|
|
Study Area
The geology of the study area has been mapped and described by Rogers (1977) and Rogers and others (1979). The
hydrogeology of the Valle de Yabucoa alluvial aquifer has been described by Anders (1971), Robison and Anders (1973), and
Troester and Richards (1996). The hills surrounding the Valle de Yabucoa as well as the bedrock underlying the alluvium in
the valley are composed of the San Lorenzo Batholith, a large, igneous intrusive body emplaced during the Late Cretaceous
(Rogers, 1977; Rogers and others, 1979). The San Lorenzo Batholith is a composite body that is composed of gabbro, diorite,
tonalite, granodiorite, and quartz monzonite.
The altitude of the hills surrounding the Valle de Yabucoa reaches a maximum of about 650 m at the head of the Río Guayanés basin. The land surface in the Valle de Yabucoa slopes gently from an altitude of about 30 m above mean sea level at the western edge of the valley to sea level where the valley meets the Caribbean Sea.
The Río Guayanés is the major stream flowing through the Valle de Yabucoa. During the Pleistocene, sea level was as much as 120 m lower than it is today. At that time the Río Guayanés cut a deep valley into the underlying bedrock as the stream eroded down toward base level. The paleovalley of the Río Guayanés and its principal tributaries was not only much deeper than the present valley, but the valley surface was much more irregular. Ridges now partly or completely buried extended far out into the valley, and isolated or semi-isolated hills rose rather abruptly from the valley floor. As sea level began to rise to its current position, the valley was gradually filled by alluvial material, possibly behind a bay-mouth bar. As a result of shifts in the positions of the ancestral streams, periodic advances and retreats of the sea, and the presence of hills and ridges, the deposited thickness of alluvium and its grain-size distribution varies appreciably over short distances. The maximum thickness of the alluvium in the Valle de Yabucoa is probably between 107 m (the depth of the deepest observation well in the valley) and 120 m (the maximum drop in of sea level during the Pleistocene).
The alluvium derived from the erosion of the batholith is composed of sand, silt, and clay with occasional pebbles, cobbles, and boulders and minor amounts of calcareous shell fragments and organic material. Most of the sand is yellowish- brown to medium-gray in color and intimately mixed with the yellowish-brown to grayish-green silt and clay. The sand is coarser and the sand beds are thicker and more common with increasing distance inland from the coast; however, there are large local variations in aggregate sand thickness and character. Beach and swamp deposits are present close to the coast. The mineralogy and chemistry of the alluvium have been studied by Collar and others (1990) and Marsh (1992a, b).
Optical and electron microscopy coupled with energy-dispersive x-ray spectroscopy have determined that the alluvium is composed of plagioclase (andesine), quartz, orthoclase, a calcium- and iron-rich amphibole, biotite, magnetite, and ilmenite grains. Microscopy also identified iron oxide coating on grains and, probably, maghemite, gamma-Fe2O3, an alteration of magnetite. Small, 2- to 5-mm crystals of chalcopyrite were also found. X-ray diffraction of the clay-size fraction indicates that the major clay minerals are kaolinite, mica, and smectite, with minor amounts of vermiculite, goethite, and hematite. Microscopy indicates that manganous ilmenite comprises about 0.2 percent of the aquifer alluvium. The ilmenite (Fe1.0-0.8,Mn0.0-0.2)TiO3 contains about 0 to 20 percent manganese substituting for iron in a solid solution with pyrophanite, MnTiO3. The ilmenite is being altered to leucoxene (a combination of rutile, TiO2; anatase, TiO2; and sphene, CaTiO(SiO4)).
Methods
Eighteen wells were drilled and installed by the USGS with a hollow stem auger. In addition, a 107-m-deep observation
well (number 47 in fig. 1) was drilled by a private contractor into the deepest part of the paleochannel to monitor the
movement of saline water into the aquifer on a long-term basis. The land surface elevation at well 47 was estimated to be
about 0.8 m above mean sea level. The well did not penetrate the consolidated granodiorite, but the well did penetrate the
saprolite on top of the bedrock. This is the deepest well in the Valle de Yabucoa. The well is constructed with a 6.1-m screen
centered at 101 m below land surface.
