This is the accepted version of the journal article: Barba, Carme; Folch, Albert; Sanchez-Vila, Xavier; [et al.]. «Are dominant microbial sub-surface communities affected by water quality and soil characteris- tics?». Vol. 237 (May 2019), p. 332-343. DOI 10.1016/j.jenvman.2019.02.079 This version is available at https://ddd.uab.cat/record/287666 under the terms of the license *Manuscript Click here to view linked References 1 Are dominant microbial sub-surface communities affected by 1 2 2 water quality and soil characteristics? 3 4 5 3 Carme Barbaa,b,* (carme.barba@upc.edu), Albert Folcha,b (folch.hydro@gmail.com), 6 7 4 Xavier Sanchez-Vilaa,b (xavier.sanchez-vila@upc.edu), Maira Martínez-Alonsoc 8 5 (maira.martinez@uab.cat), Núria Gajuc (nuria.gaju@uab.cat) 9 10 6 11 a 12 7 Department of Civil and Environmental Engineering, Universitat Politècnica de Catalunya (UPC), C/Jordi 13 8 Girona 1-3, 08034 Barcelona, Spain 14 b 15 9 Associated Unit: Hydrogeology Group (UPC-CSIC) 16 c 10 Department of Genetics and Microbiology, Universitat Autònoma de Barcelona (UAB), 08193 Bellaterra, 17 18 11 Spain 19 20 21 12 Abstract 22 23 24 13 Subsurface microorganisms must deal with quite extreme environmental conditions. 25 26 14 The lack of light, oxygen, and potentially nutrients are the main environmental stresses 27 28 15 faced by subsurface microbial communities. Likewise, environmental disruptions 29 30 16 providing an unbalanced positive input of nutrients force microorganisms to adapt to 31 32 33 17 varying conditions, visible in the changes in microbial community diversity. In order to 34 35 18 test microbial community adaptation to environmental changes, we performed a study 36 37 19 in a surface Managed Aquifer Recharge facility, consisting of a settlement basin (two- 38 39 20 day residence time) and an infiltration pond. Data on groundwater hydrochemistry, soil 40 41 42 21 texture, and microbial characterization was compiled from surface water, groundwater, 43 44 22 and soil samples at two distinct recharge operation conditions. 45 46 47 23 Multivariate statistics by means of Principal Component Analysis (PCA) was the 48 49 24 technique used to map the relevant dimensionality reduced combinations of input 50 51 25 variables that properly describe the system behavior. The methodology selected allows 52 53 54 26 including variables of different nature and displaying very different range values. Strong 55 56 27 differences in the microbial assemblage under recharge conditions were found, 57 58 28 coupled to hydrochemistry and grain-size distribution variables. Also, some microbial 59 60 29 groups displayed correlations with either carbon or nitrogen cycles, especially showing 61 62 1 63 64 65 30 abundant populations of denitrifying bacteria in groundwater. A significant correlation 1 2 31 was found between Methylotenera mobilis and the concentrations of NO3 and SO4, and 3 4 32 also between Vogesella indigofera and the presence of DOC in the infiltrating water. 5 6 7 33 Also, microbial communities present at the bottom of the pond correlated with 8 9 34 representative descriptors of soil grain size distribution. 10 11 12 35 1. Introduction 13 14 15 36 Groundwater systems are perceived as relatively stable environments as compared to 16 17 37 most aquatic ecosystems (Zhou et al., 2012). Despite that, investigations have shown 18 19 38 that soil-aquifer systems support a wide diversity of organisms (e.g., Griebler and 20 21 22 39 Lueders, 2009). Actually, the unsaturated zone, and more specifically the topsoil, 23 24 40 supports the highest microbial activity and biomass of all compartments within the 25 26 41 subsurface environment (Lapworth et al., 2012). Likewise, microorganisms are 27 28 42 responsible for most biological processes in aquifers (Stein et al., 2010). 29 30 31 43 Several studies evidence microbial adaptation to groundwater extreme environments 32 33 34 44 (thermal or hypersaline) (e.g., Rothschild and Mancinelli, 2001) or disturbed by human 35 36 45 activities (Meckenstock et al., 2015). Human activities have caused disruption in 37 38 46 aquifer dynamics to some extent (Griebler and Lueders, 2009; et al., 2017), with 39 40 47 biological implications as indigenous microorganisms can acclimate (Pett-Ridge and 41 42 43 48 Firestone, 2005) or even take advantage (Rezanezhad et al., 2014) of environmental 44 45 49 disturbances. 46 47 48 50 In fact, the water treatment industry has benefited from the adaptability and metabolic 49 50 51 capabilities of microorganisms to maximize the improvement of water quality. Several 51 52 52 laboratory experiments and engineering applications have tested the effectiveness of 53 54 55 53 microbial engineered techniques for water reclaim purposes. The former has been 56 57 54 conducted aiming at (1) describing degradation pathways of specific pollutants and 58 59 55 quantifying their degradation rates (Greskowiak et al., 2017; Regnery et al., 2015; 60 61 62 2 63 64 65 56 Rodriguez-Escales and Sanchez-Vila, 2016), (2) determining the physical and 1 2 57 hydrochemical conditions that can govern the behavior of specific microbial groups 3 4 58 (Alidina et al., 2014; Drewes et al., 2014; Freixa et al., 2015; Kolehmainen et al., 2008; 5 6 7 59 Perujo et al., 2017), or (3) understanding the role of organic matter (dissolved organic 8 9 60 carbon -DOC- or micropollutants), on the growth of microbial communities (Li et al., 10 11 61 2013, 2012). Regarding the engineered applications of microbial ecology designed to 12 13 62 improve the quality of reclaimed water, examples are constructed wetlands (Faulwetter 14 15 63 et al., 2009; Truu et al., 2009; Zhang et al., 2018) and sand filters (D’Alessio et al., 16 17 18 64 2015). 19 20 21 65 Natural and induced microbial attenuation have been studied and applied at the field 22 23 66 scale concerning groundwater related environmental issues; this includes, e.g., landfill 24 25 67 leakage affections (Röling et al., 2001; Staats et al., 2011), contaminant spills 26 27 68 (Fahrenfeld et al., 2014; Haack et al., 2004; Martínez-Pascual et al., 2010; Nijenhuis 28 29 30 69 and Kuntze, 2016), or nitrate polluted aquifers (Bellini et al., 2018, 2013). 31 32 33 70 In recent years, the number of soil aquifer treatment (SAT) facilities have increased 34 35 71 worldwide. SAT is a particular case of the Managed Aquifer Recharge (MAR) family, 36 37 72 that combines the replenishment of groundwater bodies with the treatment of water 38 39 40 73 during infiltration, by taking advantage of the potential of the soil for the degradation of 41 42 74 subsurface microbial communities (Bouwer, 2002). Studies testing the link between 43 44 75 water quality and microbial communities in MAR systems depend on the system type, 45 46 76 whether recharge wells (Ginige et al., 2013), riverbank areas (Schütz et al., 2009), or 47 48 77 surface infiltration ponds (Barba et al., 2019; Reed et al., 2008; Regnery et al., 2016). 49 50 51 78 Infiltration ponds are low-cost, low-tech, passive facilities compared to advanced water 52 53 79 treatment methods (Drewes et al., 2003; San-Sebastián-Sauto et al., 2018); for these 54 55 80 reasons, they are widely implemented, mostly in arid or semi-arid environments (Goren 56 57 81 et al., 2014; Greskowiak et al., 2006; Rodríguez-Escales et al., 2018). 58 59 60 61 62 3 63 64 65 82 Biodegradation processes in surface infiltration ponds include aerobic oxidation of DOC 1 2 83 (Maeng et al., 2011; Mermillod-Blondin et al., 2015), denitrification (Grau-Martínez et 3 4 84 al., 2018), and removal of some emerging organic compounds (Hamann et al., 2016; 5 6 7 85 Valhondo et al., 2015). Parameters modifying bioprocesses, such as DOC availability, 8 9 86 dissolved oxygen content, redox conditions, temperature, nutrient concentrations, or 10 11 87 soil moisture have been widely reported (Alidina et al., 2015; Bekele et al., 2011; Dutta 12 13 88 et al., 2015; Goren et al., 2014; Greskowiak et al., 2005; Hellauer et al., 2017; Laws et 14 15 89 al., 2011; Massmann et al., 2006; Rezanezhad et al., 2014). Apart from environmental 16 17 18 90 parameters, properties of the porous media, such as porosity and the distributions of 19 20 91 grain sizes and pore sizes, determine the spatial and temporal distributions of 21 22 92 microorganisms in soils (Chau et al., 2011); they also influence hydraulic conductivity, 23 24 93 and thus the access to nutrients, affecting in a distinct way the growth and activity of 25 26 27 94 microorganisms (Perujo et al., 2018, 2017). 28 29 30 95 Subsequently, determining how much and to what extent physical, geochemical, 31 32 96 biological and operational parameters influence a SAT system functioning is useful for 33 34 97 managing purposes (Dutta et al., 2015; Grau-Martínez et al., 2018; Hellauer et al., 35 36 98 2017; Pedretti et al., 2012a; Rodríguez-Escales et al., 2017). Yet, it is difficult to find 37 38 39 99 multidisciplinary research dealing with integrated approaches to improve understanding 40 41 100 of infiltration problems. One such example is the Llobregat MAR surface infiltration 42 43 101 system, where a number of experiments have been performed in the last decade. The 44 45 102 system is, thus, well characterized in terms of natural and induced flow regime, 46 47 48 103 including numerical modelling (Valhondo et al., 2016), DOC mapping (Valhondo et al., 49 50 104 2015) and the evaluation of nitrate attenuation (Grau-Martínez et al., 2018), and the 51 52 105 fate of several emerging compounds (Valhondo et al., 2018) along the infiltration path. 53 54 106 Finally, microbial fingerprinting was studied, and the spatial distribution of dominant 55 56 57 107 microbial phylotypes was linked to the overall recharge processes (Barba et al., 2019). 58 59 108 In that work, the microbial assemblage was characterized and discussed under an 60 61 62 4 63 64 65 109 ecological point of view. However paramount aspects, such as the effect of 1 2 110 hydrochemical composition or the grain-size distribution, in water and soil samples, 3 4 111 linked to their role in the presence of specific microbial signatures were not addressed. 5 6 7 112 We contend that a full analysis of processes occurring in any MAR facility should 8 9 10 113 involve the simultaneous study of physical, hydraulic, geochemical and microbial data; 11 12 114 therefore, it includes the joint analysis of continuous, discrete and categorical data, with 13 14 115 different ranges of values and resolution windows, calling for the use of multivariate 15 16 116 (MV) statistical techniques. Such techniques have been widely used in hydrogeology to 17 18 19 117 provide process understanding and to accompany groundwater models (El Alfy et al., 20 21 118 2017; Menció et al., 2012). In the case of microbial ecology, the development of 22 23 119 molecular analyses allowed the generation of large data sets, best treated using MV 24 25 120 statistical techniques (see Paliy and Shankar, 2016 for a review). 26 27 28 121 Here, we aim at combining ecological, hydrochemical and hydrological approaches to 29 30 31 122 understand the spatial distribution of subsurface microbial communities under changing 32 33 123 recharge conditions, with significant consequences in the management of MAR 34 35 124 facilities. For this purpose, we used a physical-bio-geochemical dataset from an 36 37 125 existing infiltration facility, and then applied PCA, aiming at statistically discriminate the 38 39 40 126 relationships among different parameters corresponding to microbial community 41 42 127 structure, geochemical variables, soil grain size distribution, and operational conditions. 43 44 45 128 This work allows providing the most relevant microbial indicators present in the system 46 47 129 and to correlate them with soil and groundwater local characteristics and feeding water. 48 49 130 Furthermore, the existing data between microbial clades both in water and soil were 50 51 52 131 also analyzed separately to obtain further relevant inter-clade correlations. 53 54 55 132 56 57 58 133 2. Material and methods 59 60 61 62 5 63 64 65 134 The present work is built on the dataset compiled and presented in Barba et al. (2019). 