- © 2016 Association of Environmental & Engineering Geologists
Stable O and H isotopes and radioisotopes show that Río Bravo water recharged prior to the construction of Elephant Butte Dam is the principal source of groundwater in the Hueco Bolson aquifer beneath Ciudad Juárez. Mixing between pre-dam and post-dam river water occurs in samples from the aquifer beneath the Río Bravo floodplain, where finite tritium is also present. Along the flanks of the Sierra de Juárez, mixing occurs between pre-dam river water and water derived from the Sierra de Juárez. Carbon-14 content of groundwater generally decreases from northwest to southeast beneath the city; corrected residence times in city supply wells range from post-bomb to 3600 years. Sulfur isotopes in groundwater of pre-dam river water origin and anion ratios (Cl/Br and Cl/SO4) are consistent with mixing of solutes from native Hueco Bolsón groundwater with river water. Stable isotopes serve to identify mixing of 9 percent or more of local water into pre-dam river-derived water. Mixing of smaller amounts of saline Hueco Bolson groundwater into river-derived groundwater can best be identified using anion ratios. The improved understanding of groundwater origins, flow paths, and quality arising from this study is crucial information for future water-supply engineering in Ciudad Juárez and for elucidating preferred pathways for contaminant movement in groundwater.
Ciudad Juárez, Chihuahua, Mexico, has until recently depended mainly on groundwater pumped from the Hueco Bolsón aquifer beneath the city for a potable water supply. Rapid urban growth has stretched the local groundwater supply to its limits, leading to the development of a large cone of depression under the city (Hibbs et al., 1997) and to increasing salinity in older wells as the fresher, shallow groundwater has been pumped (del Hierro-Ochoa et al., 2004). In urban study sites such as Ciudad Juárez, an understanding of current and pre-development groundwater flow paths and of the ambient water-quality signature derived from pre-development recharge prepares groundwater scientists and engineers for a better understanding of contemporary contaminants, the origin of salinity, and water budget transfers.
In order to improve the understanding of groundwater origin, flow paths, and residence times in the part of the Hueco Bolsón aquifer beneath Ciudad Juárez, we undertook a study of stable isotopes (O, H, S), radioactive isotopes (tritium, carbon-14), and major anion concentrations in groundwater from the aquifer in Texas and Chihuahua (Hibbs et al., 2003; Druhan et al., 2008; and Eastoe et al., 2008, 2010). This article reports on the chemistry, including isotope chemistry, of groundwater from municipal wells in Ciudad Juárez, Chihuahua, and a well-nest beneath the Río Bravo floodplain in adjacent El Paso, TX. The aims of the study are to identify groundwater origins and flow paths in the Hueco Bolson aquifer beneath Ciudad Juárez, to identify mixtures of waters of different origins and the implications of mixing for salinity, to compare the capabilities of isotopes and anion ratios in this regard, and to quantify residence times of groundwater in the aquifer.
The study area includes most of the urban area of Ciudad Juárez (Figure 1). It is bounded to the north and east by the international border with the United States along the Río Bravo (known as the Rio Grande in the United States), and to the southwest by the Sierra de Juárez, a hard-rock mountain range. The city has grown in population from 567,365 in 1980 to 1,332,131 in 2010 (INEG y GEC, 2014). The elevation is 1134 m above sea level (masl) in the city center, rising to 1700 masl at the crest of the Sierra de Juárez. The climate is semiarid, with an annual average temperature of 25°C and an average annual rainfall of 250 mm (Climatological Laboratory, Universidad Autónoma de Ciudad Juárez, 2014). Precipitation is highly variable from year to year. It is largely confined to two seasons: a winter season (November to March) of infrequent frontal storms from the west, and a summer monsoon season (late June to September) during which convective storms drop moisture originating from the southeast or southwest. On average, summer rain (June–October) constitutes 67 percent of the mean annual total of 236 mm in nearby El Paso, TX (National Weather Service, 2014, http://www.srh.noaa.gov/epz/?n=elpaso_monthly_precip). Flow in the reach of the Río Bravo adjacent to Ciudad Juárez is at present intermittent, but was usually perennial prior to development. A large dam that impounds the river at Elephant Butte, NM, was completed in 1916, and flow in the river downstream of the dam has been managed for irrigation and municipal supply since that date. At present, there is insufficient river water to meet all demands (both in the United States and in Mexico) in the Hueco Bolsón.
