2.1 Improved Analytical Methods Provided Insight into the Occurrence and Behavior of DOC

In 1970, a team of four U.S. Geological Survey hydrologists led by Jerry A. Leenheer began sampling wells throughout the United States and analyzing the groundwater for concentrations of DOC. Up to that time, there had been relatively few attempts to analyze groundwaters for DOC, primarily because it was difficult to measure the low concentrations that were typically encountered. Leenheer’s team (R.L Malcolm, P.W McKinley, and L.A. Eccles) developed a method for the “wet oxidation” of filtered groundwater samples using sodium persulfate as an oxidant in a pressurized vessel at 175 °C followed by acidification and quantification of the carbon dioxide produced with an infrared analyzer, as illustrated in Figure 5. As improvement of analytical methods allowed detection of much lower concentrations of dissolved organic carbon, better field observations and laboratory experiments were conducted so, the occurrence and behavior of DOC was better understood.

Figure showing the steps in quantifying low levels of DOC in water

Figure 5  The steps in quantifying low levels of DOC in water as described in the text. The DOC is oxidized to dissolved inorganic carbon (DIC), then the DIC is converted to dissolved CO2 by acidification and the CO2 is vented to a detector capable of quantifying total CO2 evolved (Mackay, 2022).

This analytical methodology provided a lower detection limit for DOC of 0.1 milligrams per liter (mg/L), which is 8.3 micromoles per liter (µmol/L). The team collected and analyzed 100 groundwater samples from 27 states within the USA and from aquifers of differing lithologies that included sandstones, limestones, crystalline rocks, as well as shallow (< 200 feet or < ~60 m), and deep sand and gravel (> 200 feet or > ~60 m). Their working hypotheses included the possibility that different aquifer lithologies and different sample depths might exhibit DOC concentrations.

Based on that study, Leenheer and others (1974) made four principal observations:

  1. DOC concentrations in groundwater were generally much lower than commonly found in surface waters, with median concentrations ranging between 0.5 and 0.7 mg/L (~40 to ~60 µmol/L);
  2. a shallow aquifer in Florida receiving active recharge from surface-water sources had much higher DOC concentrations (15 mg/L; ~1,250 µmol/L) than wells tapping the deeper Floridan aquifer (1.4 to 0.1 mg/L; ~120 to ~8 µmol/L) that was not immediately influenced by surface water;
  3. there were no statistically significant differences in DOC concentrations between aquifers of different lithology; and,
  4. there were clear statistical correlations between DOC concentrations, specific conductance (the ability of the water to conduct electricity, which increases with increasing concentration of dissolved ions), and alkalinity (the ability of the water to resist acidification, generally related to the content of bicarbonate ion, HCO3, in water with near neutral pH).

Leenheer and others (1974) concluded that virtually all groundwaters contained low but measurable concentrations of DOC, that there were a variety of possible allochthonous (upgradient) and autochthonous (local) sources for DOC, and that DOC seemed to be involved in geochemical reactions with minerals present in aquifers that produced both dissolved solids and alkalinity. What was not clear was: the nature of those geochemical processes; how they affected the chemical composition of groundwater; and their hydrologic and ecologic significance.

Identifying those processes became the principal focus of groundwater DOC research in the years that followed. From the beginning, however, these studies had to overcome a fundamental difficulty. There are literally thousands of possible combinations for the organic carbon compounds present in DOC, even if consideration is limited to its three main elemental components carbon, hydrogen, and oxygen (Hertkorn et al., 2006). That, in turn, limits the utility of direct elemental analysis for characterizing the chemical and biological properties of DOC. For that reason, studies of both groundwater and surface water have been based on indirect analytical methods for characterizing the properties of DOC.

The first such indirect method was categorizing DOC based on its humic and fulvic acid composition, as illustrated in Figure 6 and discussed after the figure.

Figure showing the initial steps in characterizing soil organic matter for agricultural purposes

Figure 6  Illustration of initial steps in characterizing soil organic matter for agricultural purposes. Humic and fulvic acids are extracted from soil by a strong base, then separated by acidification and settling. In the analysis of groundwater, this distinction is less useful since total DOC is generally low (Mackay, 2022).

Historically, characterizing organic matter in soils for agricultural purposes was based on extracting soils with strongly basic solutions (pH ~13) of sodium or potassium hydroxide (Achard, 1986). The resulting basic solutions were then acidified to a pH of 1 using hydrochloric acid, causing one fraction of the organic carbon, operationally defined as humic acids, to precipitate from solution. The fraction that did not precipitate was termed fulvic acids. Humic acids have a higher molecular weight (> 2000 unified atomic mass units, AMU) and are less water-soluble whereas fulvic acids have a lower molecular weight (800 to 2000 AMU) and are more water-soluble. While of historical interest, the humic-fulvic acid dichotomy is less useful for DOC present in groundwater, primarily due to the typically low concentration of DOC in groundwater (Thurman, 1985).

