Dissolved organic matter and trace element variability in a blackwater-fed bay following precipitation

Journal Pre-proof Dissolved organic matter and trace element variability in a blackwater-fed bay following precipitation M.S. Sankar, Padmanava Dash, YueHan Lu, Varun Paul, Andrew E. Mercer, Zikri Arslan, Jac J. Varco, John C. Rodgers PII:

S0272-7714(19)30325-7

DOI:

https://doi.org/10.1016/j.ecss.2019.106452

Reference:

YECSS 106452

To appear in:

Estuarine, Coastal and Shelf Science

Received Date: 29 March 2019 Revised Date:

21 October 2019

Accepted Date: 25 October 2019

Please cite this article as: Sankar, M.S., Dash, P., Lu, Y., Paul, V., Mercer, A.E., Arslan, Z., Varco, J.J., Rodgers, J.C., Dissolved organic matter and trace element variability in a blackwater-fed bay following precipitation, Estuarine, Coastal and Shelf Science (2019), doi: https://doi.org/10.1016/ j.ecss.2019.106452. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Ltd.

1

Dissolved organic matter and trace element variability in a blackwater-fed

2

bay following precipitation

3 4

M. S. Sankar1, Padmanava Dash1*, YueHan Lu2, Varun Paul1, Andrew E. Mercer1, Zikri

5

Arslan3, Jac J. Varco4, John C. Rodgers1

6 7

1

Department of Geosciences, Mississippi State University, Mississippi State, MS 39762, USA

8

2

Department of Geological Sciences, University of Alabama, Tuscaloosa, AL 35487, USA

9

3

Department of Chemistry & Biochemistry, Jackson State University, Jackson, MS 39217, USA

10

4

Department of Plant & Soil Sciences, Mississippi State University, Mississippi State, MS

11

39762, USA

12 13 14 15

*Corresponding author: Department of Geosciences, Mississippi State University, Mississippi

16

State, MS 39762. Room # 109B, Hilbun Hall, Email: [email protected], Phone: 001-662-325-

17

0364, Fax: 001-662-325-9423

18 19 20 21 1

22

Abstract

23

Dissolved organic matter (DOM) often forms complexes with trace metals. The main objective

24

of this study was to identify the sources of trace elements to a coastal bay that is fed by two

25

blackwater rivers using DOM compositions. Surface water samples from twelve sites in Weeks

26

Bay, Alabama were collected during four field trips and a bottom sediment sample was collected

27

during one of the trips. Spectroscopic measurements in tandem with parallel factor analysis and

28

multivariate statistics were used to derive DOM compositions and inductively coupled plasma

29

mass spectrometry was used for determining trace metal concentrations. DOM chemistry and

30

trace element concentrations together with precipitation and discharge, watershed land use and

31

land cover data, and physicochemical parameters were used to determine the source of trace

32

elements in the adjoining areas of the watershed to the bay and finally settling into the bay

33

sediments. Arsenic, copper, and uranium fluxes increased while fluxes of mercury, zinc,

34

manganese, and iron decreased following precipitation events. Concentration of all trace

35

elements analyzed were 10,000 times higer in the bay botton sediments than their concentrations

36

in the water colunmn. Microbially reprocessed and soil derived DOM components increased

37

along with the trace elements and nutrients in the bay as a result of precipitation whereas

38

concentration of terrestrial humic-like DOM component remained the same during both dry

39

(spring) and wet (summer and fall) periods. Principal component analysis revealed significant

40

correlations between Microbially reprocessed DOM and the trace elements including arsenic,

41

copper and uranium indicating urban, pasture, and agricultural areas in the watershed were the

42

sources for these trace metals. On the other hand, multiple sources were identified for

43

manganese, zinc, and iron. Overall it was found that DOM compositions are useful for inferring

44

the possible sources and mobility of trace elements in aquatic systems.

2

45

Keywords: Dissolved organic matter; Trace elements; Weeks Bay; Precipitation events; EEM-

46

PARAFAC; PCA

47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 3

68 69

1. Introduction It has been recently reported that shallow Mississippi deltaic sediments serve as a source

70

of iron, manganese, and arsenic to the water column and redox reactions in the deltaic marshes

71

control the availability of trace elements like molybdenum, tungsten, arsenic, vanadium in the

72

coastal waters (Joung and Shiller, 2016; Telfeyan et al., 2018, 2017). The contamination of

73

mercury in the estuaries, bays and offshore regions of northern Gulf of Mexico (NGoM) has

74

been attributed to urban and agricultural runoff as well as atmospheric deposition (Apeti et al.,

75

2012; Lewis and Chancy, 2008). Consensus is yet to be reached regarding the source of these

76

trace elements, despite the associated contamination being one of the major problems adversely

77

impacting many forms of life in these coastal waterbodies (Lafabrie et al., 2011; Lewis and