The Hueco Bolsón is an asymmetric graben of the Basin and Range Province. Beneath Ciudad Juárez, the trough is more than 1,000 m deep, and it is filled with Neogene lacustrine, fluvial, eolian, and alluvial sediment (Figure 2). The following description uses stratigraphic terminology applied locally in Texas and New Mexico. Fluvial sand and gravel of the Camp Rice Formation, up to 300 m thick, overlie lacustrine silt and clay of the Fort Hancock Formation. The transition from the Fort Hancock Formation to the Camp Rice Formation reflects the hydrologic evolution of the basin from an internal drainage with playa lakes fed by the ancestral upper Río Bravo to a basin in which the ancestral Río Bravo was through-flowing about 1.8 Ma (Stuart and Willingham, 1984; Connell et al., 2005). Until about 0.67 Ma, the river entered the Hueco Bolsón through Fillmore Gap in New Mexico (41 km north of the present point of entry) and flowed south along the eastern flanks of the Franklin Mountains and the Sierra de Juárez. The Camp Rice Formation is thickest near the western boundary fault system of the Hueco Bolsón graben, where the paleo-river was localized by rapid subsidence; in this area, it comprises braided fluvial channel deposits (Stuart and Willingham, 1984). At present, the river flows south through the Mesilla Valley, west of the Franklin Mountains, and enters the Hueco Bolsón through the Narrows, a rock-cut gorge between the Franklin Mountains and the Sierra de Juárez.
The Hueco Bolsón aquifer is hosted by the Camp Rice and Fort Hancock Formations. Potable water is limited to the Camp Rice Formation. In Texas, the Fort Hancock Formation hosts saline water (Hibbs et al., 1997; Druhan et al., 2008). The Hueco Bolsón aquifer, with pre-development water levels >30 m below the surface, is distinguished from the Río Bravo alluvial aquifer (known as the Rio Grande alluvial aquifer) hosted by recent fluvial deposits of the active floodplain. Water in the alluvial aquifer is shallow, as close as 2 m below the surface, and the aquifer extends no more than 30 m below the surface.
Potable water supplied by the Junta Municipal de Agua y Sanamiento (JMAS) at the time of sampling for this study in 2003–2005 came entirely from 143 deep wells that ranged in depth to 251 m, drilled within the Camp Rice Formation, in the upper part of the Hueco Bolsón aquifer (IMIP, 2003). Continued pumping over many years has led to drawdown of water levels. Figure 3.10 of Hibbs et al. (1997) showed maximum drawdown of about 18 m between 1987 and 1993 at the center of a cone of depression 8 to 11 km across, beneath the southern half of the urban area, and extending from the Sierra de Juárez foothills to the Río Bravo floodplain. In 2014, the JMAS was drawing water from up to 173 wells in the Hueco Bolsón (M. C. Ezequiel Rascón, JMAS, pers. comm.). In the northwestern part of the urban area, groundwater levels rebounded a few meters between 1987 and 1993. Since 2010, pumping of groundwater from the Conejos-Médanos aquifer, west of the Sierra de Juárez, has allowed the JMAS to draw less water from the Hueco Bolsón aquifer, and this has led to a small recovery of static water levels in the center of the cone of depression (M. C. M. Herrera, JMAS, pers. comm.).
Two major geological units make up the aquifer system beneath Ciudad Juárez. A lower unit of Neogene age consists of clay, siltstone, sandstone, and conglomerate, includes fluvial facies, and is approximately equivalent to the Camp Rice Formation. An upper unit, of Quaternary age, contains alluvial fan and piedmont slope deposits; fluvial and eolian units are also present. In detailed view, many lenses or thin beds of clay and silt occur where the supply wells are located (Leos Rodriguez, 2004). Together, the units form a single, semi-confined aquifer system of moderate permeability, with a depth of more than 500 m. Transmissivity has been measured at 1.8 × 10−3 m2/s. Groundwater quality varies around 600 mg/L total dissolved solids (TDS) in the urban areas (INEG y GEC, 1999).
