Interpretation of Water Chemistry
Interpretation of Water Chemistry Data from a Florida Underwater Cave for Determination of Possible Groundwater Pollution
Harris Martin, Ph.D
by Roger Herring
It was a beautiful weekday in January 1997. I had called the office and given some flimsy excuse for not coming in. I had planned to just lay around the lake and soak up some sun.
Around 8:30 that same morning, Ken Sallot of the Woodville Karst Plain Project (WKPP) team called me and asked if I had heard the news regarding WKPP science director Peter Gomez. I told him no, but that I had just spoken to him several nights ago regarding some water analysis we were doing.
“He’s dead,” Sallot somberly told me. I was stunned. For reasons that may never be known, my new friend had elected to take his own life — a permanent solution to what probably were temporary problems, as WKPP director George Irvine so aptly put it.
I had met Peter, a promising young doctoral candidate at the University of Florida, the previous year while diving support for the WKPP. Peter, myself, and several other divers soon began working together in an effort to provide water chemistry data to various universities, park officials and other state and federal entities concerned with ground water quality and its protection.
Our goal was to monitor water quality in the WKP during clear and dark water intrusions and to create a historical database that could be reviewed and studied by one and all. One of the scientific goals surrounding the research and exploration efforts in the WKP is understanding the hydrology, geology and biology of the Wakulla river drainage basin. Wakulla spring is believed to be the primary discharge point for groundwater entering the basin through either direct recharge or cross basin connections. Wakulla, Sally Ward, the Leon Sinks system and Indian are all within the WKP and represent the most famous system in Florida and potentially the world.
Additionally, we had hoped that in providing this type of data, we could somehow educate the general public on how important and sensitive this unique karst system is and that its protection is of the utmost importance.
Fortuitously, I work for an analytical laboratory that has an environmental section capable of handling organic, inorganic and microbiological water analysis. This allows me to follow the WKPP samples through the system, view all QA/QC data associated with the samples and to maintain a database of the analytical results.
Peter and I enlisted the help of several top karst and environmental scientist in determining the analytical parameters to monitor at multiple sites in the WKP on an ongoing basis.
A special note of thanks is offered to the family of Peter Gomez. After his death, his family allowed me to go through his files and computer in order to recover data that Peter had previously collected. For that the WKPP is eternally grateful.
One of the most frequent questions I receive is what parameters are we testing for and why are they significant to understanding phreatic cave systems.
I decided to enlist the help of noted karst scientist Dr. Harris Martin. I sent Dr. Martin actual analytical results of a water sample taken in a Florida cave without telling him the location (other than it was somewhere in Florida) and asked him to explain in layman terms the significance of the results and the parameters.
Following is his interpretation of the data:
Roger Herring of ABC Research Corp. in Gainesville, Florida, has asked me to describe for the non-scientist, why we analyze water chemistry in underwater caves and what some of the data mean. After briefly describing why we do all this, I will try to explain some water chemistry data that Roger has collected from an underwater cave in Florida.
A very important aspect of karst is the chemistry of karst groundwater. This is because the chemistry of the groundwater can tell us a lot about the geology, hydrology, and biology of the phreatic (below the water table) part of the karst system. In addition, some water chemistry parameters are indicative of pollution. In karst that contains divable underwater (phreatic) caves, cave divers can collect water samples for field and laboratory analysis. The things that chemists analyze for in water are called parameters. Most of these, including the dissolved minerals, are expressed as concentrations, i.e., milligrams per Liter (mg/L) which is the same as parts per million (ppm). We only use metric units for this sort of thing. Some important parameters such as turbidity, color, conductivity, and temperature, have other units. One very important parameter, pH, has no units.
Most parameters can be measured in the chemical analysis laboratory in water samples that have been collected by cave divers in the field. For some of these parameters, water samples need to be preserved in the field (right after divers remove the samples from the cave and exit the water) by adding a “fixative” such as a few drops of concentrated acid. Some parameters need to be analyzed in the field because they might change before the water sample reaches the laboratory. These include dissolved oxygen, pH, temperature, and alkalinity.
