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

Evaluation of Formation Salinity Using Borehole Geophysical Logs

Two threshold salinity values of interest in the Floridan aquifer system are 10,000 and 35,000 mg/L of dissolved-solids concentration. As previously defined, 10,000 mg/L of dissolved-solids concentration separates brackish and slightly saline water, and 35,000 mg/L of dissolved-solids concentration separates slightly saline and saline water. The depth to the top of the zones in the Floridan aquifer system that contain water with these dissolved-solids concentrations or greater can be approximately determined using bore-hole geophysical logs.

The resistivity of a nonshaley, water-bearing formation is related to the porosity and resistivity of the formation water according to the following equation (Archie, 1942):

R₀ = a ø^-m R_w (1)

where R₀ is the water-saturated formation resistivity, a is an empirical constant, ø is total or bulk formation porosity as a fraction, m is the empirical cementation factor, and R_w is the formation water resistivity. The values of R₀ and R_w are at formation temperature.

As previously described, the predominant lithology in the Floridan aquifer system in southeastern Florida is micritic to fine-grained limestone with low permeability. Rock with cavernous or vugular porosity occurs only in a few thin zones in the Floridan aquifer system, except in the Lower Floridan aquifer. Oolitic or fragmental limestone that has not been sealed by secondary calcite generally can be analyzed as though it were a clastic rock (MacCary, 1983, p. 335), and clastic rocks are analyzed using equation 1. The high total porosity in the Floridan aquifer system, particularly in the Avon Park Formation, indicates that: (1) the pore system is dominated by microporosity, which occurs between and within grains (intraparticle porosity) rather than large pores or vugs; and (2) sealing by secondary calcite has not occurred to a great extent.

Porosity was measured from whole diameter cores (4.5 in.) in the middle confining unit of the Floridan aquifer system in well PBP-I1 in Broward County (fig. 1). Six limestone plugs were taken from cores cut between 1,931 and 2,233 ft deep, and the porosity of these plugs determined by Boyle's law (Core Lab, 1974, p. 4-5) ranged from 33.6 to 46.4 percent and averaged 40.2 percent (Post, Buckley, Schuh, and Jernigan, Inc., 1987, table 7-4).

Figure 8. (above) Variations of average porosity with depth derived from a borehole-compensated neutron-density log of well C-962 showing geologic units and a gamma-ray log in eastern Collier County. Porosity is an average of neutron and density responses. Tool was calibrated to a limestone matrix. [larger image] Figure 9. (above) Relation between concentrations of dissolved solids and chloride for 19 water samples from the Floridan aquifer system in southeastern Florida. [larger image]

A borehole-compensated neutron-density log, used to determine porosity was run on well C-962 in Collier County (fig. 1). The neutron and density devices on this tool produce a signal that responds to porosity changes with depth, and the rock matrix used to compute porosity from both of these signal was limestone. The two porosity curves output from the devices were averaged for well C-962 (fig. 8). The value for porosity output from each device was usually in relatively close agreement (difference of less 1 or 2 percent porosity), indicating that the rock matrix is limestone, the clay content is low ("clean" formation), and the average porosity is a reasonable estimate of the true total porosity. Descriptions of drill cuttings indicate that the shallowest dolomite present in well C-962 in the Floridan aquifer system is in the Oldsmar Formation at a depth of about 2,270 ft. Log-derived porosity ranged from 20 to 45 percent in the upper 1,200 ft of the Floridan aquifer system and from 30 to 40 percent throughout most of this interval (fig. 8). A general tendency for porosity to decrease with depth is indicated.

On the basis of the aforementioned data for wells PBP-I1 and C-962 and other core and log data, porosity in the Upper Floridan aquifer and the middle confining unit of the Floridan aquifer system is generally high (at least 30 percent in the study area). Although sonic logs have been run in many of the wastewater injection system wells in the study area, traveltime (unlike the neutron-density log response), is difficult to calibrate to true total porosity in carbonate rocks. Calculations of R₀ will be made using a range in porosity (30 to 40 percent) because of the lack of usable porosity measurements in most wells in the study area.

