Introduction

The availability of groundwater resources for human use depends not only on the quantity and quality of water in residence at any given time, but also on the rate that water replenishes an aquifer (Healy et al. 2007). Dating of groundwater using environmental tracers can be used for both evaluating the rate at which the groundwater is replenished and to better understand flow and transport processes (Cook and Bohlke 2000). In mountainous terrain, the use of noble-gas thermometry can provide additional constraints on recharge zone elevation (Aeschbach-Hertig et al. 1999; Manning and Solomon 2003; Gardner and Heilweil 2014). Most groundwater samples from both springs and production wells with long screened intervals represent mixtures of multiple flow paths. The combination of tracers reflecting different timescales is both a powerful and necessary approach when studying residence times of groundwater following varying pathways and having a wide range of travel times (Plummer et al. 1993; Newman at al. 2010). Combining the measurements of two or more different tracers with very different dating ranges can make the presence of such mixtures obvious by showing discrepant ages (IAEA 2013). This has been particularly useful in studies for evaluating binary mixtures of recent (<60 years) and older (>1,000 years) groundwater in fractured or karstic bedrock aquifers (Yager et al. 2013; Heilweil et al. 2014).

The Rio Actopan Basin extends over 2,000 km2 in the central part of the Veracruz State, Mexico. The center of the basin is located 19°32′5.11″ latitude north and 96°41′39.89″ longitude west; between 2,110,000 and 2,180,000 UTM latitude N (Fig. 1). Its shape is elongated in an orientation northwest to southeast (approximately 90 km long) and narrower from northeast to southwest (approximately 20 km wide). It is bounded on the east by the Gulf of Mexico and to the west by the Sierra Madre Oriental. The topographic relief is variable and characterized by two main areas: a lower-elevation lower-gradient sector along the coast, and a higher-elevation steeper sector situated in the Sierra Madre Oriental Mountains (Tejeda et al. 1989). The altitudinal gradient is extremely large in this short distance, rising from sea level at the Port of Veracruz to 4,220 m above sea level (masl) at the peak of the extinct Cofre de Perote volcano in less than 70 km. As expected, the stable-isotope (δ2H, δ18O) gradient and precipitation patterns in the study area are greatly influenced by these topographic features (Pérez Quezadas et al. 2015).

Fig. 1
figure 1

Simplified geological map and cross-section (A–A′) of the Rio Actopan Basin, indicating the location of the groundwater sampling sites. Cross-section B–B′ is shown in Fig. 5

Full size image

Due to the relative abundance of surface water in the Rio Actopan Basin, previous studies on groundwater dynamics and (or) water quality aspects in the study area have been limited. A CONAGUA (2002) study found that the unconsolidated alluvial aquifer located in the coastal area in the eastern part of the basin was partially recharged by the Actopan River, as indicated by observed streamflow losses at the hydrographic stations of Ídolos, Actopan and Naranjillo. A study of several springs in the vicinity of the Emiliano Zapata municipality showed that the occurrence of springs is controlled by the lithological contrast of the different geological units (Salas-Ortega 2010).

The objective of this study was to identify recharge zones, directions of groundwater flow, and groundwater ages (transit time) in the Río Actopan Basin using water chemistry and environmental tracers. Such information would provide a better understanding of the processes controlling regional groundwater occurrence, pathways, and interactions with surface water. In particular, the data collection and interpretation in this study were focused on testing the hypotheses that (1) the groundwater within the watershed is relatively young, indicating an active flow system in highly permeable aquifers driven by large hydraulic gradients; and (2) the majority of recharge to the groundwater system occurs in the upland parts of the watershed and moves towards the coast.

Study area

Climate and precipitation

The climate within the study area varies from “tropical wet-and-dry” (average annual temperature > 18 °C and more than 150 mm precipitation during the summer months) at the Port of Veracruz to “humid subtropical” (average temperature > 10 °C during the warmest months and more than 1,200 mm of rain during the wet season) in the Sierra Madre Oriental mountains (Geo-Mexico 2014). The maximum monthly air temperature occurs during May and June. Mean annual air temperature at 3,100 masl is 9.6 °C; at 1,400 masl is 18.3 °C and at sea level is 25.6 °C (Hernández López 2012). The seasonal variation of temperature at three weather stations is 4 °C at 3,000 masl, 8 °C at 1,400 masl, and 6 °C at sea level. The primary controls on the local climate within the study area are both topography and the atmospheric circulation systems prevailing during the different seasons (Mosiño and García 1973). These factors result in the occurrence of a rainy season during summer and autumn. Although the study area is located within the tropics, the precipitation regime is also affected by mid-latitude phenomena such as extratropical cyclones (Jones et al. 2003). About 80 % of total annual rainfall occurs between June and October. It is also interesting to note that the monthly amount of precipitation during the rainy season is similar for both high- and low-elevation stations (Tembladeras and Port of Veracruz, respectively). While rainfall during the months of July and August exceeds 300 mm at both these stations, precipitation during the same months at Xalapa is less than 200 mm. It is theorized that a local vertical circulation pattern caused by a barometric pressure anomaly causes a microclimate that brings drier air to the Xalapa area during the summer months (Pérez Quezadas et al. 2015).

