Science of the Total Environment 485–486 (2014) 518–527

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Spectroscopic measurements of estuarine dissolved organic matter dynamics during a large-scale Mississippi River flood diversion Paulina E. Kolic a, Eric D. Roy b, John R. White b,⁎, Robert L. Cook a,⁎⁎ a b

Department of Chemistry, Louisiana State University, Baton Rouge, LA 70803, USA Department of Oceanography and Coastal Sciences, Louisiana State University, Baton Rouge, LA 70803, USA

H I G H L I G H T S • • • • •

We examine estuarine DOM dynamics during a large-scale freshwater diversion. DOC measurements yielded limited insight into estuarine carbon cycling. We used spectroscopic measurements to document a perturbation in DOM chemistry. The diversion provided elevated concentrations of lignin of terrestrial origin. Terrestrial DOM was rapidly processed and DOM became more microbial over time.

a r t i c l e

i n f o

Article history: Received 31 October 2013 Received in revised form 14 March 2014 Accepted 26 March 2014 Available online 16 April 2014 Editor: C.E.W. Steinberg Keywords: Hydrologic manipulation Dissolved organic carbon Absorbance Fluorescence Freshwater diversion Lake Pontchartrain Pulsing

a b s t r a c t The Mississippi River Flood of 2011 prompted the opening of the Bonnet Carré Spillway (BCS) in southeastern Louisiana to protect the City of New Orleans. The BCS diverted approximately 21.9 km3 of river water into the oligohaline Lake Pontchartrain Estuary over the course of 43 days. We characterized estuarine dissolved organic matter (DOM) dynamics before, during, and after the diversion in order to better understand the biogeochemical dynamics associated with these immense freshwater inflows. Dissolved organic carbon (DOC) exhibited a large degree of variability during and after the period of elevated primary productivity that occurred following the diversion. Furthermore, DOC analysis provides limited insight into carbon cycling during these dynamic periods. In order to overcome the limitations of DOC, spectroscopic methods were used to gain insights into chemical composition dynamics. Both ultraviolet visible (A254, A350, SUVA254, spectral slope, and normalized UV/Vis) and fluorescence spectroscopy (excitation emission matrices and fluorescence and biological indices) were used to study the compositional changes of DOM over time. Collectively, our results document a perturbation in DOM chemistry in Lake Pontchartrain due to the diversion and a subsequent return toward pre-diversion conditions. Immediate increases in A350 indicate that BCS freshwater contained elevated concentrations of lignin of terrestrial origin. Ensuing declines in A350, along with changes in the fluorescence and biological indices, indicate that DOM rapidly became more microbial in composition. Our results provide insights into estuarine DOM dynamics relevant to systems receiving flood pulses of freshwater due to either hydrologic manipulation or precipitation events. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Dissolved organic matter (DOM) in estuaries is a complex, heterogeneous mixture comprised of diverse decay products that can be allochthonous or autochthonous in origin and its composition varies due to ⁎ Correspondence to: J.R. White, 3234 Energy Coast and Environment Building, Department of Oceanography & Coastal Sciences, Louisiana State University, Baton Rouge, LA 70803, USA. Tel.: +1 225 578 8792. ⁎⁎ Correspondence to: R.L. Cook, 307 Chopin Hall, Department of Chemistry, Louisiana State University, Baton Rouge, LA 70803, USA. Tel.: +1 225 578 2980. E-mail addresses: [email protected] (P.E. Kolic), [email protected] (E.D. Roy), [email protected] (J.R. White), [email protected] (R.L. Cook).

http://dx.doi.org/10.1016/j.scitotenv.2014.03.121 0048-9697/© 2014 Elsevier B.V. All rights reserved.

differences in the parent organic matter and geochemical processes (McKnight et al., 2001). DOM plays an important role in aquatic ecosystems through binding trace metals, sorbing organic pollutants, and serving as a nutrient source to microorganisms (Cook et al., 2009; Huguet et al., 2009; McKnight et al., 2001; Murphy et al., 2008; Wu et al., 2007). Additionally, DOM absorbs ultraviolet and visible light, thus influencing the depth of sunlight penetration in the water column (Chen and Gardner, 2004; McKnight et al., 2001; Murphyet al., 2008; Ohno, 2002; Weishaar et al., 2003; Wu et al., 2007). DOM has frequently been analyzed using spectroscopic techniques such as ultraviolet visible spectroscopy (UV/Vis) and fluorescence spectroscopy due to its ability to absorb ultraviolet and visible light. A