In Roger’s example cave water sample, the calcium (Ca) concentration is 29.2 mg/L. Dissolved calcium in water exists as a cation, a positively charged ion. Most dissolved minerals exist as cations with one or more positive charges per ion. Dissolved calcium and magnesium (Mg) exist as divalent cations, meaning that each ion has two positive charges. Other cations such as sodium (Na) and potassium (K) have only one positive charge per ion. Most of these dissolved cations can be analyzed for simultaneously in the laboratory with a machine called an ICP (inductively coupled plasma emisssion spectrometer). In karst waters, calcium is nearly always the dominant cation. The 29.2 mg/L calcium concentration in this sample is actually a little low when compared to data collected from other karst systems.
Magnesium (Mg) is the second most important dissolved mineral cation in karst waters. Like calcium, dissolved magnesium exists as a divalent cation with two positive charges per ion. In karst groundwater, freshwater surface waters, and soils, calcium concentrations are almost always greater than magnesium concentrations. However, in ocean water and in saltwater caves (saline karst groundwater) such as those found in the Bahamas and the Yucatan, magnesium concentration usually exceeds calcium concentration. The magnesium concentration in this sample is 7.23 mg/L. This is fairly typical but just a little on the low side for karst waters. The calcium to magnesium ratio (Ca/Mg) in karst groundwater is an important indicator of the type of limestone the water is flowing through and provides other information on the geochemistry of the phreatic karst system. Geochemistry is the field of science that examines the interaction of rocks (geology) with water and air (chemistry), i.e., geology + chemistry = geochemistry.
The third and fourth most common dissolved mineral cations in karst groundwaters are sodium (Na) and potassium (K). The chemical symbols for these mineral elements do not match the English words because they are based on the German words for the elements, natrium and kalium. Sodium and potassium both exist as monovalent cations when dissolved in water. This means that they both have one positive charge per ion. In Roger’s cave water sample, the sodium concentration is 4.51 mg/L and the potassium concentration is 0.32 mg/L. These are fairly typical values for karst groundwater. Sodium concentration increases with increasing saltiness with seawater having over 2,000 mg/L sodium. Dissolved potassium in karst groundwater is usually quite low and is usually not very important except for completing a full inventory of dissolved ions.
In unpolluted karst groundwaters, a number of dissolved minerals exist naturally in only trace amounts. These include iron (Fe), manganese (Mn), aluminum (Al), barium (Ba), titanium (Ti), and vanadium (V) which Roger has measured for in his sample. All of these elements exist as cations when dissolved in water and all can be measured in the laboratory on an ICP machine. In Roger’s cave water sample, the iron (Fe) concentration is less than 0.39 mg/L. This means that the government-mandated quality control (QA/QC) protocols used by commercial laboratories for environmental regulatory work, do not permit the iron concentration in this particular sample to be measured at less than 0.39 mg/L. Academic research laboratories are not subject to these government-mandated QA/QC protocols so can measure dissolved iron in water at lower concentrations, down to a “limit of detection” (LOD) that is a function of the type of analytical instrumentation and the type of “interfering substances” in the water sample. Dissolved iron is usually not considered very important in karst groundwaters but may be important in the formation of goethite deposits which are largely unique to underwater caves. The dissolved aluminum concentration in Roger’s sample was 0.44 (rounded off from 0.437) mg/L. This is high for karst groundwater so Roger an I are a bit perplexed by it and are looking into why it might be this high. The chemistry of iron and aluminum in water, soil, and sediments is too complex to get into here. Both elements participate in many reactions with water, dissolved oxygen, and rock and soil surfaces. The barium (0.0082 mg/L), titanium (less than 0.006 mg/L), and vanadium (less than 0.008) concentrations in this sample are quite low and are of little consequence.
Dissolved nitrogen (N) in groundwater is important as an indicator of groundwater pollution. The various forms of nitrogen are very important in biological processes and their concentrations in water are influenced by plants, animals, and microbes. Nitrogen in water can exist in organic form (combined with carbon atoms to form organic molecules) or inorganic form. There are three dissolved inorganic forms of nitrogen in water, ammonium, nitrate, and nitrite. Nitrite (NO2-) is rare except in water or soil with very high pH values, so we usually don’t bother measuring it in karst groundwater. Ammonium is NH4+ so is a monovalent cation. When ammonia gas (NH3) is dissolved in water, it takes a hydrogen atom from water to form NH4+ and OH-. In water, we are always talking about ammonium, not ammonia. The ammonium-nitrogen concentration (only the N in the NH4+ is reported in the concentration value) in Roger’s sample was less than 0.008 mg/L so is of little consequence.