In a study of DOC concentrations in soil water, surface water, and groundwater of forested watersheds of the Adirondack Mountains, Cronan and Aiken (1985) showed that soil water collected in lysimeters from the A soil horizon (10 cm depth) had DOC concentrations of 21 to 32 mg/L (~1700 to ~2600 µmol/L), 5 to 7 mg/L (~400 to ~600 µmol/L) in soil water from the B soil horizon, and 2 to 4 mg/L (~170 to ~330 µmol/L) in shallow groundwater. They also found that the DOC consisted of both hydrophobic and hydrophilic fractions. This study clearly showed that DOC produced at the land surface from organic detritus had initially high concentrations and that those concentrations decreased markedly with depth. This study also hypothesized, as illustrated in Figure 7, that adsorption and microbial degradation processes were the major processes leading to the observed DOC removal.

Figure showing DOC concentration decreases as water infiltrates from land surface

Figure 7  DOC concentration decreases as water infiltrates from land surface to underlying groundwater due to adsorption to POC and AOC as well as biodegradation to metabolites. Mechanisms that decrease DOC concentrations as water infiltrates are: adsorption of DOC to POC, AOC and mineral surfaces; and, biodegradation to metabolites, including DIC. The weights of the arrows illustrate that adsorption may be less significant with depth because the remaining DOC is more hydrophilic with depth, and biodegradation may decrease with depth as the remaining DOC becomes less bioavailable (more resistant to biodegradation) (Mackay, 2022).

This adsorption/biodegradation hypothesis for DOC removal in groundwater was substantiated by later studies. Qualls and Haines (1992) observed that as much as 95 percent of DOC leached from leaf litter was removed as the water passed through the underlying unsaturated zone. Furthermore, the relative biodegradability of the remaining DOC was also observed to decline with depth once the water reached the underlying saturated zone. That suggested that some of the observed DOC removals were due to biodegradation processes. However, incubation experiments performed in that same study removed only about 33 percent of the DOC over 134 days. That, in turn, suggested that chemical adsorption onto mineral surfaces was also an important process contributing to the observed loss of DOC, a suggestion that was also observed by Davis (1982), Baham and Sposito (1994), and Lilienfein and others (2004).

An experimental approach examining DOC adsorption from water by soils was undertaken by Jardine and others (1989). Batch experiments were performed using water collected from a surface-water stream draining a peat deposit with an initial DOC concentration of 53 mg/L (4,400 µmole/L). Upon addition to soils, it was found that DOC concentrations decreased rapidly, coming to an equilibrium concentration in two days. The amount of DOC loss from solution at equilibrium was a function of initial DOC concentration, pH, and the mineral content of the soil. The soils used in the experiments contained significant amounts of POC and ferric oxyhydroxides coating the soil mineral grains. When much of the POC was removed from the soils by chemical extraction, the amount of DOC adsorption decreased by factors ranging from 5 to 10, demonstrating that POC itself was an important adsorbing substrate for DOC. Similarly, if the ferric hydroxides were removed by chemical extraction, the amount of DOC adsorption decreased by factors ranging from 2 to 3, indicating that ferric hydroxides were also a major adsorbing substrate. Furthermore, Davis (1982) had previously shown that clay minerals such as kaolinite were also a significant adsorbing substrate for DOC. The maximum amount of DOC adsorption in the study by Jardine and others (1989) was observed at a pH of 4.5, and adsorption was unaffected by increasing the ionic strength of the solutions.

This experimental study (Jardine et al., 1989) was consistent with the field results of Cronan and Aiken (1985) that showed rapid DOC removal as water moved through the soil zone to the water table. Furthermore, the study of Jardine and others (1989) suggested that increasing amounts of particulate organic matter and ferric oxyhydroxides in the soil increased the observed DOC adsorption. Finally, Jardine and others (1989) showed that hydrophobic DOC was more efficiently adsorbed (~80 percent) during the 2-day equilibration time as compared to hydrophilic DOC (~20 percent). That is consistent with the intuitive expectation that adsorption efficiency depends upon the specific chemical properties of each DOC component, as was suggested by Aiken (1989). Later, Jardine and others (2006) described how these processes affected the transport of DOC through soils at a variety of spatial scales.