There is also a large inter-annual variability in total rainfall within the study area. Total annual rainfall varied from 1,206 to 2,076 mm at the station in Global Network of Isotopes in Precipitation (GNIP), during 1962–1988. This station was located at coordinates 19°12′00″ latitude N and 96°07′48″ longitude west in Port of Veracruz. This variation in the precipitation is associated with global phenomena such as El Niño - Southern Oscillation (ENSO; Rasmusson and Wallace 1983), and La Niña (the negative phase), resulting in excess rain in some parts of the state of Veracruz during the summer (Magaña et al. 2003). The Pacific Decadal Oscillation (PDO; Mantua et al. 1997), in its negative phase, can also induce excessive rain in southern portion of the state of Veracruz (Méndez González et al. 2010). Lastly, the Atlantic Multidecadal Oscillation (AMO; Sutton and Hodson 2005) can increase the frequency of hurricanes, causing excessive rainfall in the state of Veracruz (Vázquez 2007).

Goldsmith et al. (2011) used environmental isotopes to carry out a water balance study in the tropical forest of the central part of the state of Veracruz. They found a marked seasonality variation in the isotopic composition of precipitation of about 130 ‰ for δ2H and 18 ‰ for δ18O, with more isotopically depleted precipitation during the rainy season. This observation is consistent with the historical isotope records of the Port of Veracruz GNIP station, which was operated from 1962 to 1987. Recently, Pérez Quezadas et al. (2015) found an δ18O isotopic gradient in precipitation within the Rio Actopan Basin of −2.1 ‰ km−1 from the sea level to about 4,200 masl. That study also identified an inverse relationship between monthly precipitation amount and δ18O, with rainfall during the wet summer season (June–October) depleted by up to −5 ‰ compared with rainfall during the dry winter season, based on historical (1962–1988) data from the Port of Veracruz.

Geology of the Rio Actopan Basin

The geological framework of the Rio Actopan Basin has been recently delineated by the Servicio Geológico Mexicano (2007a, b; 2010). The oldest rocks cropping out in the basin are of Mesozoic and Tertiary age, overlain by alkaline volcanic rocks of the Miocene (Campo Volcánico Palma Sola), followed by volcanic rocks of the Pliocene, and the more recent Quaternary magmatic rocks of the Trans-Mexican Volcanic Belt (Fig. 1).

The pre-volcanic basement of the area is formed by limestones and other carbonate rocks of the Upper Mesozoic (Cretaceous), locally known as the Guzmantla Fm (Viniegra-Osorio 1965). These formations have been assigned to the Turonian age and were deposited in a continental platform depositional environment. In the southern part of the study area, marly rocks, lutites and carbonaceous mudstones have been reported and dated as Late Cretaceous Campanian-Maastrichtian (Servicio Geológico Mexicano 2010).

In the northern part of the basin, the dominant geological feature is Palma Sola volcanic field, which is part of the Western Alkaline Province. In this region, andesitic rocks of the middle Miocene are overlain by pyroclastic flows that have been dated by Negendank et al. (1985) and López-Infanzón (1991) as Miocene. The northwestern part of the basin is dominated by upper Pliocene to early Pleistocene andesite and andesitic tuffs. Outcrops of these andesitic tuffs have a typical thickness of several tens of meters and are located at elevations ranging from 1,000 to 3,000 masl. In the center of the basin, (northeast of the city of Xalapa) there are outcrops of late Pliocene rhyolitic ignimbrite (Morales-Barrera 2009). This is overlain by Pliocene and Pleistocene andesitic-andesite tuffs, which outcrop further north.