78

Russell, 2015; Perry et al., 2015). Naturally occurring DOM is a complex mixture of different

79

organic compounds which can form complexes with trace elements, hence trace elements often

80

co-occur and covary with DOM of different origins (Yamashita and Jaffé, 2008). Based on DOM

81

composition, a substantial amount of humified terrestrial and some protein-like DOM

82

components were identified along the coastal water bodies of the southern US (Sankar et al.,

83

2018; Shank et al., 2011; Singh et al., 2010). Rivers and streams flowing into these coastal

84

waterbodies drain agricultural, forested and urbanized hinterlands and transport high amount of

85

dissolved nitrate, phosphate, carbon, and trace elements (Lafabrie et al., 2011; Reuter and

86

Perdue, 1977; Sankar et al., 2018; Stolpe et al., 2010; USEPA, 1999). There is a strong

87

correlation between the amount of water discharged and the quantity of organic matter delivered

88

through the coastal rivers of the United States (Keul et al., 2010; Reuter and Perdue, 1977; Zhou

89

et al., 2016). Specific studies have been conducted to understand the role of DOM in the

90

mobilization of trace elements (Hongxia et al., 2014; Reuter and Perdue, 1977; Stolpe et al.,

4

91

2010; Zhou et al., 2016). Strong metal binding capacity of both bacterially derived and soil

92

derived humic DOM compositions have been widely reported (Gu et al., 2011; Hongxia et al.,

93

2014; Reuter and Perdue, 1977). For instance, dissolved and colloidal trace elements, along with

94

nutrients and DOM are transported by Mississippi-Atchafalaya river system (Duan et al., 2007;

95

Stolpe et al., 2010; Zhou et al., 2016). This association might have important consequences in

96

terms of delivery and accumulation of toxic trace elements, such as arsenic, manganese, copper,

97

nickel, uranium, mercury, zinc, and nutrients including nitrate and phosphate along the

98

Mississippi deltaic sediments (Joung and Shiller, 2014, 2016; Slowey and Hood, 1971; Stolpe et

99

al., 2010).

100

Weeks Bay is a small coastal water body located near the city of Mobile (3rd most

101

populous city in Alabama) and is connected to NGoM through Mobile Bay. The freshwater input

102

into Weeks Bay is through two blackwater rivers, Fish River in the north and Magnolia River in

103

the southeast. Blackwater rivers are quite common all along the coastal areas of the southern US

104

and are connected to NGoM through estuaries and bays (Ruecker et al., 2017). They drain

105

extensive forested wetlands, thereby transporting brown colored water rich in soil-derived DOM

106

to coastal oceans (Spencer et al., 2013). Previous studies have shown that soil-derived DOM is

107

efficient in mobilizing trace elements via complexation (Benedetti et al., 1996; Tribovillard et

108

al., 2004; Weng et al., 2002). We hypothesized that these blackwater rivers are the main conduits

109

for not only DOM but also for toxic trace element transport into the coastal water bodies as a

110

result of rapid urbanization and industrialization in the southeastern US. We further hypothesised

111

that following precipitation events, the concentration of trace metals co-varied with specific

112

DOM components in the bay, and the DOM composition can be utilized to identify the source of

113

DOM and in turn, the trace elements. Hence, our objectives were to (1) determine the trace

5

114

element-DOM compositional associations, and (2) evaluate the changes in composition and

115

concentration of DOM and trace elements delivered into the Weeks Bay, NGoM in relation to

116

the average amount of precipitation in the watershed.

117

2. Materials and Methods

118

2.1 Study area, sample location and land use/ land cover analysis

119

Weeks Bay is fed by black water rivers and located on the eastern side of the Mobile Bay

120

in Baldwin County, coastal Alabama (Fig.1). It is designated as the 16th national estuarine

121

research reserve (NERR) of the US in 1986, hosting a wide variety of aquatic plants and animal

122

species. It is a shallow micro tidal tributary estuary (Camacho et al., 2014). The Weeks Bay

123

watershed covers a total area of 520.87 km2 and the Weeks Bay is a shallow water body with an

124

average depth of 2 m covering an area of 4.72 km2 (0.91% of the total watershed area). It

125

empties into the larger Mobile Bay through a narrow opening and the average water residence

126

time of Weeks Bay is 5 days (Caffrey et al., 2014). Two rivers flow into the bay; Fish River in

127

the north and Magnolia River in the south-east. To delineate the drainage basins of these two

128

rivers, hydrologic unit code-12 (HUC-12) boundary datasets for the Weeks Bay watershed were

129

downloaded from the United States Department of Agriculture (USDA) data gateway and 10 m

130

resolution digital elevation models were downloaded from the United States Geological Survey

131

(USGS) National Map download client. The digital elevation models and the HUC-12 watershed

132

boundary datasets were used in Arc-GIS 10.4.1 platform to delineate the drainage basins of the

133

Fish and Magnolia Rivers emptying to the Weeks Bay. Subsequently, the land use and land