1 2 135 This work describes in detail the fieldwork procedures and laboratory tasks performed 3 4 136 to obtain data representative of the Llobregat MAR site regarding a large set of 5 6 7 137 physical, geochemical and biological variables, and from two sampling campaigns 8 9 138 involving different recharge conditions. In this section, we provide only the essential 10 11 139 information to contextualize the statistical analysis, being the core of this work. 12 13 14 140 2.1. Study site 15 16 17 141 The field site is located in the Llobregat River Basin (Sant Vicenç dels Horts, 18 19 142 Catalonia), 10 km SW of Barcelona. The recharge system replenishes the Lower Valley 20 21 22 143 Aquifer, a strategic groundwater source for the Barcelona Conurbation. It is an alluvial 23 24 144 aquifer, composed mainly by sands and gravels with small paleo-channels of clay 25 26 145 (Pedretti et al., 2012b). Hydrogeologically, it is of high transmissivity that reaches 27 28 146 14000 m2·d-1 locally, with an average thickness of 15m. 29 30 31 147 The system is composed of two ponds. The first one (7300 m2) is fed with river water, 32 33 34 148 and acts as a sedimentation pond for fine particles, with 2-3 day residence time. Then, 35 36 149 water passes through a pipe to the second pond (6500m2), where it infiltrates to the 37 38 150 aquifer through 4-10m of vadose zone. Being devised as a research facility, in 2011 an 39 40 151 organic layer was placed at the bottom of the infiltration pond to test its efficiency for 41 42 43 152 the removal of emergent pollutants by the addition of labile DOC to the inflow water. 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 6 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 153 28 29 30 154 Figure 1 Geographical situation and cross-section of the Llobregat MAR System. Triangles and circles 31 32 155 represent the location of soil samples. Diamond and squares represent those of water and groundwater 33 156 samples, respectively. Blue symbols indicate samples taken in July 2014, after six months of continuous 34 35 157 uninterrupted recharge. Orange samples were taken after recharge was discontinued for 4 months in 36 37 158 March 2015. 38 39 40 159 The system includes a network of piezometers for monitoring and groundwater 41 42 160 sampling corresponding to different travel times after recharge (Figure 1). Detailed 43 44 161 information about the conceptual regional flow model and the local flow and transport 45 46 47 162 models can be found in Barba et al. (2019) and Valhondo et al. (2016), respectively. 48 49 50 163 2.2. Input data for the multivariate statistical approach 51 52 53 164 Water and soil samples were taken in two different periods, July 2014 and March 2015, 54 55 165 representing different recharge operation conditions. In the July campaign, sampling 56 57 166 was performed after six months of uninterrupted recharge; in the March one, recharge 58 59 60 167 had been discontinued for four months prior to sampling. In total, 21 water samples (20 61 62 7 63 64 65 168 for groundwater and 1 for recharge water) and 10 soil samples were subjected to 1 2 169 molecular analyses, in duplicate (only average values are presented in this work). 3 4 170 Furthermore, water samples were hydrochemically characterized, while granulometric 5 6 7 171 curves were obtained for the soil samples. The sampling and analysis methodologies 8 9 172 for chemistry, granulometry, DNA extraction, PCR, DGGE, diversity indices and relative 10 11 173 abundances of the microbial clades, are reported in Barba et al. (2019). 12 13 14 174 The present study divides the statistical analyses according to the nature of the 15 16 175 samples, treating those from water and soil separately, as different variables are 17 18 19 176 included in each case. In the former, the statistical analysis accounted for (1) microbial 20 21 177 diversity indexes, (2) relative microbial abundances at the taxonomical rank level of 22 23 178 class and species, and (3) hydrochemical characterization (major, minor and trace 24 25 179 elements). Instead, statistical analyses of soil samples contained (1) microbial diversity 26 27 180 indexes, (2) relative microbial abundances at the taxonomical rank level of class and 28 29 30 181 species, (3) grain-size distribution representative parameters, (4) depth, and (5) an 31 32 182 operational binary variable indicating dry/wet conditions characterized by the presence 33 34 183 or absence of recharge. Finally, a general statistical analysis was performed for all soil 35 36 184 and water samples (31 in total) and the common variables (i.e., (1) and (2) from the 37 38 39 185 previous lists), plus a binary variable indicating the sample nature (soil or water). 40 41 42 186 2.3. Rationale for the multivariate statistical approach: requirements for PCA 43 44 45 187 Multivariate statistics is a useful technique to treat large datasets involving different 46 47 188 types of variables, allowing the inclusion of quantitative and categorical data together. 48 49 189 In short, PCA transforms a set of data values of variables that in principle are 50 51 52 190 correlated into a set of values of linearly uncorrelated variables called principal 53 54 191 components (PCs). PCs are defined such that the first one accounts for as much of the 55 56 192 variability in the data as possible, and each succeeding component, in turn, has the 57 58 193 highest variance possible under the constraint that it is orthogonal to the preceding 59 60 61 62 8 63 64 65 194 ones. Thus, a multidimensional system in terms of variables is projected into a low 1 2 195 dimensional map of components. Solutions were subjected to a varimax rotation of the 3 4 196 original system corresponding to the directions of the largest variance in the dataset. 5 6 7 197 Statistical analysis was done using software SPSS (IBM SPSS Statistics 24). 8 9 10 198 The full dataset is provided as supplementary material. It also includes plots 11 12 199 corresponding to some of the PCA analyses that are included for completeness but do 13 14 200 not provide enough additional information to merit inclusion in the body of the text. 15 16 17 201 2.3.1. Selecting variables for the statistical analysis of water samples 18 19 20 202 Statistical parametric methods perform best when data follows a unimodal symmetric 21 22 23 203 distribution (Paliy and Shankar, 2016). For this reason, some variables from the initial 24 25 204 dataset were eliminated, grouped and/or transformed in order to conform better to the 26 27 205 assumptions of PCA analysis. For example, hydrochemical variables with most values 28 29 206 below the detection limit were eliminated (this is the case of Al, B, Cd, Co, Fe, Pb, and 30 31 207 P). Also, Ba concentration values were rejected for data inconsistency. Second, 32 33 34 208 hydrochemical variables that displayed a positive skewness were log transformed (Cu, 35 36 209 DOC, Ni, S and V concentrations). Finally, sample depth, electrical conductivity (EC), 37 38 210 temperature, and the concentrations of HCO3, Ca, Cl, Li, Mg, Mn, Na, NO3, pH, Si, 39 40 211 SO4, and Zn were added to the analysis without any data transformation. 41 42 43 44 212 With reference to variables of microbial abundances, they were all transformed into 45 46 213 new variables. For ecological data, displaying a large number of zero values, Legendre 47 48 214 and Gallagher (2001) recommend applying either the Chord or the Hellinger 49 50 215 transformations to the data values. In our case, we applied the latter one, because after 51 52 216 some preliminary analyses, the transformed variables presented higher correlation 53 54 55 217 coefficients (r2) than the non-transformed ones. Hellinger transformation ( ′ ) of a 56 57 218 datum pertaining to the -class and -object is given as: 58 59 60 61 62 9 63 64 65 219 ′ = ; where i+ denotes all i’s. 1 ∑ 2 3 4 220 Finally, in the case of microbial diversity indices, the input data for the PCA were the 5 6 221 Shannon, Richness and Evenness index values for each sample. 7 8 9 10 222 The next step was to perform bivariate correlation analyses with all variables taken two 11 12 223 by two, including hydrochemical variables (both raw and log-transformed; actually this 13 14 224 was the way of selecting whether the log-transformation was finally applied or not) and 15 16 225 Hellinger-transformed microbial variables in order to select the ones that would be used 17 18 226 in the final PCA analyses. Since the objective was to emphasize the correlations with 19 20 21 227 the microbial data, most redundant geochemical variables, as well as those displaying 22 23 228 very low values of the Pearson r2 coefficient to the biological ones, were removed. This 24 25 229 was the case of As, HCO3, Ca, Li, Mg, Mn, Si, Zn, and also depth, pH, and 26 27 230 temperature. In the case of microbial phylotypes, some of them were eliminated 28 29 30 231 because they did not show significant correlations with other variables; e.g., Bacilli, 31 32 232 Acidobacteria, Actinobacteria, Chlorobia, Nitrospira, Gammaproteobacteria, 33 34 233 Alphaproteobacteria, Subgroup3 sp (Acidobacteria), Stenotrophomonas sp, 35 36 234 Chryseomicrobium sp, Nitrospira1 sp, Methylobacterium sp, and Nitrospira2 sp. 37 38 39 235 The selected variables could not be incorporated all together in a single PCA because 40 41 42 236 the total variance was too high to allow proper discrimination of components. Then, 43 44 237 different PCAs (each one involving a different subset of parameters) were performed, 45 46 238 trying to maximize the amount of explained information. We performed two PCAs for 47 48 239 microbial abundances at the class level and two more at the species level. The most 49 50 51 240 informative set of variables were selected as the subset that maximized the measure of 52 53 241 sampling adequacy reported by Kaiser-Meyer-Olkin test (KMO); the variables included 54 55 242 in each analysis are provided in Table 1. 56 57 58 243 2.3.2. Selecting variables for the statistical analysis of soil samples 59 60 61 62 10 63 64 65 244 Soil samples were subjected to sieve analysis, obtaining grain-size distribution curves. 1 2 245 For each sample, the uniformity coefficient (CU) and the coefficient of curvature (CC) 3 4 ( ) 5 246 were calculated as (Terzaghi et al., 1996): CU= , CC= · , where D60, D30 and D10 6 7 247 are the diameters so that 60%, 30% and 10% of the material (in weight) pass the 8 9 10 248 corresponding sieve. has also been used as one of the variables in the statistical 11 12 249 analysis. Another variable incorporated in the analysis was the proportion (in weight) of 13 14 250 fine material (<0.074 mm) for each sample. Soil sample depth and sampling campaign 15 16 251 (operation) were also included. 17 18 19 252 However, here Hellinger transformation was not applied to relative abundances of 20 21 22 253 microbial species, because correlations were high already for the raw relative 23 24 254 abundances. Some microbial variables were removed for the final analyses, based on 25 26 255 the low bivariate r2 coefficients obtained (this was the case of Cyanobacteria, 27 28 256 Nitrospira, Gammaproteobacteria, Alphaproteobacteria, Stenotrophomonas sp, 29 30 31 257 Chryseomicrobium sp, and Nitrospira1 sp). 32 33 34 258 Here, the optimization between the number of PCAs, the measurement of sampling 35 36 259 adequacy (KMO) and the number of considered variables was done following the same 37 38 260 criteria indicated in the statistical analysis of the water samples. 39 40 41 261 2.3.3. PCA for microbial print in water and soil samples 42 43 44 262 Taking advantage of the shared variables by water and soil samples, a third statistical 45 46 47 263 analysis was performed, now including all 31 samples (soil and water) together. 48 49 264 Microbial classes and species were treated separately and were Hellinger transformed. 50 51 265 In addition to microbial data, the type of sample (water or soil) was also included as a 52 53 266 binary variable. From the bivariate correlation analysis, some variables were removed 54 55 56 267 from the final statistical analysis: Cyanobacteria, Gammaproteobacteria, 57 58 268 Alphaproteobacteria, Methylotenera mobilis, Vogesella indigofera, Stenotrophomonas 59 60 61 62 11 63 64 65 269 sp, Chryseomicrobium sp, Methylobacterium sp, and Nitrospira2 sp. The optimization 1 2 270 criteria were the same as indicated in the two previous sections. 