Water samples were collected from existing public water-supply, industrial, and abandoned wells and from the new
observations wells drilled as part of this investigation. Abandoned and new wells were sampled with submersible pumps
following the evacuation of at least three well volumes of water. Production wells were sampled close to the well head using
the available pump and before the water enters in contact with the chlorination process. Field measurements of water
temperature, specific conductance, pH, and Eh were made following the procedures described by Wood (1976). Water samples
were collected, filtered, preserved, and sent to the USGS National Water Quality Laboratory for analysis of common cations,
anions, and nutrients. Thin sections were prepared from grain mounts and analyzed petrographically with both optical and
electron microscopy coupled with energy-dispersive x-ray spectroscopy. Additional details on the geochemical techniques
used to investigate the water and alluvium samples in this study were reported by Collar and others (1990), Troester and others
(1990a, b), and Troester and Richards (1996).
Results and Discussion
Iron concentrations in the Valle de Yabucoa alluvial aquifer ranged from less than 0.003 to 28 mg/L and manganese
concentrations ranged from 0.016 to 3.6 mg/L (Troester and Richards, 1996). The oxidation states of iron and manganese of
hydrogeological significance are Fe2+, Fe3+, Mn2+, and Mn4+. The theoretical stable ranges of iron and manganese ions,
oxides, and hydroxides as a function of Eh and pH are shown in figure 2. Oxidized iron, Fe3+, is the dominant iron ion in
solution only below a pH of about 3; oxidized manganese, Mn4+, is never the dominant manganese ion in solution. The
reduced species, Fe2+ and Mn2+, are the dominant hydrated ions in solution in the Eh and pH range of natural waters. In much
of the Eh and pH range of natural waters, Fe3+ and Mn4+ form insoluble oxides or hydroxides. The Mn2+ ion is
thermodynamically stable across a wider Eh-pH range than is the Fe2+ ion, as shown in figure 2. Consequently, the Mn2+ ion
remains in solution in aqueous environments in which Fe2+ is oxidized to Fe3+ and precipitated as Fe(OH)3.
Figure 2. -- Eh-pH stability diagram for ferromagnesian oxides showing the Eh
and pH of 15 ground-water samples from the Valle de Yabucoa alluvial
aquifer, Puerto Rico. Iron and Manganese activities of 10-6 are assumed for
the construction of the Eh-pH diagram. This diagram assumes that Fe(OH)3 is the stable phase for the Fe(III) precipitate.
The Eh and pH field values and iron and manganese concentrations for water samples collected from the Valle de
Yabucoa alluvial aquifer and analyzed are plotted in figure 2 and listed in Troester and Richards (1996). All of these samples
are from wells drilled exclusively for monitoring purposes during this investigation. These chemical analyses agree with the
theoretical predictions shown in the Eh-pH diagram (fig. 2). All of the water compositions that fall in the stability field of the
Fe2+ ion have iron concentrations greater than 10 mg/L, whereas all the water compositions that fall in the stability field for
Fe(OH)3 have iron concentrations that are less than 0.2 mg/L.
None of the samples plotted in figure 2 have Eh-pH combinations that plot within the stability field of MnO2; however, it is possible that in high-production wells in the Valle de Yabucoa the redox potential is sufficiently elevated as a result of oxygenation that the water composition would plot within the stability field of manganese dioxide. In this case the Mn2+ ion should oxidize and precipitate as MnO2; however, the kinetics of this reaction are sufficiently slow that precipitation may not occur until after the water was pumped from the well. Hem (1985) reported that iron oxidation occurs on the order of minutes, whereas manganese oxidation requires several hours to a day under commonly encountered ground-water conditions. Accordingly, given the same environment, iron will be oxidized more rapidly than manganese.