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527

number of elegant but data- or hardware-intensive fluorescence-based methods taking advantage of fluorescence's four-dimensional space have been applied to DOM characterization. Among them, lifetime measurements offer one promising avenue (Marwani et al., 2009). Likewise, the excitation emission matrix technique (EEM) provides an excitation and emission map of fluorescence intensities and parallel factor (PARAFAC) analysis of the EEM data can yield information-rich data. Usually, PARAFAC fitting of EEM spectra provides three major components (Cook et al., 2009). The combination of biological and fluorescence indices, BIX (Huguet et al., 2009; Parlanti et al., 2000) and FI (McKnight et al., 2001), respectively, provides a large amount of the data embedded within the three major PARAFAC components (Cook et al., 2009). This means that single wavelength measurements allow for much of the insight that EEM and the subsequent PARAFAC analysis can provide. Single wavelength measurements are simple, rapid, inexpensive, accessible, and easily field-amenable, while still providing valuable information on the origins, composition, and age of a DOM sample (Bianchi et al., 2011; Cook et al., 2009; Hernes and Benner, 2003; Hernes et al., 2008; Spencer et al., 2008). Southeastern Louisiana is home to the Mississippi River delta, the largest river delta in North America, and is occasionally at risk of large-scale flooding during high river discharge periods in spring. The Bonnet Carré Spillway (BCS) was constructed in 1931 following the Great Flood of 1927 to divert floodwater from the Lower Mississippi River and protect the downstream City of New Orleans, LA (Barry, 1997). The BCS has been opened ten times to prevent flooding since its construction. During operation, the BCS diverts large amounts of freshwater into the Lake Pontchartrain estuary, which can substantially influence estuarine biogeochemistry (White et al., 2009; Roy and White, 2012). In the spring of 2011, the extreme flood stage of the Lower Mississippi River led to the opening of the BCS by the U.S. Army Corps of Engineers for 43 days from May 9 to June 20, 2011. During this time approximately 21.9 km3 of river water was diverted to Lake Pontchartrain, depositing sediment within the spillway (Nittrourer et al., 2012) and carrying immense nutrient loads to the estuary (e.g., N25,000 Mg NOx–N) (Roy et al., 2013). The leading edge of the sediment-rich freshwater plume was observed exiting Lake Pontchartrain's eastern outlets in ≤14 days, indicating substantial modification of the estuary's residence time, estimated to typically be approximately 60 days (Swenson, 1980; Roy et al., 2013). Introduction of river water and nutrients to Lake Pontchartrain has been shown to greatly influence primary productivity (Bargu et al., 2011). However, no studies have assessed the impact of the BCS on DOM and few studies have examined the influence of other Mississippi River diversions on DOM (e.g., Bianchi et al., 2011). The 2011 BCS opening provided an opportunity to study the dynamics of DOM in a large estuary subjected to hydrologic manipulation. Hydrologic manipulation has become a prominent feature of many large river basins around the world (Bianchi and Allison, 2009; Dynesius and Nilsson, 1994) and therefore the results presented here have implications for highly engineered coastal river systems in other locations, as well as estuaries receiving flood pulses of freshwater from heavy precipitation events (Osburn et al., 2012; Yang et al., 2013). Understanding the biogeochemistry associated with freshwater inflows is critical in coastal Louisiana due to the proposed use of river diversions for sediment delivery to mitigate land loss (State of Louisiana, 2012). Our specific objectives in this study were to: (1) investigate the dynamics of dissolved organic carbon (DOC) in Lake Pontchartrain during and after the 2011 BCS opening in relation to observations of primary production (Roy et al., 2013), (2) characterize the composition of DOM in the estuary throughout the event using UV/Vis and fluorescence analyses, (3) compare the utility of different DOM analytical techniques, and (4) synthesize results to provide insight into carbon cycling during large-scale freshwater inflows to an estuarine system.

519

2. Materials and methods 2.1. Site description The Lake Pontchartrain estuary is located in southeast Louisiana, just north of New Orleans, LA. It is a shallow system (mean depth = 3.7 m) with a surface area of 1637 km2 and a volume of approximately 6 km3 (Turner et al., 2002). Lake Pontchartrain is influenced by northern tributaries that discharge on average 6.1 km3 y−1 (Roblin, 2008; Roy et al., 2013), storm water drainage from New Orleans, freshwater discharge through the BCS, and exchange of water via the Gulf of Mexico. Salinity typically varies between 2 and 9 (Li et al., 2008). Argyrou et al. (1997) reported that the water column concentration of dissolved organic carbon (DOC) in Lake Pontchartrain during 1995–1996 (non-diversion period) ranged from 5.3 to 8.5 mg C L− 1 with an annual mean of 5.8 mg C L−1 and was not correlated to chlorophyll a concentration. DOC appeared to be mostly derived from allochthonous sources including northern tributaries which were found to discharge water containing 6.5–7.3 mg DOC L−1 and up to 28 mg DOC L−1 (Argyrou et al., 1997). 2.2. Sampling regime Water samples were collected along a 30-km 10-station transect extending from the BCS inflow to the center of Lake Pontchartrain (Fig. 1). Sample sites were labeled T-1 and T-11 to T-19 with T-1 being located geographically closest to the BCS inflow and T-19 being near the center of the estuary. The first sample set was collected on May 8, 2011 before the BCS opening. Three additional sample sets were collected while the BCS was open (May 18 to June 16, 2011) and seven sample sets were collected following BCS closure (between June 21 and August 10, 2011), for a total of eleven sample sets. Concurrently salinity and chlorophyll a were measured from samples collected at the same times and locations as samples collected for this study and reported in Roy et al. (2013). 2.3. Materials Polyethylene bottles and filters for DOC samples and analysis were obtained from Nalgene and Pall Life Sciences, respectively. Standards for DOC analysis were obtained from RICCA Chemical Company. Sterile syringes (30 mL), nylon filters (0.22 μm), and borosilicate glass scintillation vials with polyseal caps (20 mL) for spectroscopic sample storage and analysis were obtained from BD, Nalgene, and Kimble Chase, respectively. 2.4. Dissolved organic carbon (DOC) Water samples for DOC analysis were collected from approximately 10 cm below the water surface using acid-washed polyethylene bottles, placed on ice, and returned to the laboratory immediately for processing. Samples were then vacuum-filtered through 0.45 μm membrane filters and stored in combusted glass vials with Teflon coated caps at 4 °C. Prior to analysis, samples were acidified with concentrated hydrochloric acid (ca. 0.2 mL HCl in 40 mL sample) and exposed to air to remove inorganic carbon. DOC was measured on a Shimadzu TOC-V CSN, which converts DOC into CO2 via high-temperature catalytic oxidation. Quality control measures included measurement of standards along with each batch, field triplicates, laboratory duplicates, and spikes. 2.5. Spectroscopic measurements The spectroscopic measurements, indices, and matrices used in this study for DOM characterization are described in Table 1. Water samples for UV/Vis and fluorescence spectroscopy were collected by syringe and then filtered. Samples were stored in borosilicate glass scintillation vials with polyseal caps, stored on ice during transport, refrigerated at 4 °C

520

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527

Fig. 1. MODIS satellite imagery of Lake Pontchartrain on May 17, 2011 provided by the US Naval Research Laboratory at the Stennis Space Center in Mississippi. Orange and red tones depict sediment-laden Mississippi River water (measured as beam attenuation at 547 nm). Blue tones depict estuarine water and land is colored green/brown. The 10 stations shown as Δ comprise the 30-km transect used in this study.