3 4 5 271 3. Results 6 7 8 272 As a consequence of the preliminary analysis, PCAs of water samples considered 10 9 10 11 273 hydrochemical variables, 2 diversity indices and 10 microbial variables (5 classes and 5 12 13 274 species). PCAs of soil samples took into account 1 operational variable, depth, 4 grain- 14 15 275 size soil parameters, 3 diversity indices and 14 microbial variables (7 classes and 7 16 17 276 species). Finally, the statistical approach for the total of 31 samples (soil and water) 18 19 277 included 3 diversity indices, the type of sample and 13 microbial variables (9 classes 20 21 22 278 and 4 species). 23 24 25 279 We present firstly global results of the eleven PCAs performed (Table 1). We indicate, 26 27 280 in the same table and for each statistical analysis performed, the extracted components 28 29 281 with their variables and the proportion of variance represented for each component. 30 31 282 Also, the measure of sampling adequacy (KMO test value) is reported for each 32 33 34 283 analysis. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 12 63 64 65 Microbial 1 Type of clade PCAs PCA % of KMO Test 2 Extracted components 3 sample data performed number variance value 4 treated 5 C1: V-log, Cu-log, Cl, EC, S-log, Na, OthersH 46.9 6 PCA1 0.744 C2: Shannon, BetaproteobacteriaH, DehalococcoidiaH, Richness 27.2 7 Classes 2 8 C1: CyanobacteriaH, EC, Cl, Cu-log, Ni-log 52.6 PCA2 0.701 9 C2: CytophagiaH, Richness 24.4 10 Water C1: Methylotenera_mobilisH, NO3, SO4, DOC-log 61.5 11 PCA3 0.797 12 C2: Nitrospira_sp2H, EC, Na, V-log, Cu-log, Cl 15.4 Species 2 13 C1: Pontibacter_spH, Dehalogenimonas_spH, Richness 52.9 14 PCA4 0.635 C2: Vogesella_indigoferaH, Shannon, DOC-log 24.3 15 16 C1: BetaproteobacteriaH, ActinobacteriaH, Dryness, Depth, 62.5 17 PCA5 CytophagiaH 0.750 18 Classes 2 C2: CU, % Fine, AcidobacteriaH 22.1 19 20 C1: Richness, DehalococcoidiaH, ChlorobiaH 75.5 PCA6 0.727 21 Soil C2: Shannon, Evenness, ActinobacteriaH, BacilliH 12.3 22 C1: Nitrospira_2sp, Dryness, Shannon, Pontibacter_sp, 64.6 23 PCA7 0.758 24 C2: % Fine, Subgroup3_sp, CU 27.6 Species 2 25 C1: CC, D10, Methylotenera_mobilis, Methylobacterium sp, 59.5 26 PCA8 0.640 C2: % Fine, CU, Dehalogenimonas_sp, Vogesella indigofera 23.9 27 28 C1: Shannon, Richness, ActinobacteriaH, Type, Evenness 53.8 29 PCA9 C2: OthersH, BetaproteobacteiraH 16.4 0.692 30 C3: BacilliH 12.5 31 Classes 2 32 Water C1: Shannon, Evenness, Richness 44.0 33 and soil PCA10 C2: DehalococcoidesH, CytophagiaH, AcidobacteriaH 26.4 0.696 34 C3: NitrospiraH, ChlorobiaH 14.4 35 36 C1: Shannon, Richness, Evenness, Type, Nitrospira_sp1H 48.8 Species 1 PCA11 0.705 37 C2: Pontibacter_spH, Dehalogenimonas_spH, Subgroup3_spH 25.5 38 39 40 41 Table 1 Relation to main correlations among variables obtained by PCA analysis (inverse correlations in bold). Subscript “H” in microbial variables 42 43 indicates that Hellinger transformation had been applied to the data. Suffix “-log” in geochemical variables indicates that log transformation was 44 performed. 45 284 46 47 13 48 49 285 3.1. Key parameters influencing microbial classes in surface water and 1 2 286 groundwater 3 4 5 287 Two PCAs were developed; the first one (PCA1 in Table 1) included Shannon and 6 7 288 Richness indices, Cl, EC, Na, the log transformations of S, Cu, and V, as well as 8 9 10 289 BetaproteobacteriaH, DehalococcoidiaH, NitrospiraH and OthersH classes. The group 11 12 290 “Others” includes microbial classes that could not be classified. Results of this PCA 13 14 291 show that these twelve variables could be mostly explained with only two principal 15 16 292 components (combined they explained 74% of the total variance). The strength of the 17 18 19 293 analysis is supported by a value of 0.744 in the KMO test. Results of a 2D 20 21 294 representation of varifactors and components (Figure 2) show the significance of 22 23 295 hydrochemistry beyond microbial variables. All the chemical variables included in the 24 25 296 analysis are grouped in the first component, showing a high correlation amongst them; 26 27 297 Na, Cl, EC and Cu-log (“-log” indicates log-transformed) are positively correlated, 28 29 30 298 whereas V-log and S-log are situated in the opposite side of the plot (negative 31 32 299 correlation). 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 14 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 300 27 28 29 301 Figure 2 PCA1 analysis including diversity indices, microbial classes and hydrochemical data from 30 31 302 groundwater samples (squares) and one infiltration water sample (diamond) during the recharge period 32 33 303 (blue) and no-recharge period (orange). The position of samples is scaled in VF1 and VF2 axes for 34 35 304 visualization purposes. Red arrows represent the contribution of each variable projected into the varifactor 36 305 plane. Samples are labeled corresponding to the sampling point and the height (meters above sea level, or 37 38 306 masl) indicated in brackets. The symbols assignment follows the same criteria in all PCA plots. 39 40 41 307 The second component relates inversely the presence of BetaproteobacteriaH to the 42 43 308 Shannon and Richness indices. Furthermore, BetaproteobacteriaH is most significant in 44 45 46 309 the samples most affected by recharge (piezometer P8, located directly below the 47 48 310 infiltration pond). Finally, the projection of sample data in the plane composed of the 49 50 311 two varifactors (VF1, VF2) clearly separate data according to the operational period 51 52 312 (recharge/no-recharge). Moreover, the sample not affected by recharge (P1, placed 53 54 55 313 upstream) is displayed between the samples corresponding to the no-recharge and 56 57 314 recharge periods. 58 59 60 61 62 15 63 64 65 315 The second PCA (Figure S1, PCA2 in Table 1) also separates data corresponding to 1 2 316 operational periods in the VF1-VF2 plane. Furthermore, two groups of variables can be 3 4 317 distinguished. The first one relates CyanobacteriaH, Cl, EC, Cu-log and Ni-log. Apart 5 6 7 318 from hydrochemical differences that force the separation of samples between 8 9 319 operational periods, Cyanobacteria are more abundant in the groundwater samples in 10 11 320 the recharge period. Component 2 shows an inverse correlation between Richness and 12 13 321 CytophagiaH. 14 15 16 322 3.2. Key parameters influencing microbial species in water samples 17 18 19 323 PCA3 analysis (Table 1) included variables EC, Na, Cl, V-log, Cu-log, NO3, SO4, DOC- 20 21 22 324 log, and two microbial species (as opposed to microbial classes, that were the 23 24 325 objective of PCA1 and PCA2): Methylotenera mobilisH and Nitrospira sp2H. From the 25 26 326 projection on the varifactor plane (Figure 3), samples corresponding to operational 27 28 327 conditions clearly display on different areas within the plot. Furthermore, samples taken 29 30 31 328 during active operation display a disturbance degree caused by recharge (indicated in 32 33 329 qualitative terms by the blue arrow). On the other hand, the sample representing 34 35 330 background conditions (P1) is located separated from the rest, indicating that in no 36 37 331 recharge periods the areas that were recharged kept some memory of the events, i.e., 38 39 40 332 the microbial communities did not go back to the pre-recharge conditions. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 16 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 333 27 28 29 334 Figure 3 PCA3 analysis including diversity indices, microbial species and hydrochemical data from 30 31 335 groundwater samples (squares) and the infiltration water sample (diamond). The discontinuous blue arrow 32 33 336 indicates the degree of disturbance caused by recharge. 34 35 36 337 Figure 3 also shows the main relations amongst variables. The first component groups 37 38 338 mainly SO4, NO3 and DOC-log (inversely) with Methylotenera mobilisH. The behavior of 39 40 339 Nitrospira sp2H and its high correlation with conductivity were explained by the second 41 42 340 component. 43 44 45 341 PCA4 indicated a relationship between Dehalogenimonas sp.H, Pontibacter sp.H and 46 47 48 342 Vogesella indigoferaH with some hydrochemical and biological indicators (Figure S2). 49 50 343 Dehalogenimonas sp.H appears to be a key contributor to Richness and Diversity 51 52 344 indices and in turn, is inversely correlated with Pontibacter sp.H. However, Vogesella 53 54 345 indigoferaH is correlated with DOC-log (see Figure S2), placed in the vicinity of the P8 55 56 57 346 samples projection, indicating the effect of recharge on both variables. 58 59 60 347 61 62 17 63 64 65 348 3.3. Key parameters influencing microbial classes in MAR soils 1 2 3 349 PCA5 to PCA8 provide statistical analyses for samples taken from soils. Here, the 4 5 350 parameters extracted from the grain-size distribution curves are included, together with 6 7 351 microbial variables, depth and the operational variable. In all cases, the position of 8 9 10 352 samples is scaled in VF1 and VF2 axes. Red arrows represent the contribution of each 11 12 353 variable projected into the plane defined by the two varifactors (Figure 4). 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 354 Figure 4 PCA5 analysis including microbial classes, grain-size distribution curve parameters and depth 40 41 355 from soil samples during the recharge period (blue) and no-recharge period (orange). Circles are related to 42 43 356 the infiltration pond samples (E-Entrance, M-Midfield and F-Final stretch). Numbers indicate the depth the 44 45 357 sample was taken (in m below surface). Triangle corresponds to the sedimentation pond sample. Symbols 46 358 follow the same criteria in the following PCA charts. 47 48 49 359 PCA5 involved microbial soil classes (AcidobacteriaH, BetaproteobacteriaH, 50 51 52 360 ActinobacteriaH, CytophagiaH), Uniformity Coefficient (CU), fine material content (% 53 54 361 Fine), Dryness (operation condition dry-wet) and Depth (Table 1). From Figure 4, 55 56 362 BetaproteobacteriaH displays oppositely to Depth, ActinobacteriaH, CytophagiaH and 57 58 363 Dryness, suggesting that BetaproteobacteriaH is more favorable to live in wet and 59 60 364 shallow soils, while ActinobacteriaH and CytophagiaH are preferably found in dry and 61 62 18 63 64 65 365 deep soils. It should also be noted that when projected over the first two varifactors 1 2 366 plane, wet samples (blue circles) are clustered, whereas dry samples (orange circles) 3 4 367 are more scattered in space, indicating that they are affected by depth gradients. 5 6 7 368 Finally, AcidobacteriaH is positively correlated with CU and proportion of fines, well 8 9 369 represented by the sample taken in the sedimentation pond (blue triangle). 10 11 370 PCA6 clusters microbial classes with diversity indices (Shannon, Richness and 12 13 371 Evenness), ActinobacteriaH, DehalococcoidiaH, ChlorobiaH, BacilliH and OthersH (Figure 14 15 372 S3). Whereas BacilliH behavior is basically explained by the second component, 16 17 18 373 Shannon, Evenness and ActinobacteriaH are partially included in both components. 19 20 374 Same as shown in a previous analysis, ActinobacteriaH is negatively correlated with 21 22 375 Shannon and Evenness indices. ChlorobiaH and DehalococoidiaH are explained by the 23 24 376 first component and contribute visibly to the microbial print of the wet soil samples. 25 26 27 377 28 29 378 3.4. Key parameters influencing microbial species in MAR soils 30 31 32 379 In PCA7 (Figure 5, Table 1), the first two components explain 92% of the total variance. 33 34 380 Projection of samples on varifactor space provides a clear separation of samples 35 36 381 according to the operational period (during recharge and no-recharge periods), while 37 38 39 382 the sample from the sedimentation pond remains separated. This last sample (SD in 40 41 383 Figure 5) is explained mainly by the proportion of fines, CU and notably, the presence 42 43 384 of Subgroup_3 Acidobacteria. Moreover, Shannon Index and Dryness are displayed in 44 45 385 opposite sides of the plot, while the latter is placed in the same side as Pontibacter sp 46 47 48 386 and Nitrospira sp2, highlighting that both species have a high affinity to dry conditions. 