INTRODUCTION
The distribution of radiocarbon (14C) between different components in freshwater streams has been the subject of study since the early days of radiocarbon (Broecker and Walton Reference Broecker and Walton1959). The sources of different components of dissolved inorganic (DIC) and organic carbon in riverine systems has been discussed extensively by Raymond and Bauer (Reference Raymond and Bauer2001), and more recently Barnes et al. (Reference Barnes, Butman, Wilson and Raymond2018) have given an extensive survey of the origins of dissolved organic carbon (DOC). Butman et al. (Reference Butman, Wilson, Barnes, Xenopoulos and Raymond2015) also have documented changes in the 14C ages of carbon in river systems in part due to human disturbance.
The study of 14C in freshwater systems is complex due to the many pathways that can exist in a freshwater ecosystem (Meltzer and Steinberg Reference Meltzer, Steinberg and Lange1983). In the context of this paper, we define DIC, meaning mainly carbonates and bicarbonates in solution in fresh waters. These species are generally in equilibrium, depending on pH, such that:
$${\rm{C}}{{\rm{O}}_{\rm{2}}}{\rm{ + }}{{\rm{H}}_{\rm{2}}}{\rm{O = }}{{\,\,\rm{H}}_{\rm{2}}}{\rm{C}}{{\rm{O}}_{\rm{3}}}\hskip-3pt*$$
$${{\rm{H}}_{\rm{2}}}{\rm{C}}{{\rm{O}}_{\rm{3}}}{\rm{* = }}{{\,\,\rm{H}}^{\rm{ + }}}{\rm{ + HC}}{{\rm{O}}_{\rm{3}}}^{\rm{-}}{\rm{}}$$
$${\rm{HC}}{{\rm{O}}_{\rm{3}}}^{\rm{-}}{\rm{ = H + C}}{{\rm{O}}_{\rm{3}}}^{{\rm{2}}{\rm{-}}}$$
where the left-hand species are more abundant at lower pH and the bicarbonate species predominate at the highest pH. pK1 and pK2 are the log of the two important equilibrium constants (Stumm and Morgan Reference Stumm and Morgan1995). We note also that at high pH, carbonates can also precipitate depending on the cation chemistry of the waters. This equilibrium can be specified in artificial solutions, however it is not entirely correct in natural waters. It appears that at pH>8, calculated values of CO2 greatly exceed measured values. The reasons why these differences exist is in part due to interferences by organic acids and other organic constituents which disturb the purely inorganic equilibrium (Reddy Reference Reddy1975) and also the dissociation of HCO3– to CO32– or release of CO2 (Wetzel Reference Wetzel1975).
DOC defines a wide range of species, which are operationally defined here as passing a <0.45 μm filter. They include fulvic acids, which are water-soluble acids at any pH and humic acids, which are higher molecular weight than fulvic acids and are base-soluble but acid-insoluble (Thurman Reference Thurman, Aiken, McKnight, Wershaw and MacCarthy1985). Other DOC components might include low-molecular-weight acids, carbohydrates, lipids, aromatics, plant particulates, and animal particulates (Ishiwatari Reference Ishiwatari, Aiken, McKnight, Wershaw and MacCarthy1985; Malcolm Reference Malcolm, Aiken, McKnight, Wershaw and MacCarthy1985). Several different theories have been proposed as to how humic substances form, but all assume they are polymerized in some way from fulvic acids (Malcolm Reference Malcolm, Aiken, McKnight, Wershaw and MacCarthy1985). There are many different ways to extract DOC from natural waters, including: vacuum drying and subsequent combustion of the organic residues, solid-phase extraction, UV oxidation, chemical oxidation, and ultrafiltration (as summarized by Murphy et al. Reference Murphy, Davis, Long, Donahue and Jull1985; Leonard et al. Reference Leonard, Castle, Burr, Lange and Thomas2013).
Several algae and macrophytes can process dissolved CO2 or bicarbonate in different proportions in photosynthesis. Elodea (waterweeds) and Lemna (duckweed) can take up both CO2 and HCO3– in varying proportions but algae take up mainly carbon from bicarbonate and little from CO2 (Cleland Reference Cleland1967; Meltzer and Steinberg Reference Meltzer, Steinberg and Lange1983). Aquatic plants such as Potamogeton are also known to photosynthesize carbon from bicarbonate, and not just CO2 (Meltzer and Steinberg Reference Meltzer, Steinberg and Lange1983). These processes can result in DIC being converted into DOC. Lemna and Potamogeton are found in the Sabino Canyon watershed (Shreve and Wiggins Reference Shreve and Wiggins1964).
