Analysis of Natural Responses and Hydrochemical Data by Statistical Approaches as a Tool to Understanding the Hydrogeological Functioning of a Highly Karstified Evaporitic System in South Spain

Metadata of the chapter that will be visualized in SpringerLink Book Title EuroKarst 2016, Neuchâtel Series Title Chapter Title Analysis of Natural Response and Hydrochemical Data by Statistical Approaches to Characterize the Hydrogeological Functioning of a Highly Karstified Evaporitic System in South Spain Copyright Year 2017 Copyright HolderName Springer International Publishing Switzerland Corresponding Author Family Name Gil-Márquez Particle Given Name Jose Manuel Prefix Suffix Division Organization Department of Geology and Center of Hydrogeology of the University of Malaga (CEHIUMA) Address 29071, Málaga, Spain Email [email protected] Author Family Name Mudarra Particle Given Name Matías Prefix Suffix Division Organization Department of Geology and Center of Hydrogeology of the University of Malaga (CEHIUMA) Address 29071, Málaga, Spain Email [email protected] Author Family Name Andreo Particle Given Name Bartolomé Prefix Suffix Division Organization Department of Geology and Center of Hydrogeology of the University of Malaga (CEHIUMA) Address 29071, Málaga, Spain Email [email protected] Abstract At southern Cordoba Province (S Spain), an evaporitic karst plateau drained by brine springs is located. A periodic monitoring of discharge rate, EC, water temperature, and pH was performed in one of them (Lower Anzur spring) and water samples were collected for chemical analysis. Physicochemical data were used to perform a principal component analysis (PCA) to characterize the functioning of the system. Temporal evolution of the controlled parameters reveals a markedly karst behavior. PCA has defined two principal factors, one related to salinity and other linked to infiltration processes (NO3−, Ca2+). The slight thermal anomaly and the high mineralization registered indicate the existence of ascending regional groundwater flows of long residence time, which would converge into the spring, mixed with recently infiltrated water that circulated through a conduit network formed in the evaporite rocks. UNCORRECTED PROOF Layout: T2 Med_Medium Book ID: 430122_1_En Book ISBN: 978-3-319-45465-8 Chapter No.: 32 Date: 22 October 2016 18:53 Page: 335/343 335335 32.1 Introduction – 336 32.2 Site Description – 336 32.3 Methods – 337 32.4 Results – 338 32.5 Discussion – 341 32.6 Conclusions – 343 References – 343 Analysis of Natural Response and Hydrochemical Data by Statistical Approaches to Characterize the Hydrogeological Functioning of a Highly Karstified Evaporitic System in South Spain Jose Manuel Gil-Márquez, Matías Mudarra and Bartolomé Andreo Abstract At southern Cordoba Province (S Spain), an evaporitic karst plateau drained by brine springs is located. A periodic monitoring of discharge rate, EC, water temperature, and pH was performed in one of them (Lower Anzur spring) and water samples were collected for chemical analysis. Physicochemical data were used to perform a principal component analysis (PCA) to characterize the functioning of the system. Temporal evolution of the controlled parameters reveals a markedly karst behavior. PCA has defined two principal factors, one related to salinity and other linked to infiltration processes (NO3−, Ca2+). The slight thermal anomaly and the high mineralization registered indicate the existence of ascending regional groundwater flows of long residence time, which would converge into the spring, mixed with recently infiltrated water that circulated through a conduit network formed in the evaporite rocks. 32 © Springer International Publishing Switzerland 2017 P. Renard, C. Bertrand (Eds.), EuroKarst 2016, Neuchâtel, Advances in Karst Science, DOI 10.1007/978-3-319-45465-8_32 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 AQ1 Author Proof UNCORRECTED PROOF Layout: T2 Med_Medium Book ID: 430122_1_En Book ISBN: 978-3-319-45465-8 Chapter No.: 32 Date: 22 October 2016 18:53 Page: 336/343 32 336 Chapter 32 · Analysis of Natural Response and Hydrochemical Data … to obtain a better understanding on the hydrogeological functioning of the evaporite-karst plateau that it drains. Some of the most commonly hydrogeological methods used in carbonate aquifers have been applied (Drew and Goldscheider 2007): temporal evolution of natural responses and multivariate analysis (PCA), in addition to the intrinsic lithological and hydrological characteristics of the area, which are also considered. Moreover, the relationships between the observed hydrogeological variables, possible water–rock interaction processes, and the composition of the evaporitic bedrock are analyzed. 