Article pubs.acs.org/est
Changing Dynamics of Dissolved Organic Matter Fluorescence in the Northern Gulf of Mexico Following the Deepwater Horizon Oil Spill Eurico J. D’Sa,*,† Edward B. Overton,‡ Steven E. Lohrenz,§ Kanchan Maiti,† R. Eugene Turner,† and Angelina Freeman∥ †
Department of Oceanography and Coastal Sciences and ‡Department of Environmental Sciences, Louisiana State University, Baton Rouge, Louisiana 70803, United States § School for Marine Science and Technology, University of Massachusetts, Dartmouth, Massachusetts 02744, United States ∥ Coastal Protection and Restoration Authority, 450 Laurel Street, Baton Rouge, Louisiana 70801, United States S Supporting Information *
ABSTRACT: The characteristics of fluorescent components of dissolved organic matter (DOM) were examined using excitation emission matrix (EEM) fluorescence spectroscopy combined with parallel-factor analysis (PARAFAC) for seawater samples obtained from the northern Gulf of Mexico (NGoM) before, during, and after the 2010 Deepwater Horizon (DwH) oil spill. An EEMs PARAFAC modeling of samples collected within 16 km of the wellhead during the oil spill in May 2010, which included one typical subsurface sample with a PAH concentration of 1.09 μg/L, identified two humic-like and two previously reported oil-like components. Compared to prespill levels, however, there were order-ofmagnitude higher fluorescence intensities associated with these components that are consistent with an oil-spill source. The spectral decomposition of the EEMs data using individual and combined data sets from coastal and offshore waters impacted by the DwH spill further revealed the changing nature of fluorescent DOM composition. Although the PAHs concentrations were at prespill conditions after the spill in 2012 and 2013 near the DwH site, the variable and anomalous levels of fluorescence intensities and DOC concentrations three years after the spill suggest the potential long-term persistence of the oil in the DOC pool in the NGoM.
■
the enhanced dissolution by subsurface dispersants.17 PAHs, a component of crude oil that is toxic, mutagenic, and carcinogenic, formed ∼3.9% by weight of the MC252 oil,18 and an estimated ∼2.1 × 1010 g of PAHs was released into the NGoM waters during the spill.19 Similar to the aromatic components of dissolved organic matter (DOM), PAHs are photochemically reactive and have characteristic fluorescence properties. PAHs have been routinely quantified by gas chromatography and mass spectroscopy (GC−MS) or by GC−MS-selective ion monitoring (GC−MS-SIM), which offers an increased sensitivity and specificity, enabling better fingerprinting of known oil-source biomarkers.14,20 A chemical analysis is, however, costly and time-consuming, and alternate fluorescence methods such as excitation emission matrix (EEM) spectroscopy combined with parallel factor (PARAFAC) analysis have been used to support the initial discrimination of oil types.21 Following the DwH spill, PAHs9
INTRODUCTION The Deepwater Horizon (DwH) oil “spill” at the Macondo prospect (MC252), located ∼77 km southeast of the Mississippi River Delta in the northern Gulf of Mexico (NGoM), released approximately 4.4 million barrels (∼5.0 × 108 L) of oil into the Gulf between April 20 to July 15 2010.1,2 In addition, ∼2.1 million gallons (7.95 × 108 L) of chemical dispersants were released to dissipate the oil on the sea surface and at the wellhead, likely contributing to a greater dissolution of the oil components in the seawater.2 These dispersants had the potential to increase the oil’s toxicity, bioavailability, and introduction into the food web.3−5 A large midwater plume of dispersed oil extended southwest from the MC252 at a depth of ∼1100 m and was also enriched with oil-degrading bacteria.6−8 Following the spill, high levels of polycyclic aromatic hydrocarbons (PAHs) and n-alkanes were detected near the wellhead,7,9 with surface slicks covering a considerable area of the NGoM10,11 that spread to the Louisiana and Alabama coastal waters and impacted coastal ecosystems.4,12−16 Recent studies contend that there was hydrocarbon dispersion over a wider area than previously predicted or reported and that there was more oil remaining in the subsurface water column due to © XXXX American Chemical Society
Received: October 8, 2015 Revised: April 7, 2016 Accepted: April 18, 2016
A