The cementation factor, m, was measured using limestone core samples collected from well PBP-I1 at a depth between 1,931 and 2,934 ft. Using equation 1 and assuming a = 1.0, values of m ranged from 1.96 to 2.15 and averaged 2.07 (Post, Buckley, Schuh, and Jernigan, Inc., 1987, table 7-8). The values a = 1.0 and m = 2.0 are recommended for chalky limestone (Schlumberger, 1972a, p. 2). The constant a was assumed to be equal to 1 for most Tertiary carbonates in the Southeastern Coastal Plains of the United States where Kwader (1986) reported the value of m for Tertiary clean "platform type" limestones to range from 1.6 to 1.8. This range for m could be low in comparison to the range expected in southeastern Florida in the Floridan aquifer system because of the possibility of less compaction and cementation in the Coastal Plain area. The depths of the Tertiary section worked within the Coastal Plain area are less than 1,000 ft (Kwader, 1986, fig. 1). Because of the uncertainty in establishing a value for m in the study area and the probability that it does vary given the thick section of interest (Upper Floridan aquifer and middle confining unit of the Floridan aquifer system), calculations of R₀ were made using a range of values from 1.8 to 2. 1. The formation water resistivity, R_w, for a given salinity of water in the Floridan aquifer system can be determined from water analysis. From the relation between dissolved-solids and chloride concentrations, chloride concentration can be determined for a given dissolved-solids concentration; then by relating chloride concentration to specific conductance, resistivity can be determined. Finally, resistivity is corrected using formation temperature to give R_w. Chloride concentration was related to specific conductance instead of relating dissolved-solids concentration directly to specific conductance for two reasons. First, many of the analyses of water samples from the Floridan aquifer system (table 2) did not include determination of dissolved-solids concentration, whereas chloride concentration was determined. Second, a better correlation of specific conductance with chloride concentration is expected than with dissolved-solids concentration, as was the case for Floridan aquifer system water in the panhandle area of Florida (Kwader, 1986, fig. 3). Chloride concentration is used later in this report to map the distribution of salinity in the Floridan aquifer system.

Linear regression analysis of dissolved-solids concentration in relation to chloride concentration was made using USGS water-quality data collected from the Floridan aquifer system (fig. 9). Dissolved-solids concentrations from 19 water samples used in this analysis ranged from 2,000 to 40,100 mg/L (table 2). Samples from wells that are not completed in the Floridan aquifer system or for which the depth of the top of the sample interval is unknown were not used. Additionally, samples from wells PU-I1 and PU-I2 were not used because of the large sample interval, and samples from wells G-2617, G-2618, G-2619 and G-2296 (sample on January 15, 1992) were not used because earlier data from well G-2296 (in 1981) at the same location were used. The correlation coefficient for this analysis was good (r = 0.998) as shown in figure 9, and the relation determined was:

chloride concentration = 0.548 * dissolved-solids concentration -243 (2)

where chloride and dissolved-solids concentrations are in milligrams per liter.

Linear relations were also determined between chloride concentration and specific conductance. Two regression analyses were made using USGS water-quality data collected from the Floridan aquifer system (table 2). In the first analysis, 34 water sample with chloride concentrations ranging from 850 to 22,000 mg/L were used (fig. 10). In the second analysis, the data set used in the first analysis was restricted by selecting only water samples with chloride concentrations less than 5,000 mg/L (fig.11). This resulted in 24 water samples with chloride concentrations ranging from 850 to 4,500 mg/L.

Figure 10. (above) Relation between chloride concentration up to 22,000 milligrams per liter and specific conductance for 34 water samples from the Floridan aquifer system in southeastern Florida. [larger image] Figure 11. (above) Relation between chloride concentration less than 5,000 milligrams per liter and specific conductance for 24 water samples from the Floridan aquifer system in southeastern Florida. [larger image]

For the analysis with chloride concentration up to 22,000 mg/L, the relation is:

specific conductance = 2.42 * chloride concentration + 2,142 (3)

where specific conductance is in microsiemens per centimeter at 25°C. For the analysis with chloride concentration less than 5,000 mg/L, the relation is:

specific conductance = 2.95 * chloride concentration + 1,085 (4)

Each analysis provides a different regression relation, and there is a high degree of linearity with correlation coefficients of greater than 0.99 in both cases. The difference between these relations indicates there is some curvature in the relation between chloride concentration and specific conductance for natural waters (Hem, 1989, fig. 11).