Nitrate is NO3-, so is a monovalent anion. The concentration of nitrate-nitrogen in Roger’s sample was 0.282 mg/L. This is a normal nitrate concentration for unpolluted natural waters. Higher concentrations of nitrate, say more than 4 or 5 mg/L, are indicative of pollution from sewage, fertilizer runnoff from farms or suburban yards, or animal manure. More than 10 mg/L nitrate in drinking water is considered by the federal EPA and state environmental regulatory agencies to be harmful to infants because it interferes with the ability of infant blood to carry oxygen. Adults can handle much higher concentrations of nitrate in drinking water than infants can.
Ammonium and nitrate in water can not be measured with an ICP machine. They are usually measured in a machine called an autoanalyzer. Organic nitrogen in water is also an indicator of organic pollution and is measured with a process called a Kjeldahl digestion. The total Kjeldahl nitrogen in Roger’s sample was less than 0.50 mg/L and does not indicate a pollution problem. Dissolved organic nitrogen in water is usually calculated by subtracting the ammonium- and nitrate-nitrogen concentrations from the total Kjeldahl nitrogen concentration.
Phosphorus (P) concentrations are usually very low in unpolluted fresh surface and groundwaters, well below 0.1 mg/L. Higher concentrations are often indicative of nutrient enrichment and/or pollution. Elevated phosphorus concentrations in surface waters can cause “algal blooms” that can “choke out” plant and animal species that require clear, clean water. In groundwater ecosystems, the negative effects of elevated phosphorus concentrations are less severe but not well documented.
Orthophosphate, sometimes called “soluable reactive phosphorus,” consists of dissolved inorganic phosphorus and is usually measured on the same autoanalyzer machine used to measure ammonium and nitrate concentrations. Total phosphorus is determined by one of several “digestion” methods and is supposed to measure both inorganic phosphorus and organic phosphorus compounds in solution. Roger’s sample analysis indicates that this cave water has 0.24 mg/L of orthophosphorus but less than 0.12 mg/L of total phosphorus. This sounds contradictory but I have sometimes gotten similar results when I studied water in wetland soils. There is sometimes a problem with one or both of these phosphorus methods at low concentrations. The analytical chemistry researchers need to address this issue. The orthophosphate concentration of 0.24 mg/L that Roger found is a bit high for pristeen waters, so may be a reason for concern.
Roger’s result of less than 1 mg/L of total inorganic carbon (TOC) does not tell us much because the “method limit of detection” is so high. If TOC in the groundwater where high, it would be an indication of pollution.
Alkalinity is a very important parameter in karst water chemistry. It is a measure of the relative concentration of carbonate (CO32-) and bicarbonate (HCO3-) in the water. It is the carbonate concentrations that largely determine whether speleothems will form in caves and whether limestone will dissolve to form caves or enlarge existing caves under water. In combination with other water chemistry parameter, alkalinity is critical in calculating a variable called saturation index for the mineral calcite (CaCO3) and dolomite (CaMg)2CO3)). Alkalinity is best measured in the field at the water’s edge with a mobile field titration set-up. If water samples are shipped to the lab for alkalinity analysis, shipping must be rapid and analysis done immediately. Roger’s sample had an alkalinity of 89.2 units. This is normal for karst groundwater.
Chlorine (Cl) exists as the chloride anion (Cl-) when in solution. High chlorine concentrations are an indication of brackish water but not necessarily of pollution. It is particularly important to measure chloride concentrations in coastal and anchialine (with a halocline) caves. Roger’s sample contained less than 1 mg/L chloride, so this water is very fresh. Fluorine (F) also exists as the fluoride anion (Fl-) when in solution. Most karst waters are very low in fluoride, so a high fluoride concentration could be an indication of industrial contamination. Roger’s sample contained less that 0.10 mg/L fluoride so no fluoride contamination is indicated.
Sulfate (SO42-) is the oxide of sulfur (S) and is a divalent anion. It can come from dissolved gypsum (CaSO4.2H2O) or anhydrite (CaSO4) that exist as impurities in some dolomites (dolomite is a very high-magnesium limestone) found deep in the Florida aquifer or can be produced by microbial oxidation of sulfide. High sulfate concentrations in karst groundwaters are usually due to natural causes, not pollution, although sewage waters can be high in sulfate. The sulfate concentration of 13 mg/L in Roger’s sample is unusually high for fresh karst groundwaters. This sulfate concentration is similar to that we have found at Orange Spring, one of Florida’s best known “sulfur springs.” Karst groundwaters with this much sulfate are usually rich in white sulfur bacteria, the “cave slime” which I use as one of my email addresses. The cave Roger sampled is probably naturally high in sulfate but if it is near a chemical factory, sewage treatment plant, cattle feed lot, or dairy farm, organic pollution may be suspected.
