Axis2) and diverse other aspects (0.840-Axis1, -0.493-Axis2); in Sumé the main indicators were odor (-0.757- Axis1, -0.504-Axis2) and water color changes (-0.560-Axis1, 0.818-Axis2); in the Traíras community, the correlated indicators were odors (0.678-Axis1, 0.350-Axis2), changes in water color (-0.693-Axis1, 0.403-Axis2), and changes in water taste (-0.806-Axis2); in Sabugí the indicators with the highest correlations were water color changes (0.951-Axis2) and odor (-0.964-Axis1) (Fig. 3, Table 2).
Fig. 3 Principal Components Analysis showing the main local indicators of water quality in each community studied; O (Odor), AP (Presence of aquatic plants), AA (Presence of aquatic animals), WC (Color), DP (Disappearance of aquatic plants), DA (Disappearance of aquatic animals), TW (Taste of water), AW (Apperance of the water) and OT (Diverse). A = Poções, B = Sumé, C = Traíras, and D = Sabugí.
46 Table 2 Correlation values for axes 1 and 2 of Principal Component Analysis (PCA). Bold numbers indicate correlation> 0.30
3.2. Relationships of local indicators of water quality parameters with socioeconomic variables
When correlating local ecological indicators for evaluating water quality with socioeconomic variables, only the ages of the interviewees showed a positive and significant correlation with water color changes (r = 0.175, p = 0.04) (Supplementary material).
3.3. Macrobenthic fauna as an indicator of reservoir water quality
The analysis of the macrobenthic fauna using SIMPER indicated that the organisms that contributed most to the aquatic community composition were Oligochaeta (with 67%
contribution in the Traíras reservoir, 54% in Poções, 41.93% in Sabugí, and 39% in Sumé) and the Gastropoda Melanoides tuberculata (representing 53.13% of the fauna in Sumé, 43%
in Poções, 26% in Traíras, and 7.25% in Sabugí) (Table 3).
Local Ecological Indicators / Communities
Poções Sumé Traíras Sabugí
1º Axis 28.3%
2º Axis 24.7%
1º Axis 42.9%
2º Axis 24.9%
1º Axis 43.5%
2º Axis 23%
1º Axis 53.3%
2º Axis 17%
Odor water -0.490 -0.759 -0.757 -0.504 0.678 0.350 -0.964 -0.073 Presence aquatic
plants
0.011 0.079 0.000 0.000 -0.062 0.069 0.005 0.023
Presence aquatic animals
-0.139 0.253 -0.148 -0.071 -0.025 0.048 0.034 0.019
Color water -0.137 0.107 -0.560 0.818 -0.693 0.403 -0.022 0.951
Disappearance aquatic plants
-0.001 0.019 0.000 0.000 0.000 0.000 0.009 -0.055
Disappearance aquatic animals
0.061 -0.023 0.000 0.000 0.000 0.000 0.011 0.002
Taste of water -0.112 -0.244 0.270 0.240 0.005 -0.806 -0.099 -0.133 Appearance of the
water
-0.026 0.195 -0.115 -0.067 0.128 0.037 -0.046 0.071
Divers 0.840 -0.493 0.070 -0.098 0.196 0.239 0.237 -0.250