134

cover dataset of 2015 was downloaded from the USDA data gateway and cropped using the

135

delineated watershed boundary for determining the areas covering each major land use and land

136

cover. It was found that the bay and associated river channels are surrounded by emergent 6

137

herbaceous wetlands (6.46 km2, 1.24 %), woody wetlands (67.62 km2, 12.98 %) and evergreen

138

forest (72.88 km2, 13.99 %) (Table 1). Cultivated cropland and pasture make up the major part of

139

the watershed (127.52 km2, 24.48 % and 98.94 km2, 19.00 %). Urbanized areas (68.38 km2,

140

13.13 %) were located on the northwestern and northern parts along Fish River and along the

141

southeastern areas of Magnolia River.

142

Four field campaigns were conducted to measure in situ water quality parameters and

143

collect water samples. Twelve water samples were collected during the campaigns on

144

05.09.2016, 06.19.2016 and 09.30.2016, while thirteen samples and a bay bottom sediment were

145

collected during the campaign on 06.26.2016 (Fig. 1). Water samples were collected from

146

locations distributed all along the bay in clean, acid-washed 1-liter high density polyethylene

147

(HDPE) bottles following the protocol explained in Dash et al. (2015). Sediment from the bottom

148

of the bay was collected using an extendable grab sampler and was stored in an O2-impermeable

149

Remel®bag. Water quality parameters including pH, oxidation-reduction potential (ORP, mV),

150

dissolved oxygen (DO, mg/L), total dissolved solids (TDS, mg/L), salinity (PSU) and water

151

temperature (°C) were collected from each sampling location during every field campaign by

152

using a HANNA® extendable probe. The collected water and sediment samples were stored in an

153

airtight cooler filled with ice and transported back to the lab. Water samples were filtered using

154

0.2 µm Whatman® nuclepore track-etch membrane filters, stored in amber colored glass bottles,

155

and refrigerated at 4°C until spectroscopic analysis of colored dissolved organic matter (CDOM).

156

The samples for dissolved organic carbon (DOC), total dissolved nitrogen (TDN), and anions

157

(nutrients) analyses were stored in acid washed 20 mL HDPE bottles and kept frozen in a -80 °C

158

freezer until analysis. The water samples for cations (trace elements) were acidified using

159

Fisherbrand® ultrapure nitric acid (0.2 %) and stored in acid washed and pre-combusted 20 mL

7

160

glass bottles and stored in room temperature until analysis following the U.S. Geological Survey

161

and U. S. Environmental Protection Agency protocols (USGS, 1999; USEPA, 1983).

162

The field campaigns were planned in accordance with precipitation events and the

163

corresponding river water discharge into the bay (Fig.2). Discharge data were obtained from the

164

United States Geological Survey (USGS) national water information system web interface. The

165

USGS discharge gauges were located at Magnolia River (USGS 02378300) and Fish River

166

(USGS 02378500). Daily average precipitation data for the Weeks Bay watershed was obtained

167

from National Weather Service (NWS) rainfall plots.

168

2.2 Water sample analysis, CDOM indices and PARAFAC modeling

169

Trace elements; arsenic (As), manganese (Mn), zinc (Zn), copper (Cu), mercury (Hg)

170

uranium (U) and iron (Fe) were analyzed using inductively coupled plasma mass spectrometry

171

(ICP-MS). Nutrients including phosphate (PO43-), nitrate (NO3-), ammonium (NH4+) and nitrite

172

(NO2-) anions were analyzed using a continuous flow auto-analyzer (Skalar Analytical Inc.,

173

Buford, GA) at Dauphin Island Sea Lab. The DOC and TDN content of the water samples were

174

measured on a SHIMADZU® total organic carbon-total nitrogen analyzer (TOCv-TNM1)

175

equipped with an ASI-V autosampler following the method described in Shang et al., 2018.

176

Toxicity of trace elements and nutrients were expressed in terms of the Unites States

177

Environmental Protection Agency (US EPA) recommended water quality criteria.

178

CDOM absorption was analyzed in a PerkinElmer® Lambda 850 UV/VIS double-beam

179

spectrophotometer (Waltham, MA, USA) following the analytical procedure described in Singh

180

et al., 2017. Absorption spectra of the samples and blank (deionized water, 18.2 mΩ purity) were

181

collected between 200nm and 750nm wavelengths at 2 nm interval in a clean 4 mL quartz

182

cuvette with 1cm path length. Effects of background and scattering were removed by subtracting 8

183

the absorption spectra of the samples from the blank. The blank subtracted sample absorbance

184

(A) was used to calculate Napierian absorption coefficient (a) at each wavelength (λ) (Fichot and

185

Benner, 2012; Singh et al., 2017; Zhang et al., 2013). Absorption coefficients were used to

186

calculate specific UV absorbance (SUVA254) by dividing absorption coefficient at 254nm by

187

DOC concentration (Lmg-1m-1). The slope ratio (SR) is the ratio of absorption coefficient

188

between 275 and 295 nm to that between 350 and 400 nm. Values of SUVA254 are directly

189

proportional to aromaticity of DOM, while SR values are inversely related to the molecular

190

weight of DOM in the water sample (Hansen et al., 2016; Kothawala et al., 2014; H. Lin et al.,

191

2016; Singh et al., 2017).

