Home Archived October 29, 2018
(i)

South Florida Information Access (SOFIA)


publications > open file report > OFR 94-276 > methods


Summary
Introduction
Methods
Geologic Setting
Results
Rock Analysis
Water Chemistry
Ground Water
Contamination
QC/QA
Conclusions
Future Studies
Acknowledgments
References
Appendices
Tables and Figures

Methods

During the summer of 1992, personnel of the USGS St. Petersburg Coastal Center core drilled 21 wells (24 completed as monitoring wells) in three areas of the Keys. The wells were arranged in transects, one off north Key Largo, one off central Key Largo and one in the Saddlebunch Keys in the Lower Keys (Fig. 1). The wells were cored using the USGS hydraulic drill (MacIntyre, 1975; Shinn et al., 1977) equipped with standard 5-ft NX wire-line core barrels and drill rods. Most of the wells were drilled underwater by scuba divers (Fig. 2). Well depths ranged from 10 to 70 ft (3-20 m) and were drilled both on land and offshore in water depths up to 20 ft (6 m). The cores were drilled into the Pleistocene limestone, which receives the effluents of more than 600 injection wells and thousands of septic-tank drain fields. Most of the offshore monitoring wells penetrated several meters of Holocene sediment or coral reef before entering the underlying Pleistocene limestone. Rock cores with a diameter of 1 and 7/8 inch (48 mm) were examined in the field and later described in detail at the USGS Coastal Center in St. Petersburg, Florida. Selected porosity measurements were also conducted using the fluid-volume displacement method.

map of Florida Keys and well locations
Figure 1. Map of south Florida and Florida Keys showing location of the three major offshore monitoring-well transects. Onshore wells that complete the transects are not shown due to scale. Detail maps of the three areas are shown in Figures 5, 6 and 7. Locations of an additional 15 wells in Florida Bay, not used in this study, are also shown. [larger image]

Each hole drilled was completed as a water-quality monitoring well in the following manner. A standard 4-ft-long 1-inch-ID-diameter slotted PVC well screen, glued to a 20-ft-length of schedule 40 1-inch-ID-diameter PVC pipe, was lowered to the bottom through the NX drill rod. For wells more than 20 ft (6 m) deep (average depth was 35 ft, 10.7 m) an additional one or two 20-ft lengths of PVC pipe were added and then inserted into the well bore. With the screened section resting on the bottom of the well bore, the NX drill rod casing was raised 5 ft (1.5 m) to expose the screened section to the formation. A 5-ft section was then unscrewed at the top and removed.

photo of scientist drilling underwater photo of scientists drilling in nearshore shallow water
photo of a completed well (OR-5) on offshore reef photo of a scientist purging a well into a 5 gallon bucket
Figure 2. Typical underwater drilling operation using a hydraulic drill is shown in (A). Hoses lead to boat above, which has hydraulic power source and water pump. (B) Drilling operation in nearshore shallow water. Location is site of well KL-1. (C) A completed well (OR-5) on offshore reef. (D) Well being purged into 5-gal plastic bucket on stern of USGS boat Halimeda. Note portable GPS unit on transom and diver in clear blue water. Color of well water was usually a different color than that of the ambient sea water. [click on the photos above to view larger images]
Approximately 2 gallons of coarse quartz sand were poured into the annulus to fill the space between the screen and formation. Two gallons were sufficient, assuming the well encountered no large cavities. Where cavities were indicated by the recovered core, additional sand was added. The sand was too coarse to clog well screen slots and allowed unrestricted passage of fluid from the porous limestone to the screen. The sand also served to hold the PVC pipe in place during extraction of the NX drill rod casing.

After raising and removing an additional 5-ft section of drill rod casing, a slurry of Portland cement was poured down the annulus. This was accomplished by first placing the slurry in large plastic bags aboard the boat and once under water, the diver cut a hole in the comer of the bag and squeezed the cement down the hole like cake icing. The amount of cement varied but was calculated to fill approximately 5 ft (1.5 m) of the annulus above the sand pack. Cement did not penetrate the quartz sand but filled voids and irregularities in the rock, thus preventing water in the annulus, higher in the well, from entering the screened zone. After placement of the cement, the remaining NX rod was removed, leaving the PVC pipe and screen in the hole. After the NX rod was removed, a few feet of PVC pipe was left protruding from the hole.