A major feature of the Quaternary deposits in the Rio Actopan Basin is an extensive formation of massive, recrystallized Pleistocene travertine. This formation is discordantly positioned over the Cretaceous limestones. In the west and central parts of the basin, the most prominent and geologically recent feature is the monogenetic volcanic field of Xalapa, which is formed by 59 volcanic structures extending over 2,400 km2 (Fig. 1). Holocene alluvial deposits and colluvial sediments along the main river channels are the most recent formations in the basin (Rodríguez et al. 2010).

Hydrogeological setting

Synthesizing published geological information, five main hydrogeologic units are proposed for the Rio Actopan Basin groundwater system: (1) Quaternary alluvial deposits up to about 300 m in thickness (CONAGUA 2002), (2) Quaternary volcanics (including the volcanic field of Xalapa, with variable thickness up to tens of meters; Rodríguez et al. 2010), (3) Pliocene ignimbrites about 80 m in thickness (Morales-Barrera 2009), (4) Pliocene and Pleistocene andesitic tuffs and andesites up to about 500 m in thickness (Gómez 2002), and (5) a unit composed of both Pleistocene travertine (several meters in thickness) and Cretaceous limestone (more than 500 m in thickness; Santamaría-Orozco et al. 1990).

Most groundwater extraction occurs in the coastal area, usually from wells with total depths less than 150 m. These high-yielding wells are screened in the unconfined Quaternary alluvial aquifer comprised of gravels and sands; nearest the coast, these deposits also contain clays. The transmissivity of this unit ranges from 1 to 95 × 10−3 m2/s (CONAGUA 2002). In contrast to wells located primarily at lower elevations, most of the spring discharge emanates higher up in the basin from the Quaternary volcanics, which have high permeability resulting from both fractures and lava-flow structures. These springs have flow rates of up to several m3/s. In the center of the basin, especially in areas with abundant andesitic tuffs, some springs likely discharge from perched aquifers. The Pliocene ignimbrites are less fractured, resulting in springs with smaller flow rates, often less than 5 L/s. The deepest well in the basin (V29, 200 m depth) is located in this formation. No springs with significant flow rates have been identified in the areas dominated by the Pliocene and Pleistocene andesitic formations. Major springs V28 and V20 (discharging up to 1 m3/s), along with well V54, are located in the combined Pleistocene travertine and Cretaceous limestone formations (Table 1). The limestone has differing amounts of karstification, caused by variable fracturing and dissolution features.

Table 1 Site information for groundwater sampling during 2013 in the Rio Actopan Basin, Veracruz, Mexico. Analysis Types: C/I indicates chemistry and stable isotope (δ2H, δ18O) results presented in Table S1 of the ESM; T/C indicates tritium (3H) and carbon isotope (14C, δ13C) results presented in Table S2 of the ESM; CFC indicates CFC-11, CFC-12, and CFC-113 results also presented in Table S2 of the ESM; NG indicates noble gas (He, Ne, Ar, Kr, Xe) results presented in Table S3 of the ESM
Full size table

Major tectonic structures (fault lineaments) within the Rio Actopan Basin are aligned along the SE–NW direction (Fig. 2). Many springs are located either along these structures or in areas of steep hydraulic gradients, usually discharging into rivers and streams. In some cases, major springs are located in the headwaters of smaller catchments, contributing to the formation of perennial stream flow. The predominant groundwater flow direction in the Quaternary alluvial aquifer near the coast is from west to east, as shown by the potentiometric contours of Fig. 2. These contours are based on water-level data provided by the Comisión Nacional del Agua.

Fig. 2
figure 2

Spatial relationship observed between major tectonic structures (lineaments) and the location of springs and wells in the Río Actopan Basin. The location of springs is partly controlled by the presence of these lineaments and locally steep topographic slopes (>30°). Hydraulic groundwater-level contours near the coast show the predominant west–east direction of groundwater flow, controlled by both the general topography gradient and well production near José Cardel. Modified from Morales-Barrera (2009)

Full size image

Methods

Hydrochemical and stable isotope contents were determined from a number of samples collected from springs, wells, rivers and sea water as part of an initial reconnaissance study during 2012. Based on the results of the previous study, 15 springs and 5 wells were selected for hydrochemistry and stable isotopes (Table 1). From this group, a subset of seven springs and five wells were selected for analysis of tritium (3H), carbon-14 (14C) and carbon-13 (13C), and a subset of five springs and two wells were selected for chlorofluorocarbon (CFC) analysis (CFC-11, CFC-12, CFC-113 and noble gases (He, Ne, Ar, Kr, Xe), but the noble-gas samples from three sites were contaminated. The selection criterion was based on sampling groundwater discharging from different hydrogeological units. In addition, 14 precipitation samples were collected during 2007 and 2008 for 3H analysis.