The resistivity of sample water, in ohm-meters can be calculated from specific conductance, in microsiemens per centimeter at 25°C, by use of the following expression:

resistivity = 10,000/specific conductance (5)

The resistivity of Floridan aquifer system formation water for the two threshold salinity values, 10,000 and 35,000 mg/L of dissolved-solids concentration, were calculated and the results are given in table 4. For the 10,000 mg/L dissolved-solids salinity value, the chloride concentration was calculated using the regression equation developed for chloride in relation to dissolved-solids concentration (eq. 2). Specific conductance was then calculated using the regression equation developed for specific conductance in relation to chloride concentration, with chloride concentration less than 5,000 mg/L (eq. 4, second analysis). Equation 4 was used instead of equation 3 because the chloride concentration determined for this salinity (5,240 mg/L) was only slightly higher than 5,000 mg/L. For the 35,000 mg/L dissolved-solids salinity value, after the chloride concentration was determined using equation 2, specific conductance was calculated using the first analysis of specific conductance in relation to chloride concentration (eq. 3).

The resistivity of water that is a sodium chloride type, such as Floridan aquifer system water in southeastern Florida (Sprinkle, 1989, pl. 9), can be adjusted for a change in temperature using a resistivity graph for sodium chloride solutions (Schlumberger, 1972b, chart Gen-9). The calculated resistivity of Floridan aquifer system formation water for the two salinity values of interest can be adjusted from 25°C to the formation temperature to give R_w.

Formation temperature in the Upper Floridan aquifer and the middle confining unit does not vary greatly in the study area. The water temperatures reported for 20 water analyses measured during flow at the wellhead were analyzed. These samples came from 14 wells in the study area with the depth of the intervals sampled, ranging from 900 to 2,475 ft. Temperature ranged from about 20°C to 26°C and averaged 23.8°C. The lack of geothermal temperature increase between that near the ground surface (26°C) and the Floridan aquifer system is because of the cooling effect of cold deep seawater that probably enters the Boulder zone along the southeastern coast in the Straits of Florida (Meyer, 1989, fig. 24).

The formation resistivity (R₀) was computed for the two threshold salinity values using equation 1 for the expected ranges in porosity (ø), cementation factor (m), and formation temperature (tables 5 and 6). The values for the constant, a, was assumed to be 1.0 (Kwader, 1986, p. 11), and the values used for formation water resistivity at 25°C came from table 4. These computations indicate that the variation in porosity results in the greatest uncertainty in R₀. As salinity increases or decreases with depth in the Floridan aquifer system, a range in depth at which salinity equals 10,000 or 35,000 mg/L of dissolved-solids concentration can be determined using the results given in tables 5 and 6, if variation of true formation resistivity with depth is known and the porosity can be estimated.

Geophysical resistivity devices used in determining the depth of salinity boundaries in this study were the short normal (16 in.), long normal (64 in.), 18-ft 8-in. lateral, spherically focused device, laterlog 8, medium induction, and deep induction. Generally, the devices that have the greatest depth of penetration into the formation were given preference. Conventional electric logging tools usually include short normal (16 in.), long normal (64 in.) and 18-ft 8-in. lateral devices (listed in order of increasing depth of penetration). Although the lateral device from this tool has the greatest depth of penetration, the long normal curve seemed to give a more reasonable response in the Floridan aquifer system in the study area and was usually favored in determining the salinity boundaries.

If the resistivity tool run in a well included a deep induction device and the depth of invasion by drilling fluid was relatively small, salinity boundaries were determined using only the curve from this device without correction. However, in some cases, deep induction resistivity was corrected for borehole and invasion effects. Experience gained in working with invasion correction charts designed for dual induction logs (Schlumberger Educational Services, 1988, p. 89-91) in the Floridan aquifer system showed that correction for invasion generally was important if the ratio of the medium induction reading to the deep induction reading was greater than about 1.4. This applies only if the borehole fluid is not heavily contaminated with salt (salinity of the borehole fluid less than that of the formation water).