49 50 51 52 53 54 55 56 57 58 59 60 61 62 19 63 64 65 387 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 388 25 26 27 389 Figure 5 PCA7 analysis including microbial species, grain-size distribution curve parameters and the 28 29 390 operational period from sediment samples during the recharge period (blue) and no-recharge period 30 31 391 (orange). Codes and symbols are reported in Figure 4. 32 33 34 392 PCA8 indicates the affinity of Vogesella indigofera to silty soils (Figure S4). Despite 35 36 393 Dehalogenimonas sp. behavior is explained by the second component, this species 37 38 394 does not show correlation with V. indigofera or CU. However, the bivariate correlations 39 40 395 matrix (not showed) revealed the correlation of Dehalogenimonas sp. with fine particles 41 42 43 396 proportion. On the other hand, both Methylotenera mobilis and Methylobacterium sp. 44 45 397 are correlated positively with two other soil characteristics, CC and D10. 46 47 48 398 3.5. Interclass and interspecies relationships 49 50 51 399 PCA9 – PCA11 involves the statistical analysis of all 31 samples together. In PCA9, soil 52 53 400 and water samples are distinctly located in the varifactor plane involving the first two 54 55 56 401 components (explaining 70% of the total variance, Figure 6). From the analysis, soil 57 58 402 samples are the most diverse, equally-distributed and rich environments sustaining 59 60 403 microbial growth, as the location of Shannon, Evenness and Richness indices indicate 61 62 20 63 64 65 404 in the PCA. On the other hand, BetaproteobacteriaH shows higher affinity for water 1 2 405 samples. BaciliiH becomes an independent third component explaining 13% of the total 3 4 406 variance (Table 1). 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 407 35 36 37 408 Figure 6 PCA9 analysis including diversity indices, microbial classes and sample type from soil (circles), 38 39 409 groundwater (squares) and surface water (diamonds and triangle) samples during the recharge period 40 410 (blue) and no-recharge period (orange). 41 42 43 411 PCA10 was performed to analyze the intra-relationships among microbial classes. 44 45 412 Three components were needed to explain 85% of the total variance. The first one 46 47 48 413 clusters the three microbial diversity indices. The second one explains the positive 49 50 414 correlation between DehalococcoidiaH and AcidobacteriaH and the negative one with 51 52 415 CytophagiaH. Finally, the third component showed a negative correlation between 53 54 416 NitrospiraH and ChlorobiaH. Samples distribution in VF axes did not show any relevant 55 56 57 417 association among samples (Figure S5). 58 59 60 61 62 21 63 64 65 418 Finally, PCA11 emphasizes microbial species. Similar to PCA10, the first component 1 2 419 groups all three diversity indices, and is representative of soil samples in the varifactor 3 4 420 plane (Figure 7). Nitrospira sp1H was also positively correlated with Shannon and 5 6 7 421 Evenness indices. Regarding the second component, Pontibacter sp.H is inversely 8 9 422 correlated with Dehalogenimonas sp.H and in turn, the latter correlated positively with 10 11 423 Subgroup3 spH, pointing out that these two species seem to be more associated to 12 13 424 water and groundwater samples during active recharge period. 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 425 44 45 46 426 Figure 7 PCA11 analysis including diversity indices, microbial species, and type of sample from soil 47 427 (circles), groundwater (squares) and surface water (diamond and triangle) during the recharge period 48 49 428 (blue) and no-recharge period (orange). 50 51 52 429 53 54 55 430 4. Discussion 56 57 58 59 60 61 62 22 63 64 65 431 The most significant result from the statistical analysis is that the projection of samples 1 2 432 on the planes defined by the principal components shows a clear influence of the 3 4 433 operational conditions, this being the parameter that is subject to optimization in a MAR 5 6 7 434 facility management scheme. On one hand, most of the figures show the polarization of 8 9 435 samples according to the recharge conditions (blue and orange marks); on the other 10 11 436 hand, the variable “Dryness”, indicating soil samples during non-recharge conditions, 12 13 437 remains always beside the samples obtained in dry conditions. These results together 14 15 438 redundantly imply that recharge operation drives significant changes in microbial 16 17 18 439 communities. Recharge results in shifts in the microbial community structure, in terms 19 20 440 of relative abundance and composition of dominant taxonomic groups. 21 22 23 441 Cyanobacteria, the only diazotrophs that produce oxygen as a by-product of the 24 25 442 photosynthetic process, were found most abundantly in the surface water, and the 26 27 443 measured concentrations correlated with several hydrogeochemical variables (EC, Cl, 28 29 30 444 Cu-log and Ni-log) (recall Figure S1). Correlations among EC and cyanobacteria 31 32 445 abundance have been already reported, for example, in a Neotropical urban lake (Frau 33 34 446 et al., 2018) or in a eutrophicated reservoir in Brazil (Chellappa and Mederios Costa, 35 36 447 2003). Cu-log is positively correlated with cyanobacteria; this follows Dwivedi et al. 37 38 39 448 (2006), who described the simultaneous presence of cyanobacteria and some metals 40 41 449 (including Cu) in river water samples enriched by fly-ash originated in a coal thermal 42 43 450 power station. Indeed, some cyanobacterial species have the ability to immobilize 44 45 451 metals, and this is used in some water treatment industries (de-Bashan and Bashan, 46 47 48 452 2010). 49 50 51 453 The highest relative abundance of Betaproteobacteria was found in the groundwater 52 53 454 samples most affected by recharge (e.g., conforming more than 50% of the total 54 55 455 microbial abundance in piezometer P8, located directly below the infiltration pond). This 56 57 456 microbial group correlates inversely with Richness and Shannon indices (recall Figure 58 59 60 457 2) suggesting that members of the Betaproteobacteria class could have an 61 62 23 63 64 65 458 opportunistic behavior, displacing other microbial populations when recharge is active, 1 2 459 and becoming the leading microorganisms degrading organic matter. 3 4 5 460 We also evaluated the role of DOC, nitrate and sulfate on the abundance of 6 7 461 Methylotenera mobilis in surface and subsurface water environments (see PCA3). M. 8 9 10 462 mobilis may carry out heterotrophic denitrification by using different substrates, such as 11 12 463 methanol (Kalyuhznaya et al., 2009; Sun et al., 2016) or methylamine (Kalyuzhnaya et 13 14 464 al., 2006) in aerobic environments. In fact, M. mobilis is the first methylotroph using 15 16 465 nitrate as an electron acceptor, thus being a potential explanation of the positive 17 18 19 466 correlation observed between M. mobilis abundance and nitrate concentration in water. 20 21 467 Denitrification is an anaerobic process, but in some cases occurs in the presence of 22 23 468 oxygen. In these situations, anoxic microenvironments (called microsites) may develop 24 25 469 due to the heterogeneity in flow paths that would lead to local denitrification (Knowles, 26 27 470 2005; Modin et al., 2007) even in a zone with overall aerobic conditions. One surprising 28 29 30 471 finding from PCA3 is that correlation of M. mobilis and nitrate seems associated with 31 32 472 the non-recharge period. Therefore, denitrification conditions are not completely 33 34 473 reversed even after recharge has been discontinued for a significant amount of time. 35 36 37 474 On the other hand, it is not surprising that abundance of M. mobilis correlates with 38 39 40 475 Methylobacterium sp. presence in the soil environment (see PCA8); since they belong 41 42 476 to different proteobacterial classes, they are both obligate methylotrophs (Kumaresan 43 44 477 et al., 2018). Indeed, correlations between both species of methylotrophs have been 45 46 478 reported by Wright et al. (2017) in a groundwater dichloromethane contaminated site in 47 48 479 the US. Unlike M. mobilis, Methylobacterium sp. is a dichloromethane degrader (Gisi et 49 50 51 480 al., 1998; Vuilleumier et al., 2009), capable to dechlorinate in aerobic conditions. 52 53 481 Following Wright et al. (2017), this suggests that M. mobilis could take advantage by 54 55 482 using the byproducts of dichloromethane degradation carried out by Methylobacterium 56 57 483 sp. However, methane has not been measured at the site, so we are not assuming its 58 59 60 484 presence via dichloromethane degradation. 61 62 24 63 64 65 485 In any case, it is not clear whether M. mobilis abundance correlates with some redox 1 2 486 species; while the bivariate correlation coefficients with DOC-log and SO4 were not 3 4 487 significant (data not shown), M. mobilis grouped in the same component with NO3 and 5 6 7 488 SO4. Regarding this last point, there is no evidence that SO4 might be part of the 8 9 489 metabolic pathway of M. mobilis. 10 11 12 490 Vogesella indigofera positively correlates with DOC-log (PCA5). Members of this genus 13 14 491 are important denitrifiers in groundwater systems (Bellini et al., 2018), capable of 15 16 492 oxidizing few monosaccharides in low-oxygen concentration conditions (Grimes et al., 17 18 19 493 1997). Contrarily, DOC-log was negatively correlated to NO3 concentration (PCA3). In 20 21 494 fact, the isotopic study performed by Grau-Martínez et al. (2018) at the same facility 22 23 495 and on the same dates confirmed that recharge induced denitrification. 24 25 26 496 Dryness (a binary dry/wet variable) is the most significant one controlling the 27 28 497 distribution of microbial communities in soils. Similar to what was observed in water 29 30 31 498 samples, soil samples taken during recharge periods displayed largest diversity and 32 33 499 highest Betaproteobacteria abundance. Recharge of river water ensured the 34 35 500 continuous supply of nutrients and DOC to the system, favoring Betaproteobacteria 36 37 501 growth, as it is one of the main bacterial groups biodegrading DOC (Li et al., 2013). In 38 39 40 502 fact, in a work carried out in column models (Li et al., 2012), Betaproteobacteria were 41 42 503 found mainly close to the inlet, associated to the high DOC concentrations fed; they 43 44 504 also found that water content in soils (and extrapolating, recharge conditions) 45 46 505 contribute to microbial diversity in soils. Along the same line, decreasing proportions of 47 48 506 Betaproteobacteria were found in dry soils (as compared to wet ones) in an experiment 49 50 51 507 performed with streambed fresh sediments (Pohlon et al., 2013). 52 53 54 508 Distribution of microbial populations in soils is also strongly affected by granulometry 55 56 509 (PCA8). M. mobilis and Methylobacterium sp. correlated with CC and . There is not 57 58 510 much literature on this subject. The work performed by Madhaiyan et al. (2007) 59 60 61 62 25 63 64 65 511 associated the variability in the distribution of pink-pigmented facultative metylotrophs, 1 2 512 such as Methylobacterium sp, with soil type and moisture among other environmental 3 4 513 variables. In fact, soil texture conditions is of paramount importance for nutrient 5 6 7 514 transference into flooded soils (Perujo et al., 2017), also influencing methane 8 9 515 consumption rates (Jäckel et al., 2001). In the same way, the presence of Vogesella 10 11 516 indigofera and Subgroup3 sp (Acidobacteria) in soils is linked to CU (an indicator of 12 13 517 grain-size heterogeneity) and proportion of fine particles, resulting in these two 14 15 518 populations being most abundant in sandy-clayey soils. 16 17 18 19 519 Members of the phylum Acidobacteria, such as Subgroup 3 sp are widespread 20 21 520 abundant in soils and sediments, and capable to tolerate moisture fluctuations (Ward et 22 23 521 al., 2009). This tolerance can be related to the capability to form biofilms, highly 24 25 522 hydrated structures (Kielak et al., 2016), allowing them to survive under stressful 26 27 523 dryness conditions (Ward et al., 2009); this could be the reason why they are found 28 29 30 524 regardless the recharge operational conditions. Moreover, it is also important to know 31 32 525 the interactions of this group with other microbial classes present in the soil. In this 33 34 526 regard, we found a significant statistical positive correlation between Subgroup3 sp 35 36 527 (Acidobacteria) and Dehalogenimonas sp (Dehalococcoidia class); we associate this 37 38 39 528 correlation to both groups being environmentally important, as they are involved in 40 41 529 different contaminant degradation processes (Chen et al., 2018; Song et al., 2016). 