upon return to the laboratory, and shielded from light until analysis. UV/Vis absorbance spectra were collected on a Cary 100 Spectrophotometer. Spectra were collected from 240 to 750 nm using a 1 nm band pass with a 1 cm quartz cell. The spectral slope (S275–295) was calculated using the following equation (Fichot and Benner, 2012): −Sð295–275Þ

ag ð295Þ ¼ ag ð275Þe

where ag is the absorption coefficient (Hernes et al., 2008) at the specified wavelength, and S is the spectral slope in the range of 275–295 nm. Fluorescence spectra were collected on a Spex Fluorolog-3 spectrofluorometer using a 1 cm quartz cell. Fluorescence excitation–emission matrices (EEMs) were obtained using excitation wavelengths of 240 nm to 700 nm with 5 nm increments and emission wavelengths of 300 nm to 750 nm with 5 nm increments. The data from the EEM analysis were blank subtracted using 18 MΩ-cm water and corrected for inner filter effects (Ohno, 2002). The fluorescence index (FI) was determined by dividing the emission intensity at 450 nm by that at 500 nm using

an excitation wavelength of 370 nm (McKnight et al., 2001) and the biological index (BIX) was determined by dividing the emission intensity at 380 nm by that at 430 nm using an excitation wavelength of 310 nm (Huguet et al., 2009). The samples for the first time period were analyzed in triplicate, with standard deviations of b3%, with the majority being b 1%. Therefore, the analysis continued primarily on one sample per station per each time period. To ensure continued quality control, duplicate samples were analyzed at two sites per sampling time period throughout the study. Duplicates also had a standard deviation of b 3%. 2.6. Salinity data and primary production time periods Salinity and phytoplankton dynamics during the 2011 BCS opening are described in detail by Roy et al. (2013) and can be summarized as follows. The introduction of Mississippi River water at a rate of up to nearly 9000 m3 s−1 during the diversion event (Fig. 2) resulted in the expansion of a freshwater plume across the majority of the sample transect during the period of May 18 to June 16, 2011. Salinity decreased

Table 1 Spectroscopic measurements, indices, and matrices used in this study for DOM characterization. Measurement

Description

A254

Non normalized absorbance at wavelength of 254 nm (quantitative; concentration): allows for concentration determination of aromatic fraction of DOM (Bianchi et al., 2011). Non normalized absorbance at wavelength of 350 nm (quantitative; concentration): allows for concentration determination of lignin fraction of DOM (Hernes et al., 2008; Spencer et al., 2008). Specific ultraviolet absorption at 254 nm determined by dividing A254 by dissolved organic carbon (DOC) concentration. Provides qualitative assessment of aromaticity (Weishaar et al., 2003). Maps the composition of chromophores in a DOM sample (Bianchi et al., 2011). Spectral slope in the range of 275–295 nm. It is an indicator of lignin molecular weight and photobleaching (Fichot and Benner, 2012). Maps the composition of fluorophores in a DOM sample.

A350 SUVA254 Normalized UV/Vis s275–295

EEM (Excitation emission matrix) FI

BIX

The fluorescence index indicates DOM origin (terrestrial vs. microbial) and reflects DOM composition. FI is calculated using an excitation wavelength of 370 nm and dividing the emission intensity at 450 nm (biological) by 500 nm (terrestrial). Thus, a higher value of FI is more microbial (McKnight et al., 2001). The biological index indicates the “freshness” of DOM or biological activity and reflects DOM composition. BIX is calculated by using an excitation wavelength of 310 nm and dividing the emission intensity at 380 nm (protein) by 430 nm (biological). Larger values of BIX indicate more recently produced DOM (Huguet et al., 2009).

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527

521

Fig. 3. Dissolved organic carbon (DOC) concentrations as a function of salinity for the pre-diversion (May 8), diversion pre-bloom (May 18–May 28), bloom (June 16–June 25), and post-bloom (June 30–August 10) periods during the 2011 BCS event. Salinity data and primary production information are from Roy et al. (2013). Fig. 2. Bonnet Carré Spillway (BCS) freshwater discharge rates (m3 s−1) during the 2011 diversion. Data is from the United States Army Corps of Engineers. Modified from Roy et al. (2013).

from 2.6–4.9 pre-diversion to 0.15 at all stations except those furthest (T-18 and T-19) from the spillway (0.15–2.72) during the diversion period. Chlorophyll a measurements indicated relatively low phytoplankton biomass pre-diversion (mean ± 1 standard deviation = 3.6 ± 1.1 μg L−1), which then increased to up to 35 μg L−1 during the diversion period at the sites further from the BCS inflow, where greater water transparency and reduced turbulence likely enabled phytoplankton growth (Roy et al., 2013). Following the BCS closure, the post-diversion period from June 21 to August 10, 2011 was characterized by gradually increasing salinity at all stations that remained below 1.5 across the transect (Roy et al., 2013). Chlorophyll a concentration peaks ranging from 18.6 to 45.1 μg L−1 occurred at all transect stations between June 16 and June 25. At these times the phytoplankton community was most frequently dominated by chlorophytes. The term “bloom” is used in this paper to denote this period of elevated chlorophyll a as a measure of overall phytoplankton biomass and is not intended to reference any particular species. Phytoplankton biomass declined after June 25, with chlorophyll a concentrations similar to pre-diversion values through July and August (Roy et al., 2013). Based on these observations by Roy et al. (2013), the results in this study are presented for 4 time periods in 2011: (1) the pre-diversion sampling on May 8, (2) the diversion pre-bloom period from May 18 to May 28, (3) the bloom period from June 16 to June 25, and (4) the post-bloom period from June 30 to August 10. 3. Results and discussion 3.1. Dissolved organic carbon (DOC) Concentrations of DOC for this study ranged between 3.1 and 16.8 mg C L−1 (Fig. 3) with a majority of samples within the reported range of DOC in Lake Pontchartrain during the non-diversion years of 1995 and 1996 (5.3–8.5 mg DOC L− 1; Argyrou et al., 1997). Prior to the diversion on May 8, mean DOC concentration ± 1 standard deviation across all transect stations was 5.6 ± 0.2 mg C L−1 (n = 10). DOC concentration decreased to 4.9 ± 0.3 mg C L−1 (n = 20) during May 18 to May 28 following the BCS opening as Mississippi River water inundated the majority of transect stations. During both the pre-diversion and diversion pre-bloom periods, DOC concentrations were relatively constant across a range of salinities (Fig. 3). DOC increased and became much more variable during the bloom period including chlorophyll a maxima for all stations between June 16 and June 25, with a mean DOC