Turbidity is caused by suspended particles in the water, not dissolved minerals. Cave divers are acutely aware of turbidity because it increases when we kick up any silt. Turbidity is measured optically by how much visible light can pass through a water or gas sample. In the field, this can be done with measured line and a secci disk. Peter Horne of the South Australian Underwater Speleological Society described how to do this in his Research Manual for Cave Divers. Natural turbidity in some underwater cave systems is hydrologically significant because it indicates that sediment is being stirred up in the cave naturally, somewhere upstream of the point of observation or the cave system is receiving turbid surface water at an upstream sinkhole (karst window — if the surface water input into a sinkhole is substantial, that sinkhole is referred to as a swallet hole). The turbidity in Roger’s sample was less than 0.10 NTU (turbidity units). This is low, indicating fairly clear water.
Water color can be effected by a number of factors including dissolved and suspended organic compounds such as “tannins” which are abundant in many surface waters found in Florida and other coastal plain areas around the world. Water color is measured optically and reported with color units. The color of Roger’s sample was less than 5.0 color units. The EPA standard for color in drinking water is 15 color units, so 5.0 is well below the current limit.
Total suspended solids is similar to turbidity but is measured gravimetrically (suspended solids are separated from the water and weighed). Roger’s sample contained less than 2.50 mg/L of suspended solids so again, this water is fairly clear. This value could be interpreted better if a lower limit of detection had been achieved by the measurement.
Total dissolved solids consists of all dissolved ions and molecules in the water. This value compliments those of color, total suspended solids, and conductivity. In Roger’s sample, total dissolved solids was 140 mg/L. Compared to data reported in McColloch’s 1986 Springs of West Virginia, this is a fairly typical value for karst groundwaters.
Conductivity is caused by dissolved minerals (salts) in water. As the dissolved mineral concentration in water increases, the amount of electricity that water can pass increases. The opposite (inverse) of electrical conductivity is electrical resistance, traditionally measured in ohms. An electrical current is passed through a small water sample and rather than measuring conductivity, resistance is measured and corrected for temperature effects. Traditionally and in commercial laboratories, this resistance value is converted to a unit called micromhos per centimeter (uMohs/cm). In academic laboratories, a newer unit adhering to the metric Systeme Internationale (SI) units, deci-semens per meter, is used to express conductivity. Like chloride and sodium concentrations, high conductivity indicates brackish water. Seawater has a very high conductivity. Roger’s sample had a conductivity of 210 micromohs per centimeter, a fairly typical value for karst groundwaters. Most of the springwater chemistries in Rosenau et al.’s Springs of Florida show conductivities in the 200 to 350 uMohs/cm range. Higher values are indicative of salt water intrusion or geothermal water sources.
Hydrogen sulfide (H2S) is only present in measurable amounts in “sulfur springs” and sulfurous karst groundwaters which are scattered throughout Florida and other karst regions. These springs usually have a marine or geothermal influence or are receiving shallow groundwater from surface wetlands. Sulfurous karst groundwaters can also be the result of organic pollution from sewage or animal manure. In karst groundwater, most dissolved sulfide exists as the HS- anion rather than dissolved H2S gas. Measurement of hydrogen sulfide concentrations in water (which includes detection of the HS- anion) requires special preservative to be added to the sample in the field before transport to the lab, to prevent H2S from de-gassing or H2S or HS- being transformed by bacteria to sulfate (a biological oxidation process). Sulfide causes the “rotten egg” smell sometimes observed in coastal marshes. Dissolved sulfide can stain divers’ brass gear black and prolonged contact can damage regulators. Very high concentrations of dissolved sulfide can cause divers to feel nauseous and in extreme cases, can be dangerous to divers. The hydrogen sulfide concentration in Roger’s sample was 0.48 mg/L which is compatible with the 13 mg/L sulfate concentration. Depending on the sulfur input source and type, the “Oxidation-reduction potential” of the water and the level of biological activity, the sulfide can be transformed by oxidation to sulfate or the sulfate can be transformed by reduction to sulfide.