42 43 44 530 In the soil samples, the relative abundance of Actinobacteria was directly correlated 45 46 531 with dry conditions (PCA5), similar to the observations of Pohlon et al. (2013). Different 47 48 532 members of this phylum have the ability to grow developing mycelia structures and 49 50 51 533 forming spores (Bhatti et al., 2017), providing them an advantage in the colonization of 52 53 534 soil with limited water availability. Moreover, the combined analysis of soil and water 54 55 535 samples showed the widespread presence of this phylum, especially abundant in soil 56 57 536 samples, and could be associated to their important role in the cycling of organic matter 58 59 60 537 due to their decomposition capabilities (Bhatti et al., 2017; Polkade et al., 2016). This 61 62 26 63 64 65 538 behavior is consistent with results observed in our system (PCA9), where 1 2 539 Actinobacteria is linked to microbial diversity indexes and dryness conditions. These 3 4 540 last results are not surprising, since soils probably are one of the most diverse and rich 5 6 7 541 environments supporting microbial life in the world (Curtis et al., 2002; Or et al., 2007; 8 9 542 Torsvik et al., 2002; Young and Crawford, 2004). 10 11 12 543 Characterization of microbial communities may have strong implications in MAR 13 14 544 facilities management when better understood. A particular combination of soil and 15 16 545 inflow water characteristics at the infiltration pond would result in a particular microbial 17 18 19 546 population signature. Such population would degrade preferably some particular 20 21 547 organic compounds. The spatial distribution of microorganisms can be used for 22 23 548 mapping the area influenced (even partially) by recharge. From another point of view, 24 25 549 such detailed mapping can be eventually used in the future for optimal design of biofilm 26 27 550 growth and composition in order to promote degradation of different target compounds. 28 29 30 551 Furthermore, indicator species could be included in monitoring networks in order to 31 32 552 characterize environmental conditions that take place at longer temporal scales than 33 34 553 the one represented by punctual hydrochemical data. 35 36 37 554 5. Conclusions 38 39 40 555 The work presented in this paper led us to draw conclusions regarding both the 41 42 43 556 statistical analysis proposed and the intrinsic results obtained. 44 45 46 557 Multivariate statistical methods allow the simultaneous treatment of interdisciplinary 47 48 558 data (physical, chemical, and biological), from different nature (categorical, binary and 49 50 559 continuous) and spanning different ranges, in a formal statistical way. This is potentially 51 52 560 very useful for many experimental studies in soil science or freshwater systems as a 53 54 55 561 way to integrate data in a subset of statistical components that despite being small in 56 57 562 dimensional terms, explains a large amount of the variance of the dataset. 58 59 60 61 62 27 63 64 65 563 The application of this methodology in the dataset obtained in the Sant Vicenç 1 2 564 recharge facility enabled to conclude that MAR influences significantly microbial 3 4 565 communities in soil, surface and subsurface water. Therefore, they can be used to 5 6 7 566 assess the area of the soil and the volume of the aquifer influenced by recharge. 8 9 567 Microbial communities are directly influenced by a combination of biohydrogeochemical 10 11 568 processes and vice versa. Aquifer recharge via infiltration ponds has a significant 12 13 569 impact upon carbon and nitrogen cycles in the topsoil, since most of the microbial 14 15 570 species detected in this study have a role in the aerobic degradation of organic matter 16 17 18 571 and in denitrification processes. Thus, a strong correlation was found between 19 20 572 hydrochemical compounds and some microbial communities, both in the soil and 21 22 573 groundwater samples; two examples are the correlation between Vogesella indigofera 23 24 574 with DOC and Methylotenera mobilis with NO3. 25 26 27 575 The statistical analysis of both groundwater and soil samples clearly discriminate 28 29 30 576 operational conditions. Recharge drives distinct populations to become dominant 31 32 577 (Betaproteobacteria, Dehalococcoidia, and Nitrospira). Moreover, groundwater 33 34 578 samples can be graphically represented according to the degree of disturbance (in 35 36 579 terms of the proportion of infiltrating water versus the native one in the corresponding 37 38 39 580 water samples). 40 41 42 581 In the case of soil samples, there are two most set of significant variables affecting the 43 44 582 distribution of microbial communities: (1) grain-size distribution (affecting Vogesella 45 46 583 indigofera, Subgroup3 sp (Acidobacteria), Methylotenera mobilis, Dehalogenimonas sp 47 48 584 and Methylobacterium sp),; and (2) operation conditions, as wet soils contain more 49 50 51 585 microbial diversity than dry ones. However, there are some classes (e.g., 52 53 586 Acidobacteria) that are less affected by recharge conditions. In the future, this could 54 55 587 provide ideas for the affection of the climate change consequences in stream-flows, 56 57 588 wetlands, and other drought-vulnerable aquatic systems. 58 59 60 61 62 28 63 64 65 589 Multivariate statistical analysis indicates that DOC was another determinant factor 1 2 590 shaping microbial community structure. Among them, members of Betaproteobacteria, 3 4 591 Dehalococcoidia and Acidobacteria are involved in pollutant bioremediation and can be 5 6 7 592 considered as potential bioindicators for recharge monitoring. 8 9 10 593 11 12 13 594 Acknowledgments 14 15 16 595 This investigation was financially supported by the European Union project MARSOL 17 18 596 FP7-ENV-2013-WATER-INNO-DEMO, Generalitat de Catalunya via FI scholarship 19 20 597 program (FI-DGR 2014) and the Spanish Ministry of Economy and Competitiveness 21 22 23 598 and EU (project ACWAPUR PCIN-2015-239). The authors would like to acknowledge 24 25 599 Comunitat d’Usuaris d’Aigües de la Vall Baixa i del Delta del Riu Llobregat (CUADLL) 26 27 600 for their cooperation. 28 29 30 601 31 32 33 602 References 34 35 36 603 Alidina, M., Li, D., Drewes, J.E., 2014. Investigating the role for adaptation of the 37 38 39 604 microbial community to transform trace organic chemicals during managed aquifer 40 41 605 recharge. Water Res 56, 172–180. doi:10.1016/j.watres.2014.02.046 42 43 44 606 Alidina, M., Shewchuk, J., Drewes, J.E., 2015. Effect of temperature on removal of 45 46 607 trace organic chemicals in managed aquifer recharge systems. Chemosphere 47 48 608 122, 23–31. doi:10.1016/j.chemosphere.2014.10.064 49 50 51 52 609 Barba, C., Folch, A., Gaju, N., Sanchez-Vila, X., Carrasquilla, M., Grau-Martínez, A., 53 54 610 Martinez-Alonso, M., 2019. Microbial community changes induced by Managed 55 56 611 Aquifer Recharge activities: Linking hydrogeological and biological processes. 57 58 612 Hydrol Earth Syst Sci., 23, 139-154, doi: 10.5194/hess-23-139-2019 59 60 61 62 29 63 64 65 613 Bekele, E., Toze, S., Patterson, B., Higginson, S., 2011. Managed aquifer recharge of 1 2 614 treated wastewater: Water quality changes resulting from infiltration through the 3 4 615 vadose zone. Water Res 45, 5764–5772. doi:10.1016/j.watres.2011.08.058 5 6 7 616 Bellini, M.I., Gutiérrez, L., Tarlera, S., Scavino, A.F., 2013. Isolation and functional 8 9 10 617 analysis of denitrifiers in an aquifer with high potential for denitrification. Syst Appl 11 12 618 Microbiol 36, 505–16. doi:10.1016/j.syapm.2013.07.001 13 14 15 619 Bellini, M.I., Kumaresan, D., Tarlera, S., Murrell, J.C., Fernández-Scavino, A., 2018. 16 17 620 Identification of active denitrifiers by DNA-stable isotope probing and amplicon 18 19 621 sequencing reveals Betaproteobacteria as responsible for attenuation of nitrate 20 21 22 622 contamination in a low impacted aquifer. FEMS Microbiol Ecol 94, fix181-fix181. 23 24 25 623 Bhatti, A.A., Haq, S., Bhat, R.A., 2017. Actinomycetes benefaction role in soil and plant 26 27 624 health. Microb Pathog. doi:10.1016/j.micpath.2017.09.036 28 29 30 625 Bouwer, H., 2002. Artificial recharge of groundwater: hydrogeology and engineering. 31 32 626 Hydrogeol J 10, 121–142. doi:10.1007/s10040-001-0182-4 33 34 35 627 Chau, J., Bagtzoglou, A., Willig, M., 2011. The effect of soil texture on richness and 36 37 38 628 diversity of bacterial communities. Environ Forensics 12, 333–341. 39 40 629 doi:10.1080/15275922.2011.622348 41 42 43 630 Chellappa, N.T., Mederios Costa, M.A., 2003. Dominant and co-existing species of 44 45 631 Cyanobacteria from a Eutrophicated reservoir of Rio Grande do Norte State, 46 47 48 632 Brazil, in: Acta Oecologica. Elsevier Masson, pp. S3–S10. doi:10.1016/S1146- 49 50 633 609X(03)00005-5 51 52 53 634 Chen, J., Wang, P.-F., Wang, C., Liu, J.-J., Gao, H., Wang, X., 2018. Spatial 54 55 635 distribution and diversity of organohalide-respiring bacteria and their relationships 56 57 636 with polybrominated diphenyl ether concentration in Taihu Lake sediments. 58 59 60 637 Environ Pollut 232, 200–211. doi:10.1016/J.ENVPOL.2017.08.124 61 62 30 63 64 65 638 Curtis, T.P., Sloan, W.T., Scannell, J.W., 2002. Estimating prokaryotic diversity and its 1 2 639 limits. Proc Natl Acad Sci 99, 10494–10499. doi:10.1073/pnas.142680199 3 4 5 640 D’Alessio, M., Yoneyama, B., Kirs, M., Kisand, V., Ray, C., 2015. Pharmaceutically 6 7 641 active compounds: Their removal during slow sand filtration and their impact on 8 9 10 642 slow sand filtration bacterial removal. Sci Total Environ 524–525, 124–135. 11 12 643 doi:10.1016/j.scitotenv.2015.04.014 13 14 15 644 de-Bashan, L.E., Bashan, Y., 2010. Immobilized microalgae for removing pollutants: 16 17 645 Review of practical aspects. Bioresour Technol 101, 1611–1627. 18 19 646 doi:10.1016/j.biortech.2009.09.043 20 21 22 23 647 Drewes, J.E., Li, D., Regnery, J., Alidina, M., Wing, A., Hoppe-Jones, C., 2014. Tuning 24 25 648 the performance of a natural treatment process using metagenomics for improved 26 27 649 trace organic chemical attenuation. Water Sci Technol 69, 628–633. 28 29 650 doi:10.2166/wst.2013.750 30 31 32 651 Drewes, J.E., Reinhard, M., Fox, P., 2003. Comparing microfiltration-reverse osmosis 33 34 35 652 and soil-aquifer treatment for indirect potable reuse of water. Water Res 37, 3612– 36 37 653 3621. doi:10.1016/S0043-1354(03)00230-6 38 39 40 654 Dutta, T., Carles-Brangarí, A., Fernàndez-Garcia, D., Rubol, S., Tirado-Conde, J., 41 42 655 Sanchez-Vila, X., 2015. Vadose zone oxygen (O2) dynamics during drying and 43 44 656 wetting cycles: An artificial recharge laboratory experiment. J Hydrol 527, 151– 45 46 47 657 159. doi:10.1016/j.jhydrol.2015.04.048 48 49 50 658 Dwivedi, S., Tripathi, R.D., Rai, U.N., Srivastava, S., Mishra, S., Shukla, M.K., Gupta, 51 52 659 A.K., Sinha, S., Baghel, V.S., Gupta, D.K., 2006. Dominance of algae in ganga 53 54 660 water polluted through fly-ash leaching: Metal bioaccumulation potential of 55 56 661 selected algal species. Bull Environ Contam Toxicol 77, 427–436. 57 58 59 662 doi:10.1007/s00128-006-1083-y 60 61 62 31 63 64 65 663 El Alfy, M., Lashin, A., Abdalla, F., Al-Bassam, A., 2017. Assessing the 1 2 664 hydrogeochemical processes affecting groundwater pollution in arid areas using 3 4 665 an integration of geochemical equilibrium and multivariate statistical techniques. 5 6 7 666 Environ Pollut 229, 760–770. doi:10.1016/j.envpol.2017.05.052 8 9 10 667 Fahrenfeld, N., Cozzarelli, I.M., Bailey, Z., Pruden, A., 2014. Insights into 11 12 668 Biodegradation Through Depth-Resolved Microbial Community Functional and 13 14 669 Structural Profiling of a Crude-Oil Contaminant Plume. Microb Ecol 68, 453–462. 