Water samples were directly collected from the springs and wells at the first access point (before the addition of chlorine). These waters are commonly used for domestic supply and agricultural uses; the wells are typically pumped on a daily basis. Although the springs are perennial, the flow rate usually decreases up to 50 % during the dry season. Peristaltic pumps were used for the collection of spring samples. Samples for cation analyses were filtered using 0.45-μm filters and acidified with concentrated HNO3. Temperature, pH, alkalinity and electrical conductivity were measured at all sampling points using standard field methods.

Chemical analyses were carried out at the Laboratorio de Química Ambiental, Universidad Nacional Autónoma de Mexico (UNAM), by ICP-OES for cations and ion chromatography for anions. δ18O and δ2H isotopic ratios, expressed as per mil deviations versus VSMOW, were determined at the Laboratorio de Espectrometría de Isótopos Estables del Instituto de Geología, UNAM, using a Gas Bench II coupled to a mass spectrometer stable isotope Thermo Finnigan MAT 253. Final δ18O and δ2H results were obtained after normalization in the VSMOW- SLAP scale, following the procedure outlined by Coplen (1988) and Wegner and Brand (2001). Radiocarbon and δ13C analyses of total dissolved inorganic carbon (TDIC) were carried out at the Groningen University, Centre for Isotope Research (The Netherlands), by accelerator mass spectrometry (AMS). Measurement accuracy by AMS is typically in the range of 0.2–0.5 percent modern carbon (pmC) for 14C and about ±1.0 ‰ for δ13C. Tritium analyses were carried out by HYDROSYS, Hungary, and the Isotope Hydrology Laboratory of the International Atomic Energy Agency (IHL-IAEA), Vienna (Austria), using electrolytic enrichment and a low-level liquid scintillation analyser, with a detection limit of 0.1 tritium units (TU) and uncertainty of ± 0.1–0.2 TU.

Analysis of dissolved noble gases from water samples was done at IHL-IAEA using dedicated static mass spectrometers for the determination of noble gas concentration and isotope abundances of noble gases dissolved in groundwater samples. Samples were collected in the field using standard 80-cm long × 1-cm diameter copper tube samplers. CFCs were collected in specifically designed brown glass bottles and also analyzed at IHL-IAEA in triplicate samples by a dedicated gas chromatographic system.

Results

Field parameters and major-ion chemistry

Table S1 of the electronic supplementary material (ESM) summarizes the results of field parameters and major-ion chemistry from samples collected from wells and springs in April and May 2013. Water temperature in springs ranged from 13 to 38 °C. The coolest temperature was measured in Spring V52, located at an elevation of 2,613 masl. The warmest temperature was measured in the thermal spring V28, located at 200 masl, indicating the existence of deep groundwater circulation. Most wells in the coastal plain had temperatures ranging from 23 to 30 °C, similar to the long-term mean annual air temperature in the area. Five hydrogeochemical samples have charge balances above 10 % which could be due to (1) analytical errors in the concentration determination or (2) ionic species having significant concentration levels that were not included in the analysis (Freeze and Cherry 1979).

The lowest specific conductance (SC) values are found in springs located at higher elevations (e.g. 52 μS/cm at V52) and the highest value (1,550 μS/cm) is from the thermal spring V28 closer to the coast, likely indicative of the degree of rock/water interaction due to varying flowpaths and geology. SC in the wells in the coastal plain ranges from 142 μS/cm in V19 located 10 km southeast of Xalapa to 890 μS/cm in V36, located at a short distance from the Actopan River and the coast. A similar trend is observed with chloride, ranging from less than 2 mg/L at high-elevation springs such as V51, V52, and V53 to 51 mg/L at V28. Cl concentrations in wells range from 2.6 mg/L in V19 (southeast of Xalapa) to 138 mg/L in V36 (along the coast). Nitrate concentrations in water from most wells and springs are below the 10 mg/L drinking-water standard, but two springs (V15 and V54) and one well (V37) exceed this with concentrations up to 44 mg/L; all three locations are near or downgradient of urban centers.