192

Fluorescence properties of the water samples (Excitation emission matrix, EEM) were

193

analyzed using Horiba® Jobin Yvon Inc., Fluoromax-4 spectrofluorometer (Edison, NJ, USA)

194

following the analytical procedure explained in Sankar et al., 2018. After applying company

195

specified instrumental corrections and calibrations, the EEMs of the samples and blanks

196

(deionized water,18.2 mΩ purity) were measured with excitation wavelengths of 240-450nm at

197

every 10 nm increment and emission wavelengths of 300-550nm at every 2nm increment. After

198

removing inner filter effects; the blank corrected and Raman normalized EEMs were used for

199

PARAFAC modeling (Kothawala et al., 2013; Singh et al., 2017; Stedmon and Markager, 2005).

200

Fluorescence Index (FI), Humification Index (HIX) and Biological Index (BIX) or Freshness

201

Index were calculated using the corrected EEMs. The calculation steps for these indices can be

202

found in Hansen et al., 2016 and elsewhere. The FI index indicates the source of DOM, which

203

can either be microbial (with high FI, between 1.6-2.0, usually derived from extracellular release

204

and leachate from bacteria and algae) or terrestrial (with low FI, between 1.2-1.5, usually from

205

terrestrial plants and soil organic matter). HIX can be used as a proxy for humification of natural

9

206

organic matter. An increase in HIX values indicates the condensation of fluorescent molecules

207

and decrease in the H/C ratio of DOM. The BIX is an indicator of the freshness of DOM. High

208

BIX values (>1) indicate freshly released DOM as a result of autochthonous production in the

209

aquatic system. On the other hand, low values of BIX (0.6-0.7) indicate less autochthonous

210

production of DOM in the water body (Huguet et al., 2009)

211

A total of 49 sample EEMs were used for PARAFAC modeling. After removal of the

212

outliers, the sample size was reduced to 46. Out of the 46 sample EEMs, 11 were following the

213

spring sampling event (9th May 2016), 12 each following summer-1 (19th June 2016) and

214

summer-2 (26th June 2016) sampling events, and 11 were following the fall sampling event (30th

215

September 2016). To avoid abnormalities during the modeling, the EEMs were normalized using

216

Raman integrated area at 350 nm excitation and 397 nm emission over 5 nm bandpass. Prior to

217

PARAFAC modeling, the excitation wavelengths were interpolated at 5 nm increments, and

218

emission wavelengths were interpolated at 4nm increments. A four component PARAFAC

219

model which explained 99.96 % variance of the data was validated by split half analysis using

220

Tucker congruence coefficients and random initialization. The scores of the PARAFAC

221

components (fluorescence intensity maximum, Fmax) were expressed in Raman units (R.U). The

222

DOMFluor tool box in MATLAB®- R2018a computing platform was used to create the

223

PARAFAC model (Stedmon and Bro, 2008).

224

2.3 Sediment characterization and geochemical analysis

225

The mineralogical characterization of the bay bottom sediment was carried out using a

226

Rigaku Ultima III X-ray diffractometer with a graphite monochromator and scintillation detector

227

at I2AT laboratory, Mississippi State University. The sample was pulverized and homogenized

228

using an agate mortar and piston and analyzed over 0° to 90° (2θ) angle. 10

229

Trace element analysis of the bay bottom sediment sample in duplicates were carried out

230

for As, Mn, U, Cu, Zn, Hg, and Fe. 0.5 gm of homogenized sediment sample and its duplicate

231

were digested using concentrated trace metal grade nitric acid and 30 % hydrogen peroxide on a

232

hot plate at 90°C following the EPA protocol (method SW-846 3050B). The digestate was later

233

refluxed using concentrated trace metal grade hydrochloric acid and then diluted to 50 mL using

234

deionized water before the ICP-MS analysis for As, Mn, U, Cu, Zn and Fe. The concentration of

235

Hg in bay bottom sediment was analyzed separately by cold vapor atomic absorption (CVAA)

236

spectroscopy (SW-846 Method 7471B). The soil standard reference material (PR-LCSSRM-

237

00018 B/ERA-550/http.eraqc.com) was also analyzed to estimate recovery.

238

2.4 Statistical Analysis

239

The R-version 3.3.2 (R® Core Team) programming platform was used for the statistical

240

analysis of all the measured and observed data. Principal Component Analysis (PCA) was

241

performed on PARAFAC components, fluorescence indices, absorption indices, water quality

242

parameters, and fluxes of DOC, TDN, nutrients, and trace elements to understand their

243

relationship. Bootstrap resampling statistics at confidence intervals of 2.5%, 50%, and 97.5%

244

was applied to trace element concentrations to estimate whether they were significantly different

245

statistically following different precipitation events by taking 1000 means (1000 replicates).