15 16 670 doi:10.1007/s00248-014-0421-6 17 18 19 671 Faulwetter, J.L., Gagnon, V., Sundberg, C., Chazarenc, F., Burr, M.D., Brisson, J., 20 21 22 672 Camper, A.K., Stein, O.R., 2009. Microbial processes influencing performance of 23 24 673 treatment wetlands: A review. Ecol Eng 35, 987–1004. 25 26 674 doi:10.1016/j.ecoleng.2008.12.030 27 28 29 675 Frau, D., Pinto, P.D.T., Mayora, G., 2018. Are cyanobacteria total, specific and trait 30 31 676 abundance regulated by the same environmental variables? Ann Limnol 54, 3. 32 33 34 677 doi:10.1051/limn/2017030 35 36 37 678 Freixa, A., Rubol, S., Carles-Brangarí, A., Fernàndez-Garcia, D., Butturini, A., 38 39 679 Sanchez-Vila, X., Romaní, A.M., 2015. The effects of sediment depth and oxygen 40 41 680 concentration on the use of organic matter: An experimental study using an 42 43 44 681 infiltration sediment tank. Sci Total Environ 540, 20–31. 45 46 682 doi:10.1016/j.scitotenv.2015.04.007 47 48 49 683 Ginige, M.P., Kaksonen, A.H., Morris, C., Shackelton, M., Patterson, B.M., 2013. 50 51 684 Bacterial community and groundwater quality changes in an anaerobic aquifer 52 53 685 during groundwater recharge with aerobic recycled water. FEMS Microbiol Ecol 54 55 56 686 85, 553–567. doi:10.1111/1574-6941.12137 57 58 59 687 Gisi, D., Willi, L., Traber, H., Leisinger, T., Vuilleumier, S., 1998. Effects of bacterial 60 61 62 32 63 64 65 688 host and dichloromethane dehalogenase on the competitiveness of methylotrophic 1 2 689 bacteria growing with dichloromethane. Appl Environ Microbiol 64, 1194–1202. 3 4 5 690 Goren, O., Burg, A., Gavrieli, I., Negev, I., Guttman, J., Kraitzer, T., Kloppmann, W., 6 7 691 Lazar, B., 2014. Biogeochemical processes in infiltration basins and their impact 8 9 10 692 on the recharging effluent, the soil aquifer treatment (SAT) system of the Shafdan 11 12 693 plant, Israel. Appl Geochemistry 48, 58–69. 13 14 694 doi:10.1016/j.apgeochem.2014.06.017 15 16 17 695 Grau-Martínez, A., Folch, A., Torrentó, C., Valhondo, C., Barba, C., Domènech, C., 18 19 696 Soler, A., Otero, N., 2018. Monitoring induced denitrification during managed 20 21 22 697 aquifer recharge in an infiltration pond. J Hydrol 561, 123–135. 23 24 698 doi:10.1016/j.jhydrol.2018.03.044 25 26 27 699 Greskowiak, J., Hamann, E., Burke, V., Massmann, G., 2017. The uncertainty of 28 29 700 biodegradation rate constants of emerging organic compounds in soil and 30 31 701 groundwater – A compilation of literature values for 82 substances. Water Res 32 33 34 702 126, 122–133. doi:10.1016/j.watres.2017.09.017 35 36 37 703 Greskowiak, J., Prommer, H., Massmann, G., Johnston, C.D., Nützmann, G., 38 39 704 Pekdeger, A., 2005. The impact of variably saturated conditions on 40 41 705 hydrogeochemical changes during artificial recharge of groundwater. Appl 42 43 44 706 Geochemistry 20, 1409–1426. doi:10.1016/j.apgeochem.2005.03.002 45 46 47 707 Greskowiak, J., Prommer, H., Massmann, G., Nützmann, G., 2006. Modeling Seasonal 48 49 708 Redox Dynamics and the Corresponding Fate of the Pharmaceutical Residue 50 51 709 Phenazone During Artificial Recharge of Groundwater. Environ Sci Technol 40, 52 53 710 6615–6621. doi:10.1021/es052506t 54 55 56 711 Griebler, C., Lueders, T., 2009. Microbial biodiversity in groundwater ecosystems. 57 58 59 712 Freshw Biol 54, 649–677. doi:10.1111/j.1365-2427.2008.02013.x 60 61 62 33 63 64 65 713 Grimes, D.J., Woese, C.R., MacDonell, M.T., Colwell, R.R., 1997. Systematic study of 1 2 714 the genus Vogesella gen. nov. and its type species, Vogesella indigofera comb. 3 4 715 nov. Int J Syst Bacteriol 47, 19–27. doi:10.1099/00207713-47-1-19 5 6 7 716 Haack, S.K., Fogarty, L.R., West, T.G., Alm, E.W., McGuire, J.T., Long, D.T., 8 9 10 717 Hyndman, D.W., Forney, L.J., 2004. Spatial and temporal changes in microbial 11 12 718 community structure associated with recharge-influenced chemical gradients in a 13 14 719 contaminated aquifer. Env Microbiol 6, 438–448. doi:10.1111/j.1462- 15 16 720 2920.2003.00563.x 17 18 19 721 Hamann, E., Stuyfzand, P.J., Greskowiak, J., Timmer, H., Massmann, G., 2016. The 20 21 22 722 fate of organic micropollutants during long-term/long-distance river bank filtration. 23 24 723 Sci Total Environ 545–546, 629–640. doi:10.1016/j.scitotenv.2015.12.057 25 26 27 724 Hellauer, K., Karakurt, S., Sperlich, A., Burke, V., Massmann, G., Hübner, U., Drewes, 28 29 725 J.E., 2017. Establishing sequential managed aquifer recharge technology 30 31 726 (SMART) for enhanced removal of trace organic chemicals: Experiences from field 32 33 34 727 studies in Berlin, Germany. J Hydrol. doi:10.1016/j.jhydrol.2017.09.044 35 36 37 728 Jäckel, U., Schnell, S., Conrad, R., 2001. Effect of moisture, texture and aggregate size 38 39 729 of paddy soil on production and consumption of CH4. Soil Biol Biochem 33, 965– 40 41 730 971. doi:10.1016/S0038-0717(00)00248-0 42 43 44 731 Kalyuhznaya, M.G., Martens-Habbena, W., Wang, T., Hackett, M., Stolyar, S.M., Stahl, 45 46 47 732 D.A., Lidstrom, M.E., Chistoserdova, L., 2009. Methylophilaceae link methanol 48 49 733 oxidation to denitrification in freshwater lake sediment as suggested by stable 50 51 734 isotope probing and pure culture analysis. Environ Microbiol Rep 1, 385–392. 52 53 735 doi:10.1111/j.1758-2229.2009.00046.x 54 55 56 736 Kalyuzhnaya, M.G., Bowerman, S., Lara, J.C., Lidstrom, M.E., Chistoserdova, L., 2006. 57 58 59 737 Methylotenera mobilis gen. nov., sp. nov., an obligately methelamine-utilizing 60 61 62 34 63 64 65 738 bacterium within the family Methylophilaceae. Int J Syst Evol Microbiol 56, 2819– 1 2 739 2823. doi:10.1099/ijs.0.64191-0 3 4 5 740 Kielak, A.M., Barreto, C.C., Kowalchuk, G.A., van Veen, J.A., Kuramae, E.E., 2016. 6 7 741 The ecology of Acidobacteria: Moving beyond genes and genomes. Front 8 9 10 742 Microbiol. doi:10.3389/fmicb.2016.00744 11 12 13 743 Knowles, R., 2005. Denitrifiers associated with methanotrophs and their potential 14 15 744 impact on the nitrogen cycle. Ecol Eng 24, 441–446. 16 17 745 doi:10.1016/J.ECOLENG.2005.01.001 18 19 20 746 Kolehmainen, R.E., Tiirola, M., Puhakka, J.A., 2008. Spatial and temporal changes in 21 22 23 747 Actinobacterial dominance in experimental artificial groundwater recharge. Water 24 25 748 Res 42, 4525–4537. doi:10.1016/j.watres.2008.07.039 26 27 28 749 Kumaresan, D., Stephenson, J., Doxey, A.C., Bandukwala, H., Brooks, E., Hillebrand- 29 30 750 Voiculescu, A., Whiteley, A.S., Murrell, J.C., 2018. Aerobic proteobacterial 31 32 751 methylotrophs in Movile Cave: genomic and metagenomic analyses. Microbiome 33 34 35 752 6, 1. doi:10.1186/s40168-017-0383-2 36 37 38 753 Lapworth, D.J., Baran, N., Stuart, M.E., Ward, R.S., 2012. Emerging organic 39 40 754 contaminants in groundwater: A review of sources, fate and occurrence. Environ 41 42 755 Pollut. doi:10.1016/j.envpol.2011.12.034 43 44 45 756 Laws, B. V, Dickenson, E.R. V, Johnson, T.A., Snyder, S.A., Drewes, J.E., 2011. 46 47 48 757 Attenuation of contaminants of emerging concern during surface-spreading aquifer 49 50 758 recharge. Sci Total Environ 409, 1087–1094. doi:10.1016/j.scitotenv.2010.11.021 51 52 53 759 Legendre, P., Gallagher, E.D., 2001. Ecologically meaningful transformations for 54 55 760 ordination of species data. Oecologia 129, 271–280. doi:10.1007/s004420100716 56 57 58 761 Li, D., Alidina, M., Ouf, M., Sharp, J.O., Saikaly, P., Drewes, J.E., 2013. Microbial 59 60 762 community evolution during simulated managed aquifer recharge in response to 61 62 35 63 64 65 763 different biodegradable dissolved organic carbon (BDOC) concentrations. Water 1 2 764 Res 47, 2421–2430. doi:10.1016/j.watres.2013.02.012 3 4 5 765 Li, D., Sharp, J.O., Saikaly, P.E., Ali, S., Alidina, M., Alarawi, M.S., Keller, S., Hoppe- 6 7 766 Jones, C., Drewes, J.E., 2012. Dissolved organic carbon influences microbial 8 9 10 767 community composition and diversity in managed aquifer recharge systems. Appl 11 12 768 Env Microbiol 78, 6819–6828. doi:10.1128/aem.01223-12 13 14 15 769 Madhaiyan, M., Poonguzhali, S., Sa, T., 2007. Influence of plant species and 16 17 770 environmental conditions on epiphytic and endophytic pink-pigmented facultative 18 19 771 methylotrophic bacterial populations associated with field-grown rice cultivars. J 20 21 22 772 Microbiol Biotechnol 17, 1645–1654. 23 24 25 773 Maeng, S.K., Sharma, S.K., Lekkerkerker-Teunissen, K., Amy, G.L., 2011. Occurrence 26 27 774 and fate of bulk organic matter and pharmaceutically active compounds in 28 29 775 managed aquifer recharge: A review. Water Res 45, 3015–3033. doi: 30 31 32 776 10.1016/j.watres.2011.02.017 33 34 35 777 Martínez-Pascual, E., Jiménez, N., Vidal-Gavilan, G., Vinas, M., Solanas, A.M., 2010. 36 37 778 Chemical and microbial community analysis during aerobic biostimulation assays 38 39 779 of non-sulfonated alkyl-benzene-contaminated groundwater. Appl Microbiol 40 41 780 Biotechnol 88, 985–995. doi:10.1007/s00253-010-2816-8 42 43 44 781 Massmann, G., Greskowiak, J., Dünnbier, U., Zuehlke, S., Knappe, A., Pekdeger, A., 45 46 47 782 2006. The impact of variable temperatures on the redox conditions and the 48 49 783 behaviour of pharmaceutical residues during artificial recharge. J Hydrol 328, 50 51 784 141–156. doi: 10.1016/j.jhydrol.2005.12.009 52 53 54 785 Meckenstock, R.U., Elsner, M., Griebler, C., Lueders, T., Stumpp, C., Aamand, J., 55 56 786 Agathos, S.N., Albrechtsen, H.J., Bastiaens, L., Bjerg, P.L., Boon, N., Dejonghe, 57 58 59 787 W., Huang, W.E., Schmidt, S.I., Smolders, E., Sørensen, S.R., Springael, D., Van 60 61 62 36 63 64 65 788 Breukelen, B.M., 2015. Biodegradation: Updating the Concepts of Control for 1 2 789 Microbial Cleanup in Contaminated Aquifers. Environ Sci Technol 49, 7073–7081. 3 4 790 doi:10.1021/acs.est.5b00715 5 6 7 791 Menció, A., Folch, A., Mas-Pla, J., 2012. Identifying key parameters to differentiate 8 9 10 792 groundwater flow systems using multifactorial analysis. J Hydrol 472–473, 301– 11 12 793 313. doi:10.1016/j.jhydrol.2012.09.030 13 14 15 794 Mermillod-Blondin, F., Simon, L., Maazouzi, C., Foulquier, A., Delolme, C., Marmonier, 16 17 795 P., 2015. Dynamics of dissolved organic carbon (DOC) through stormwater basins 18 19 796 designed for groundwater recharge in urban area: Assessment of retention 20 21 22 797 efficiency. Water Res 81, 27–37. doi:10.1016/j.watres.2015.05.031 23 24 25 798 Modin, O., Fukushi, K., Yamamoto, K., 2007. Denitrification with methane as external 26 27 799 carbon source. Water Res 41, 2726–2738. doi:10.1016/j.watres.2007.02.053 28 29 30 800 Nijenhuis, I., Kuntze, K., 2016. Anaerobic microbial dehalogenation of organohalides— 31 32 801 state of the art and remediation strategies. Curr Opin Biotechnol 38, 33–38. 33 34 35 802 doi:10.1016/j.copbio.2015.11.009 36 37 38 803 Or, D., Smets, B.F., Wraith, J.M., Dechesne, A., Friedman, S.P., 2007. Physical 39 40 804 constraints affecting bacterial habitats and activity in unsaturated porous media – 41 42 805 a review. Adv Water Resour 30, 1505–1527. doi: 43 44 806 10.1016/j.advwatres.2006.05.025 45 46 47 48 807 Paliy, O., Shankar, V., 2016. Application of multivariate statistical techniques in 49 50 808 microbial ecology. Mol Ecol. doi:10.1111/mec.13536 51 52 53 809 Pedretti, D., Barahona-Palomo, M., Bolster, D., Fernàndez-Garcia, D., Sanchez-Vila, 54 55 810 X., Tartakovsky, D.M., 2012a. Probabilistic analysis of maintenance and operation 56 57 811 of artificial recharge ponds. Adv Water Resour 36, 23–35. doi: 58 59 60 812 10.1016/j.advwatres.2011.07.008 61 62 37 63 64 65 813 Pedretti, D., Barahona-Palomo, M., Bolster, D., Sanchez-Vila, X., Fernàndez-Garcia, 1 2 814 D., 2012b. A quick and inexpensive method to quantify spatially variable infiltration 3 4 815 capacity for artificial recharge ponds using photographic images. J Hydrol 430– 5 6 7 816 431, 118–126. doi: 10.1016/j.jhydrol.2012.02.008 8 9 10 817 Perujo, N., Romaní, A.M., Sanchez-Vila, X., 2018. Bilayer Infiltration System Combines 11 12 818 Benefits from Both Coarse and Fine Sands Promoting Nutrient Accumulation in 13 14 819 Sediments and Increasing Removal Rates. Environ Sci Technol 52, 5734–5743. 15 16 820 doi:10.1021/acs.est.8b00771 17 18 19 821 Perujo, N., Sanchez-Vila, X., Proia, L., Romaní, A.M., 2017. Interaction between 20 21 22 822 Physical Heterogeneity and Microbial Processes in Subsurface Sediments: A 23 24 823 Laboratory-Scale Column Experiment. Environ Sci Technol 51, 6110–6119. 