DOI: 10.1021/acs.est.5b04924 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology and the more advanced fluorescence techniques using EEMs with PARAFAC modeling were applied in the NGoM, thereby providing additional insights into the composition and state of DOM and oil components.22−27 The absorption by the dissolved or chromophoric component of DOM (CDOM) generally decreases in the offshore direction in the NGoM.28−31 Although the CDOM optical properties of absorption have been well-documented, observations of fluorescence properties, especially analyses using EEMs spectroscopy, have been lacking for samples taken before the DwH spill. EEMs spectroscopy, however, is widely used to characterize CDOM in oceanic and coastal environments for applications such as source identification of fluorophores in CDOM, differentiating between terrestrial and marine CDOM, and detecting anthropogenic activity and environmental pollutants.32−40 Different fluorescent components of DOM, such as humic-like or protein-like material, have been identified, with the humic-like materials displaying broad fluorescence peaks with emission wavelengths typically above 400 nm, whereas the protein-like material is characterized by narrower peaks at emission wavelengths typically below 400 nm.32,41 Fractions within the humic group include compounds with fluorescence peaks labeled as “A” and “C”, typical of terrestrial sources that are often associated with river discharge in coastal environments.32 These peaks are red-shifted because of their greater aromaticity in comparison to marine humics or “M”-type compounds, a characteristic that is attributed to biological activity and microbial reworking of plankton-derived DOM.42 The spectral characteristics of the protein-like fluorescent fraction of DOM are similar to those of pure tryptophan and tyrosine, which are common in living organisms and also found in natural waters.41 These materials are derived from freshly produced DOM or may be the result of bacterial activity.34,43 Because EEMs are due to various fluorophores with overlapping spectra, statistical modeling approaches such as PARAFAC are commonly used to decompose EEMs data into their individual fluorescent components as well as to assess the quantitative and qualitative changes in the DOM pool.35 Generally, due to limitations of instrument noise at the lower UV wavelengths and because of the effect of background fluorescence,34,43 most EEMs studies (this study included) have used the ∼250−500 nm excitation spectral range.33−35,37−40 However, due to the enhanced sensitivity of the PAHs at the shorter wavelength range, studies for PAH detection, including those reported following the DwH spill, have presented EEMs results with PARAFAC modeling of the components in the ∼220−450 nm excitation spectral range.22−27,44 For consistency, the results from this study were compared to the above studies only for PARAFAC model components derived in the overlapping region of the excitation and emission (ex and em) spectra. The results from fluorescence measurements taken after the DwH spill accident, supported in many instances by GC−MS results, have identified some oil-related fluorescent components linked to PAHs dissolved in seawater.22,24−26 Field and laboratory studies indicate that there were increased levels of dissolved organic carbon (DOC) concentrations and changes in fluorescence properties of CDOM linked to the oil spill.22,23 More recent studies (Bianchi et al.)27 that combined biogeochemical and fluorescence measurements suggest a longer-term effect of the DwH oil on the optical properties and composition of dissolved organic matter in the offshore waters of the NGoM. Although the DwH oil spill also strongly
impacted the coastal waters and ecosystems, 4,15,16 the application of CDOM fluorescence analyses for oil spill detection has been sparse in coastal waters because of the strong fluorescence associated with riverine- and wetlandsderived organic material39 that could mask the oil-related signatures. Major questions, therefore, remain on the fate of spilled oil, especially related to its impact on the DOM pool in the coastal and offshore waters of the NGoM. Here, we present the first comparisons of prespill baseline fluorescence EEMs data combined with PARAFAC analysis with postspill observations that allow for an assessment of short- and long-term impacts (2010−2013) of the DwH spill on the coastal and open-ocean DOM fluorescence pool.