246

Mann-Whitney-Wilcoxon Test was used to explain the significant increase or decrease in

247

PARAFAC components from spring with low amount of precipitation to fall with high amount

248

of precipitation. The SUVA254 values showed minimum variation between the seasons hence,

249

One-way ANOVA was performed to understand the variance between and among the SUVA254

250

values during four sampling events. Tukey-HSD pairwise multiple comparisons post-hoc

251

analysis was also performed to find the significance in differences between SUVA254 values 11

252

following precipitation events. All of the statistical analyses were performed at 95% confidence

253

level (α=0.05). Maximum, minimum, average (mean) and standard deviation were calculated for

254

all the measured parameters. Interquartile range (IQR) was calculated for expressing the

255

variability in the observations. Data values greater than 3 times the IQR or above the third

256

quartile in the box plot were presented as outliers. Both p-value (t-test of regression, α=0.05) and

257

r value or correlation coefficient (0.5>r≥0.3 weak, 0.7>r≥0.5 moderate and r≥0.7 strong

258

relationship) were estimated to represent the significance, strength, and direction of the

259

relationship among the variables.

260

3. Results

261

3.1 Temporal changes in water quality and chemistry in response to seasonal meteorological

262

events

263

Monthly meteorological data for the year 2016 revealed that average precipitation in the

264

Weeks Bay watershed was higher during September 2016 (mean=8.49 mm, max=49.53 mm,

265

min=0 mm) followed by June 2016 (mean=5.60 mm, max=47.50 mm, min=0 mm) and May

266

2016 (mean=3.05 mm, max= 27.43 mm, min=0 mm). The precipitation events in the watershed

267

were marked by a sudden increase in water discharge into the bay through both rivers (Fig.2).

268

The average discharge of Fish and Magnolia Rivers during the months of September, June and

269

May 2016 were 2.11±0.79 m3/s, 2.38±0.82 m3/s, 3 .20±2.57 m3/s and 1.19±1.42 m3/s, 0.74±0.34

270

m3/s, 0.77±0.32 m3/s. The Fish River is larger yet displayed a greater increase in discharge than

271

the Magnolia River. As such, our sampling campaigns represented four seasonal precipitation

272

conditions, we coined spring for 9th May 2016, summer-1 for 19th June 2016, summer-2 for 26th

273

June 2016 and fall for 30th September 2016. The average amount of precipitation in the

274

watershed over five days prior to the sampling events is used for describing the variations during 12

275

each of the sampling events (Table 2). It is also important to note that the soil profile of entire

276

region remained wet with rainwater following 26th of June 2016 (summer-2) due to the

277

prevalence of rain events from the month of June 2016 onwards. Similarly, several precipitation

278

events were observed in September 2016.

279

Water quality parameters (pH, ORP, TDS, Salinity, DO, water temperature) of the bay

280

exhibited changes in accordance with seasonal precipitation and corresponding discharge events

281

(Table 3). The pH, TDS, and salinity showed an increasing trend, while both ORP and DO

282

values decreased with precipitation and discharge. pH values following fall precipitation event

283

were higher than the pH following spring precipitation event. The steady increase in pH from

284

spring to fall through summer precipitation events in the bay indicates freshwater input following

285

precipitation events. Net negative ORP values of the bay during all field sessions show the

286

prevalence of reducing condition. The water temperature was variable following spring

287

precipitation event compared to the summer and fall. It is important to note that high tides were

288

coinciding during water sampling following the summer-1 (0.52 m), summer-2 (0.40 m), and fall

289

precipitation events (0.40 m), while low tides coincided during sampling following the spring

290

precipitation (0.3 m) event (www.tidesandcurrents.noaa.gov).

291

The flux of trace elements, nutrients, DOC and TDN were calculated by multiplying their

292

concentration with five-day average river discharge to evaluate their variability between three

293

different seasons. While their actual concentrations were used to explain their toxicity if they had

294

exceeded EPA recommended water quality criteria for saltwater (EPA, 2018). The average flux

295

of As (20.75±1.29 mgs-1 to 43.46± 10.09 mgs-1), U (1.62±0.39 mgs-1 to 2.41±0.62 mgs-1) and Cu

296

(1.82±0.33 mgs-1 to 73.30±78.81 mgs-1) in the bay showed an increase in trend following the

297

spring precipitation event to fall precipitation event (Fig. 3). On the contrary, the average flux of 13

298

Hg (0.42±0.13 mgs-1 to 0.21± 0.63 mgs-1), Fe (1432.19±563.30 mgs-1 to 444.15±146.15 mgs-1),

299

Mn (268.53±83.02 mgs-1 to 156.50±31.16 mgs-1), and Zn (9.04± 8.69 mgs-1 to 3.45±1.45 mgs-1)

300

showed a general decrease in trend from spring precipitation event to fall precipitation event.