Quick-setting hydraulic cement, composed of 1 part molding plaster (plaster of Paris) and 7 parts type II Portland cement (Hudson, 1979), was mixed with water to form a stiff ball about 15 cm in diameter. The ball of cement was quickly taken to the bottom and hand-molded into the annulus around the PVC pipe. Hydraulic cement sets in approximately 5 minutes and is very hard in a few hours. Next, the excess PVC pipe was sawed off with a hacksaw leaving 15 to 30 cm protruding above the surface. A tight-fitting PVC end cap sealed the wells. A typical installation is shown in Figures 2C and 3. When the hydraulic cement was sufficiently hard, wells were developed by pumping until the water ran clear. Purging was accomplished by fitting a PVC end cap (equipped with 3/4-inch 50-ft-long, 15 m, tygon hose) over the 1-inch-diameter PVC wellhead. The other end of the hose was attached to a small 12-VDC electric-powered rubber impeller pump aboard the boat. The pump, with a discharge rate of approximately 5 gallons/minute, was run for 5 to 10 minutes or until the water ran clear (see Fig. 2D).

Latitude and longitude were determined at each site with a portable GPS unit (Fig. 2D). Latitude and longitude, well name, depth, and other data are provided in Table I. All wells were left to stabilize at least 30 days before the first sampling run.

schematic diagram of well installations Figure 3. Schematic diagram of well installations. (A) Typical offshore well with single PVC liner located below unconformity. (B) A multi-well completion used on land. Screen in upper well is situated just below water table. At SB-1 and OR-1, two wells were drilled ~6 ft (1.8 m) apart and consisted of a deep well (Fig. 3A) and a shallow well (C). In the offshore OR transect, wells were completed as shown in (D). The lime mud was allowed to slump in and form a seal. [larger image]

One unexpected difficulty, especially at well sites within 2 nmi of shore, was caused by tidal pumping. These wells could not be completed when the tide was falling because outflow prevented introduction of quartz sand or Portland cement. Nearshore wells could be completed only when the tide was rising and flow was into the wells. Outflowing water was often so strong that the quartz sand would not settle to the bottom of the well even when the top of the drill casing was several feet above sea level.

Although GPS readings were obtained at all sites, care was taken to locate well sites where there were visual objects onshore or on the bottom as well. Objects such as navigational markers, telephone poles, and so forth, were lined up with other objects to facilitate relocation. This method of relocation was more efficient than using GPS, which presently is only accurate to within 65 to 100 ft (20-30 m). Well sites were kept as unobtrusive as possible to avoid molestation and creation of eye sores.

Water sampling protocol

Sampling was accomplished during the weeks of February 22, May 8, August 9, and November 15, 1993, and will hereafter be referred to in the above order as sampling rounds 1, 2, 3 and 4, respectively. After locating a well site and anchoring the boat, a diver would locate the wellhead, remove the end cap, and fix the 5/8-inch ID hose to the wellhead. Each well was purged for 5 minutes (approximately 5 casing volumes for a 35-ft, 10.7 m well). After purging, temperature and conductivity were measured in the field using, an Orion model 122 conductivity meter. The measurements were made in a 1-liter plastic beaker while the pump was running. After purging and temperature and conductivity measurements were completed, the hose was disconnected from the impeller pump and attached to a 1/4-inch-diameter silicone tubing using a brass coupling. The silicone tubing is an integral part of a portable 12-VDC peristalic pump. The outlet end of the same length of silicone tubing was attached to an acrylic filter unit containing a 142-mm-diameter cellulose nitrate Millipore filter. The pore diameter of the filter is 0.45 um. Water was first pumped through the filter to remove air. After 500 ml of water had been flushed through the filter and discarded, samples were filtered directly into pre-cleaned, pre-labeled bottles. A total of 7 bottles was filled, three of which were not filtered. Sizes, types and purpose for each are described below:

  1. Dissolved nutrients, 125-ml amber polyethylene bottle (for USGS laboratory).
  2. Dissolved nutrients, 125-ml amber polyethylene bottle (for NOAA Undersea Research Laboratory).
  3. Total nutrients, 125-ml amber polyethylene bottle (water for this sample was not filtered through the cellulose nitrate filter).
  4. Dissolved solids and chloride, 500-ml clear polyethylene bottle.
  5. Total organic carbon (TOC), 125-ml glass bottle with teflon-lined cap (not filtered).
  6. Dissolved organic carbon (DOC), 125-ml glass bottle with teflon-lined cap (filtered using a 0.45um silver filter).
  7. Fecal coliform and fecal strep (one bottle), 250-ml sterile clear plastic bottle.