Isotope data (O and H stable isotopes and tritium) for groundwater reported to be from irrigation wells in the Juárez Valley, without specific location data, were published by Payne (1976). Payne noted that the data plotted on an evaporation trend, and that salinity and tritium were higher for more evaporated samples. He interpreted the data as indicating concurrent evaporation, salinity increase, and exposure to atmospheric tritium in irrigated fields. Streeter (1990) analyzed heavy metals in groundwater samples from central Ciudad Juárez in a study focusing on public heath, but also potentially provided tracers of groundwater movement since industrial development in the area. Elevated antimony concentrations were detected in well water from beneath the Río Bravo floodplain. Eastoe et al. (2008) showed that groundwater from most Ciudad Juárez supply wells had originated as surface water from the Río Bravo, and that most had infiltrated prior to the completion of the Elephant Butte Dam. Eastoe et al. (2010) presented stable O and H isotope data and MODFLOW modeling in a 300 m vertical section of the Hueco Bolsón aquifer beneath the Río Bravo between the cities of El Paso and Juárez. Within 5 km of the point of entry of the river, pre-dam river water could be identified as far as 300 m below the surface. Further downstream, the stable isotopes and modeling indicated a zone in which native saline Hueco Bolsón groundwater was being drawn southward beneath the river and toward the cone of depression beneath Ciudad Juárez.
SAMPLE COLLECTION AND METHODS
Groundwater samples were collected in 2003 and 2004 from public supply wells of the JMAS of Ciudad Juárez, from springs in the Sierra de Juárez, and from a U.S. Geological Survey (USGS) well nest and from Río Bravo surface water in 2003 and 2004. The JMAS well samples represent groundwater from the entire screened interval of each well, while the well-nest samples represent discrete screen intervals listed in Table 1. Precipitation samples were collected during 2004 and 2005 in a rain gauge at an elevation of 1120 masl in central Ciudad Juárez and combined to make seasonal aggregates weighted for amount. Isotope measurements (except for carbon-14) were performed at the Environmental Isotope Laboratory, University of Arizona, Tucson, AZ. Stable O, H, and C isotopes were measured on a Finnigan Delta S® dual-inlet mass spectrometer equipped with an automated CO2 equilibrator (for O) and an automated Cr-reduction furnace (for H). Stable S isotopes were measured on a Thermo Electron Delta Plus XL® continuous-flow mass spectrometer equipped with a Costech® elemental analyzer for preparation of SO2. Tritium was measured in a Quantulus 1220® Spectrometer by liquid scintillation counting after electrolytic enrichment. Carbon-14 was measured by accelerator mass spectrometry at the National Science Foundation–Arizona Accelerator Facility, University of Arizona. Analytical precisions (1σ) are 0.08‰ (O), 0.9‰ (H), 0.15‰ (C), and 0.15‰ (S). Detection limits are 0.6 tritium units (TU; 3H) and 0.2 percent modern carbon (pMC, 14C). Major anions were analyzed by ion chromatography at the Texas A&M Research and Extension Center at El Paso, Texas A&M AgriLife Research, The Texas A&M University System, and at the Department of Soil, Water and Environmental Sciences at the University of Arizona. The data are listed in Table 1, and sample sites are shown in Figure 3A.