Piper and Stiff diagrams (Fig. 3) show that most groundwaters are classified as bicarbonate type (Na, K, Mg, or Ca), except thermal spring V28, which can be classified as calcium-sulfate type, and well V36, which has a sodium-chloride component. Spring V28 emanates from an outcropping limestone formation, but likely represents discharge from the deep regional groundwater flow based on its high temperature and sulphate concentration (700 mg/L). Well V36, located near the coast, seems to be a mixture of both calcium-bicarbonate and sodium-chloride types, perhaps indicating saltwater intrusion.

Fig. 3
figure 3

a Map showing the spatial distribution of the hydrochemcial facies of groundwater in the Rio Actopan Basin using Stiff diagrams. b Piper diagram for the samples collected in 2013

Full size image

Stable isotopes

Stable-isotope analyses of springs and wells sampled in 2013 indicate a meteoric source of groundwater within the Rio Actopan Basin. Figure 4b (δ18O versus δ2H) illustrates that these samples plot along both the global meteoric water line (GMWL), δ2H = 8 δ18O + 10 (Craig 1961) and the local meteoric water line (LMWL), δ2H = 7.44 δ18O + 7.3 (Pérez Quezadas et al. 2015). The δ18O values of the 15 springs range from −7.1 to −9.6 ‰; δ2H values range from −45.3 to −64.4 ‰ (Table S1 of the ESM). The isotopic values are more enriched for the five wells in the coastal plain, ranging from −5.5 to −7.3 ‰ for δ18O and −36.5 to −48.5 ‰ for δ2H (Table S1 of the ESM).

Fig. 4
figure 4

a δ18O contents of springs and wells in the Rio Actopan Basin compared to the isotopic elevation gradient of precipitation estimated by Pérez Quezadas et al. (2015). b The relationship between δ18O and δ2H values of spring and well samples and their position with respect to the global and local meteoric water lines (GMWL, LMWL, respectively)

Full size image

Figure 4 shows the relation between δ18O and elevation of the site sampled, which is generally consistent with a previously published isotopic elevation gradient for precipitation in the Rio Actopan Basin (Pérez Quezadas et al. 2015). All of the springs above 1,400 masl, however, are slightly enriched compared with precipitation. This indicates some preferential recharge of isotopically enriched precipitation that is characteristic of the drier winter months, perhaps because of reduced evapotranspiration during this cooler period. Conversely, many of the springs and wells at elevations below 1,200 masl are slightly depleted compared with precipitation; this may be explained by recharge to these sites occurring at higher elevations in the watershed.

Carbon-14

Measured 14C from 12 springs and wells ranged from about 12 to 96 pmC; δ13C ranged from −3 to −19 ‰ VPDB (Table S2 of the ESM). Figure 5 shows 14C (along with tritium) at several springs and wells along a NW–SE cross-section. Typical values are greater than 80 pmC in the upper part of the basin (spring V51) and less than 20 pmC in the coastal plain (well V54 and spring V28, both located in limestone formations). The Fontes and Garnier (1979) model was used to calculate the age using both 14C and δ13C. This is a hybrid of the Ingerson and Pearson (1964) model and Tamers (1975) model, combining both chemical and isotopic data to correct for exchange reaction effects on 14C activity. It considers a two-stage evolution of recharge waters, using δ13C to account for isotopic exchange with CO2 in the unsaturated zone and isotopic exchange with carbonate rocks in the saturated zone. A chemical mass balance is performed similar to Tamers, with a provision for base exchange, to define the mass of carbon dissolved from inorganic sources (Plummer et al. 1994).

Fig. 5
figure 5

Cross section along a NW–SE line showing the tritium and radiocarbon activities, expressed as tritium units (TU) and pmC, respectively. See Fig. 1 for location of the cross-section