Bottles and caps for dissolved nutrients were rinsed twice with the water sample and filled with 100 ml of sample water. USGS samples were preserved with one 1/2-ml ampule of mercuric chloride/sodium chloride. The duplicate sample for the NOAA laboratory was not preserved with mercuric chloride. Both bottles were immediately placed on ice in an ice chest.

After disconnecting the tubing from the filter unit, unfiltered samples for total nutrients were placed in 125-ml amber bottles after rinsing bottle and cap twice in the sample. One 1/2-ml ampule of mercuric chloride/sodium chloride was added and the bottle was sealed and placed on ice.

For total organic carbon (TOC), the 125-ml glass bottle was filled (bottle not rinsed) and a 1-ml ampule Of H2SO4 was added before sealing the bottle. Ph of the water after addition of the acid was less than 2.

Samples for dissolved organic carbon (DOC) were filtered through a 0.45-um silver filter. The replaceable filter was sealed inside a pressure-proof stainless steel filter unit. After rinsing the unit in deionized water with a new filter seated in the bottom, the top of the filter unit was unscrewed and the filter unit was filled three-quarters full with sample water. The top portion was then reattached and the silicon tubing from the peristalic pump connected to the unit. The peristalic pump was used only to create a positive pressure to force the water out through the silver filter. The first 10-25 ml of sample were discarded. The 125-ml glass bottle was then filled with the filtered water leaving enough head room for addition of acid as described above for TOC samples.

H2S in the ground waters sampled produced silver sulfide, which darkened the silver filter. Discoloration produced colors depending on H2S concentration, ranging from silver (low H2S) through light gold to dark gold then to various shades of gray and finally coal black. High concentrations of H2S encountered in many wells turned the filters black. Field notes describing the degree of discoloration were kept as a means of estimating H2S concentration. During the last quarterly sampling round, a simple field kit was used to determine H2S concentrations. The method uses a plastic container with a perforated lid holding a replaceable copper sulfate saturated paper disk. The color of the disk is compared to a standard color chart provided by the manufacturer (HACH model HS-C). These tests confirmed and provided numerical values for our previous impressions, which were based on odor, discoloration of drilling rods and staining of the silver filters used for DOC analyses.

Surface sea water was collected at selected drill sites using the same protocols used for the well water. Surface seawater samples were coded with the same field identifications used for well water but included "SW" in the identification. A plastic screen was placed over the intake hose for surface samples to avoid intake of seaweed or other debris in the water column.

Duplicate samples were taken at selected sites after changing filters. The same identification numbers were used for these samples except that "DUP" was added to the identification.

Eight equipment blanks and 2 field blanks were run using the same DI water used for field cleaning. The equipment blanks tested the 5/8-inch-ID plastic tubing, silicone tubing and filter units by taking a sample of de-ionized (DI) water using the same procedures as used for environmental samples. The field blanks were a test of the DI water. Field blank procedure consisted of pouring DI water directly from its container in the field into a sample bottle of the same type as used for environmental samples. Samples were treated using the same procedures as for well and sea water described above. Results of blank sample analyses will be discussed later.

Duplicate samples for nutrient analysis were provided to the NOAA/National Underwater Research Center field station on Key Largo and were stored frozen until analyzed.

Bacterial analysis

Fecal coliform and fecal streptococcal bacteria analyses were conducted in the field. Analyses were conducted in the afternoon or evening of the sampling day within 6 hrs of the time each sample was collected. The membrane-filter method as described in Greeson et al. (1977) was used. The membrane-filter method is the standard used by the American Public Health Association and others (1976). The method estimates the number of bacteria filtered from 100 ml of sample and is based on counting colonies, which grow on a special medium after 24 hrs of incubation for fecal coliform or 48 hrs for fecal streptococcal bacteria. The tests were performed by a different technician for each of the 4 sampling runs.

Laboratory analysis

Analyses were performed in the USGS analytical laboratory in Ocala, Florida, whose CompQAP number is 910161G with annual amendments approved on 12/3/92. Results are expressed as mg/L, the standard used in groundwater investigations. The parameters analyzed and the analytical methods used are listed in Table II. The methods in Table II are detailed in Fishman and Friedman (1989).