ISOTOPES IN PRECIPITATION AND RECHARGE SEASONALITY
The δ18O and δD values of the seasonal aggregate rain samples plot close to the global meteoric water line (GMWL) of Craig (1963), and they are compared in Figure 4 with data for Tucson, AZ (400 km west of the study area), adjusted for altitude. Tucson, with an average ratio of summer (May–October) to winter precipitation of 1.7 (National Weather Service, 2015), is a better match for the study area (ratio 3.4, http://www.srh.noaa.gov/epz/?n=elpaso_monthly_precip) than the nearest International Atomic Energy Agency (IAEA) database site at Ciudad Chihuahua (ratio 7.9, http://www.chihuahua.climatemps.com/). In semiarid places, data for a single year are not good estimates of long-term mean δ18O and δD. Therefore it is instructive to compare the Ciudad Juárez data with the long-term data set from Tucson. Rain in Ciudad Juárez appears to show a similar strong seasonal variation in isotope composition. Groundwater discharging from springs in the Sierra de Juárez provides a longer-term estimate of the isotope composition of rainwater infiltrating the mountain block. The values of δ18O and δD are lower than those in 2004–2005 rain and plot close to winter means for Tucson, adjusted to altitudes appropriate for the present study area. Therefore winter recharge appears to predominate over summer in the Sierra de Juárez, the small amount of winter rain notwithstanding. Rainwater from Ciudad Juárez contained 5.1 to 9.3 TU in 2004–2005 (Table 1), i.e., higher than the post-bomb mean of 5.2 TU for Tucson Basin since 1992 (Eastoe et al., 2010).
Hydrogen and Oxygen Isotopes
Figure 5 shows O and H isotope data for the JMAS wells and the springs. Well locations are classified into three geographic zones: near the Sierra de Juárez, near the Río Bravo, and “central” for all other JMAS wells as indicated in Figure 3A. The data are compared with data clusters proposed in Eastoe et al. (2008) for the Hueco Bolsón regional aquifer, and they are plotted relative to the GMWL and the Río Bravo evaporation line (based on data from Phillips et al., 2003). Cluster A, mixtures of pre-dam and post-dam river water, is found in the Río Bravo floodplain (Figure 3B). Cluster B corresponds to pre-dam river water. It is found in all three geographic zones of Ciudad Juárez. Cluster C, resembling local winter precipitation recharged along the western boundary of the Hueco Bolsón (Eastoe et al., 2008), is present only in the springs of the Sierra de Juárez.
Samples from near the river are mixtures of pre-dam and post-dam river water, plotting close to the present-day evaporation trend of the river. Samples from near the Sierra de Juárez appear to form a linear array in Figure 5, indicating mixing between pre-dam river water having δ18O < −11.5‰ and mountain-derived water. Three other samples, from well 134 (near the river), from well 176 (towards the mountains), and from a Sierra de Juárez spring, also plot on the mixing trend. In the case of well 134, the cluster C water involved is more likely to be from the Hueco Bolsón north of the Río Bravo than from the Sierra de Juárez. The spring sample cannot be a natural mixture containing river-derived water because recharge in this case is at elevations higher than the river. An alternative explanation could be human-caused addition of water from the city supply.
Samples classified as central are almost all cluster B water with δ18O ranging from −11.9‰ to −10.2‰; these plot along an evaporation trend slightly below the present river trend. The spatial distribution of O isotopes is illustrated in Figure 3B. Among central samples, those in the southeastern part of the city have the lowest δ18O values.
The data are presented as plots of inverse SO4 concentration versus δ34S (Figure 6A and B) and as a map (Figure 3C). Values of δ34S in Río Bravo surface water (see Table 1) are negative upstream of, and positive downstream of, the Elephant Butte reservoir. Data for the central group of samples form a horizontal array with δ34S values from +2‰ to +8‰, with the exception of two outlying samples with negative δ34S values. Near-river samples form an array extending between present-day river water and the high-δ34S end of the central group. Samples from near the Sierra de Juárez have a δ34S range of +5.8‰ to +7.1‰, and those from springs in the mountains have δ34S values of +2.4‰ and +3.1‰. Deeper samples at the USGS well nest (with δ34S > +7‰) resemble the central group or groundwater from the Hueco Bolsón north of the Río Bravo (data of Druhan et al., 2008; shown in Figure 6B). Shallower samples resemble river water. Wells 3Z and 120 have δ34S values like those of present-day river water.