The iron concentration in the ground water increases rapidly with distance from the boundaries of the aquifer and with depth. After a rapid increase, the concentrations then decrease near the deeper portions in the center of the valley. Low iron concentrations were detected at wells 13 and 20 near the boundary of the aquifer, where recharge occurs. Low iron concentrations were also measured in water samples collected at shallow wells 36, 39, and 40 in the center of the valley; however, at these sites the iron concentration increased greatly with depth. Water from wells in the center of the valley with low iron concentrations frequently has a slight hydrogen sulfide odor. If hydrogen sulfide is present in the water, the iron has probably precipitated out of the ground water as an iron sulfide mineral.
The geographic distribution of manganese is similar to that of iron. Low concentrations occur near the boundary of the aquifer and in shallow wells in the center of the valley. High manganese concentrations were measured in the deeper wells in center of the valley.
The three-dimensional distribution of both iron and manganese is more complicated than a two-dimensional distribution derived from fully penetrating wells. The observation wells at Martorell (numbers 34 and 35 in fig. 1), Limones (numbers 36 to 38), and Juan Martín (numbers 39 to 41) were carefully constructed to be open to only certain depths in the aquifer. At these sites, the Eh of the ground water decreases with depth, whereas iron and manganese concentrations increase with depth in the aquifer. Similar chemical changes with depth probably occur throughout the aquifer, but detailed information on these changes with depth is only available for these well sites.
Electron and optical microscopy, energy-dispersive x-ray spectroscopy, and x-ray diffraction have identified several iron-containing minerals that are weathering and releasing iron into the ground water. These minerals include plagioclase, calcian amphibole, biotite, magnetite, hematite, ilmenite, and iron oxide coatings on grains. Mechanisms of iron introduction into the ground water include both the dissolution of ferrous iron-bearing minerals and the reduction of ferric oxides. Reducing conditions [Eh ranges from +439 to -30 millivolts], in a great portion of the aquifer maintain much of the ground water within the region of thermodynamic stability of the Fe2+ ion, which enables the dissolution of iron from the iron bearing minerals in the aquifer. Loss of ferrous iron from these minerals is probably responsible for most of the observed elevated iron concentrations.
The largest contributor of ferrous iron to the ground water appears to be an iron oxide phase (probably maghemite, gamma-Fe2O3, an alteration of magnetite). Petrographic observations indicate that maghemite has been removed leaving behind a skeleton of hematite on the (111) planes of the original magnetite crystal. This agrees with the saturation indexes calculated by the WATEQ4F water-quality computer model (Ball and Nordstrom, 1991). The results from WATEQ4F model indicates that the reducing ground water in most of the aquifer is slightly under saturated with maghemite and oversaturated with both magnetite and hematite.
Only two mineral phases in the aquifer have been found to contain detectable manganese: amphibole and ilmenite. Trace amounts of manganese were measured in the amphibole; however, the low concentrations and the fact that amphibole is not extensively corroded make it unlikely that this mineral is the primary source of the relatively high manganese concentrations measured in the water. Ilmenite, (Fe,Mn)TiO3, constitutes approximately 0.2 percent of the sediment, contains from 0 to 20 percent manganese substituting for iron in a solid solution with pyrophanite, MnTiO3, and is probably the sole significant source of manganese. Individual ilmenite grains contain patches and streaks of leucoxene (a combination of rutile, TiO2; anatase, TiO2; and sphene, CaTiO(SiO4)), indicating that leaching of iron and manganese from the original ilmenite has resulted in the formation of the iron and manganese-free alteration product and the release of the manganous and ferrous ions into the reducing ground water without oxidation or reduction reactions. The liberation of iron and manganese during the progressive alteration of ilmenite to leucoxene (possibly including the intermediate alteration product pseudo-rutile) has been reported by Anand and Gilkes (1984) and Morad and Aldahan (1985).
A mass-balance geochemical model was derived using methods described by Plummer and others (1983) with the aid of the NETPATH computer program (Plummer and others, 1991) to account for the chemical evolution of water in the aquifer. In the mass-balance model andesine, amphibole, ilmenite, and biotite dissolve; organic carbon is oxidized; dissolved carbon dioxide gas is supplied from an external source; calcium is exchanged for sodium; and silica, goethite, kaolinite, and chalcopyrite are precipitated. Results of geochemical mass-balance modeling of the water chemistry agree with the optical and electron microscopy data (Troester and others, 1990a; b).