The parameters analyzed in the laboratory and field and provided in the following order in Table IV were:

  1. Specific conductance (µS/cm), measured in the field.
  2. Dissolved solids (ROE at 180oC expressed as mg/L). Dissolved solids can be expressed as salinity (ppt) by moving the decimal point 3 places to the left.
  3. Dissolved chloride (mg/L).
  4. Water temperature (oC), measured in the field.
  5. MBAS total (mg/L). MBAS is an analysis for detecting a component in washing detergents. MBAS analysis was conducted for sampling round 1 only.
  6. Dissolved organic carbon (mg/L as C).
  7. Total organic carbon (mg/L as C).
  8. Dissolved phosphorous (mg/L as P).
  9. Dissolved orthophosphate (mg/L as P).
  10. Total phosphorous (mg/L as P), last two sampling rounds only
  11. Dissolved NO2 (mg/L as N).
  12. NO2+NO3 (mg/L as N).
  13. NH4+ORG-N (mg/L as N).
  14. Total NH4+ORG-N (mg/L as N), last two sampling rounds only
  15. Dissolved NH4-N (mg/L as N).
  16. Fecal coliform bacteria (colonies/100 ml), determined in the field.
  17. Fecal streptococcal bacteria (colonies/100 ml) determined in the field.

Duplicate samples for dissolved nutrient analyses (NO3+NO2, NH4 and PO4) from round 1 were run within 30 days at the NOAA National Underwater Research Center (NURC) field facility on Key Largo. Rounds 2, 3, and 4 were run (after round 4 was collected) at the Florida International University laboratory, which has a cooperative agreement with the NOAA/NURC facility. Results from this laboratory are expressed in molar units, the standard used in oceanographic and biological investigations. Results can be converted to mg/L or vice versa. These analyses are given in Appendix C.

Rock chemistry

Selected subsamples of cores were analyzed for key elements to test the possibility that phosphates could be precipitating in subsurface limestone and thus could be removing phosphorous from the ground water. The analyses were performed at Pennsylvania State University using the induction coupled plasma spectrography method (ICP). Analyses were made on small samples @ 10 grams. Samples were from internal or secondary sediments that had either infilled or precipitated in voids. The bulk of the samples was from discolored or otherwise altered rock, internal sediments or material associated with unconformities. Samples of unstained white grainstone and coral were analyzed for comparison. Thirty representative samples were analyzed. Data presented as either weight percent or parts per million (ppm) are provided in Table III. SPB in Table III is a core from Sprigger Bank in Florida Bay.

Core description and porosity analysis methods

All cores were described in the St. Petersburg laboratory using the combined carbonate classifications of Dunham (1962), Scholle (1978), and that used by Perkins (1977) in his study of Pleistocene limestone in south Florida. Graphic core logs for each well are presented in Appendix B.

Porosity of selected core sections was determined by the water displacement method (Gilbert, 1984). It should be pointed out that the bulk of these Pleistocene limestones is the most porous and permeable type of rock on the planet. When the core bit encounters zones of high porosity, where the leached voids approach or exceed the diameter of the core bit, core recovery is practically nil. In such zones the samples either are not recovered (voids cannot be sampled) or loss of integrity causes the rock to break and be pulverized to fragments too small for recovery. It is not uncommon, therefore, for an entire 5-ft (1.5-m) run of the core barrel to come up empty. Often the 5-ft core barrel retrieves only 1 to 2 ft (0.3-0.6 m) of sample. The only uncemented material encountered was quartz sand. The drill and the coarse carbide bit used for these wells make it easy to "feel" the difference between cemented limestone and uncemented sediment during drilling. In these leached Pleistocene limestones, poor or no recovery is therefore considered a direct indication of extreme porosity and permeability as opposed to uncemented sediment that would have lower permeability. Zones of no recovery are indicated in the core descriptions in Appendix B.

When relatively dense zones are encountered, a full 5-ft core may be recovered. Porosity can be determined only on cores recovered in the barrel, thus creating bias toward lower porosity analyses. This bias, however, is not considered a deterrent in this study since virtually all the rocks in the Florida Keys are very porous and permeable by most standards. The only impermeable rocks are thin, laterally extensive zones associated with unconformities. Investigation of the confining effect of these extremely low-porosity and permeability zones on fluid flow led directly to initiation of this study. Selected porosity measurements are provided in the graphic core logs in Appendix B.





| Disclaimer | Privacy Statement | Accessibility |

U.S. Department of the Interior, U.S. Geological Survey
This page is: http://sofia.usgs.gov/publications/ofr/94-276/methods.html
Comments and suggestions? Contact: Heather Henkel - Webmaster
Last updated: 10 September, 2014 @ 12:20 PM (KP)