Home Introduction Hydrogeology of the Floridan Aquifer System Collection and Analysis of Salinity Data Evaluation of Formation Salinity Distribution of Salinity in the Floridan Aquifer System >Origin of Salinity Summary and Conclusions References Cited Appendix: Inventory of Wells PDF Version

Origin of Salinity

The origin of salinity in the brackish-water zone of the Floridian aquifer system in southern Florida, salinity due mostly to the sodium chloride content of the water, has been explained by two contrasting theories. One will be referred to as the convention cell theory and the other as the residual salinity theory. Evidence for and against these theories in southeastern Florida, some of which results from this study, will discussed in this section. Of special importance to the theories are the processes that could be controlling the depth of the base of the brackish-water zone, and the processes will be examined. In the last part of this section, the residual salinity theory will be used to explain the areas of high salinity found in the Upper Floridian aquifer along the coast.

The conventional cell theory, proposed by Kohout (1965), calls for the upward movement of saline water from the Boulder zone through the middle confining unit of the Floridian aquifer system and mixing of this water with seaward-flowing freshwater in the upper part of the aquifer system. The impetus in this theory is geothermal heating under the Florida Plateau. The lower part of the flow cell is the inland movement of seawater into the Boulder zone along the Straits of Florida where the Floridian aquifer system outcrops below sea level at the edge of the plateau. The residual salinity theory explains that the original saline pore water of deposition or later invaded seawater has not been completely flushed out by meteoric water. In southern Florida, salinity in the Upper Floridian aquifer is, at least in part, interpreted to be residual (Sprinkle, 1989, p. 48).

There is evidence presented by Meyer (1989) that supports the convection cell theory, chief among which is the presence of measurable carbon-14 activity (as high as 6.5 percent of modern carbon) in water from the Upper Floridian aquifer in southern Florida (Meyer, 1989, fig.11). Estimated transit times in the Upper Floridian aquifer from the aquifer's recharge area (Meyer, 1989, table 7) are much longer than the 40,000-year useful range for carbon-14 dating, and therefore the percent of modem carbon should approximate zero in southern Florida. The presence of the measurable carbon-14 in southern Florida suggests another source for the water in the aquifer of an age younger than 40,000 years--probably the upwelling of saline water from the Boulder zone into the Upper Floridian aquifer (Meyer, 1989).

The convection cell theory implies that the saline water in the Boulder zone is heated to a point that its density becomes less than that of water in the Upper Floridian aquifer, causing its upward migration. The maximum temperature of water in the Boulder zone of the Lower Floridian aquifer in southern Florida is 43.4°C (Meyer, 1989, fig. 7) at the northernmost extent of the zone, 70 mi northwest of the study area. (The temperature ranges from 10°C to 24°C in the study area.) Using oceanographic tables that give the density of seawater at various temperatures and salinities, the density of standard seawater at 43.4°C was extrapolated to be 1.017 g/cm³ (Unesco, 1987). This density is still substantially higher than the density of water in the brackish-water zone at 25°C and with a chloride concentration of 3,000 mg/L (density equals 1.0035 g/cm³). This density/temperature factor would seem to preclude the upward movement of saline water into brackish-water zone by thermal convection in the study area and perhaps in most of southern Florida.

If the brackish-water and saline-water zones occur within one hydraulic unit, the thickness of brackish-water zone might result from a state of density equilibrium that exists between the two zones. Although the permeability of rocks present in the brackish-water zone in the study area is generally low, highly impermeable rock (dense dolomite or clay) is uncommon.

Pressure measured at different depths in a well in the Floridian aquifer system can help determine if vertical hydraulic connectivity exists. A plot of pressure measurements compared to depth was made well G-2296 (Meyer, 1989, fig. 10). These measurements were made mostly during packer testing or from completed intervals (Meyer, 1989, table 3). A pressure gradient for one measurement made within the brackish-water zone at a depth interval of 1,030 to 1,154 ft was calculated to be 0.4325 lb/in²/ft on the basis of estimated fluid density and representative depth. A line representing this gradient, drawn on the plot through the measurement, closely coincided with eight other measurements made in the brackish-water zone at depths ranging from 895 to 1,618 ft (Meyer, 1989, figure 10). This indicates that the interval represented by these measurements is hydraulically connected.

If a state of equilibrium exists between the brackish-water and saline-water zones, an estimate of the depth to the base of the brackish-water zone can be made according to the Ghyben-Herzberg approximation (Bear, 1979, p. 385). Assuming steady-state horizontal flow in the freshwater (brackish) region and no flow in the saltwater region, the altitude below sea level of a sharp freshwater-saltwater interface in a confined aquifer can be calculated to the following equation (Bear, 1979, p. 385):

h_s = [_f / (_s - _f)] h_f (6)

where h_s is the depth to the interface below sea level, _f is the density of the freshwater, _s is the density of the seawater, and h_f is the freshwater head.