As ground water flows through the Valle de Yabucoa alluvial aquifer towards the Caribbean Sea, chemical reactions between the water and the aquifer slowly increase the dissolved-solids concentrations in the water and change the chemical characteristics of the water. Dissolved-solids concentrations (in milligram per liter) can be estimated by multiplying the specific conductance value measured in the field (in microsiemens per centimeter) by a factor of 0.59 (Hem, 1985, p.67). For example, water with a specific conductance of 12,000 µS/cm would have a dissolved solids concentration of about 7,000 mg/L. Seawater has a dissolved solids concentration of about 34,000 mg/L (Hem, 1985).
The specific conductance measurements of the ground water ranged from less than 200 µS/cm in a shallow well at the western edge of the aquifer (number 34 in fig. 1) to more than 12,000 µS/cm in the deep observation well (number 47 in fig. 1). Prior to drilling the deep observation well, the maximum measured specific conductance of water in the aquifer was 875 µS/cm (at well 45 in fig. 1), and the specific conductance measurements averaged about 500 µS/cm. Ground water with a specific conductance value of 875 µS/cm typically will have a dissolved-solids concentration near the U.S. Environmental Protection Agency secondary drinking water standard of 500 mg/L (U.S. House of Representatives, 1996).
Little time-series data exist that indicate what changes in ground-water chemistry occur as a well is effected by saline intrusion. Usually, shortly after a well has been taken out of service because of high dissolved-solids concentrations, the well is abandoned and, subsequently, destroyed. For example, when water from the Sun Oil #1 well (the closest industrial supply well to the ocean) was sampled on August 23, 1988, the water had a specific conductance of 800 µS/cm. Subsequently, this well was taken out of service and plugged.
Complete chemical analyses are plotted in figure 3. The large arrows indicate the general trends in the chemistry of the aquifer. In the up-gradient portion of the aquifer the water is typically rich in calcium, magnesium, and bicarbonate. As the water moves through the aquifer and approaches the ocean, it becomes rich in sodium, potassium, and chloride. More details can be found in Troester and Richards (1996).
Figure 3. -- Piper diagram showing overall composition of sampled
water in the Valle de Yabucoa alluvial aquifer, Puerto Rico.
Conclusions
Water in the Valle de Yabucoa alluvial aquifer has
iron and manganese concentrations as high as 28 and 3.6
mg/L, respectively. The largest contributor of ferrous
iron to the ground water appears to be an iron oxide
phase (probably maghemite, gamma-Fe2O3, an alteration of
magnetite). Ilmenite, (Fe,Mn)TiO3, is probably the sole
significant source of manganese. Although the elevated
iron and manganese concentrations in the aquifer pose no
health threat, they frequently exceed the secondary
standards for drinking water of 0.3 mg/L for iron and
0.05 mg/L for manganese. In addition to the iron and
manganese problems, the ground water in the aquifer
locally contains concentrations of dissolved solids that
exceed the secondary drinking water standard of 500
mg/L. The high concentrations of dissolved solids have
caused some water-supply wells to be taken out of
service. In the up-gradient portion of the aquifer the
water is typically rich in calcium, magnesium, and
bicarbonate. As the water moves through the aquifer and
approaches the ocean, it becomes rich in sodium,
potassium, and chloride.
References
Anand, R. R. and R. J. Gilkes, 1984. Weathering of ilmenite in a lateritic pallid zone. Clays and Clay Minerals. v. 32, n. 5, pp. 363-374.
Anders, R. B., 1971. Electrical analog model study of water in the Yabucoa valley, Puerto Rico--Phase 1, collecting preliminary data and assembling available data. U.S. Geological Survey Open-File Report, 47 pp.
Ball, J. W. and D. K. Nordstrom, 1991. User's manual for WATEQ4F, with revised thermodynamic data base and test cases for calculation speciation of major, trace, and redox elements in natural waters. U.S. Geological Survey Open-File Report 91-193, 189 pp.