The river data, linked sequentially, show the effect of sulfate reduction in Elephant Butte reservoir, resulting in increased δ34S in the reservoir and downstream. Pre-dam river water entering the Hueco Bolsón, particularly during spring floods, would have resembled river water now found upstream of the Elephant Butte reservoir. Pre-dam river water entering the basin at times of low flow may have resembled the present-day river samples shown in Figure 6A, which were taken at times of very low flow and were dominated by sulfate added at the terminus of Mesilla Valley. Figure 6B shows mixing trends between both types of water and native Hueco Bolsón groundwater (data of Druhan et al., 2008). Most central samples plot along the pre-dam flood mixing trend, with dominant sulfate from native Hueco Bolsón groundwater.
The two central group samples with negative δ34S values were taken beneath the oldest part of Ciudad Juárez, and may represent contamination. Alternatively, they could represent sulfate from the river far upstream, concentrated by evaporation, or oxidation of sulfide in floodplain sediment. No pyritic sediment has been observed in the Hueco Bolsón. However, in Albuquerque Basin, Plummer et al. (2004) presented evidence of sulfide oxidation induced by pumping of oxygenated groundwater into previously anoxic, pyritic sediment.
Figure 3D shows that finite tritium values > 0.6 TU are limited to the Río Bravo floodplain, with the exception of well 53R, near the Narrows. Tritium levels comparable to those of contemporaneous river water (7 to 8 TU) were found only at the USGS well nest and at JMAS well 3Z (Table 1); elsewhere in the floodplain, tritium levels were no higher than 3 TU. Southwest of the floodplain, a few groundwater samples have finite tritium at levels < 0.6 TU (Table 1), indicating infiltration of a small amount of surface water, as occurs elsewhere in aquifers in Basin and Range alluvium, e.g., Tucson Basin (Eastoe et al., 2004). Figure 7A shows that water bearing no tritium has a δ18O range from −11.9‰ to −10.5‰. In water containing finite tritium, δ18O correlates generally with tritium content, substantiating the mixing relationship between post-dam and pre-dam river water, also indicated by O and H isotopes. River water from 2003–2004 had lower tritium than predicted by the groundwater mixing trend, consistent with higher tritium in Río Bravo surface water at the time of the bomb tritium pulse.
These data suggest that part of the groundwater present beneath the Río Bravo floodplain has infiltrated in the last few decades, but the samples for this study are mixtures of such water with pre-bomb water. Pre-bomb recharge supplied most of the groundwater southwest of the floodplain.
Figure 3E is a map of 14C distribution in Ciudad Juárez groundwater. The highest values (70–92 pMC) occur beneath the floodplain, with an exception in well 134. Outside the floodplain, values range from 14 to 70 pMC. Figure 7B shows that 14C and δ18O generally decrease together for values of pMC > 50. Below that value, δ18O increases with decreasing pMC.
Correction of 14C Data
The procedure for correction of 14C data was adapted from that in Clark and Fritz (1997, p. 210). In recharge water, 14C is commonly derived from soil gas, which contains CO2 derived from decomposing plant matter and has a 14C concentration similar to that of the contemporaneous atmosphere. This 14C is diluted when the infiltrating water containing dissolved CO2 dissolves rock calcite, as in Eq. 1:
If the rock calcite has a consistent δ13C value (δ13Ccarb) and contains no 14C, the added bicarbonate will have δ13C near δ13Ccarb, and a 14C content of 0 pMC. A dilution factor, q, for the 14C may be defined as in Eq. 2:
where the a14C terms are activities of 14C expressed, for instance, as pMC; rech is recharge; and DIC is dissolved inorganic carbon. Using an isotope balance based on δ13C data, q may also be expressed as in Eq. 3:
The value δ13Crech is commonly considered to derive from that of soil CO2, which in turn originates from decaying plant matter. However, if recharge occurs from the saturated bed of a large perennial stream (as the Río Bravo appears to have been prior to development), an alternative approach to estimating δ13Crech arises. Saturated alluvium contains little soil gas. In such a case, δ13Crech is δ13C of the river water, and a14Crech is the typical 14C content of the river water. This approach is used here, invoking the following assumptions.