The depth to a freshwater-saltwater interface in the Floridian aquifer system was calculated using the Ghyben-Herzberg approximation (table 8). The density of water in the brackish-water zone used in these calculations was estimated using the chloride concentration of Upper Floridian aquifer water produced from a well and a linear relation between chloride concentration and density from freshwater to seawater. The density of seawater used was 1.0268 g/cm³. This value is the average specific gravity of seawater in the Miami area, referred to distilled water at 25°C/25°C (Parker and others, 1955, p. 573). The Upper Floridian aquifer head data (table 8) were recorded as freshwater heads. They were adjusted using the density of the actual water produced before being used in equation 6. These head data were assumed to not differ greatly from predevelopment heads.

The calculated altitude of a saltwater interface is comparable to that of the base of the brackish-water zone determined in this study. Table 8 indicates that the difference between the altitudes is 56 ft or less in five cases. Some of the difference between the two could be because of the presence of the salinity transition zone rather than a sharp interface. However, as discussed earlier, the thickness of the transition zone is commonly only 80 to 120 ft.

The increase in salinity with depth in the salinity transition zone suggests mixing or diffusion associated with a brackish-water/saline-water interface. The rapid, steady nature of this increase (fig. 12) indicates that diffusion might be a more important process than mechanical mixing or hydrodynamic dispersion in explaining the origin of this zone.

The assumption of horizontal flow above the interface is no longer valid near the coast, and the Ghyben-Herzberg approximation gives a depth that is less than the actual depth to the interface (Bear, 1979, p. 385). The reverse is true for well MO-130 on Key Largo (table 8) with the calculated depth for the interface being much greater than the measured depth. The brackish-water zone is only 24-ft thick in this well and could be underlain by rocks of very low permeability, impeding hydraulic connection between the zone and deeper rocks.

The interface is apparently at or near an equilibrium position in inland areas (table 8, wells G-2296, G-3234, and G-3239). This could be somewhat unexpected in view of the low hydraulic conductivity in the middle confining unit of the Floridian aquifer system (described earlier). The position of this interface should have changed in response to sea-level fluctuations during the Pleistocene. For example, during the last sea level low stand just before the Holocene transgression, sea level was more than 300 ft below present-day sea level. Reestablishment of equilibrium in the relatively short period of time since these changes indicates that vertical permeability is not too low in at least the upper part of the middle confining unit (probably higher than that of unconsolidated marine clay with a hydraulic conductivity of 3 x 10^-4 ft/d).

Table 8. Calculated altitudes of a saltwater interface in southeastern Florida using the Ghyben-Herzberg approximation and comparison with altitudes of the base of the brackish-water zone [Well locations shown in figure 1; head data from Bush and Johnston (1988), except for well S-1533, which is from Beaven and Meyer (1978); calculated altitude of saltwater interface uses density of Upper Floridan aquifer water produced; altitude of base of brackish-water zone determined in this study]
Well number	Freshwater head, Upper Floridan aquifer (feet)	Calculated altitude of saltwater interface (feet below sea level)	Altitude of base of brackish-water zone (feet below sea level)	Altitude of saltwater interface minus altitude of base of brackish-water zone (feet)
FTL-I1	37	1,623	1,469	154
G-2296	57	2,209	2,153	56
G-3234	46	1,804	1,684	120
G-3239	49	1,922	1,935	-13
MDS-I12	41	1,608	1,640	-32
MO-130	40	1,716	1,148	568
PBP-M1	43	l,734	1,752	-18
S-1533	38	1,564	1,527	37

An observation which lends support to the residual salinity theory is that the sediments of the brackish- water zone are generally fine grained, low in permeability, and high in porosity. These characteristics could make complete flushing difficult, even over long periods of time. This theory is also supported by comparing salinity of water from the intermediate confining unit (fig. 17) with that in the upper interval of the brackish-water zone (fig. 16). In two wells (G-2296 and G-3062) in the northern part of the study area, salinity is higher in the intermediate confining unit. Because of the lower permeability of the intermediate confining unit, less complete flushing by freshwater can be expected, resulting in higher salinity.

Areas of higher salinity in the upper interval of the brackish-water zone are present near or along the coast (fig. 16). The higher salinity in these areas is probably also residual, but its occurrence could have resulted from a different process than that for the inland areas. These coastal areas of higher salinity seem to occur where there is higher permeability in the Upper Floridian aquifer, suggesting that inland migration of saline water has occurred somewhat recently in these areas.