Full size image

It is assumed that the corrected initial 14C of infiltrating precipitation was 100 pmC (no bomb-pulse water) because measured values were less than or equal to 96 pmC, even for those having modern (post 1950s recharge) concentration of other environment tracers. Regarding isotopic exchange in the saturated zone, no published information was available for δ13C of aquifer carbonates within the study area. A range of values from −2.0 to +2.0 % was tested, with values of +1.5 ‰ producing 14C corrected ages most consistent with high tritium concentration (>0.5 TU) and the presence of CFCs at four springs (V15, V09, V51, and V08). For isotopic exchange in the vadose zone, soil zone δ13Cco2 was estimated by assuming that the predominant natural vegetation in the upland areas contributing to all of the springs and two higher elevation wells (V54 and V19) is primarily C3 plants (respiring more isotopically depleted δ13CCO2, with end-member values ranging from about −17 to −30 ‰). The δ13Cco2 of the four young springs was individually adjusted within a range of about −17 to −30 ‰ until the corrected 14C ages were modern (less than 100 years). Then the average δ13C value for these four sites (−19.4 ‰) was used for three other springs (V28, V05, V18) and the two higher elevation wells (V19, V54), which all have low tritium (<0.3 TU). Similar to the group of springs with modern water, the group of lower elevation wells (V33, V36 and V37) all have high tritium (>0.5 TU) and, therefore, individual δ13C values for each site were determined so that the corrected 14C ages were modern (less than 100 years). The resulting “calibrated” δ13C values were less depleted for these three wells (−11.7 to −15.7 ‰), consistent with the dominance of C4 plants (δ13CCO2 of −10 to –16 ‰) such as sugar cane in agricultural areas in the lower part of the watershed. With these assumptions, corrected piston-flow 14C ages of the three springs and two wells with low tritium (<0.3 TU) are all Holocene (about 5,000 years for spring V28, about 3,000 years for spring V05; about 500 years for spring V18; about 11,000 years for well V19; and about 7,000 years for well V54).

Tritium

Measured 3H from 12 springs and wells ranged from about 0 to 1.9 TU (Table S2 of the ESM; Fig. 5). 3H from precipitation samples collected in 2007 and 2008 ranged from 1.4 to 2.5 TU, with a mean value of 1.8 TU (n = 14). Comparison of groundwater concentration with corrected 14C age (Fig. 6) shows that the 3H results are consistent with 14C results, with two distinct groups: (1) older waters having 14C ages of about 3,000–11,000 years (sites V05, V19, V28, V54) with 3H values of less than 0.3 TU; and (2) younger waters having 14C ages of 30–500 years (sites V08, V09, V15, V18, V33, V36, V37, V51) with 3H values of 1.9–0.5 TU.

Fig. 6
figure 6

3H compared with corrected 14C age (log scale) for groundwater sites in the Río Actopan Basin. Maximum analytical uncertainty in the tritium activity is ± 0.2 TU

Full size image

Noble gases

Noble gas concentrations from three springs and one well (Table S3 of the ESM) were analyzed to evaluate recharge conditions, estimate excess air, and for dating using the tritium/helium method (3H/3He) and terrigenic helium production (4Heterr). Both the unfractionated air (UA; Stute and Schlosser 2000) and closed equilibrium (CE; Aeschbach-Hertig et al. 1999) models were used to fit concentrations of Ne, Ar, Kr, and Xe to theoretical values. These models evaluate recharge temperature governed by Henry’s Law solubility (plus excess air) of dissolved gases entering the water table in the recharge zone using the χ 2 goodness of fit test with the excel-based “CalcAge” program (provided by Dr. Kip Solomon of the University of Utah, USA). Results of the simpler UA model are reported for this study because it produced good fits of less than a χ 2 probability threshold (P < 5 %) of 6.0 based on 2 degrees of freedom assuming four measured gases and the two parameters of recharge temperature and excess air.

While stable isotopes (δ18O, δ2H) indicate the source elevation of precipitation, they are unable to differentiate between recharge occurring at higher elevation versus surface-water runoff and recharge at lower elevation. Because the solubilities of noble gases are temperature-dependent, they can record the water temperature at which infiltration enters the aquifer, providing additional information on the elevation of the recharge zone. Because of the co-dependence of recharge elevation and recharge temperature, noble-gas recharge temperatures were calculated for both the elevation of sampling site (T Rmax) and the maximum possible elevation of the water table within the watershed of 2,500 masl (T Rmin). The maximum water-table elevation is based on the elevation of the highest springs within the watershed. The most-probable recharge elevation for each site was then determined as the intersection of the (T Rmin)/(T Rmax) line with the groundwater temperature lapse rate (Manning and Solomon 2003). Recharge temperatures ranged from about 11–20 °C (Fig. 7; Table S3 of the ESM). These recharge temperatures are consistent with high-elevation (>1,100 m) recharge. Also, excess neon (ΔNe), an indicator of trapped air caused by water-table fluctuations in the recharge zone, ranged from 0 to about 30 %, indicating small to moderate amounts of additional gas dissolution during recharge.