Collar, P. D., J. W. Troester, R. G. Deike and P. D. Robinson, 1990. Iron and manganese in alluvial ground water, Yabucoa, Puerto Rico. 22nd Congress of the Interamerican Association of Sanitary and Environmental Engineering, AIDIS, San Juan, Puerto Rico, v. 1, pp. 708-732.
Hem, J. D., 1985. Study and interpretation of the chemical characteristics of natural water (3d ed.). U.S. Geological Survey Water-Supply Paper 2254, 264 pp.
Marsh, S. P., 1992a. Analytical results for stream sediment and soil samples from the Commonwealth of Puerto Rico, Isla de Culebra, and Isla de Vieques. U.S. Geological Survey Open-File Report 92-353A, 8 pp.
_____ 1992b. Analytical results for stream sediment and soil samples from the Commonwealth of Puerto Rico, Isla de Culebra, and Isla de Vieques. U.S. Geological Survey Open-File Report 92-353B, 5.25-inch diskette.
Morad, Sadoon and A. A. Aldahan, 1985. Leucoxene-calcite-quartz aggregates in sandstones and the relation to decomposition of sphene. Neues J. für Mineralogie, v. 10, pp. 458-468.
Plummer, L. N., D. L. Parkhurst and D. C. Thorstenson, 1983. Development of reaction model for ground-water systems. Geochimica et Cosmochimica Acta, v. 47, pp. 665-686.
Plummer, L. N., E. C. Prestemon and D. L. Parkhurst, 1991. An interactive code (NETPATH) for modeling net geochemical reactions along a flow path. U.S. Geological Survey Water-Resources Investigations Report 91- 4078, 227 pp.
Robison, T. M. and R. B. Anders, 1973. Electrical analog model study of the alluvial aquifer in the Yabucoa valley, Puerto Rico--Phase 2, The planning, construction, and use of the model. U.S. Geological Survey Open-File Report 73-1, 22 pp.
Rogers, C. L., 1977. Geologic map of the Punta Guayanés quadrangle, southeastern Puerto Rico. U.S. Geological Survey Miscellaneous Investigations Series Map I-998.
Rogers, C. L., C. M. Cram, M. H. Jr. Pease and M. S. Tischler, 1979. Geologic map of the Yabucoa and Punta Tuna quadrangles, Puerto Rico. U.S. Geological Survey Miscellaneous Investigations Series Map I-1086, scale 1:20,000.
Troester, J. W. and R. T. Richards, 1996. Geochemical properties and saline-water intrusion in the Valle de Yabucoa alluvial aquifer, Southeastern Puerto Rico. U.S. Geological Survey Water-Resources Investigations Report 96-4188, 39 pp.
Troester, J. W., P. D. Collar, P. D. Robinson and R. G. Deike, 1990. Geochemical processes controlling manganese and iron concentrations in the ground water of the Valle de Yabucoa alluvial aquifer, Puerto Rico. EOS, Transactions, American Geophysical Union, v. 71, pp. 506.
Troester, J. W., R. G. Deike, P. D. Robinson and P. D. Collar, 1990. Petrographic analyses and geochemical modeling of aqueous/solid phase reactions in the Valle de Yabucoa alluvial aquifer, Puerto Rico. Geological Society of America Annual Abstracts and Program, v. 22, no. 7, pp. 295.
U.S. House of Representatives, 1996. Secondary maximum contaminant levels. Code of Federal Regulations, Title 40, Chapter I, Subchapter C, Part 143.
Wood, W. W., 1976. Guidelines for collection and field analysis of ground-water samples for selected unstable constituents. Techniques of Water-Resources Investigations of the United States Geological Survey, Book 1, Chapter D2, 24 pp.
Acknowledgements
The author would like to thank Sucesores Roig, Inc. and the Puerto Rico Land Authority for allowing access to their
land and permitting the installation of observation wells on their property. He would also like to thank Puerto Rico Sun Oil,
Inc. and the Central Roig for their assistance. Lastly, the author is deeply indebted to the workers from the Central Roig and
the plantain groves who on numerous occasions helped extricate his vehicle from the clay-rich surficial deposits.