25 26 824 doi:10.1021/acs.est.6b06506 27 28 29 825 Pett-Ridge, J., Firestone, M.K., 2005. Redox fluctuation structures microbial 30 31 826 communities in a wet tropical soil. Appl Environ Microbiol 71, 6998–7007. 32 33 34 827 doi:10.1128/AEM.71.11.6998-7007.2005 35 36 37 828 Pohlon, E., Fandino, A.O., Marxsen, J., 2013. Bacterial community composition and 38 39 829 extracellular enzyme activity in temperate streambed sediment during drying and 40 41 830 rewetting. PLoS One 8, e83365. doi:10.1371/journal.pone.0083365 42 43 44 831 Polkade, A. V, Mantri, S.S., Patwekar, U.J., Jangid, K., 2016. Quorum sensing: An 45 46 47 832 under-explored phenomenon in the phylum Actinobacteria. Front Microbiol. 48 49 833 doi:10.3389/fmicb.2016.00131 50 51 52 834 Reed, D.A., Toze, S., Chang, B., 2008. Spatial and temporal changes in sulphate- 53 54 835 reducing groundwater bacterial community structure in response to Managed 55 56 836 Aquifer Recharge. Water Sci Technol 57, 789. doi:10.2166/wst.2008.172 57 58 59 60 837 Regnery, J., Li, D., Roberts, S., Higgins, C., Sharp, J.O., Drewes, J.E., 2016. Linking 61 62 38 63 64 65 838 Trace Organic Chemical Attenuation to Microbiome metabolic Capabilities: 1 2 839 Insights from Laboratory- and Full-scale Managed Aquifer Recharge Systems, in: 3 4 840 ACS Symposium Series. pp. 163–187. doi:10.1021/bk-2016-1241.ch011 5 6 7 841 Regnery, J., Wing, A.D., Alidina, M., Drewes, J.E., 2015. Biotransformation of trace 8 9 10 842 organic chemicals during groundwater recharge: How useful are first-order rate 11 12 843 constants? J Contam Hydrol 179, 65–75. doi:10.1016/j.jconhyd.2015.05.008 13 14 15 844 Rezanezhad, F., Couture, R.M., Kovac, R., O’Connell, D., Van Cappellen, P., 2014. 16 17 845 Water table fluctuations and soil biogeochemistry: An experimental approach 18 19 846 using an automated soil column system. J Hydrol 509, 245–256. 20 21 22 847 doi:10.1016/j.jhydrol.2013.11.036 23 24 25 848 Rodríguez-Escales, P., Canelles, A., Sanchez-Vila, X., Folch, A., Kurtzman, D., 26 27 849 Rossetto, R., Fernández-Escalante, E., Lobo-Ferreira, J.-P., Sapiano, M., San- 28 29 850 Sebastián, J., Schüth, C., 2018. A risk assessment methodology to evaluate the 30 31 851 risk failure of managed aquifer recharge in the Mediterranean Basin. Hydrol Earth 32 33 34 852 Syst Sci 22, 3213–3227. doi:10.5194/hess-22-3213-2018 35 36 37 853 Rodríguez-Escales, P., Fernàndez-Garcia, D., Drechsel, J., Folch, A., Sanchez-Vila, 38 39 854 X., 2017. Improving degradation of emerging organic compounds by applying 40 41 855 chaotic advection in Managed Aquifer Recharge in randomly heterogeneous 42 43 44 856 porous media. Water Resour Res 53, 4376–4392. doi:10.1002/2016WR020333 45 46 47 857 Rodriguez-Escales, P., Sanchez-Vila, X., 2016. Fate of sulfamethoxazole in 48 49 858 groundwater: Conceptualizing and modeling metabolite formation under different 50 51 859 redox conditions. Water Res 105, 540–550. doi:10.1016/j.watres.2016.09.034 52 53 54 860 Röling, W.F., van Breukelen, B.M., Braster, M., Lin, B., van Verseveld, H.W., 2001. 55 56 861 Relationships between microbial community structure and hydrochemistry in a 57 58 59 862 landfill leachate-polluted aquifer. Appl Environ Microbiol 67, 4619–29. 60 61 62 39 63 64 65 863 Rothschild, L.J., Mancinelli, R.L., 2001. Life in extreme environments (nature).PDF. 1 2 864 Nature 409, 1092–1101. doi:10.1038/35059215 3 4 5 865 San-Sebastián-Sauto, J., Fernández-Escalante, E., Calero-Gil, R., Carvalho, T., 6 7 866 Rodríguez-Escales, P., 2018. Characterization and benchmarking of seven 8 9 10 867 managed aquifer recharge systems in south-western Europe. Sustain Water 11 12 868 Resour Manag 1–23. doi:10.1007/s40899-018-0232-x 13 14 15 869 Schütz, K., Nagel, P., Vetter, W., Kandeler, E., Ruess, L., 2009. Flooding forested 16 17 870 groundwater recharge areas modifies microbial communities from top soil to 18 19 871 groundwater table. FEMS Microbiol Ecol 67, 171–182. doi:10.1111/j.1574- 20 21 22 872 6941.2008.00608.x 23 24 25 873 Song, M., Jiang, L., Zhang, D., Luo, C., Wang, Y., Yu, Z., Yin, H., Zhang, G., 2016. 26 27 874 Bacteria capable of degrading anthracene, phenanthrene, and fluoranthene as 28 29 875 revealed by DNA based stable-isotope probing in a forest soil. J Hazard Mater 30 31 32 876 308, 50–57. doi:10.1016/J.JHAZMAT.2016.01.009 33 34 35 877 Staats, M., Braster, M., Röling, W.F.M., 2011. Molecular diversity and distribution of 36 37 878 aromatic hydrocarbon-degrading anaerobes across a landfill leachate plume. 38 39 879 Environ Microbiol 13, 1216–1227. doi:10.1111/j.1462-2920.2010.02421.x 40 41 42 880 Stein, H., Kellermann, C., Schmidt, S.I., Brielmann, H., Steube, C., Berkhoff, S.E., 43 44 881 Fuchs, A., Hahn, H.J., Thulin, B., Griebler, C., 2010. The potential use of fauna 45 46 47 882 and bacteria as ecological indicators for the assessment of groundwater quality. J 48 49 883 Environ Monit 12, 242–254. doi:10.1039/B913484K 50 51 52 884 Sun, Y., Shen, D., Zhou, X., Shi, N., Tian, Y., 2016. Microbial diversity and community 53 54 885 structure of denitrifying biological filters operated with different carbon sources. 55 56 886 Springerplus 5, 1752. doi:10.1186/s40064-016-3451-3 57 58 59 60 887 Terzaghi, K., Peck, R.B., Mesri, G., 1996. Soil mechanics in engineering practice, Third 61 62 40 63 64 65 888 ed. ed. Wiley-Interscience. 1 2 3 889 Torsvik, V., Øvreås, L., Thingstad, T.F., 2002. Prokaryotic diversity - Magnitude, 4 5 890 dynamics, and controlling factors. Science (80- ). doi:10.1126/science.1071698 6 7 8 891 Truu, M., Juhanson, J., Truu, J., 2009. Microbial biomass, activity and community 9 10 11 892 composition in constructed wetlands. Sci Total Environ 407, 3958–3971. doi: 12 13 893 10.1016/j.scitotenv.2008.11.036 14 15 16 894 Valhondo, C., Carrera, J., Ayora, C., Tubau, I., Martinez-Landa, L., Nödler, K., Licha, 17 18 895 T., 2015. Characterizing redox conditions and monitoring attenuation of selected 19 20 896 pharmaceuticals during artificial recharge through a reactive layer. Sci Total 21 22 23 897 Environ 512–513, 240–50. doi:10.1016/j.scitotenv.2015.01.030 24 25 26 898 Valhondo, C., Martinez-Landa, L., Carrera, J., Ayora, C., Nödler, K., Licha, T., 2018. 27 28 899 Evaluation of EOC removal processes during artificial recharge through a reactive 29 30 900 barrier. Sci Total Environ 612, 985–994. doi:10.1016/j.scitotenv.2017.08.054 31 32 33 901 Valhondo, C., Martínez-Landa, L., Carrera, J., Hidalgo, J.J., Tubau, I., De Pourcq, K., 34 35 902 Grau-Martínez, A., Ayora, C., 2016. Tracer test modeling for local scale residence 36 37 38 903 time distribution characterization in an artificial recharge site. Hydrol Earth Syst 39 40 904 Sci Discuss 1–17. doi:10.5194/hess-2016-197 41 42 43 905 Vuilleumier, S., Chistoserdova, L., Lee, M.C., Bringel, F., Lajus, A., Yang, Z., Gourion, 44 45 906 B., Barbe, V., Chang, J., Cruveiller, S., Dossat, C., Gillett, W., Gruffaz, C., 46 47 48 907 Haugen, E., Hourcade, E., Levy, R., Mangenot, S., Muller, E., Nadalig, T., Pagni, 49 50 908 M., Penny, C., Peyraud, R., Robinson, D.G., Roche, D., Rouy, Z., Saenempechek, 51 52 909 C., Salvignol, G., Vallenet, D., Zaining, W., Marx, C.J., Vorholt, J.A., Olson, M. V, 53 54 910 Kaul, R., Weissenbach, J., Médigue, C., Lidstrom, M.E., 2009. Methylobacterium 55 56 911 genome sequences: A reference blueprint to investigate microbial metabolism of 57 58 59 912 C1 compounds from natural and industrial sources. PLoS One 4. 60 61 62 41 63 64 65 913 doi:10.1371/journal.pone.0005584 1 2 3 914 Ward, N.L., Challacombe, J.F., Janssen, P.H., Henrissat, B., Coutinho, P.M., Wu, M., 4 5 915 Xie, G., Haft, D.H., Sait, M., Badger, J., Barabote, R.D., Bradley, B., Brettin, T.S., 6 7 916 Brinkac, L.M., Bruce, D., Creasy, T., Daugherty, S.C., Davidsen, T.M., DeBoy, 8 9 10 917 R.T., Detter, J.C., Dodson, R.J., Durkin, A.S., Ganapathy, A., Gwinn-Giglio, M., 11 12 918 Han, C.S., Khouri, H., Kiss, H., Kothari, S.P., Madupu, R., Nelson, K.E., Nelson, 13 14 919 W.C., Paulsen, I., Penn, K., Ren, Q., Rosovitz, M.J., Selengut, J.D., Shrivastava, 15 16 920 S., Sullivan, S.A., Tapia, R., Thompson, S., Watkins, K.L., Yang, Q., Yu, C., Zafar, 17 18 19 921 N., Zhou, L., Kuske, C.R., 2009. Three genomes from the phylum Acidobacteria 20 21 922 provide insight into the lifestyles of these microorganisms in soils. Appl Environ 22 23 923 Microbiol 75, 2046–2056. doi:10.1128/AEM.02294-08 24 25 26 924 Wright, J., Kirchner, V., Bernard, W., Ulrich, N., McLimans, C., Campa, M.F., Hazen, 27 28 925 T., Macbeth, T., Marabello, D., McDermott, J., Mackelprang, R., Roth, K., 29 30 31 926 Lamendella, R., 2017. Bacterial Community Dynamics in Dichloromethane- 32 33 927 Contaminated Groundwater Undergoing Natural Attenuation. Front Microbiol. 34 35 36 928 Young, I.M., Crawford, J.W., 2004. Interactions and self-organization in the soil- 37 38 929 microbe complex. Science (80- ). doi:10.1126/science.1097394 39 40 41 930 Zhang, Y., Sun, R., Zhou, A., Zhang, J., Luan, Y., Jia, J., Yue, X., Zhang, J., 2018. 42 43 44 931 Microbial community response reveals underlying mechanism of industrial-scale 45 46 932 manganese sand biofilters used for the simultaneous removal of iron, manganese 47 48 933 and ammonia from groundwater. AMB Express 8, 2. doi:10.1186/s13568-017- 49 50 934 0534-7 51 52 53 935 Zhou, Y., Kellermann, C., Griebler, C., 2012. Spatio-temporal patterns of microbial 54 55 56 936 communities in a hydrologically dynamic pristine aquifer. FEMS Microbiol Ecol 81, 57 58 937 230–242. doi:10.1111/j.1574-6941.2012.01371.x 59 60 61 62 42 63 64 65 938 Zhu, B., Wang, X., Rioual, P., 2017. Multivariate indications between environment and 1 2 939 ground water recharge in a sedimentary drainage basin in northwestern China. J 3 4 940 Hydrol 549, 92–113. doi:10.1016/j.jhydrol.2017.03.058 5 6 7 941 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 43 63 64 65 Supplementary material Click here to download Figure: Supplementary figures and tablesl.docx Supplementary material 1. Input data for Principal Component Analyses Table S1 Summary of physico-chemical variables of water samples (values in bold represent concentrations under detection limit) Table S2 Summary of diversity indices of water samples Table S3 Summary of relative abundance (%) of microbial community classes of water samples Table S4 Summary of relative abundance (%) of microbial community species of water samples Table S5 Summary of diversity indices and depth of soil samples Table S6 Summary of soil parameters from grain-size distribution curves Table S7 Summary of relative abundance (%) of microbial community classes of soil samples Table S8 Summary of relative abundance (%) of microbial community species of soil samples 2. Supplementary figures Figure S1 PCA2 analysis with microbial classes and hydrochemical data from groundwater samples and infiltration water sample Figure S2 PCA4 analysis with microbial species and hydrochemical data from groundwater samples and infiltration water sample Figure S3 PCA6 analysis with microbial classes data from soil samples Figure S4 PCA8 analysis with microbial species data from soil samples Figure S5 PCA10 analysis with microbial classes data from soil and water samples 1. Input data for Principal Component Analyses Table S1 Summary of physico-chemical variables of water samples (values in bold represent concentrations under detection limit) Operating Recharge (July 2014) No Recharge (March 2015) conditions Infiltration Sample P1(-0.6) P2 (3) P2 (-5) P3 (0.8) P3 (-4.2) P5 (2.3) P5 (-2.2) P8 (1) P8 (-3) P9(3.5) P9(-2) P10(-1) P1(-0.6) P2 (-2) P3 (0.8) P5 (2.3) P8 (1) P8 (-3) P9(-2) P10(-1) water Relative depth 14.00 -0.59 4.80 -5.00 0.80 -3.20 2.31 -2.19 0.96 -3.04 3.53 -1.91 -0.90 -0.59 -2.00 0.80 2.31 0.96 -3.04 -1.97 -0.90 (m) Temperature 25.10 16.87 22.72 18.66 17.77 17.71 24.84 22.05 18.00 17.50 18.86 16.99 15.56 16.50 18.50 17.10 24.50 16.30 17.80 14.50 13.70 (⁰C) Conductivity 1.28 1.56 1.51 1.52 1.65 1.68 1.56 1.54 1.48 1.54 1.57 1.55 1.65 1.33 1.27 1.29 1.37 1.24 1.31 1.22 1.32 (μS/cm DOC (mg/l) 6.60 1.65 3.17 4.70 3.22 3.15 3.61 3.76 13.65 8.02 16.31 11.71 2.57 1.93 1.80 2.33 1.53 3.96 1.65 1.74 1.61 pH 8.47 7.05 7.30 7.28 7.15 7.15 7.39 7.42 7.45 7.39 7.30 7.32 7.28 7.36 7.34 7.15 6.98 7.13 7.27 7.12 7.35 Cl- (ppm) 245.86 224.75 267.00 212.00 241.00 241.66 274.62 269.68 267.52 262.78 256.58 249.53 265.49 