■
MATERIALS AND METHODS Study Sites and Sample Collection. We collected water samples in the NGoM shelf waters in 2009 and in the vicinity of the DwH spill site in May 2010, April 2012, and April 2013. Samples were also collected at four other coastal sites in May, June, and July 2010. The prespill cruise in January 2009 included a cross-shelf transect off Mobile Bay, with samples collected at an offshore station that was located just 34 km from the Macondo wellhead (MC252) (Figure 1). Seawater samples were collected May 15−17 2010, during the DwH spill, from approximately 1−16 km from the DwH wellhead with Niskin bottles. These samples were brought back to the laboratory and
Figure 1. Top: sampling locations in the northern Gulf of Mexico. Blue dots indicate January 9−17 2009, shelf and slope stations; red dots indicate May, June, and July 2010 stations off of Mobile Bay (MB), the Mississippi delta (D), Barataria Bay (BB), and south of Terrebonne Bay (TB) and Atchafalaya Bay (AB). The yellow box indicates expanded area of Figure 1 near the DwH site (bottom). Red stars indicate May 15−17 2010; the black cross indicates the DwH site; yellow pluses indicate April 2012; and black triangles indicate April 2013. The labeled stations are discussed in the text. B
DOI: 10.1021/acs.est.5b04924 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 2. Typical EEMs spectra (shown in Raman units of intensity) of samples at station a2 (off of Mobile Bay) and a6 (offshore, ∼34 km from the DwH spill site; Figure 1) in 2009 before the spill (top left). Fluorescence EEM spectra of an MC252 oil sample (oil drop mixed in MQ water and filtered through a 0.2 μm filter, ex−em peaks of 275 and 320 nm, respectively, FL scale: 0.2 × 5) and example fluorescence EEM spectra of seawater samples obtained during the DwH oil spill (May 15−17, 2010) at locations near the MC252 wellhead: B27A (surface, 8 km from wellhead; 270 and 325 nm); (middle panel) B23B (5 km from wellhead, 5 m depth, 1.09 μg/L PAHs, 270 and 315 nm); B28B (16 km from the wellhead, 5 m depth, weathered oil, 270 and 315 nm); B26A (1 km from wellhead, surface, match to MC252, ∼229.9 μg/L PAHs, 270 and 315 nm), B26J (1 km from wellhead, 1290 m depth; 275 and 320 nm); (lower panel) typical seawater coastal samples BB (Barataria Bay) and mb5 (near Fort Morgan, Mobile Bay; FL scale: 0.1 × 2) collected during the oil spill in 2010. Typical EEMs of samples collected near the DwH spill site in April 2012 (d12_3 at 100 m depth, 270 and 315 nm) and April 2013 (d13_3 at 75 m depth in Figure 1 bottom, 275 and 325 nm).
processed for PAHs using GC−MS-SIM and also filtered through 0.2 μm filters for CDOM EEMs analysis. We collected additional samples from coastal sites impacted by the oil spill offshore of the Delta at Pass a Loutre on May 27 and 28 2010, in Barataria Bay on June 6, in Mobile Bay in early July 2010, and from transects off the Terrebone and Atchafalaya Bays between June 12 and 14 2010 (Figure 1). Samples were collected after the spill in April 2012 and 2013 in the vicinity of the DWH wellhead and processed for EEMs fluorescence, DOC concentrations, and dissolved PAHs (Figure 1). Fluorescence EEMs and PARAFAC Model Analysis. We filtered water samples through 0.2 μm nucleopore polycarbonate membrane filters under low-vacuum conditions. Most samples were immediately filtered on board or after return from boat trips on the same day and stored at 4 °C in the dark in acid-cleaned, precombusted amber bottles before laboratory analysis. EEMs were recorded using a FluoroMax-4 (Jobin Yvon Horiba) fluorometer by scanning the emission spectra from 290 to 550 at 5 nm intervals over excitation wavelengths between 250 and 450 at 5 nm increments. The EEMs spectra were obtained after correction of the fluorescence spectra for instrument bias and the water Raman normalization of the fluorescence intensity. The correction for inner filter effects (IFE) using the absorbance-based approach45,46 was applied to samples obtained in the bays where absorbance values were elevated. Some samples collected in the delta and bays with high absorbance values that likely exceeded the linear range of the spectrophotometer and required dilution46 were excluded from PARAFAC analysis. For the coastal and offshore samples,