301

Arsenic concentrations were highly variable between sites following the fall precipitation event

302

(IQR=10.99) with an average value of 23.31±5.41 µgL-1 (SI Fig. 1). The maximum variability of

303

Cu flux between sites was observed following fall precipitation event (IQR= 79.72) and for U it

304

was following summer-1 precipitation event (IQR=1.25). Cu concentrations exceeded the EPA

305

recommended water quality criteria of 4.8 µgL-1 for acute toxicity and 3.1 µgL-1 for chronic

306

toxicity (EPA, 2018) following the fall precipitation event with an average concentration of

307

39.30±42.26 µgL-1 and at one site near the mouth of Fish River following the summer-1

308

precipitation event with a concentration of 14.39 µgL-1. The average concentration of Zn was

309

2.91±6.97 µgL-1 with maximum variation of flux between sites following the spring precipitation

310

event (IQR= 4.96). Fluxes of Hg, Fe, and Mn showed highest variation between sites following

311

the spring precipitation event (IQR=0.21; 624.27; 116.77). The average concentration of Mn was

312

always high and with an average concentration of 92.98±27.01 µgL-1. Concentration of Fe

313

following summer-1, summer-2, and fall precipitation events were lower (avg=235.57±73.03

314

µgL-1) compared to its concentration following the spring precipitation event

315

(avg=455.39±179.11 µgL-1). On the contrary, the Hg concentration was always below its toxicity

316

limits (EPA recommended water quality criteria for salt water 1.8 µgL-1 for acute toxicity and

317

0.94 µgL-1 for chronic toxicity (EPA, 2018)) with an average value of 0.08±0.11 µgL-1 except of

318

one site where the concentration was 1.17 µgL-1 near the mouth of the Weeks Bay Estuary to the

319

Mobile Bay (SI Fig. 1).

14

320

There was an increase in the average flux of both NO3- (214.84± 155.78 mgs-1 to

321

4211.34± 1550.98 mgs-1) and PO43- (109.27± 39.74 mgs-1 to 389.37± 122.09 mgs-1) from spring

322

to fall precipitation events. However, PO43- showed slight decrease in flux following the

323

summer-2 precipitation event. On the other hand, both NH4+and NO2- showed no such trend in

324

flux (Fig. 4). The greatest variation of PO43- flux was following the high precipitation event

325

(IQR= 184.65), and for NO3-, it was following the summer-1 precipitation event (IQR= 1724.69).

326

Both NH4+ and NO2- showed maximum variation following the low precipitation event (IQR=

327

469.37 and 439.25). The average concentrations of NO3-, NO2-, and NH4+ were 1267.41±478.78

328

µgL-1, 40.84±71.11 µgL-1 and, 24.93±28.26 µgL-1, respectively. PO43- concentrations were high

329

in all sampling locations following the high precipitation event (mean= 208.78± 65.46 µgL-1,

330

max= 303.91 µgL-1) and at one location following the moderately high precipitation event

331

(112.07 µgL-1) (SI Fig. 2).

332

Average flux and variation of both DOC and TDN were high following the spring

333

precipitation event compared to other sampling seasons (IQR= 2207.004; 2111.396) (Fig. 4).

334

During other three field campaigns, the flux of both DOC and TDN remained nearly same with

335

5582.79± 309.14 mgs-1; 636.61± 149.93 mgs-1, respectively, yet DOC showed a slight increase in

336

flux (mean= 7414.85± 352.73 mgs-1) following the fall precipitation event compared to the

337

summer-2 precipitation event (3979.49± 286.07 mgs-1). The actual concentration of DOC and

338

TDN is given in the supplementary material (SI. Fig. 2).

339

The fluxes of trace elements such as As, Cu, U and nutrients such as PO4-, NO3-, and

340

NO2- showed an increasing trend, while the fluxes of Fe, Mn, Zn, NH4+, DOC and TDN showed

341

a decreasing trend with salinity of the bay (SI. Fig. 4 & 5).

342

3.2 Seasonal variation of spectroscopic properties and development of PARAFAC model 15

343

The mean values of SUVA254 were high following both spring (6.68±1.43 L mg-1m-1) and

344

summer-1 precipitation events (7.50±0.63 L mg-1m-1) compared to both summer-2 (5.31±0.43 L

345

mg-1m-1) and fall precipitation events (5.89±0.39 L mg-1m-1) (Fig.5). On the other hand, average

346

values of SR showed a general increase in trend from spring (1.27±0.04) to fall precipitation

347

event (1.51±0.35) with maximum variation following the fall precipitation event (IQR= 0.43).