212.39 211.80 193.19 191.65 186.14 193.46 192.34 189.11 NO3- (ppm) 6.14 5.02 3.13 3.64 6.06 8.73 3.41 4.08 2.75 2.88 4.29 6.57 5.88 17.79 12.90 18.20 17.70 2.82 15.49 11.54 16.28 SO4-2 (ppm) 123.43 173.04 140.79 138.37 168.60 168.66 133.80 147.84 120.04 152.06 152.89 131.52 179.98 181.48 139.62 177.03 200.31 141.23 199.54 137.00 174.81 HCO3- (mg/l) 144.50 354.30 261.20 272.60 399.00 385.80 256.10 255.20 241.30 253.80 301.00 293.90 327.10 382.60 314.20 294.30 408.10 379.30 408.10 333.10 396.40 Ca (ppm) 67.59 135.23 94.24 103.49 144.15 141.27 97.40 96.73 92.46 99.88 102.87 105.21 123.29 142.63 121.41 133.29 141.88 122.96 72.31 114.70 132.74 Fe (ppm) 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 Si (ppm) 0.12 3.28 1.86 1.99 4.14 4.27 1.73 1.73 1.64 1.89 1.94 1.94 2.10 3.00 2.64 3.49 3.21 3.85 3.30 2.85 2.46 B (ppm) 0.20 0.26 0.23 0.22 0.22 0.21 0.20 0.20 0.21 0.21 0.23 0.22 0.22 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 Mg (ppm) 22.99 33.73 27.44 29.81 35.99 35.15 29.06 28.77 26.67 28.47 29.13 29.33 31.44 36.28 31.86 33.02 38.17 30.81 38.34 31.85 37.22 Na (ppm) 147.89 138.91 153.85 158.19 142.48 143.92 163.36 165.02 159.96 161.18 158.53 157.98 162.79 116.28 113.52 108.76 118.03 110.09 116.74 117.43 117.49 S (ppm) 42.22 52.21 41.72 42.55 47.64 47.90 42.93 42.78 40.45 47.19 46.72 46.27 51.40 63.51 58.08 58.73 63.37 64.34 51.23 53.88 59.89 K(ppm) 28.01 19.63 30.15 30.99 24.18 24.41 38.60 31.86 29.97 29.39 33.13 30.60 26.32 23.76 22.90 23.92 24.44 23.02 20.73 23.57 27.52 P (ppm) 0.38 1.14 0.96 1.63 0.71 0.50 0.50 0.66 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 0.50 Ba (ppb) 50.00 74.89 70.28 70.62 59.40 58.68 66.94 66.11 74.87 74.62 58.33 656.87 56.72 76.99 56.13 71.99 74.42 58.55 39.87 46.10 60.64 Li (ppb) 14.54 46.28 23.96 34.39 23.78 23.26 24.15 24.57 20.88 21.89 25.22 24.84 22.80 15.64 15.46 16.27 18.62 16.49 18.54 16.69 19.00 Mn (ppb) 4.76 28.93 10.79 8.94 3.67 1.74 2.17 1.46 2.46 2.77 4.46 2.26 1.58 0.87 5.51 0.81 0.80 145.30 0.80 3.90 0.80 Zn (ppb) 10.67 57.28 69.22 19.92 23.94 22.56 15.62 9.70 24.55 30.71 18.21 12.78 15.73 53.76 15.13 24.75 13.36 17.37 6.61 32.25 13.68 Al (ppm) 10.09 12.08 18.17 86.17 4.11 3.70 3.96 99.48 4.93 8.20 2.36 10.52 0.50 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20 As (ppb) 3.51 3.18 2.64 2.43 1.20 1.29 3.56 3.47 3.77 3.64 1.49 1.52 1.46 3.28 1.78 1.56 3.00 2.02 1.89 1.70 1.67 Pb (ppb) 0.80 1.29 1.12 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 2.09 0.80 0.80 0.80 0.80 0.80 0.80 0.80 Cu (ppb) 4.95 6.94 8.28 5.22 4.91 5.16 5.05 4.50 6.47 6.93 3.82 3.70 3.17 1.32 1.61 1.14 1.13 0.80 0.80 1.96 1.11 U (ppb) 1.23 1.80 1.11 1.27 2.18 2.11 1.01 1.04 0.91 1.08 1.44 1.45 1.81 2.32 2.02 2.49 2.35 1.46 1.97 2.07 2.42 Cd (ppb) 9.08 6.49 17.04 0.80 10.06 4.44 1.45 0.65 0.80 13.41 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 V (ppb) 2.70 29.78 2.40 2.50 2.17 2.20 2.77 2.68 2.51 2.63 1.71 2.07 2.10 8.76 8.64 8.50 8.99 9.47 9.80 9.30 9.06 Ni (ppb) 10.62 35.34 6.97 5.51 6.16 5.37 6.24 4.63 5.66 8.26 4.48 4.20 4.50 2.53 1.89 2.00 1.91 2.59 1.32 1.62 1.82 Co (ppb) 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 Table S2 Summary of diversity indices of water samples Operating Recharge (July 2014) No Recharge (March 2015) conditions Infiltration P1(- P3 P3 (- P5 P5 (- P1(- P3 P5 Sample P2 (3) P2 (-5) P8 (1) P8 (-3) P9(3.5) P9(-2) P10(-1) P2 (-2) P8 (1) P8 (-3) P9(-2) P10(-1) water 0.6) (0.8) 4.2) (2.3) 2.2) 0.6) (0.8) (2.3) Shannon 2.78 2.57 2.29 2.8 2.65 2.65 1.91 2.22 1.43 1.69 2.18 2.07 2.51 2.73 2.67 2.01 2.41 1.73 2.03 3.01 2.49 Richness 29.00 22.00 19.00 23.00 21.00 20.00 15.00 13.00 10.00 11.00 11.00 13.00 17.00 19.00 18.00 11.00 14.00 11.00 10.00 23.00 16.00 Evenness 0.65 0.6 0.54 0.66 0.62 0.62 0.45 0.52 0.34 0.4 0.51 0.49 0.59 0.64 0.63 0.47 0.57 0.41 0.48 0.71 0.59 Table S3 Summary of relative abundance (%) of microbial community classes of water samples Operating Recharge (July 2014) No Recharge (March 2015) conditions Infiltration P1(- P3 P3 (- P5 P5 (- P1(- P3 P5 P10(- Sample P2 (3) P2 (-5) P8 (1) P8 (-3) P9(3.5) P9(-2) P10(-1) P2 (-2) P8 (1) P8 (-3) P9(-2) water 0.6) (0.8) 4.2) (2.3) 2.2) 0.6) (0.8) (2.3) 1) Betaproteobacteria 13.81 9.48 7.29 9.09 7.24 10.49 21.88 22.88 69.94 54.76 33.71 37.80 23.50 22.69 4.83 38.52 14.29 3.02 57.02 19.60 2.69 Cyanobacteria 12.85 11.13 6.25 2.86 13.57 13.64 9.09 7.63 5.83 8.44 8.57 6.71 21.31 8.40 4.42 5.74 7.94 1.51 0.00 3.71 8.31 Bacili 1.53 23.92 17.53 12.63 10.86 6.29 42.90 11.02 1.84 1.73 21.71 0.61 6.56 8.40 2.00 18.03 15.87 0.00 11.57 2.47 4.64 Dehalococcoidia 3.74 4.54 29.86 14.98 17.19 16.43 1.42 0.00 0.00 0.00 2.86 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.28 0.00 Acidobacteria 7.57 4.95 9.03 4.55 11.31 9.09 1.99 0.00 0.00 0.00 2.29 9.15 4.92 4.20 3.64 0.00 0.00 13.57 0.00 0.00 0.00 Actinobacteria 8.53 1.24 0.52 4.21 3.17 5.24 0.00 0.00 0.92 1.08 0.00 0.00 0.00 3.36 8.14 6.15 0.00 0.00 7.44 2.20 1.81 Cytophagia 0.19 0.00 0.00 0.00 0.00 0.00 12.22 18.64 6.75 3.90 12.00 6.71 9.29 16.81 4.91 0.00 4.76 5.53 7.02 1.51 0.00 Chlorobia 13.52 10.52 6.77 2.53 0.00 0.00 2.84 0.00 2.76 10.61 0.00 0.00 0.00 0.00 4.09 0.00 0.00 5.03 0.00 0.00 17.18 Nitrospira 4.03 0.00 0.00 4.38 9.50 8.39 1.14 11.02 0.00 0.00 2.86 7.32 8.20 2.52 0.00 0.00 3.17 0.00 0.00 4.27 0.00 GAmmaproteobacteria 5.08 0.82 4.17 5.89 0.00 0.00 4.55 20.34 5.83 7.79 3.43 6.71 0.00 11.76 5.65 13.93 18.25 0.00 0.00 1.86 11.61 Alphaproteobacteria 4.89 1.44 1.74 0.17 2.26 0.00 0.28 1.69 0.61 0.65 0.00 0.00 2.73 2.52 0.00 1.64 9.52 0.00 0.00 8.03 7.15 Others 24.26 31.96 16.84 38.72 24.89 30.42 1.70 6.78 5.52 11.04 12.57 25.00 23.50 19.33 62.32 15.98 26.19 71.36 16.94 47.08 46.61 Table S4 Summary of relative abundance (%) of microbial community species of water samples Operating Recharge (July 2014) No Recharge (March 2015) conditions Infiltration Sample P1(-0.6) P2 (3) P2 (-5) P3 (0.8) P3 (-4.2) P5 (2.3) P5 (-2.2) P8 (1) P8 (-3) P9(3.5) P9(-2) P10(-1) P1(-0.6) P2 (-2) P3 (0.8) P5 (2.3) P8 (1) P8 (-3) P9(-2) P10(-1) water Methylotenera 0.48 0.00 0.17 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 10.92 0.00 32.79 14.29 0.00 31.40 0.00 0.00 mobilis Pontibacter sp 0.19 0.00 0.00 0.00 0.00 0.00 12.22 18.64 6.75 3.90 12.00 6.71 9.29 16.81 4.91 0.00 4.76 5.53 7.02 1.51 0.00 Dehalogenimonas 3.74 4.54 29.86 14.98 17.19 16.43 1.42 0.00 0.00 0.00 2.86 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 9.28 0.00 sp Subgroup 3 sp 7.57 4.95 9.03 4.55 11.31 9.09 1.99 0.00 0.00 0.00 2.29 9.15 4.92 4.20 3.64 0.00 0.00 13.57 0.00 0.00 0.00 (Acidobact.) Vogesella 3.36 9.48 4.17 2.86 3.17 3.50 2.84 1.69 60.12 48.70 13.14 32.93 0.00 4.20 3.23 2.46 0.00 0.00 6.20 0.00 2.69 indigofera Stenotrophomonas 5.08 0.82 4.17 5.89 0.00 0.00 4.55 20.34 5.83 7.79 3.43 6.71 0.00 11.76 5.65 13.93 18.25 0.00 0.00 1.86 11.61 sp Chryseomicrobium 1.53 23.92 17.53 12.63 10.86 6.29 42.90 11.02 1.84 1.73 21.71 0.61 6.56 8.40 2.00 18.03 15.87 0.00 11.57 2.47 4.64 sp Nitrospira sp1 3.36 0.00 0.00 0.84 1.81 2.80 0.00 4.24 0.00 0.00 0.00 5.49 4.37 0.00 0.00 0.00 3.17 0.00 0.00 4.27 0.00 Methylobacterium 4.89 1.44 1.74 0.17 2.26 0.00 0.28 1.69 0.61 0.65 0.00 0.00 2.73 2.52 0.00 1.64 9.52 0.00 0.00 8.03 7.15 sp Nitrospira sp2 0.67 0.00 0.00 3.54 7.69 5.59 1.14 6.78 0.00 0.00 2.86 1.83 3.83 2.52 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Table S5 Summary of diversity indices and depth of soil samples Operating Recharge (July 2014) No Recharge (March 2015) conditions Sample SD E_0.1 M_0.1 M_0.2 F_0.1 E_0.2 E_0.8 M_0.3 F_0.1 F_0.3 Depth (cm) 0.10 0.10 0.10 0.20 0.10 0.20 0.80 0.30 0.10 0.30 Shannon 2.89 3.22 3.36 3.14 3.29 2.93 2.86 2.90 2.81 2.40 Richness 25.50 35.50 37.00 30.50 34.00 25.00 24.00 21.00 19.50 16.00 Evenness 0.68 0.76 0.79 0.74 0.77 0.69 0.67 0.68 0.66 0.57 Table S6 Summary of soil parameters from grain-size distribution curves Operating Recharge (July 2014) No Recharge (March 2015) conditions Sample SD E_0.1 M_0.1 M_0.2 F_0.1 E_0.2 E_0.8 M_0.3 F_0.1 F_0.3 CC 0.04 2.01 1.43 0.84 0.53 3.31 0.54 0.93 0.55 1.27 CU 159.39 53.52 28.32 20.03 13.28 17.72 19.23 32.59 19.69 8.13 % Fine 10.12 2.17 2.92 0.85 2.60 0.54 1.67 1.78 1.94 0.43 D10 0.07 0.27 0.54 0.20 0.31 1.08 0.25 0.28 0.28 1.42 Table S7 Summary of relative abundance (%) of microbial community classes of soil samples Operating Recharge (July 2014) No Recharge (March 2015) conditions Sample SD E_0.1 M_0.1 M_0.2 F_0.1 E_0.2 E_0.8 M_0.3 F_0.1 F_0.3 Betaproteobacteria 23.01 21.37 16.97 18.29 18.89 21.32 8.06 14.28 11.31 14.01 Cyanobacteria 8.66 5.31 8.26 6.26 3.54 11.09 8.36 3.40 4.76 2.72 Bacili 6.39 5.17 4.35 8.07 7.25 7.74 7.56 4.53 6.85 21.50 Dehalococcoidia 4.40 1.77 2.10 2.97 3.37 0.00 1.93 0.00 0.00 0.00 Acidobacteria 9.80 2.31 1.95 0.99 1.69 0.00 2.09 2.83 0.00 0.00 Actinobacteria 3.27 2.18 4.35 3.79 2.36 4.96 15.63 9.49 7.89 16.11 Cytophagia 0.85 2.45 0.60 3.29 4.72 4.53 13.21 10.51 13.99 11.77 Chlorobia 0.00 2.31 2.55 2.97 4.22 0.00 0.32 0.00 0.00 0.00 Nitrospira 5.40 1.91 7.06 6.75 4.05 14.72 10.47 5.07 5.65 4.53 GAmmaproteobacteria 2.13 3.95 1.80 0.00 4.05 4.34 2.26 3.97 0.00 0.00 Alphaproteobacteria 0.99 3.81 2.10 1.65 7.76 0.57 3.06 6.71 6.43 7.88 Others 35.09 47.47 47.90 44.98 38.11 30.74 27.04 39.21 43.13 21.50 Table S8 Summary of relative abundance (%) of microbial community species of soil samples Operating Recharge (July 2014) No Recharge (March 2015) conditions Sample SD E_0.1 M_0.1 M_0.2 F_0.1 E_0.2 E_0.8 M_0.3 F_0.1 F_0.3 Methylotenera 0.00 0.00 0.60 0.00 0.00 2.64 0.00 0.00 0.00 0.91 mobilis Pontibacter sp 0.85 2.45 0.60 3.29 4.72 4.53 13.21 10.51 13.99 11.77 Dehalogenimonas 4.40 1.77 2.10 2.97 3.37 0.00 1.93 0.00 0.00 0.00 sp Subgroup 3 sp 9.80 2.31 1.95 0.99 1.69 0.00 2.09 2.83 0.00 0.00 (Acidobact.) Vogesella indigofera 8.10 7.49 4.80 4.94 4.22 4.53 2.42 2.83 3.57 0.00 Stenotrophomonas 2.13 3.95 1.80 0.00 4.05 4.34 2.26 3.97 0.00 0.00 sp Chryseomicrobium 6.39 5.17 4.35 8.07 7.25 7.74 7.56 4.53 6.85 21.50 sp Nitrospira sp1 2.70 0.54 5.56 3.62 2.53 14.72 7.41 1.39 4.17 0.00 Methylobacterium sp 0.99 3.81 2.10 1.65 7.76 0.57 3.06 6.71 6.43 7.88 Nitrospira sp2 2.70 1.36 1.50 3.13 1.52 0.00 3.06 3.68 1.49 4.53 2. Supplementary figures Figure S1. PCA2 analysis with microbial classes and hydrochemical data from groundwater samples (squares) and the infiltration water sample (diamond) during the recharge period (blue) and during no-recharge period (orange). Position of samples is scaled in VF1 and VF2 axes. Red arrows represent the contribution of each variable in both Varifactors. Figure S2 PCA4 analysis with microbial species and hydrochemical data from groundwater samples (squares) and the infiltration water sample (diamond) during the recharge period (blue) and during no-recharge period (orange). Position of samples is scaled in VF1 and VF2 axes. Red arrows represent the contribution of each variable in both Varifactors. Figure S3 PCA6 analysis with microbial classes data from soil samples during the recharge period (blue) and during no-recharge period (orange). Circles are related to the infiltration pond samples (Entrance, Midfield and Final strecht according the capital letter close to each symbol). Triangle is referred to the sedimentation pond sample. Position of samples is scaled in VF1 and VF2 axes. Red arrows represent the contribution of each variable in both Varifactors scaled by components. Figure S4. PCA8 analysis with microbial species data from soil samples during the recharge period (blue) and during no-recharge period (orange). Circles are related to the infiltration pond samples (Entrance, Midfield and Final strecht according the capital letter close to each symbol). Triangle is referred to the sedimentation pond sample. Position of samples is scaled in VF1 and VF2 axes. Red arrows represent the contribution of each variable in both Varifactors scaled by components. Figure S5. PCA10 analysis with microbial classes data from soil and water samples during the recharge period (blue) and during no-recharge period (orange). Circles are related to soil samples (Entrance, Midfield and Final stretch according the capital letter close to each symbol). Triangle is referred to the sedimentation pond soil sample. Squares represent groundwater samples and the diamond is for infiltration water. Position of samples is scaled in VF1 and VF2 axes. Red arrows represent the contribution of each variable in both varifactors.