The pH of water is a measure of its acid and base content. Technically, pH is the negative base-10 logarithm of hydrogen ion (H+) activity (concentration) so pH has no units. Low pH means acidic; high pH means basic; pH 7.0 is completely “neutral.” Karst groundwaters are usually “circumneutral” (around neutral pH of 7.0) and fall in the in the 6.7 to 8.0 range. The pH of 7.81 in Roger’s sample is definately in the high end of the normal range. This may be due to active biological (bacteria mostly) “reduction” of sulfate to sulfide, a reaction which tends to raise water pH.
Temperature of karst waters must be measured in-situ (in the field, in the cave, and underwater) and for science work, is reported in degrees centigrade or celcius. In-situ water temperature measurements are very usefull for identifying infeeder tunnels with differing water sources. Until cave diver-deployable remote water chemistry sondes came along, all we had for measuring water temperature in-cave were small hand held mercury thermometers which are generally only accurate to plus or minus 2 degrees fahrenheit, or digital dive watches with built-in thermometers. As far as I know, no information is available on the accuracy or precision of these dive watch thermometers. The temperature of 24.4 degrees C of Roger’s cave water is equal to 75.9 degrees fahrenheit (degrees C X 9/5 + 32 = degrees fahrenheit). This is unusually warm for a Florida spring or underwater cave because normally, groundwater temperature is within one degree centigrade of the average annual temperature for that area. Most north-central Florida caves and springs have temperatures in the 70 to 73 fahrenheit range (21.1 to 22.8 degrees centigrade) while panhandle Florida caves and springs generally range from 67 to 70 degrees fahrenheit (19.2 to 21.1 degrees centigrade). Cave water temperatures greater than 73 degrees fahrenheit indicate either some geothermal influence or summer-time input of warmer surface waters through an upstream sinkhole.
Dissolved oxygen in groundwater is crucial to the survival of troglobitic (specially cave-adapted) organisms such as cave crayfish, amphipods, and isopods. Aquatic (water-dwelling) cave organisms can tolerate fairly low oxygen concentrations in the water but if it gets too low, they will try to move to waters with higher oxygen concentration or will die. Oxygen concentration in water can be expressed as mg/L concentration or as percent saturation which is a function of water temperature and oxygen concentration (colder water can dissolve more oxygen than warmer water). While shallow groundwater in wetlands and saturated soils is usually very low or devoid in dissolved oxygen (“anaerobic”), deeper groundwater is usually “aerobic” and contains more than one mg/L dissolved oxygen. Borderline waters are called hypo-oxic, containing low but measurable concentrations of dissolved oxygen. Dissolved oxygen in karst groundwater is best measured in-situ with a remote water chemistry sonde. The second best method is to measure dissolved oxygen with an oxygen meter at the water’s edge, as soon as the diver gets out of the water with the water sample. Laboratory measurements of dissolved oxygen in water samples are not valid because dissolved oxygen concentration of a sample can change during transport to the lab and storage in the lab prior to analysis. The 1.8 mg/L dissolved oxygen concentration in Roger’s cave water is a little low but adequate for survival of most aquatic trolobites.
Summary and Conclusions
The data from Roger’s water sample, particularly the relatively high concentrations of aluminum, orthophosphate, sulfate, and sulfide, along with the high pH, high temperature, and relatively low dissolved oxygen concentration, indicate that one of three things is happening to this underwater cave. Either this cave is 1) receiving a little natural geothermal water from very deep in the Florida aquifer, 2) it is receiving natural “bad water” from a surface stream or shallow groundwater in adjacent wetlands, or 3) it is being polluted by sewage or agricultural manure runnoff. To determine which one of the possibilities is correct, the geology of the area around the cave should be examined to look for faults that might allow up-welling of deep geothermal water; coliform bacteria concentration of the water should be determined to confirm or rule out pollution by sewage or farm manure; the cave drainage basin should be examined for the presence of upstream sinkholes or sinking streams (swallet holes); and possible upstream organic pollution sources such as sewage treatment plants, industrial facilities, dairy farms, or cattle feed lots should be identified.
This is a classic example of how cave divers can collect valuable environmental data that conventional investigations would miss. I hope that Roger and his collaborators collect additional data to confirm or disprove my hypotheses, then turn all of these data over to the Florida DEP for regulatory action if my hypotheses are not disproved. In addition, this type of underwater cave water chemistry data should be published in the peer-reviewed scientific literature (journals, etc.).