348

The mean values of both FI and BIX also increased from spring precipitation

349

(FI=1.44±0.02, BIX=0.65±0.02) to fall precipitation (FI=1.47±0.02, BIX=0.71±0.03) event (Fig.

350

5). The maximum variation of FI was following the summer-1 precipitation event (IQR= 0.043),

351

while for BIX it was the same following the fall-precipitation and summer-1 precipitation events

352

(IQR= 0.04). However, BIX values were always less than 1 during all field campaigns

353

(avg=0.69±0.04, max=0.77, min=0.60). The mean values of HIX remained nearly the same

354

following all the four field campaigns (0.76±0.05). The variation of HIX was maximum

355

following both fall and summer-2 precipitation events (IQR=0.08 and 0.07).

356

The PARAFAC analysis of the Weeks Bay water samples collected during the four field

357

campaigns in three different seasons resulted in a four-component model (C1, C2, C3, and C4).

358

The PARAFAC component C1 was a terrestrial, humic-like component; C2 was a microbial,

359

humic-like component; C3 was an allochthonous, terrestrial component or soil-derived

360

component; and C4 was a microbial tryptophan-like component (Fig. 6, Table 4). Components

361

C2, C3, and C4 showed a general increase of 7.61%, 11.19%, and C4=27.61%, respectively, in

362

average fluorescence intensity (Fmax) in the bay from spring precipitation to fall precipitation

363

events (Fig. 7). However, component C2 showed a slight decrease (2.18 %) in fluorescence

364

intensity from summer-2 to fall precipitation events. On the other hand, component C1 showed

365

limited change in average fluorescence intensity and had less variability during all seasons. The 16

366

components C2, C3, and C4 showed slightly higher variation in Fmax values with IQR of 0.03,

367

0.04 and 0.03, respectively, following the summer-1 precipitation event compared to other

368

precipitation events (Fig. 7).

369

3.3 Bottom sediment characterization and chemistry

370

Total digestion of Weeks Bay bottom sediment sample revealed the presence of Mn

371

(110.00±0.00 mg/Kg); Zn (30.00±1.41 mg/Kg); As (7.75±0.21 mg/Kg); Cu (4.05±0.07 mg/Kg);

372

U (2.35±0.07 mg/Kg); Hg (0.03±0.00 mg/Kg) and Fe (19500.00±707.11 mg/Kg). The recovery

373

of the above trace elements from the standard reference material (PR_LCSSRM_00018) was

374

above 95%. However, the XRD analysis of the bottom sediment revealed complete absence of

375

As, Cu, U, Mn, Zn, and Hg bearing minerals, with the presence of two predominant minerals:

376

Quartz (SiO2: 97.2%), and Pyrite (FeS2: 2.8%) (SI. Fig.3).

377

3.4 Statistical Results

378

To evaluate the change in water chemistry of the bay in response to precipitation in the

379

watershed and corresponding river water discharge, we used PCA analysis of PARAFAC

380

components, optical indices, DOC flux, TDN flux, inorganic ions (trace elements and nutrients)

381

flux, and water quality parameters (Fig. 8). The two principal components (PC); PC1 and PC2,

382

together explained 49.50 % of the total variance of the data. The presence of both microbial

383

PARAFAC components C2, C4 and the soil derived component C3 along with high FI, BIX, SR

384

on the positive side of the PC1 indicates allochthonous anthropogenic and some terrestrial soil

385

input. Whereas the presence of high SUVA254 and HIX along with PARAFAC component C1

386

along the negative side of the PC1 indicates terrestrial input. The HIX is inversely related to

387

component C3. The TDS, salinity, pH, nutrients (PO4, NO3-, NO2-), and the trace elements As, U,

17

388

and Cu were associated with microbial humic-like or protein-like PARAFAC component C4.

389

The ammonium ion (NH4+) and the trace elements Mn and Zn were associated with terrestrial

390

humic-like PARAFAC components C1. Hg and Fe showed no association with any PARAFAC

391

components, yet both were strongly associated with DOC and TDN. The ORP and DO were

392

present along the negative side of the PC1 axis and also directly related to both Fe and Hg.

393

Among the samples collected following all the precipitation events, all fall precipitation events

394

were located on the positive side of the PC1 axis, while all the samples representing the spring

395

precipitation events were positioned on the negative side of the PC1 axis. Whilst both summer

396

samples were located on the negative side of the PC2 axis. Additionally, positive relationship of

397

SUVA254 against Zn flux and FI against U flux and the negative relationship of component C4

398

vs Mn were also seen (SI. Fig.6)

399

Bootstrap resampling method at confidence intervals 2.5%, 50%, 97.5% was applied to

400

the flux of trace elements (As, Mn, U, Cu, Zn, Hg, Fe) following four field campaigns (Fig. 9).