Fig. 7
figure 7

Comparison of elevation to measured spring temperatures and calculated noble-gas recharge temperatures (T R ). The groundwater temperature lapse rate was calculated from measured spring temperatures for the sites not influenced by warm geothermal circulation. The most probable T R is the interaction of the groundwater temperature lapse and the line between T Rmin and T Rmax for each site

Full size image

The groundwater temperature lapse rate (about 6 °C per km) in this study is higher than one previously reported based on spring temperatures in the Wasatch Mountains of Utah, USA (Manning and Solomon 2003) but lower than those reported for other tropical areas, including the Cape Verde Islands, West Africa (Heilweil et al. 2012); it is within the range of atmospheric temperature lapse rates for the Cascade Mountains of Washington, USA (Minder et al. 2010). While groundwater lapse rates measured from springs do not account for the conversion of gravitational potential energy to heat through viscous dissipation of about 2 °C per km (Manga and Kirchner 2004), the lapse rate for the Rio Actopan watershed is consistent with the mean annual air temperature at the port of Veracruz of about 25.6 °C (Hernández López 2012).

3H/3Hetrit and 4Heter

The noble gas analyses also provided measurements of total helium (He) and helium isotope ratios (3He/4He; Table S3 of the ESM). Using the equilibrium (plus excess air) solubility of the noble gases (Ne, Ar, Kr, Xe), along with measured He and 3He/4He ratio, terrigenic helium (4Heter) was calculated using “CalcAge” for the four sites with noble-gas data. 4Heter accumulation occurs with time as 4He is produced as a byproduct of uranium and thorium decay in the aquifer matrix. Although this is only a qualitative indicator of groundwater ages between 103 and 106 years, older groundwater in an aquifer having uniform 4He production rates would typically have higher 4Heter than younger groundwater (Solomon 2000). Calculations made for this study assume a terrigenic 3He/4He ratio of 2.8 × 10−8, a value commonly used for crustal helium production during U/Th decay (Solomon 2000). While the actual ratio may be lower in the Río Actopan study area because of volcanism, there are no published data to support this. Using 3He/4He ratio of 2.8 × 10−8, 4Heter concentrations at the four sites ranged from 6.7 × 10−11 to 4.9 × 10−9 ccSTP/g (Table S3 of the ESM), relatively small values that would be even less if a larger 3He/4He ratio were used. The lower two values from V09 and V51 (2.2 × 10−10 and 6.7 × 10−11 cc(STP)/g, respectively) are insignificant (essentially zero). All four sites are assumed to contain primarily young water based on their corrected 14C ages of only 30 to 60 years. Sites having older 14C ages in the Rio Actopan Basin such as at V19, V28, and V54, may contain higher amounts of terrigenic helium, but were not sampled for noble gases during this study.

Chlorofluorocarbons

Using the estimated recharge temperature, elevation, and excess air calculated from noble gases, measured chlorofluorocarbon (CFC) concentrations were converted to equivalent atmospheric concentrations (pptv) for age determination using the US Geological Survey (USGS) Reston Groundwater Dating Lab’s CFC age calculator (Excel). The excess air used for four sites with noble gases (V09, V15, V36, V51) ranged from 0.02 to 2.6 cc/kg. The average value of 1 cc/kg was used for the other three sites without noble gases (V08, V19, V28). Calculated piston-flow ages were not very sensitive to the amount of excess air. For example, at the three sites without noble-gas derived excess air, an order of magnitude increase (to 10 cc/kg) changed the age by less than 0.5 years. Piston-flow years range from pre-1950s to the late 1980s (Table S2 of the ESM). The tracer-tracer plot of CFC-12 versus CFC-113 (Fig. 8) confirms piston-flow recharge conditions for springs V08 and V09, which were likely recharged in the late 1980s. Slightly older piston-flow ages for V08 and V09 determined from CFC-11, and for V15 determined from CFC-113 may be due to preferential bacterial anaerobic sulfate-reducing biodegradation of CFC-11 and CFC-113 (Plummer and Busenberg 2000). The tracer-tracer plot indicates that well V36 and spring V051 contain mixtures of young and older water. Assuming a binary mixing model, V36 contains about 25 % modern water recharged between 1985 and 1995 and 75 % recharge occurring more than 50 years ago. V51 contains about 80 % recent recharge (2013) and 20 % recharge occurring more than 50 years old. Young corrected 14C ages for both V36 and V51 (Table S2 of the ESM) indicates that the “old” groundwater component at these sites is less than hundreds of years old. The tracer-tracer plot shows that both V19 and V28 have primarily pre-modern water. This is consistent with older 14C ages (about 12,000 and 5,000 years, respectively) and low tritium (0.2 TU) for these sites.