32.2 Site Description South of the Cordoba Province (Andalusia, Spain), an evaporite-karst plateau belonging to the CSC, is placed between Genil (southward) and Anzur (northward) rivers (. Fig. 32.1). The outcrop (80 km2) is characterized by gentle hills, not well-defined drainage network, and the presence of numerous endorheic areas. The prevailing climate is temperate Mediterranean, with a marked seasonal pattern in the annual distribution of rainfalls (mainly in autumn and winter). The mean historic annual precipitation and air temperature are 412 mm and 18 °C, respectively (Consejería de Medio Ambiente 2005). The research period (September 2014 to February 2016) could be considered slightly dry, with an annual precipitation of 361 mm (hydrological year 2014–2015) and a mean annual temperature of 17.1 °C, recorded in a meteorological station installed on the area (. Fig. 32.1). Poljes and karst depressions are well represented in the pilot site and some of them are often intersected by the water table, constituting wetlands of variable size during flooding periods (. Fig. 32.1). The altitude of these karst features is ranged from 410 to 430 m a.s.l., at the center of the plateau, to 300 m a.s.l. northward, where other endorheic areas exist, aligned according to N-S direction (along 3 km). The drainage of these last zones is produced through swallow holes, sometimes obstructed, causing temporary flooding of depressions. The most significant wetlands in the area are Jarales and Amarga (Andreo et al. 2016), which conform, together with other ephemeral lakes, a complex of wetlands placed at the central part of the outcrop (between 360 and 425 m a.s.l.). Besides direct infiltration of surface water into swallow holes, recharge also takes place by diffuse infiltration of rainwater through dry depressions and gypsum and calcareous outcrops. Natural discharge mainly occurs northward, via springs located beneath the Anzur River or by a diffuse way, although some outflow occurs toward the Genil River as well (. Fig. 32.1, Andreo et al. 2016). The most significant discharge point (Lower Anzur spring) 32.1 Introduction Research on Karst Hydrogeology has been performed by the application of several specific approaches, which have provided basic information about the structure and dynamics of karst aquifers (Ford and Williams 1989; Drew and Goldscheider 2007). Traditionally, these investigations have been focused on carbonate media, whereas researches on evaporitic karst areas have been mainly motivated by other aspects such as karstology, natural impacts, or human-induced geohazards (Klimchouk et al. 1999; Gutiérrez et al. 2008; Cooper and Gutiérrez 2013), rather than by its hydrogeological functioning. Despite having little interest from the water supply standpoint, it is necessary to achieve an acceptable knowledge about the hydrogeological functioning of karst systems developed within evaporite rocks. Groundwater resources stored in them have normally high saline content and the water drained by brine springs can provoke the natural deterioration of the water quality in many rivers and reservoirs (Memon et al. 1999). In Andalusia (southern Spain), the northern sector of the Subbetic Domain in the Betic Cordillera is formed by an olistostrome unit known as Chaotic Subbetic Complex (CSC). This megabreccia is basically made up by Triassic (Keuper) clays and evaporites (gypsum, anhydrite, and halite) as well as blocks of other lithologies: limestones, sandstones, etc. (Vera and Martin-Algarra 2004). Despite low permeability and aquitard behavior have been traditionally assumed for these materials, groundwater flow and storage within them are possible due to dissolution/ karstification processes affecting evaporite rocks, causing a great development of secondary porosity and permeability. As a consequence, unstable karst conduits and cavities appear, giving place to subsidence phenomena, sinkholes, surface depressions, and springs (Calaforra and Pulido-Bosch 1999), typical features of karst aquifers. Likewise than in carbonate aquifers (Lastennet 1994; Perrin et al. 2003), a detailed monitoring of natural responses of springs, jointly with the use of quantitative approaches, such as statistical methods (principal component analysis—PCA), applied to evaporitic media should provide basic information about the behavior of the systems that they drain. This would make possible to distinguish the location of storage (soil–epikarst, unsaturated, or saturated zones), the degree of