The rock carbonate, in this case, calcite in the basin-fill sediments, has average δ13Ccarb values near –4‰ (Monger et al., 1998).
The value of δ13C of DIC in river water can be estimated by examining the range of δ13C in DIC of groundwater, as a function of pMC (Figure 8). The range −9.6‰ to −6.4‰ appears largely to be independent of pMC, and to indicate that dissolution of rock carbonate results in increases in δ13C in DIC of the infiltrating water. By this reasoning, the average δ13C of river water would be −9.6‰ or less.
Acceptable results cannot exceed the peak pMC of seasonally averaged atmospheric CO2 during the bomb spike of the 1960s, approximately 190 pMC (Burchuladze et al., 1989). The corrected pMC values in Table 1 include only one unreasonable value (196 pMC for a sample from the USGS well nest) when calculated using a value of −9.6‰ for δ13Crech. The use of lower values of δ13Crech generates higher values of corrected pMC. The calculation of bulk groundwater residence ages from the pMC values requires an estimate of a14Crech. No direct measurement was made on river water, but sample JL-49-21-324, from the top 11.5 m of alluvium next to the river, contained 110 pMC in 2003. C4 plant material in rural Arizona in 2002–2003 contained 108 pMC (Eastoe, unpublished data), for which the corresponding atmospheric pMC is 109.2 (Saliège and Fontes, 1984), i.e., close to the 110 pMC of recharge just before the sampling date. Therefore, observation suggests that the 14C level in river water is close to that in the atmosphere and was close to 100 pMC before the bomb 14C spike. Ages shown in Table 1 were calculated for a14Crech = 100 pMC and range from post-bomb near the river to 3600 years at the southern end of the study area.
The relationship between 14C and δ18O in samples with pMC < 50 can be explained by mixing between type B water and very old type C water with δ18O values of −10‰ or greater. This set of samples includes wells 142, 176, and 221, all of which are close to the Sierra de Juárez, and all of which show independent evidence of mixing of type B water with type C water from the mountain range (Figure 5). In the case of well 134, the mixing is most likely with type C water from north of the Rio Bravo; this is discussed further below. The source of ancient water for well 194 is unclear. Correction of 14C data is not possible for such mixed waters using the method and assumptions outlined here; therefore no corrected values are given for them in Table 1 and Figure 3E.
Only SO4, Cl, and Br are considered here, because of their generally conservative behavior. Figure 9 shows anion ratio behavior in the present-day river for average flow conditions in winter 2001. Both Cl/SO4 and Cl/Br increase downstream as a result of progressive evaporation, addition of deep basin groundwater, and leaching of salt from irrigated land (Phillips et al., 2003). Elephant Butte reservoir has a strong effect, particularly on Cl/SO4, which increases sharply at the reservoir and remains about constant between the dam and Ciudad Juárez. Under low-flow conditions, represented in a sample set (RG1 to RG8 in Table 1) taken in El Paso during this study, salinity and the anion ratios increase as a result of evaporation and the effect of additions of saline water at the southern terminus of the Mesilla Valley.
Concentrations of all three anions and their ratios vary widely in Ciudad Juárez groundwater (Table 1) and follow no clear spatial patterns (e.g., Figure 3F). Figure 10A and B show anion ratio variations as functions of Cl concentration. The ratios are commonly higher than those of 2003–2004 river water. Cl/Br ratios are >1,000 in many cases, consistent with halite as the ultimate source of the halide anions.
These findings suggest that mixing of river water and native Hueco Bolsón groundwater can be discussed in terms of three end members:
Pre-dam river water under flood conditions. As indicated by O, H, and S isotopes, this would have resembled present-day river water from near Albuquerque, NM. The composition used here (Cl 9.2, SO4 58, Br 0.037 mg/L) is typical of reaches just north of Albuquerque (Phillips et al., 2003; Plummer et al., 2004), with the location chosen in order to avoid perturbations due to the city.