Sabino Creek, Arizona
Sabino Creek is an ephemeral stream in Pima County, Arizona (http://web.sahra.arizona.edu/sabinocanyon/intro.html). It is the only stream in the Tucson region that flows for much of the year. Our rationale for sampling this stream is that we noted that there are very few reported 14C measurements on this stream water and this warranted further investigation. Bennett (Reference Bennett1965) reported a value for DIC collected in 1965 of 164 ± 9 percent modern carbon at close to the peak of the bomb pulse, about 175 percent modern for NH zone 2 (Hua et al. Reference Hua, Barbetti and Rakowski2013). Eastoe et al. (Reference Eastoe, Gu, Long, Hogan, Phillips and Scanlon2004) reported two values for Sabino Creek water downstream from our sample location.
Sabino Creek is characterized by flow during most of the year, with flow peaks in the mid-winter (Jan.–Feb.) and the summer rainy season (Jul.–Sept.). The average annual precipitation data for this region is shown in Figure 1. The flow station is located 9 m upstream from the Lower Sabino Dam (32°19′00″N, 110°48′35″W), which is shown in Figure 2. The mean flow rate is 0.15 m3/s with intermittent extreme flows of over 30 m3/s. The highest flow rate ever recorded was on July 31, 2006, with a transient flow of 437 m3/s, during a flash flooding event, as documented in detail by Webb et al. (Reference Webb, Magirl, Griffiths and Boyer2008). The flow rate over the period 2009–2016 is shown in Figure 3.
Annual average precipitation data for Tucson, for the period 2009–2017 derived from National Weather Service data.
A relatively high flow event (1.37 m3/s) at Lower Sabino Dam on February 5, 2017.
Discharge flow pattern for Sabino Creek, Arizona, derived from USGS (2018) data. Note that the peak events reflect major precipitation events.
Sampling Plan
We collected freshwater samples from Sabino Creek at various times over the last 7 years, to try to understand the relationship between DIC and atmospheric 14C and also the relation between DIC and DOC. Apart from the samples collected on June 30, 2009, all samples were collected from water in the pool exiting the Lower Sabino Dam. We sampled at times with different flow rates over this period. Results are shown in Table 1 and compared to historical flow rates (USGS 2018), noting that zero flow means that there is no surface flow at the USGS sensor, not that there is no water.
Dissolved inorganic and organic carbon in Sabino Creek, Arizona.
METHODS
Samples of fresh water from Sabino Creek were collected in 2-L plastic containers and sealed. The containers were flushed several times with the water before the final sample. Samples were then transferred to the AMS Laboratory at the University of Arizona and processed as rapidly as possible. Samples were refrigerated before processing. We handled dissolved inorganic carbon (DIC) samples by transferring about 250 mL of water to a sample vessel, evacuated of the head space and acidified the water with phosphoric acid. The evolved CO2 was collected cryogenically, the volume is measured and the gas converted to graphite using our standard protocols (Jull et al. Reference Jull, Burr, Beck, Hodgins, Biddulph, McHargue, Lange and Povinec2008). We processed DOC samples using the procedure on similar-sized samples as described in the wet chemistry approach of Leonard et al. (Reference Leonard, Castle, Burr, Lange and Thomas2013) using KMnO4 as the oxidizing agent, after the DIC component has been removed. The evolved CO2 is then measured volumetrically and converted to graphite for AMS measurement. 14C is measured using the NEC Pelletron AMS machine at the University of Arizona, running at 2.5 MV. Measurements for δ13C were done off-line using a split of the sample gas, and run on a VG Optima stable-isotope mass spectrometer.
RESULTS AND DISCUSSION
We found a distinct relationship between our 14C data and flow rates. We have plotted F(DIC) and F(DOC) against flow rate in Figure 4, where F is the fraction of modern carbon (Donahue et al. Reference Donahue, Linick and Jull1990). Higher-flow events are characterized by F (DIC) values closer to atmospheric and less negative δ13C. The trends for δ13C in both DIC and DOC are shown in Figure 5. Periods of lower flow, particularly after the summer rainy season (in July–September) show more negative δ13C values for DIC, while the DOC δ13C values are generally unchanged. DIC tends to have depressed 14C at these times, yet DOC remains close to contemporary. Similar trends are observed for the April–May values, which is the driest season of the year (see Figure 1). These trends are conspicuous in the data for the years 2012–2013 (Table 1). This suggests that the summer rainy season dislodges older organic carbon, perhaps tied up in soils, which are then released during large flow events. These results can be understood in terms of different sources of DOC originating from terrestrial sources (Raymond and Bauer Reference Raymond and Bauer2001; Barnes et al. Reference Barnes, Butman, Wilson and Raymond2018).