karstification, and the participation of each zone in the functioning of the system (Hunkeler and Mudry 2007; Mudarra and Andreo 2011). In this work, data series of discharge, electrical conductivity, pH, water temperature, and hydrochemistry (major hydrochemical components) from a brine spring located in southern Spain (. Fig. 32.1) have been coupled 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 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 Author Proof UNCORRECTED PROOF Layout: T2 Med_Medium Book ID: 430122_1_En Book ISBN: 978-3-319-45465-8 Chapter No.: 32 Date: 22 October 2016 18:53 Page: 337/343 337 32 ±1 µS/cm for EC, ±0.1 °C for temperature, and ±0.1 for pH units. A gauging station with a 60 °V-notch weir was equipped in the spring for recording outflow data. In addition, precipitation was recorded hourly at a rain gauge located on the CSC outcrops (395 m a.s.l., . Fig. 32.1), near Jarales wetland. During the study period, sampling periodicity was adapted to the different hydrological conditions (recharge, recession, depletion): daily in high flow conditions and fortnightly during periods of depletion. Water samples (59) were taken with the same periodicity in 150-ml amber glass bottles with no headspace to avoid degassing and, therefore, calcite is situated in the northern sector of the system, at 258 m a.s.l. (. Fig. 32.1). 32.3 Methods From September 2014 to February 2016, field measurements of discharge (OTT C2 Flow Meter), electrical conductivity (EC), and water temperature (WTW Cond 3310), together with pH lectures (Hach HQ40d) were recorded in water drained by Lower Anzur spring. The accuracy of measurements was ±2 % for discharge, . Fig. 32.1 Geological and hydrogeological sketch of the study area. Modified from Andreo et al. (2016) 32.2 · Site Description 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 Author Proof UNCORRECTED PROOF Layout: T2 Med_Medium Book ID: 430122_1_En Book ISBN: 978-3-319-45465-8 Chapter No.: 32 Date: 22 October 2016 18:53 Page: 338/343 32 338 Chapter 32 · Analysis of Natural Response and Hydrochemical Data … precipitation. They were stored at 4 °C and processed within 48 h. Hydrochemical parameters considered in this study were analyzed in the laboratory of the Centre of Hydrogeology of the University of Malaga (CEHIUMA). Total alkalinity (Alk) was determined by volumetric titration with 0.02 N H2SO4 to pH 4.45. Chemical analyses of major components (Ca2+, Mg2+, Na+, K+, Cl−, SO42−, NO3−) were performed using high-pressure liquid chromatography (Metrohm 792 Basic IC and Metrohm Compact 881 IC pro) with ±0.1 mg/l accuracy. Samples were diluted to 1 mS/cm and filtered (0.45-µm Millipore filters) before entering the system (filter in line and precolumn filter). PCA has been applied to quantitatively corroborate the qualitative description displayed from temporal evolution and to interpret the hydrochemical data and their relevant sources and processes (Bakalowicz 1977; Mudry 1987). Data were normalized by Pearson correlation coefficient (variance 1/n). 32.4 Results .Table 32.1 summarizes the most significant statistical parameters corresponding to the physicochemical components analyzed in present work. The average value of EC (147.5 mS/cm) and the clear predominance of Cl− and Na+ facies (up to 105.1 and 61.0 g/l, respectively) determine the hypersaline nature of the water drained by Lower Anzur spring, which reflects the prevailing evaporite nature of system rocks. Spring water also shows high mean values of SO42− (5.9 g/l), Ca2+ (2.3 g/l), Mg2+ (0.4 g/l), and K+ (0.2 g/l). Alkalinity ranged from 189.4 to 268.4 mg/l, while the mean value of NO3− was 18.8 mg/l, presenting the highest coefficient of variation (0.70). Water temperature varied from 19.4 to 21.3 °C (few degrees over the mean annual air temperature −17.8 °C), whereas pH ranged from 6.6 to 7.8. The average outflow value from Lower Anzur spring during the study period was 20.8 l/s. .Figure 32.2 shows the temporal evolution of discharge and the hydrochemical parameters listed in .Table 32.1, together with the rainfall recorded in the area during the study period. The analysis of the data obtained in Lower Anzur spring enables to observe relatively quick variations in flow rate (from 0.7 to 91.9 l/s) in response to recharge events. The main changes in discharge rate were accompanied by sharp falls of EC (up to 39 mS/cm, November 2015), which later progressively rose during recession periods (. Fig. 32.2). Previous to the main dilution in November 2015, a slight increase in EC (6 mS/cm) was detected. Variations in water mineralization were caused by corresponding