Pre-dam or post-dam river water under low-flow conditions, strongly influenced by salty water emerging at the southern terminus of Mesilla Valley. Mixing between present-day river sample RG6, the sample with highest observed total dissolved solids (Cl 768, SO4 1,028, Br 1.08 mg/L), and Albuquerque river water yields curves labeled M1 in Figure 10A and B. Other present-day river samples (including river water from Elephant Butte reservoir) fall close to these curves.
Native Hueco Bolsón groundwater represented by the sample from JMAS well 134 (Cl 327, SO4 116, Br 0.19 mg/L). Mixing curves between river water sampled near Albuquerque and well JMAS 134 are labeled M2.
The anion data are consistent with derivation of solutes from flood-stage pre-dam river water mixed with low-flow stage river water and saline native Hueco Bolsón water. Several samples have elevated salinities (100–350 mg/L Cl) and anion ratios that plot near the M1 trends of Figure 10A and B, indicating that recharge under conditions of low flow in the river is important in addition to recharge of water when the river is in flood. Near-river samples with Cl > 265 mg/L are mixtures of low-flow river water and native Hueco Bolsón groundwater. Samples of lower salinity, both near the river and central, are mixtures of all three end members.
Some inconsistencies are observed between Figure 10A and B. Two samples (JMAS wells 130 and 180) that plot above trend M2 of Figure 10A might be explained by dissolution of halite, but the effect of surplus Cl is not evident in the Cl/SO4 ratios. Furthermore, large-scale dissolution of evaporite salts from sediment seems unlikely from the fluvial deposits beneath Ciudad Juárez. Samples JMAS 42R and 141 also plot in inconsistent fashion; in Figure 10A, they appear to be mixtures of river water and native groundwater, but in Figure 10B, they resemble low-flow river water.
Complementary Uses of Isotope and Ion Analyses
Isotope evidence indicates the Río Bravo as the principal source of recharge to the Hueco Bolsón aquifer beneath Ciudad Juárez, with pre-dam river water predominant, except in a few wells near the Río Bravo. Clusters C (native Hueco Bolsón groundwater) and B differ in δ18O and δD by maximum amounts of 2.5‰ and 28‰, respectively (Figure 5). For pairs of samples that differ in δ18O or δD by d‰, and for which the isotope analytical precision is σ‰ (one standard deviation), the difference has a relative error RE, calculated as
For σ = 0.08‰ (the analytical uncertainty for O isotopes), RE = 9 percent for d = 2.5‰ and 22% for d = 1‰. Similar RE values apply in the case of H isotopes. In the data set from Ciudad Juárez, O and H isotopes may therefore be incapable of detecting mixtures of less than 20 percent of type C to type B water. The anion concentration data for Ciudad Juárez groundwater traces smaller additions given the strong salinity contrast between the two water types.
As tracers, therefore, water isotopes and major anions perform complementary functions. Stable O and H isotopes make the clearest argument for the importance of pre-dam Río Bravo water in the potable water beneath Ciudad Juárez, for the presence of post-dam Río Bravo water in limited areas near the river, and for large-scale mixing with native Hueco Bolsón groundwater. The anion data provide the clearest evidence for the mixing of smaller volumes of high-salinity water, both native groundwater and river water, into the samples taken for this study. In the northeastern part of Ciudad Juárez, such mixing results from the southward encroachment of salty groundwater from the Hueco Bolsón regional aquifer on the north side of the river, as shown in Eastoe et al. (2010). In other cases, the mixing reflects the presence of native groundwater of high salinity beneath the potable groundwater throughout the city. Solutes from the native groundwater appear in the sample set where pumping of potable water has brought about upwelling of salty water, and where well screens extend to the original base of the freshwater zone.