absorbance values at 250 nm were generally close to or less than ∼0.05, and no IFE correction was applied. Thus, errors of about 6% would be expected at the lowest excitation and emission wavelengths.46 The final EEMs fluorescence values are reported in equivalent Raman units (RU).39,40 The resulting EEM fluorescence observations were evaluated by PARAFAC analysis using the DOM-Fluor toolbox35 with the model constrained by non-negativity and run with three to seven components. Model validation was carried out using split-half analysis, residual analysis, and random starts.35 The PARAFAC analyses of the EEMs data were done separately for the prespill samples obtained in 2009, for the coastal samples (Barataria Bay, Delta, Terrebonne and Atchafalaya Bays, and Mobile Bay), and the DwH spill site in May 2010, April 2012, and April 2013, as well as using the combined (NGoM) data set (Figure 2, 3, 4; Table 1). This approach was taken on the basis of previous work47−49 and to obtain a better understanding of the peak shifts and overlaps observed in the individual models in comparison to components identified in the regional NGoM model composed of the various individual data sets. Also, residual EEMs (difference between sample and model derived EEMs) were found to be generally smaller for individual than the NGoM model (see Supporting Information S1 and S2), suggesting better characterization of the fluorescent components by the individual models, especially those linked to oil-related fluorescent signatures in the seawater. Thus, in this study, the outputs of the individual PARAFAC models were used to examine the relative variations in fluorophore composition and concentration36,49 for the different subsets C
DOI: 10.1021/acs.est.5b04924 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 3. Comparison of PARAFAC components (excitation to the left of emission spectra) for sampling conducted prespill in the shelf (shelf09, red dashes), during the spill in May 2010 near the DwH spill site (dwh10, cyan dashes), bays and coastal waters during the spill in May and June 2010 (bays10, blue dashes), and post-spill near the DwH site in April 2012 (dwh12, green dashes) and April 2013 (dwh13, pink dashes). Also shown are PARAFAC components derived using the NGoM or combined EEMs data set used in all the individual models (G1−G6, black lines).
Figure 4. Fluorescent signatures (RU) of PARAFAC components (C1−C4; Table 1) for samples collected near the DwH spill site (Figure 1) in May 2010 (top), April 2012 (middle), and April 2013 (bottom).
MS-SIM) for samples collected at the DwH site in May 2010 and at Mobile Bay in July 2010, and for PAHs at the DWH site locations in 2012 and 2013. The samples were processed and analyzed following the method outlined in Adhikari et al. 2015.51 Further, oil source-fingerprinting techniques (e.g., comparison of sample oil biomarker m/z chromatograms to
of EEM data and the NGoM model, providing an overview of the spectral variability in the fluorescent components identified in the complex oil-impacted NGoM waters. GC/MS Analysis. The analysis of seawater samples for PAH components and n-alkanes50 was performed by gas chromatography−mass spectroscopy with selective ion monitoring (GC− D
DOI: 10.1021/acs.est.5b04924 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
in acid-cleaned, precombusted amber bottles with Teflon-lined caps. Laboratory measurements of DOC were made on a Shimadzu TOC 5000A (with ASI-5000A autosampler) using a high-temperature combustion method to convert carbon compounds to carbon dioxide, including using the consensus reference material (CRM) for quality assurance and quality control.40,52,53
Table 1. Broad Description of the Various Components Identified by PARAFAC Analysis of EEMs Fluorescence Data for the Different Sampling Sites and Times: shelf09, NGoM Shelf and Slope Waters in 2009; bays10, Coastal Sites in May, June, and July 2010; dwh10, dwh12, and dwh13, Deepwater Horizon Site and Vicinity in May 2010, April 2012, and 2013a components (location and year)
ex and em maxima (nm)
description and probable source (references)
G1 C1(bays10) C1(dwh10) G2 C1(shelf09)