concentration ± 1 standard deviation equal to 7.2 ± 2.6 mg C L−1 (n = 30). During this time, no correlation between chlorophyll a and DOC concentrations was observed (r2 = 0.01). An elevated mean DOC concentration above pre-diversion and pre-bloom periods coupled with the increased variability observed during the bloom period continued during the post-bloom period from June 30 to August 10 (mean DOC concentration ± 1 standard deviation = 6.5 ± 2.2 mg C L−1, n = 50). Once again, no correlation was observed between DOC and chlorophyll a (r2 = 0.004). Additionally, no clear relationship between DOC and salinity was observed during the bloom and post-bloom periods (Fig. 3). These results illustrate that DOC can exhibit large fluctuations during and after spikes in estuarine primary productivity following large-scale, nutrient-rich freshwater inflows. Furthermore, DOC measurements do not provide any insight into the composition of DOM, especially during dynamic conditions. The spectroscopic measurements made during this study can aid in moving beyond these limitations to provide insights into carbon cycling during freshwater inflows to estuaries. 3.2. Ultraviolet–visible analysis 3.2.1. Absorbance at 254 nm While DOC is a bulk carbon measurement, UV/Vis analysis provides specific details into the composition of chromophores in a DOM sample. The UV/Vis spectra of DOM are generally featureless with absorbance increasing with shorter wavelengths, thus several descriptors have been used to analyze this data (Weishaar et al., 2003). The absorbance at a wavelength of 254 nm (A254), which can be utilized as an indicator of DOM concentration (Bianchi et al., 2011), was determined for each sampling site for all of the sampling dates. The A254 values, for the complete data set, ranged from 0.103 to 0.183 with a mean ± 1 standard deviation of 0.129 ± 0.014 (n = 110). As a general trend, the sampling sites closer to the BCS inflow tended to have higher A254 values for all the sampling dates (Fig. 4a). This could be due to the higher density of vegetation at the BCS inflow leaching organic matter into surrounding waters. There was an overall decrease of A254 over the course of the three-month study, with a mean A254 ± 1 standard deviation of 0.147 ± 0.004 (n = 10), 0.133 ± 0.003 (n = 20), 0.132 ± 0.015 (n = 30), and 0.122 ± 0.013 (n = 50) for pre-diversion, diversion pre-bloom, bloom, and post-bloom, respectively. In the first four sites, in closest proximity to the BCS inflow, an immediate and pronounced decrease in A254 was found after the BCS opened (Fig. 4a). This can be explained by the lower concentration of DOM in BCS freshwater compared to estuarine water in Lake Pontchartrain or a dilution of the DOM in Lake Pontchartrain by BCS freshwater. In general, A254 decreases with time with the two sites closest to the BCS experiencing fluctuations over time from leached organic matter; from vegetation and/or sediment due to background BCS flow (Fig. 4a).

522

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527

Fig. 4. Variation of (a) A254 (b) A350 and (c) S275–295 by sampling date and distance from the opening of the Bonnet Carré Spillway (BCS).

3.2.2. Absorbance at 350 nm Another optical measurement that can be used to study the composition of DOM is its absorbance at 350 nm or A350, which has been correlated with lignin content (Hernes et al., 2008; Spencer et al., 2008). The A350 values ranged from 0.015 to 0.039 during the course of the study with a mean ± 1 standard deviation of 0.024 ± 0.005 (n = 110). Higher A350 values and thus higher lignin content were found at sampling sites closer to the BCS inflow and consequently closer to vegetation, for all sampling dates (Fig. 4b). The values of A350 generally decreased as one moved into deeper water (away from the BCS), and is consistent with the dilution of the lignin components. During the pre-diversion sampling A350 values had a mean ± 1 standard deviation of 0.025 ± 0.002 (n = 10). The mean of A350 increased to 0.030 ± 0.002 (n = 20) when the BCS was opened (diversion prebloom). This indicates that there was an increase in concentration of lignin, which can be explained by the terrestrial origin of DOM from the Mississippi River and/or the BCS waters releasing stored DOM from vegetation and sediments in the spillway area between the river and Lake Pontchartrain. During the bloom and post-bloom time periods, the A350 values decreased to 0.025 ± 0.005 (n = 30) and 0.021 ± 0.004 (n = 50), respectively (Fig. 4b). This decrease in the A350 values upon closing of the BCS suggests that the freshwater loaded into Lake Pontchartrain had a higher lignin content than resident estuarine waters. Additionally, the post-bloom A350 values were lower than the pre-diversion values indicating that lignin-like moieties in DOM may have become photobleached.

3.2.3. Spectral slope The spectral slope (S275–295) has been shown in past studies to be related to the molecular weight of lignin and the degree of photobleaching.