changes of most of the analyzed ions (. Fig. 32.2), especially Cl−, Na+, SO42−, and Mg2+, as well as alkalinity, which decreased progressively at the same time as flow rate rose. In general, Ca2+ and K+ concentrations behave in a similar pattern as EC but, unlike previous parameters, their falls were less accentuated or even slight rises of Ca2+ values can be observed during high water conditions. With respect to NO3− contents, they generally increased and reached higher values (up to 54 mg/l) coinciding with some recharge events, and they fell during the subsequent low water periods. Nevertheless, the lowest NO3− values (under detection limit) were recorded in November 2015, coinciding with the highest EC. Temperature of water drained by Lower Anzur spring presented no clear pattern in its temporal evolution at most some significant increases, as consequence of the main recharge events that took place in autumn (. Fig. 32.2). Simultaneous to the rainfall registered at November 2015, a rise of 0.6 °C was observed, followed by a decrease in equal magnitude. Temporal evolution of pH shows an opposite pattern to that of flow rate, with lowest values coinciding with maximum discharge rate. The former qualitative analysis deduced from . Fig. 32.2 has statistical significance as can be inferred from principal PCA of . Fig. 32.3 and .Table 32.2. PCA has been performed using all variables listed in .Table 32.1, except discharge. .Table 32.2 shows the factor loadings of each parameter for the first three principal components obtained from the PCA. The two main axes of the PCA account for 72 % of the total variance. Factor 1 is determined by EC, Cl−, and Na+ and in a lower degree by Alk, Mg2+, K+, and SO42− (. Fig. 32.3a, .Table 32.2), all of them at the positive part. Axis II (18 %) mainly addresses water temperature at the positive part and NO3− and also Ca2+ at the negative. Finally, pH appears related to the factor 3 (11 %, .Table 32.2). The group of variables grouped in the positive part of axis I includes parameters associated with the dissolution of halite and gypsum, which could be considered as indicators of residence time of the water throughout the system: The longer groundwater flows within the system, the greater water–rock interaction is possible and, therefore, higher water mineralization would be reached. Changes in temperature, positively related to factor 2, would be linked to the pushing and movement of water previously stored in the saturated zone during infiltration events. On the contrary, NO3−, plotted in the negative part of axis II, would have its origin in the soil–epikarst–unsaturated zone of the system drained by the spring and, therefore, it would be associated with recently infiltrated water. In factor plane I-II of the statistical units (. Fig. 32.3b), four water groups can be distinguished. Group 1, mainly constituted by cases located in the negative part of axis I, is made up of water samples taken at high water 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 2 11 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 Author Proof UNCORRECTED PROOF Layout: T2 Med_Medium Book ID: 430122_1_En Book ISBN: 978-3-319-45465-8 Chapter No.: 32 Date: 22 October 2016 18:53 Page: 339/343 339 32 . Table 32.1 Statistical summary of discharge rate, electrical conductivity (EC), water temperature (T), pH, and major components from water drained by Lower Anzur spring σ standard deviation, CV coefficient of variation (σ/mean), MAX maximum, MIN minimum, Q discharge rate Q (l/s) EC (mS/cm) T (°C) pH Alk (mg/l) Cl− (g/l) NO3− (mg/l) SO42− (g/l) Ca2+ (g/l) Mg2+ (g/l) Na+ (g/l) K+ (g/l) No. of data 56 59 59 47 59 59.0 59.0 59.0 59.0 59.0 59.0 59.0 MEAN 20.8 147.5 20.6 7.0 238.3 84.9 18.8 5.9 2.3 0.4 53.3 0.2 σ28.8 11.8 0.4 0.2 20.1 9.8 13.2 0.5 0.1 0.1 6.0 0.0 CV 1.39 0.08 0.02 0.03 0.08 0.12 0.70 0.08 0.06 0.19 0.11 0.14 MAX 91.9 163.5 21.3 7.8 268.4 105.1 53.7 7.1 2.6 0.5 61.0 0.3 MIN 0.7 120.9 19.4 6.6 189.4 63.4 0.0 4.9 2.0 0.2 39.5 0.2 MEDIAN 4.9 151.8 20.7 7.0 244.0 85.7 17.9 5.9 2.3 0.4 54.9 0.2 32.4 · Results Author Proof UNCORRECTED PROOF Layout: T2 Med_Medium Book ID: 430122_1_En Book ISBN: 978-3-319-45465-8 Chapter No.: 32 Date: 22 October 2016 18:53 Page: 340/343 32 340 Chapter 32 · Analysis of Natural Response and Hydrochemical Data … highest values of mineralization and water temperature, and the lowest NO3− contents. Finally, cases included in Group 3 are scattered between the others, being approximately centered on the null values of both axes. These samples had moderate grade of mineralization and correspond to intermediate water conditions, at the end of