Practical Results of this Research
That almost all of the potable groundwater underlying Ciudad Juárez is recharged from the Río Bravo means that previous interpretations of groundwater origin and flow are incorrect. There are technical, engineering, and policy issues that make these findings relevant to management and administration of the water resources of the Hueco Bolsón. The relatively young potable groundwater beneath most of Ciudad Juárez (> 45 pMC) compared to counterparts north of the river, where locally recharged potable groundwater typically contains less than 20 pMC, implies a more dynamic flow system and shorter residence times in the aquifer beneath Ciudad Juárez. Undoubtedly, the relatively dynamic recharge flux beneath Ciudad Juárez was due to flows from the Río Bravo that were intermittently very high during spring runoff from the highlands in Colorado and northern New Mexico. The annual cycle of scour and fill in the Río Bravo channel combined with stream avulsion probably created highly permeable streambed materials that united with high stream hydraulic head to allow ample volumes of recharge to the Hueco Bolsón aquifer beneath Ciudad Juárez in pre-development times. More recently, the channelization of the Río Bravo streambed, the grout lining of segments of the channel along the urban corridor, and the control of spring runoff by reservoirs in New Mexico have probably substantially reduced this important natural source of groundwater recharge. The over-pumping of the Hueco Bolsón, especially beneath Ciudad Juárez, has been exacerbated by human-caused changes to the natural recharge processes.
Natural recharge along the Río Bravo channel near the El Paso Narrows is now eclipsed by pumping-induced infiltration from unlined areas of the Río Bravo and from Río Bravo alluvium in areas below the Narrows. The natural recharge still operates to some extent to affect natural groundwater flow paths and fluxes. These flow paths converge on areas known to contain industrial and agricultural contaminants. Natural recharge, to the extent that it is allowed near the Narrows, may therefore act as a diluent of contaminated areas in the aquifer, including contemporary and legacy groundwater contaminant plumes.
Finally, our findings resolve a disagreement over the source of groundwater beneath Ciudad Juárez. According to previous interpretations, the groundwater originated as pre-development recharge of internal runoff along the mountain fronts north of the international border. By this account, the groundwater moved south across the border and flowed beneath Ciudad Juàrez before turning southeast and eventually upwelling and discharging along the Río Bravo. Furthermore, by this account the post-development cones of depression beneath El Paso in the United States capture water that would have flowed to Mexico under natural circumstances. Our study shows that such a hypothesis can no longer be defended.
Stable O and H isotopes show that potable groundwater in the Hueco Bolsón aquifer beneath Ciudad Juárez originated mostly as pre-dam (pre-1916) surface water from the Río Bravo. This water mixes with post-dam river near the Río Bravo, with water originating from the Sierra de Juárez locally around the base of the range, and with native Hueco Bolsón groundwater in the northeastern corner of the city.
Measurable tritium is largely limited to the Río Bravo floodplain, indicating residence times on the order of decades in the floodplain.
Southwest of the floodplain, 14C data indicate bulk groundwater residence times of up to 3600 years in samples consisting mainly of pre-dam river recharge. Corrections are not possible in samples for which the 14C and δ18O data together indicate mixing of pre-dam river water with native Hueco Bolsón groundwater.
Sulfur isotopes are consistent with widespread mixing of sulfate derived from native Hueco Bolsón groundwater into pre-dam and post-dam river recharge.
Anion rations (Cl/Br and Cl/SO4) indicate mixing between three end-member water types: pre-dam river water of low salinity recharged under high-flow conditions, pre-dam or post-dam river water of high salinity recharged under low-flow conditions, and high-salinity native Hueco Bolsón groundwater.
The anion concentration data complement the stable O and H isotope data and allow the identification of mixing of high-salinity water into low-salinity river recharge in amounts too small to detect with the isotopes.
In the sample set for this study, such mixing is human-caused and results either from the way wells are constructed, or from southward encroachment of salty water into the cone of depression beneath Ciudad Juárez.
The findings described in this paper have implications for the long-term management of the groundwater in the Hueco Bolsón aquifer beneath Ciudad Juárez.
This work was funded by Sustainability of SemiArid Hydrology and Riparian Areas (SAHRA) under the Science and Technology Centers (STC) Program of the National Science Foundation, Agreement No. EAR9876800, and by Center for Environmental Analysis, Centers for Research Excellence in Science and Technology (CEA-CREST) under National Science Foundation Cooperative Agreement No. HRD0317772. The authors acknowledge the kind support of the Junta Municipal de Agua y Sanamiento de Ciudad Juárez, who provided access to their wells.