Lower values of S275–295 are characteristic of river environments and indicate higher molecular weight lignin and lower instances of photobleaching (Fichot and Benner, 2012). The values of spectral slope (S275–295) ranged from 0.015 to 0.022 during the course of the study with a mean ± 1 standard deviation of 0.019 ± 0.002. Fig. 4c shows S275–295 for all sites and sampling dates. During the pre-diversion sampling, S275–295 had a mean ± 1 standard deviation of 0.020 ± 0.001. During the diversion pre-bloom period the mean ± 1 standard deviation of S275–295 decreased to 0.016 ± 0.001. The value of S275–295 steadily increased throughout the bloom and post-bloom periods with a mean ± 1 standard deviation of 0.017 ± 0.001 and 0.020 ± 0.001, respectively. The decrease in S275–295 from pre-diversion to diversion pre-bloom indicates the large and immediate impact of the diversion on the Pontchartrain Estuary. Increasing values of S275–295 during the bloom and post-bloom period indicate lower molecular weight lignin and photodegradation, which is consistent with intense sunlight during the summer months of this study. The data in Fig. 4c also show that during the study period the S275–295 values returned to their pre-diversion range. 3.2.4. Normalized UV/Vis absorbance UV/Vis data can be normalized to study the composition of chromophores by eliminating effects of concentration (Bianchi et al., 2011; Ohno, 2002). In this study, UV/Vis absorbance spectra were normalized to 1.00 at 290 nm (an arbitrary wavelength). Very little spatial variability was observed during the course of the study with regard to normalized UV/Vis spectra. The normalized spectra for each sampling date could be overlaid almost identically, for all sites, except T-19, as the freshwater plume took more time to reach this sampling site. Both sites T-1 and T-17 exhibited a similar trend regardless of their distance from the BCS. Compared to the aggregate UV/Vis data, there was an appreciable temporal change in the composition of the chromophores

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527

during the course of this study, especially in the aromatic region (240–290 nm), illustrating the influence of BCS freshwater on the DOM in Lake Pontchartrain. This is highlighted in Fig. 5, which presents the normalized UV/Vis spectra for sites T-1 and T-17 on May 8 (prediversion), May 18 (diversion pre-bloom), June 25 (bloom) and July 20 (post-bloom). The most significant changes to the composition of the chromophores occurred during the diversion period. Spectra obtained for pre-diversion and post-bloom are similar, although not identical, indicating that the composition of chromophores was returning to preevent conditions. 3.2.5. Specific ultraviolet absorption at 254 nm Past studies have indicated that the specific ultraviolet absorption at 254 nm (SUVA254) of DOM can be used to determine the aromaticity of a sample, with the value of SUVA254 increasing with increasing aromaticity (Weishaar et al., 2003). SUVA254 is determined by dividing the absorbance at 254 nm by DOC (Weishaar et al., 2003). In this study, SUVA254 values ranged from 0.717 to 3.527, with a mean ± 1 standard deviation of 2.188 ± 0.520. SUVA254 increased slightly from pre-diversion to diversion pre-bloom and subsequently decreased with time, indicating that the earlier samples had a higher aromatic content. Because the DOC values are used to calculate SUVA254, values of SUVA254 are highly variable during the bloom and post-bloom periods as dictated by variability in the DOC data. The mean values of SUVA254 ± 1 standard deviation were 2.653 ± 0.086, 2.719 ± 0.126, 1.989 ± 0.483, and 2.003 ± 0.487 for pre-diversion, diversion pre-bloom, bloom, and post-bloom periods, respectively. On aggregate, it is evident from the UV/Vis data that the DOM in Lake Pontchartrain was perturbed by the

523

BCS diversion and even after nearly three months had not completely recovered with regard to concentration, lignin content, and composition of other chromophores. 3.3. Fluorescence analysis 3.3.1. Excitation emission matrices Fluorescence spectroscopy provides a more specific way to interrogate DOM in that only fluorophores (emits light) are analyzed, unlike UV/Vis spectroscopy in which any compound that absorbs light (chromophore) is studied. Fluorescence excitation emission matrices (EEMs) allow a complete excitation and emission mapping of the fluorescent behavior of a sample for the sampled wavelengths. Several overriding trends were found from the complete data set of 110 EEMs. First, the overall composition of fluorophores exhibited little spatial variation among sites on each sampling date. This trend continued for each of the twelve sampling dates. For example, the EEMs were nearly identical at sites T-1 and T-17 on May 8 (Fig. 6a). Upon subtraction of normalized EEMs (highest value normalized to 1) of T-1 and T-17 from May 8, the resulting signal was essentially at the level of noise and scattering (no signal was N 5%) indicating that the two EEMs were essentially the same. Another trend found was that EEMs obtained for the first three sampling trips (May 8, 18, and 28) had a similar composition of fluorophores, whereas EEMs for the remaining eight trips were similar. The major difference between the EEMs for the first three samplings and last eight sampling trips is the large decrease in the distribution of the “amino acid-like” or “protein-like” component, which has excitation

Fig. 5. Normalized UV/Vis absorbance spectra illustrating perturbation of DOM chemistry at wavelengths between 240 and 400 nm for sites (a) T-1 and (b) T-17 and between 240 and 290 nm for sites (c) T-1 and (d) T-17 on May 8 (pre-diversion), May 18 (diversion pre-bloom), June 25 (bloom), and July 20 (post-diversion).

524

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527

3.3.2. Fluorescence index (FI) The FI serves as a simple indicator of DOM origin. DOM with terrestrial inputs has a lower FI, approximately in the range of 1.3–1.4, because of the higher content of aromatic groups and more conjugation (degraded lignin-like moieties), compared with microbially derived DOM, which has values of FI closer to 1.9 and little conjugation (Cook et al., 2009; Cory and McKnight, 2005; McKnight et al., 2001). Additionally, river samples were found to have an FI mainly between 1.4 and 1.5 due to terrestrial DOM inputs (McKnight et al., 2001; Murphy et al., 2008). Values of FI ranged from 1.380 to 1.559 for this study indicating a mainly terrestrial origin of the DOM samples. During pre-diversion conditions, the FI was at its lowest (most terrestrial in nature) with a mean ± 1 standard deviation of 1.392 ± 0.008 (n = 10) (Fig. 7a). Values of FI increased to 1.511 ± 0.019 (n = 20) during the diversion pre-bloom period and to 1.519 ± 0.017 (n = 30) during the bloom period. This increase in FI indicates that the DOM became more microbial over time. This can be explained by increased microbial activity resulting from the change in DOM composition as a result of the introduction of Mississippi River water. Post-bloom values of FI appear to decrease with a mean ± 1 standard deviation of 1.510 ± 0.016 (n = 50). Spatial variations of FI were observed in this study however no clear trend was observed. 3.3.3. Biological index (BIX) As with FI, BIX can be calculated using a simple fluorescence measurement. The BIX is used to determine the presence of the β fluorophore, which is characteristic of autochthonous biological activity (Huguet et al., 2009; Parlanti et al., 2000). Larger values of BIX correspond to an increase in biological productivity and the presence of recently produced organic matter, which is characteristic of marine environments, whereas lower values represent the presence of more