Fig. 8
figure 8

Comparison of CFC-12 and CFC-113 equivalent atmospheric concentrations for groundwater in the Rio Actopan Basin, Veracruz, Mexico

Full size image

Discussion: comparison of 3H/3He and CFC ages

For the sites such as V36 and V51 containing binary mixtures of younger and older groundwater, the 3H/3He age should theoretically be similar to the age of the young component determined by CFCs and younger than the CFC piston-flow age. This is because the 3H/3He age is determined by the amount of parent (3H) relative to its daughter (3He), the product of radioactive decay. This ratio remains unchanged even when diluted by older tritium-free water. In contrast, an “apparent” piston-flow CFC age would get older when diluted by older CFC-free water. This is true for spring V51, with a 3H/3He age of 4–2 ± 0.5 years and young component of the binary CFC12/113 mixing model recharged within 1 year of sampling, compared with piston-flow CFC ages of 25–28 years. But this is not true for well V36, which has a 3H/3He age of 39 years (recharge in 1974) which is similar to the piston-flow CFC age of 33–42 years, yet older than the young component age of 18–28 years for the binary mix based on CFC-12 and CFC-113. It is likely, therefore, that some tritiogenic 3He was lost, either due to degassing at the spring orifice or during the sampling process. Such gas loss may also explain the discrepancies for springs V09 and V15, both having piston-flow recharge rather than binary mixtures with younger 3H/3He ages (10 and 4 years, respectively) than the piston-flow CFC-12 ages (mid-1980s recharge).

Conclusions

In order to better understand groundwater resources and groundwater quality within the Rio Actopan Basin, samples from 20 sites were collected for major-ion chemistry and stable isotopes. Carbon-14, tritium, chlorofluorocarbons, and noble gases were analyzed at selected sub-groups of these sites. Major-ion chemistry indicates that both salinity and water temperature generally increases in the lower part of the basin. Total dissolved solids in springs ranged from < 40 mg/L in the upper part of the watershed to >1,000 mg/L near the coast. Assuming most recharge occurs in the upper part of the basin, this increase likely indicates increased rock/water interaction associated with longer groundwater flowpaths. The geology within the basin (a mixture of volcanics, carbonates, and unconsolidated alluvial deposits) also controls the chemical signature of groundwater, resulting in calcium-bicarbonate, calcium-sulfate, and sodium-bicarbonate type waters. One well located near the coast has a mixture of both calcium-bicarbonate and sodium-chloride types, perhaps indicating some saltwater intrusion. Also, while most wells and springs have nitrate concentrations below the 10 mg/L drinking-water standard, two springs and one well near or downgradient of urban centers have concentrations between 20–44 mg/L.

Stable isotopes plot along a local meteoric water line previously defined by rainwater samples, indicating that recharge comes from precipitation falling at various elevations within the basin. All of the springs above 1,400 masl are slightly enriched compared with precipitation, indicating preferential recharge during the drier winter months. Many sites below 1,200 masl are slightly isotopically depleted, indicating their recharge occurs at higher elevations in the watershed. This is supported by cooler noble-gas recharge temperatures, which indicate that most of the recharge to the groundwater system occurs in the higher-elevation upland part of the basin.

The ages of groundwater in the Río Actopan Basin generally are Holocene to modern, indicating active groundwater recharge and relatively rapid movement through permeable aquifers. The wide range in ages (1 to >10,000 years) determined from environmental tracers indicates a variety of groundwater flowpaths and travel times within the watershed. While mixtures containing such old groundwater are unexpected in such a wet tropical setting, similar mixtures containing old groundwater (>1,000 years) were found in a tropical lowland rain forest in Costa Rica (Solomon et al. 2010).

Groundwater ages in the Río Actopan Basin derived from different tracers were generally consistent. For example, all the groundwater older than about 500 years based on 14C has low tritium (<0.5 TU), whereas all of the younger groundwater has tritium between 0.5 and 2 TU. Similarly, ages determined with 3H/3He are generally consistent with piston-flow CFC ages, but because the sampling sites included both springs and production wells with long screened intervals (rather than monitoring wells with short screen intervals), some groundwater samples are mixtures of two or more flow paths. This is supported by a CFC-12 versus CFC-113 tracer-tracer plot, with some samples located along a piston-flow line, while others are along binary mixing lines of young and older groundwater. In conclusion, the data collection and interpretation confirmed the study’s hypothesis that the groundwater system within the Rio Actopan Basin is relatively active, with the majority of recharge to the groundwater system occurring in the upland parts of the watershed and moving towards the coast.