Dependence of F(DOC) and F(DIC) versus flow rate (m3/s). DOC is given as blue circles and DIC as the red triangles.
Dependence of δ13C on flow rate (m3/s). DOC is given as blue circles and DIC as the red triangles. Note the opposite trends for DOC and DIC.
We have plotted these results in Figure 6 as a function of F vs. δ13C. Figure 6 can be understood as a mixing diagram, where there are modern DIC and DOC components with characteristic δ13C and an older component extrapolated from the data. We expect that the “modern” DOC component has higher F due to residence time of the DOC which would have higher F due to higher level of bomb 14C in subsurface water. We also include two previous F(DIC) data from Eastoe et al. (Reference Eastoe, Gu, Long, Hogan, Phillips and Scanlon2004). Surprisingly, both F(DIC) and F(DOC) show opposite trends. Another observation is that DIC tends to decrease in F with more negative δ13C, with a weak correlation coefficient, R = 0.30, whereas DOC tends to move towards less negative δ13C with decreasing F. Further, there is a high correlation between F and δ13C for DOC, where there is a correlation of R=0.96. This suggests that there is a source of “old carbon,” that we can extrapolate from both trends, which is generating both the changes in DIC and DOC and has a δ13C closer to –12 to –15‰. An explanation for the trend in DIC is also explicable from studying the flow rates in Table 1. High flow rates generally have δ13C closer to –4‰ and 14C closer to the atmospheric value, whereas lower flow rates appear to track much lighter values of δ13C and lower 14C. This can be explained by reprocessing of older organic material in sediments into carbon which would have lighter δ13C and also some residence time in the sediment. DIC seems to be incorporating modern DOC-related carbon, accounting for the spread of samples with F values > 1. We could also consider photosynthesis of DIC to DOC, although this would not result in a lower 14C value for DOC.
Isotope plot of fraction of modern 14C (F) vs δ13C for Sabino Creek water samples. Values for DOC are shown as blue circles, DIC as red triangles. Two additional DIC points (Eastoe et al. Reference Eastoe, Gu, Long, Hogan, Phillips and Scanlon2004) are included as green squares. The “old carbon” component is extrapolated from the data.
The trend from January 20 to September 15, 2016, illustrates this transformation. Where paired samples are available, it appears that the DOC has higher values of F(14C) than the DIC. This can be understood by the DOC sampling organic material from sediments with a residence time of 5–10 years, since the level of bomb 14C has been declining during that time (Hua et al. Reference Hua, Barbetti and Rakowski2013). Barnes et al. (Reference Barnes, Butman, Wilson and Raymond2018) suggest that “aged” DOC with a longer residence time is consistent with deeper flow paths, particularly in arid environments. Other interesting trends can be observed. For example, DOC shows a stronger correlation with flow rate than DIC, with a Pearson correlation coefficient R = 0.23. However, if the one highest flow rate on March 21, 2010, is excluded, R = 0.44. This can be explained as due to the fact that since DIC is effectively sampling the atmosphere, it is less dependent on flow rate. However, values of DIC where F<1.0 are occasionally observed. A longer-term monitoring of these values and a closer investigation of the organic geochemistry of the DOC at this site is planned for the future.
CONCLUSIONS
We observe three consistent end-member dissolved carbon components in groundwater from Sabino Canyon, collected from 2009 to 2016. These are (1) modern DOC component (2) modern DIC component, and (3) an older DOC-DIC component. Our results suggest variable degrees of mixing between all three components in the case of DIC, and two-component mixing only for DOC. Although consistent with wider studies (e.g. Barnes et al. Reference Barnes, Butman, Wilson and Raymond2018), we believe these interesting trends warrant a more detailed monitoring of 14C and the geochemistry this stream.
We propose to continue these measurements using other DOC oxidation methods to compare using both chemical and UV approaches. It would also be advantageous to expand 14C measurements to aquatic flora and sediment sampling in this region.
ACKNOWLEDGMENTS
We thank D. S. Cooney for assistance with sample collection and the staff at the AMS laboratory for assistance with sample processing. We also thank M. Molnar for discussions. We are grateful to the helpful suggestions of the reviewers and Associate Editor Christine Hatté, which resulted in an improved manuscript. The work was partly supported by NSF grant EAR1313588. AJTJ was also supported in part by the European Union and the State of Hungary, co-financed by the European Regional Development Fund in the project GINOP-2.3.2.-15-2016-00009 “ICER.”
Open access