Fig. 6. Fluorescence excitation emission matrices (EEMs) for T-1 and T-17 on May 8 (pre-diversion) (a) and T-1 on May 8 (pre-diversion) and T-1 on June 16 (bloom) (b).

wavelengths between 240 to 325 nm and emission wavelengths of 300 to 350 nm (Coble, 1996; Cook et al., 2009; Cory and McKnight, 2005; Murphy et al., 2008; Wu et al., 2007). For example, a comparison of site T-1's EEM spectra on May 8 and June 16 can be seen in Fig. 6b. When the normalized signal for T-1 (June 16) is subtracted from T-1 (May 8), the result is a positive signal in the amino acid-like region and a negative signal for quinone A-like moieties (240 to 325 nm for excitation, 375 to 475 nm for emission), which are associated with microbial origins (Cook et al., 2009). This shows that early samples had a higher distribution of amino acid-like moieties compared to latter samples and the microbial distribution increased with time. This amino acid-like peak results from the fluorescence of aromatic amino acids, particularly tryptophan and tyrosine (Coble, 1996; Cook et al., 2009). The amino acid-like moiety may serve as a nutrient source for the microbial community in the Lake Pontchartrain estuary. A decrease in the amino acid-like peak, in time, may also suggest a dilution of these components via the introduction of BCS freshwater. The increase in distribution of the quinone-A fluorophore signal, in time, can be associated with a more microbial nature to the DOM, which in turn would suggest that the BCS introduced DOM is being utilized as a nutrient source.

Fig. 7. Variation of (a) fluorescence index (FI) and (b) biological index (BIX) by sampling date and distance from the opening of the Bonnet Carré Spillway (BCS).

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527

humic-like fluorophores. The values of BIX obtained in this study range between 0.527 and 0.692, which is indicative of a low autochthonous component in the overall DOM signal (Huguet et al., 2009). BIX data from this study are summarized graphically in Fig. 7b. The BIX values increase over the course of the opening and closing of the spillway gates with a mean ± 1 standard deviation of 0.560 ± 0.015 (n = 10), 0.570 ± 0.022 (n = 20), 0.627 ± 0.020 (n = 30), and 0.654 ± 0.017 (n = 50) for the pre-diversion, diversion pre-bloom, bloom, and postbloom periods, respectively. There was a sharp increase in BIX values during the bloom period, a continued increase in BIX during the closing of the BCS and bloom termination, and finally a leveling off of BIX values in the later sampling dates. There was some spatial variability of BIX values, with sites closer to the opening of the BCS having lower BIX values and those further away having higher BIX values. As an increase in BIX is a result of increased biological productivity of DOM, the BIX data strongly indicate that the introduced Mississippi River water served as a nutrient source and DOM was transformed during the bloom period. 3.4. Synthesis of spectroscopic measurements Comparisons between the different spectroscopic measurements reported here yield several insights into estuarine carbon cycling before, during, and after the 2011 BCS diversion. First, the correlation between A254 and A350 provides insights into the relationship between lignin content and DOM concentration (Fig. 8a). Considering all the samples collected and their temporal and spatial variability, A254 and A350 were well correlated (r2 = 0.637; n = 110). This correlation was also observed during the bloom (r2 = 0.870; n = 20) and post-bloom periods (0.869; n = 50). However, poor correlations between A254 and A350 during the pre-diversion (r2 = 0.385; n = 10) and diversion prebloom (r2 = 0.020; n = 20) periods indicate that the lignin content of DOM behaved independently of the total DOM concentration during these times. Higher lignin content was associated with a lower DOM concentration during the diversion pre-bloom period, indicating that

525

lignin-like moieties accounted for a higher percentage of the DOM composition, which is consistent with terrestrial sources of carbon loaded via the BCS and also leaching of organic matter from the adjacent vegetated wetland fringe. During the course of the study, A350 and S275–295 had a negative relationship (r2 = 0.544, n = 110), with a decreasing lignin concentration (A350) corresponding to an increase in photodegradation and decrease in lignin molecular weight (Fig. 8b). This relationship was more pronounced during the pre-diversion (r2 = 0.938, n = 10) and diversion pre-bloom (r2 = 0.805, n = 20) periods. During the bloom (r2 = 0.065, n = 30) and post-bloom periods (r2 = 0.648, n = 50), this trend was less apparent. Poor correlations during the bloom period may be due to the rapidly changing composition of DOM. Additional metrics suggest that DOM rapidly became more microbial in terms of its composition. Immediately following the BCS opening, FI values increased substantially (Fig. 7a), indicating that DOM became more microbial in origin. In conjunction, the increase in BIX shows an increase in biological activity (Fig. 7b). The use of terrestrial BCSloaded DOM as the parent material for transformation into microbial DOM is supported by the relationship between BIX and A350 (Fig. 8c). In general, lignin content (A350) decreases as the biological activity increases (r2 = 0.522, n = 110). Correlations between BIX and A350 were found for the pre-diversion (r2 = 0.751, n = 10) and diversion pre-bloom (r2 = 0.715, n = 20) time periods. However, during the bloom period, BIX and A350 were not as well correlated (r2 = 0.281, n = 30), this could be due to the rapidly changing environment. Finally during the post-diversion time period BIX and A350 show moderate correlation (r2 = 0.505, n = 50). It was found that S275–295 and BIX correlated well during this study for the pre-diversion period (r2 = 0.806, n = 10) and the combination of diversion pre-bloom, bloom, and post-diversion periods (r2 = 0.793, n = 100) (Fig. 8d). S275–295 and BIX were not well correlated for the entire study (r2 = 0.370, n = 110) due to the existence of two distinct clusters of data (pre-diversion versus diversion pre-bloom, bloom, and post-diversion). The values of pre-diversion samples were isolated

Fig. 8. Relationships between (a) A350 and A254, (b) A350 and S275–295, (c) the biological index (BIX) and A350, and (d) BIX and S275–295 for the pre-diversion (May 8), diversion pre-bloom (May 18–May 28), bloom (June 16–June 25), and post-bloom (June 30–August 10) periods during the 2011 Bonnet Carré Spillway (BCS) event.