The issue of DOC bioavailability (the proportion of DOC that can be readily biodegraded) in groundwater systems was initially investigated in the context of carbon flow and the ecology of stream ecosystems (Hornberger et al., 1994; Findlay and Sobczak, 1996; Boyer et al., 1997; Baker et al., 2000). These studies showed that DOC transported to streams by shallow groundwater flow supported populations of heterotrophic bacteria. Those bacteria, in turn, formed the lower trophic levels of stream ecosystems. While the focus of these studies was primarily ecological, they also provided direct evidence of the dynamic interactions between the DOC, POC, and AOC compartments. For example, these studies revealed marked annual cycles of rapid DOC delivery to streams during spring snowmelt “flush” followed by periods of less DOC delivery during the dryer parts of the year (Hornberger et al., 1994; Boyer et al., 1997; Baker et al., 2000). That implied for much of the year, DOC was actively being produced by biodegradation of plant detritus at land surface but was largely sequestered in soils, in part by adsorption processes. During the spring “flush”, when snowmelt provided increased recharge to the shallow groundwater systems, DOC was mobilized from AOC and delivered to stream ecosystems. While not the stated purpose of these studies, these data showed that DOC adsorption could be a reversible process and that the DOC, POC, and AOC compartments were interconnected at some level.

During the 1990s, the analytical techniques available for studying groundwater DOC improved steadily. Grøn and others (1996) used a variety of analytical techniques including light absorbance, molecular weight distribution, 13C-NMR (13-carbon nuclear magnetic resonance) spectroscopy, elemental composition, and measurements of hydrolyzable amino acids and carbohydrates to characterize DOC in groundwater from three different aquifers in Denmark. Many of these analytical approaches had previously been developed for application to DOC in ocean water (Ogawa et al., 2001, Hertkorn et al., 2006, Sleighter and Hatcher, 2008) and river water (Hedges et al., 1994). Volk and others (1997) observed that 75 percent of the DOC present in Pennsylvania stream water, with concentrations ranging from 0.8 to 10.4 mg/L (67 to 866 µmol/L), was composed of humic substances. In that study, humic substances were measured by macroreticular XAD-resin chromatography (Thurman, 1985) defined as the difference between DOC concentrations prior to acidification and the effluent from the XAD-8 resin. It was also found that carbohydrates composed 13 percent and amino acids 2 percent of the DOC and were predominantly bound to humic acids.

The issue of DOC bioavailability began to attract wider attention in the 1990s. There already was a large literature devoted to assessing DOC bioavailability in the context of drinking water treatment technology (Servais et al., 1989). In addition, because the world’s oceans contain much of the DOC present on earth, chemical oceanographers extensively investigated the issue of DOC bioavailability (Dauwe et al., 1999; Amon et al. 2001; Benner, 2003; Davis and Benner, 2007). One of the observations made by chemical oceanographers was that not all the organic matter present in the oceans is carbon and that other elements, notably oxygen, nitrogen, and phosphorous are also present. So, beginning in about the year 2000, the term “DOC” in the oceanographic literature was replaced by the more general term “dissolved organic matter” (DOM). In groundwater studies, however, the term DOC continued to be used (Meredith et al., 2019), and is generally considered as being synonymous with DOM (though strictly speaking DOM is about 50 percent DOC, as discussed earlier).

Kalbitz and others (2003) evaluated the bioavailability of soil-derived DOC using laboratory incubations that measured DOC loss, carbon dioxide (CO2) production, changes in ultraviolet absorbance as well as emission fluorescence over time, and NMR spectroscopy before and after incubation. Hartog and others (2004) used incubation experiments measuring the consumption of dissolved oxygen over time as an indicator of aquifer sediment organic carbon bioavailability. Chapelle and others (2008) used cell counts, the ratio of DOC to total nitrogen, total hydrolyzable neutral sugars, total hydrolyzable amino acids, and dissolved inorganic carbon (DIC) production to compare the bioavailability of DOC present in a shallow sandy aquifer and a fractured-rock aquifer.

Beginning with the study of Leenheer and others (1974), the body of research discussed in this section leads to the view that there are three main compartments of organic carbon, DOC, POC, and AOC present in groundwater systems. Furthermore, extensive experimental evidence (Davis, 1982; Jardine et al., 1989) and field evidence (Qualls and Haines,1992; Baker et al., 2000; Jardine et al., 2006; Shen et al., 2015) indicate these compartments are not static but actively interact with each other.

License

Dissolved Organic Carbon in Groundwater Systems Copyright © 2022 by Francis H. Chapelle. All Rights Reserved.