The origin of the northeast-trending area of high chloride concentration in northeastern Broward County in the upper interval of the brackish-water zone (fig. 16) might be related to higher than normal permeability in the Upper Floridian aquifer for the study area. This higher permeability could have allowed for the migration of high salinity water laterally from the coast to the northeast. The upper monitoring zone in well CS-I1, open at a depth from 1,193 to 1,222 ft, was heavily contaminated with salty drilling water. This interval is more than 100 ft below the top of the rocks of Eocene age (1,082 ft) and probably below the upper flow zone, which is developed in association with this top. The salinity of samples collected from the interval took 2 years to decrease to a background level, suggesting that there was a large volume of invaded drilling fluid and relatively high permeability in the rocks. Another flow zone could be present in this interval. The base of the zone of slightly saline water in the upper part of the brackish-water zone in well CS-I2 at 1,470 ft (fig. 14) could coincide with a decrease in rock permeability.

It is possible that the zone of slightly saline water in the upper part of the brackish-water zone in well CS-I2 resulted from invasion with salty drilling fluid, migrating from nearby well CS-I1 to CS-I2 drilled 4.5 years after well CS-I1. A dual induction resistively log run on well CS-I1 indicates deep invasion with a drilling fluid of higher salinity than the formation water. This saline invasion was present from the top of the Floridian aquifer system to a depth of least 2,050 ft.

Several observations indicate that the zone of anomalous high salinity in the upper part of the brackish-water zone in well CS-I2 is natural in origin. If this thick zone of higher salinity is naturally present, some density equilibration with the saline-water zone would be expected. Thus, the depth to the brackish water/saline-water interface would probably be greater because of the higher density of water in the brackish water zone. The base of the brackish-water zone in well CS-I2 is much deeper than expected, given its proximity to the coast (fig. 15).

An analysis of chloride concentrations made for well CS-I1 in the upper monitoring zone at a depth of 1,193 to 1,222 ft. Chloride concentrations were 9,300 mg/L in May 1985 and decreased to 5,640mg/L in May 1987. From May 1987 to March 1990, monthly water samples collected from this well indicate no further significant decrease in chloride concentration. This stabilized value (about 5,600 mg/L) is similar to the chloride concentrations from well CS-I2 in the upper part of the brackish-water zone (fig. 14, 5,400-6,550 mg/L). This similarity in salinity would be unlikely if the high salinity of the CS-I2 samples resulted from migration of salty drilling fluid from CS-I1 to CS-I2. A plume of invaded drilling fluid from well CS-I1 would be expected to migrate downgradient (east or east-southeast). However, well CS-I2 is 750 ft to the north of well CS-I1, and this distance could be enough to prevent the plume from having an effect on well CS-I2.

The area of high salinity in the upper interval of the brackish-water zone in southern Dade County and northern Key Largo (fig. 16) correlates with an area of higher permeability in the Upper Floridian aquifer. Data for wells NP-100, S-1532, S-1532 and S-1533 indicate that the transmissivity of the aquifer in this area is high (31,000 ft²/d in wells S-1532 and S-1533). Additionally, Puri and Winston (1974, fig. 35) mapped a belt of high transmissivity in the Upper Floridian aquifer in this area based on drilling characteristics.

Transmissivity of the Upper Floridian aquifer is low at the Miami-Dade South District Wastewater Treatment Plant site (2,700 ft²/d in well MDS-M2). Salinity is also relatively low at this site in the aquifer as shown by a chloride concentration of 900 mg/L in well MDS-M3 (fig. 16, upper interval of the brackish-water zone). Permeability in this area was possibly too low to allow for influx of high salinity water from the east.

High salinity in the Upper Floridian aquifer occurs along the St. Johns River Valley in northeastern Florida (Sprinkle, 1989, fig. 22). This high salinity might have resulted from high sea-level stands during the Pleistocene Epoch, causing the influx of seawater into the aquifer (Stringfield, 1966, p. 172). A high stand of sea level of about 23 ft above present sea level occurred about 140,000 years before present (Cronin, 1983). According to this theory, flushing of this saline water by the present freshwater flow system has been incomplete. The areas of high salinity near or along the coast in this study area might have resulted from a similar process.

The incomplete flushing of the invaded seawater in areas of higher permeability along the coast can be explained by the distribution of permeability in the Floridian aquifer system. Upgradient of these areas, only a few thin permeable zones are present in the Upper Floridian aquifer (the normal distribution of permeability in the study area). Because of this low transmissivity of the aquifer upgradient in the inland areas and the long distance from the recharge area of the aquifer, the rate of freshwater flushing in the invaded areas could be low. Additionally, flushing could have been impeded if the zones of higher permeability in the areas along the coast were not well connected to the thin permeable zones in upgradient areas. If the invasion were relatively recent (such as 140,000 years before present), the time available would not have allowed for complete flushing.