526

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527

from other time periods due to lower BIX and larger S275–295 values. The correlation between S275–295 and BIX indicates that photodegradation of lignin components may have played a role in the availability of DOM for biological degradation and that terrestrial BCS-loaded DOM lignin parent materials strongly influenced biological activity during the postdiversion period. When comparing chlorophyll a data presented in Roy et al. (2013) to UV/Vis and fluorescence data, phytoplankton bloom impacts on DOM changes observed from spectroscopic studies become evident. The introduction of Mississippi River DOM provided a parent DOM material in addition to inorganic nitrogen and phosphorus loads documented by Roy et al. (2013), and by the evidence presented here contributed to the spike in phytoplankton biomass observed between June 16 and June 25. Amino acid-like moieties and photodegraded lignin parent materials may have served as a nutrient source for phytoplankton and other microbial communities, as seen by the disappearance of the amino acid peak in the fluorescence EEMs (Fig. 6). In addition, the photodegradation and disappearance of lignin-like moieties is evident by the decrease and increase in the A350 and S275–295 values, respectively. This, in turn, may have led to a higher biological activity, as indicated by the increase of the BIX values, and the DOM becoming more microbial in nature, as illustrated by the increase in FI over time via the transformation of terrestrially derived DOM from the BCS. It is important to note that the insights provided by these simple single wavelength spectroscopic methods were not available from the more classic water analysis methods, such as DOC, and that it is these spectroscopic data that allow for the understanding of carbon utilization and cycling at the chemical composition level. This study shows that in order to gain essential biogeochemical information, such as a broad understanding of the impacts of large river diversions have on the receiving water bodies, one must utilize spectroscopic tools in concert with more classic chemical and biological monitoring methods. 4. Conclusions Historically hydrologic manipulation of the lower Mississippi River using the BCS results in the large-scale introduction of freshwater, nutrients, and DOM into the Lake Pontchartrain Estuary. In 2011, the nutrients loaded by the BCS stimulated a spike in primary productivity in the estuary (Roy et al., 2013). During this dynamic period of elevated primary productivity, DOC measurements proved to be highly variable and did not facilitate understanding of DOM in terms of origin and transformation. Spectroscopic measurements consisting of A254, A350, spectral slope, SUVA254, Normalized UV/Vis, excitation emission matrices, and fluorescence based indices (FI, BIX) are relatively inexpensive and field amenable methods that collectively allow for a significantly more detailed understanding of DOM dynamics during freshwater diversions. During the three-month study, values of A254, A350, spectral slope, FI, and BIX changed from pre-diversion values, documenting a perturbation in estuarine DOM chemistry associated with the diversion and subsequent recovery, which remained incomplete at the end of this study. This perturbation was marked by changes in the composition of the chromophores and fluorophores in the estuary. The diversion provided lignin of terrestrial origin to the Lake Pontchartrain water column where the DOM rapidly became more microbial in nature as a result of increased biological activity. Results indicate that DOM within the BCS diversion plume of 2011 likely contributed, as an additional source of nutrients, to the elevated primary production following the diversion. Our results have direct implications for coastal systems receiving flood pulses of freshwater and future monitoring efforts. We have shown that several relatively inexpensive and accessible spectroscopic measurements can be employed in a rapid sampling effort to gain substantial understanding of flood pulse impacts on DOM chemistry. This is useful in the context of planned river diversions for restoration purposes in coastal Louisiana. We suggest that the measurements employed in this study be used for monitoring future flood events and restoration-orientated diversions to