recharge events. conditions, showing lower values of EC, Cl−, Na+, Alk, Mg2+, K+, and SO42−, and higher NO3−contents. The majority of cases included in Group 4 appear in the positive part of axis II and negative part of axis I. These samples were taken in low flow conditions; they have higher contents of most of the chemical parameters and lower temperature values. Group 2 consists of samples with the . Fig. 32.2 Temporal evolution of discharge rate, EC, water temperature, pH, and principal chemical components of the water drained by Lower Anzur spring, compared to precipitation events 248 249 250 251 252 253 254 255 256 257 258 259 260 261 Author Proof UNCORRECTED PROOF Layout: T2 Med_Medium Book ID: 430122_1_En Book ISBN: 978-3-319-45465-8 Chapter No.: 32 Date: 22 October 2016 18:53 Page: 341/343 341 32 lag) in NO3− values (G1 in PCA, . Fig. 32.3b). This last component derives from the surface and the soil, and it has been traditionally considered as tracer of infiltration in carbonate aquifers (Perrin et al. 2003). In this case, the mean value of NO3− contents (18.8 mg/l) indicates the existence of pollution, which would be related to agriculture practices (extensive olive trees) that are carried out in the recharge area of Lower Anzur spring. When groundwater flow decreases, the water mixing became more homogeneous, temperature fell, and more saline water was drained by the spring (G3 in PCA, . Fig. 32.3b), as a consequence of changes in the dissolution rate of evaporite rocks: mainly halite which constitutes the materials in depth, in the saturated zone. Although gypsiferous materials are also present in the saturated zone, unlike halite (absent at surface, due to its high solubility), massive gypsum blocks and other smaller fragments constitute most of the evaporite rocks existing in the soil, epikarst, and unsaturated zone of the system (Andreo et al. 2016). Consequently, water inputs from concentrated infiltration (but also from slow diffuse infiltration through massive gypsum outcrops) can be stored in the soil and epikarst, and within conduits and fractures of the unsaturated zone, where gypsum dissolution occurs, contributing to the development of karst conduits. The excess in Ca2+ would provoke the oversaturation vs calcite and, hence, its precipitation, particularly in low water conditions when pH values were higher. This interpretation could explain the relatively low alkalinity in the water drained by the spring. Therefore, available information reflects the existence of highly developed karst conduits in the system 32.5 Discussion Joint analysis (qualitative and quantitative) of the data recorded in Lower Anzur spring, during different hydrological conditions, has allowed determining the hydrogeological functioning of the system drained by this brine spring, as well as understanding some hydrogeological processes that take place within it. Thus, quick variations in discharge rate following recharge events reflect the existence of a highly developed karst network, which makes possible fast flows and short transit time of water within the system, from concentrated points of infiltration at surface to the spring. This is in agreement with the observed changes in EC and in most of the related chemical components, whose values were generally lower after recharge episodes (. Fig. 32.2) and higher at low water conditions. However, the small increases in EC, Cl−, and Na+ recorded before dilutions (i.e., November 2015, . Fig. 32.2) could be caused by a “piston-flow” effect, where concentrated recharge through swallow holes pushed groundwater previously stored in the saturated zone, with greater residence time. This provoked mobilization, toward the spring, of water with higher salinity and temperature values (up to 3.5 °C higher than the mean annual air temperature), although with lower NO3− contents (G2 in PCA, . Fig. 32.3b), suggesting the existence of ascending flows from the deepest part of the saturated zone. Later, recently infiltrated water coming from the surface rapidly arrived to the spring via karst conduits, leading to the mixing of both types of water. Consequently, general dilution and slight decreases in water temperature took place, as well as a relative and rapid increase (1 day . Fig. 32.3 Principal component analysis (PCA) performed with the main physicochemical parameters and major water chemistry obtained from the Lower Anzur spring during the study period. Plots of the variables (a) and statistical units (b) 32.5 · Discussion 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 3 11 312 313 314 315 316 317 318 319 320 321 322 323 324 Author Proof