help evaluate carbon dynamics and ecological consequences at the base of aquatic food webs. Conflict of interest We declare no conflict of interest in this manuscript. Acknowledgments This material is based upon work supported by the National Science Foundation under grants CHE-0547982, CHE-1045973, DEB-0833225 and EAR-1139997 as well as from the USDA (CSRESS: 2009-3520105819 and CSRESS: 2009-65107-05926) and Louisiana Sea Grant. Any opinions, findings, and conclusion or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the any of the funding agencies. E.D.R. was supported by a Louisiana Board of Regents Fellowship during the course of this study. P.E.K. was supported by a Louisiana Board of Regents' Economic Development Assistantship. This research is supported in part (PEK) by the Department of Energy Office of Science Graduate Fellowship Program (DOE SCGF), made possible in part by the American Recovery and Reinvestment Act of 2009, administered by ORISE-ORAU under contract no. DE-AC05-06OR23100. We extend our thanks to Anthony Nguyen and Emily Smith for field and laboratory assistance and Brittany Pritchard, Loice O'jwang, and Caroline Schneider for lab assistance. We thank two anonymous reviewers for inputs that allowed for the development of a more insightful paper. References Argyrou M, Bianchi T, Lambert C. Transport and fate of dissolved organic carbon in the Lake Pontchartrain estuary, Louisiana, U.S.A. Biogeochemistry 1997;38:207–26. Bargu S, White JR, Li C, Czubakowski J, Fulweiler RW. Effects of freshwater input on nutrient loading, phytoplankton biomass, and cyanotoxin production in an oligohaline estuarine lake. Hydrobiologia 2011;661:377–89. Barry JM. Rising tide: the Great Mississippi Flood of 1927 and how it changed America. New York: Simon & Shuster; 1997. Bianchi TS, Allison MA. Large-river delta-front estuaries as natural “recorders” of global environmental change. Proc Natl Acad Sci U S A 2009;106:8085–92. Bianchi TS, Cook RL, Perdue EM, Kolic PE, Green N, Zhang Y, et al. Impacts of diverted freshwater on dissolved organic matter and microbial communities in Barataria Bay, Louisiana, U.S.A. Mar Environ Res 2011;72:248–57. Chen RF, Gardner GB. High-resolution measurements of chromophoric dissolved organic matter in the Mississippi and Atchafalaya River plume regions. Mar Chem 2004;89: 103–25. Coble PG. Characterization of marine and terrestrial DOM in seawater using excitation– emission matrix spectroscopy. Mar Chem 1996;51:325–46. Cook RL, Birdwell JE, Lattao C, Lowry M. A multi-method comparison of Atchafalaya basin surface water organic matter samples. J Environ Qual 2009;38:702–11. Cory RM, McKnight DM. Fluorescence spectroscopy reveals ubiquitous presence of oxidized and reduced quinones in dissolved organic matter. Environ Sci Technol 2005;39:8142–9. Dynesius M, Nilsson C. Fragmentation and flow regulation of river systems in the northern third of the world. Science 1994;266:753–62. Fichot CG, Benner R. The spectral slope coefficient of chromophoric dissolved organic matter (S275–295) as a tracer of terrigenous dissolved organic carbon in riverinfluenced ocean margins. Limnol Oceanogr 2012;57:1453–66. Hernes PJ, Benner R. Photochemical and microbial degradation of dissolved lignin phenols: implications for the fate of terrigenous dissolved organic matter in marine environments. J Geophys Res Oceans 2003;108:3291. Hernes PJ, Spencer RGM, Dyda RY, Pellerin BA, Bachand PAM, Bergamaschi BA. The role of hydrologic regimes on dissolved organic carbon composition in an agricultural watershed. Geochim Cosmochim Acta 2008;72:5266–77. Huguet A, Vacher L, Relexans S, Saubusse S, Froidefond JM, Parlanti E. Properties of fluorescent dissolved organic matter in the Gironde Estuary. Org Geochem 2009;40: 706–19. Li CY, Walker N, Hou AX, Georgiou I, Roberts H, Laws E, et al. Circular plumes in Lake Pontchartrain estuary under wind straining. Estuar Coast Shelf Sci 2008;80:161–72. Marwani H, Lowry M, Xing B, Warner IM, Cook RL. Frequency-domain fluorescence lifetime measurements via frequency segmentation and recombination as applied to pyrene with dissolved humic materials. J Fluoresc 2009;19:41–51. McKnight MD, Boyer WE, Westerhoff KP, Doran TP, Kulbe T, Andersen TD. Spectrofluorometric characterization of dissolved organic matter for indication of precursor organic material and aromaticity. Limnol Oceanogr 2001;46:38–48. Murphy KR, Stedmon CA, Waite TD, Ruiz GM. Distinguishing between terrestrial and autochthonous organic matter sources in marine environments using fluorescence spectroscopy. Mar Chem 2008;108:40–58.

P.E. Kolic et al. / Science of the Total Environment 485–486 (2014) 518–527 Nittrouer JA, Best JL, Brantley C, Cash RW, Czapiga M, Kumar P, et al. Mitigating land loss in coastal Louisiana by controlled diversion of Mississippi River sand. Nat Geosci 2012;5:534–7. Ohno T. Fluorescence inner-filtering correction for determining the humification index of dissolved organic matter. Environ Sci Technol 2002;36:742–6. Osburn CL, Handsel LT, Mikan MP, Paerl HW, Montgomery MT. Florescence tracking of dissolved and particulate organic matter quality in a river-dominated estuary. Environ Sci Technol 2012;46:8628–36. Parlanti E, Wörz K, Geoffroy L, Lamotte M. Dissolved organic matter fluorescence spectroscopy as a tool to estimate biological activity in a coastal zone submitted to anthropogenic inputs. Organomet Chem 2000;31:1765–81. Roblin R. Water quality modeling of freshwater diversions in the Pontchartrain Estuary. New Orleans: University of New Orleans; 2008 [M.S. Thesis]. Roy ED, White JR. Nitrate flux into the sediments of a shallow oligohaline estuary during large flood pulses of Mississippi River water. J Environ Qual 2012;41:1549–56. Roy ED, White JR, Smith EA, Bargu S, Li C. Estuarine ecosystem response to three large-scale Mississippi River flood diversion events. Sci Total Environ 2013;458: 374–87. Spencer RGM, Aiken GR, Wickland KP, Striegl RG, Hernes PJ. Seasonal and spatial variability in dissolved organic matter quantity and composition from the Yukon River basin Alaska. Glob Biogeochem Cycles 2008;22:GB4002.

527

State of Louisiana. Louisiana's comprehensive master plan for a sustainable coast; 2012 [Available online: http://www.coastalmasterplan.louisiana.gov/2012]. Swenson EM. General hydrography of Lake Pontchartrain, Louisiana. In: Stone JH, editor. Environmental analysis of Lake Pontchartrain, Louisiana, its surrounding wetlands, and selected land uses. U.S. Army Corps of Engineers, New Orleans District, contract no. DACW29-77-C-02523. Baton Rouge: Coastal Ecology Institute, Center for Wetland Resources, Louisiana State University; 1980. p. 57–155. Turner RE, Dortch Q, Justic D, Swenson EM. Nitrogen loading into an urban estuary: Lake Pontchartrain (Louisiana, U.S.A.). Hydrobiologia 2002;487:137–52. Weishaar JL, Aiken GR, Bergamaschi BA, Fram MS, Fujii R, Mopper K. Evaluation of specific ultraviolet absorbance as an indicator of the chemical composition and reactivity of dissolved organic carbon. Environ Sci Technol 2003;37:4702–8. White JR, Fulweiler RW, Li CY, Bargu S, Walker ND, Twilley RR, et al. Mississippi River flood of 2008: observations of a large freshwater diversion on physical, chemical, and biological characteristics of a shallow estuarine lake. Environ Sci Technol 2009;43:5599–604. Wu FC, Kothawala DN, Evans RD, Dillon PJ, Cai YR. Relationships between DOC concentration, molecular size and fluorescence properties of DOM in a stream. Appl Geochem 2007;22:1659–67. Yang L, Guo W, Chen N, Hong H, Huang J, Xu J, et al. Influence of a summer storm event on the flux and composition of dissolved organic matter in a subtropical river China. Appl Geochem 2013;28:164–71.