401

The bootstrap results revealed that there was a significantly high flux of As, Cu and U in the bay

402

following the fall precipitation event compared to summer-1, summer-2 and spring precipitation

403

events (Table 5). In contrast, the flux of Hg, Mn, and Fe were significantly higher statistically

404

following the spring precipitation event than the summer-1, summer-2 and fall precipitation

405

events. The flux of Mn to the bay showed a statistically significant decrease from spring

406

precipitation event to summer-2 precipitation event. The flux of Zn to the bay was statistically

407

similar but higher following both spring and summer-1 precipitation events compared to its flux

408

following summer-2 and fall precipitation events.

409 410

The one-way ANOVA analysis indicated that variability of SUVA254 between the four sampling events (mean square=11.30) and the variability within each precipitation event (mean 18

411

square=0.69) are significantly different (F=16.46), therefore mean SUVA254 of each sampling

412

event is significantly different (p<0.05) from one another. Tukey-HSD pairwise multiple

413

comparison post-hoc analyses indicated similar mean SUVA254 following the summer-2 and fall

414

(p>0.05) precipitation events. The mean SUVA254 following the summer-2 and spring

415

precipitation events were the same and following the spring and fall precipitation events were

416

also the same (p>0.05). Mann-Whitney-Wilcoxon Test confirmed a significant (p<0.05) increase

417

in PARAFAC components C2, C3, and C4 from spring to fall precipitation events with U values

418

of 118, 104, and 118, respectively, while the component C1 showed no significant change

419

(p>0.05, U=72) following the precipitation events.

420

4. Discussion

421

4.1 Water and sediment chemistry

422

The concentration of Cu in Weeks Bay exceeded the EPA recommended water quality

423

criteria both for acute and chronic toxicity for aquatic life following the fall precipitation event.

424

On the other hand, the concentrations of other trace elements including U, Zn, Hg, As, Mn, and

425

Fe, and nutrients including NO3- and NO2- in the bay were lower than the EPA recommended

426

water quality criteria (EPA, 2018). Yet it is important to note that even at such low

427

concentrations, their regular input and concurrent deposition can lead to their enrichment in the

428

bay bottom sediments over time and also bioaccumulation/magnification of these trace metals in

429

aquatic organisms.

430

The flux of As, U, and Cu along with PO4- and NO3- into the bay increased following

431

increase in precipitation in the watershed (Figs. 3 and 4). Their co-occurrence and the

432

simultaneous increase in flux were not surprising as the predominant land use land cover of the

19

433

watershed surrounding Weeks Bay is agricultural land (cultivated crops-24.48%, hay and

434

pasture-19.00%). The application of phosphate-based fertilizers, micronutrients containing Cu

435

and animal manure in those agricultural areas can be one of the possible sources for the

436

excessive influx of PO4-, NO3- and Cu (Bandyopadhyay et al., 2014; Hudak, 2000; Miller et al.,

437

2011). Moreover, phosphate-based fertilizers can be a source of As or it can enhance the release

438

of adsorbed As from the soil fraction, especially from iron oxyhydroxides (Jayasumana et al.,

439

2015; T. Y. Lin et al., 2016; Molina et al., 2009; Signes-Pastor et al., 2007), which is in

440

agreement with the significant strong positive relationship (p<0.05, r= 0.74) of As and PO43- flux

441

in the bay. Similarly, the concentration of As (7.75±0.21 mg/Kg), Cu (4.05±0.07 mg/Kg), and U

442

(2.35±0.07 mg/Kg) were also very high in the bottom sediments. While naturally occurring

443

pyrite containing As (S- and Fe2+ atoms were replaced by As) can be a source for As (Blanchard

444

et al., 2007; Le Pape et al., 2017), the absence of As-bearing pyrite (SI Fig. 3) and increase in

445

the flux of As along with U, Cu, PO43-, and NO3- into the bay (Fig. 3) clearly indicates an

446

external or anthropogenic source for As. Previous studies in the adjacent watersheds have

447

attributed enhanced concentrations of trace elements to external sources. Lafabrie et al. (2011)

448

reported high amount of As and Hg in the Fish River sediments and identified Warrior Coal

449

Field (coal enriched in As and Hg) located at northwestern corner of Mobile Bay watershed and

450

the chemical industrial plants located between Mobile Bay and Mobile-Tensaw Delta as the

451

possible sources. Zielinski et al. (2007) have also reported the presence of As, Hg, and U in the

452

feed coal and its derivative fly ash in the Warrior Coal Field. The increase in the flux of U along

453

with the flux of As and Cu in the bay in response to seasonal precipitation and discharge (spring

454

to fall) suggests additional sources of U from the watershed. There was no nuclear reactor near

455

the watershed. Although there are a couple of nuclear power plants in the state of Tennessee,

20

456

along the Tennessee River, the river is not connected to Weeks Bay. However, there are exposed

457

out crops of black shale formations present along the north-eastern Alabama, but no streams

458

flowing to Weeks Bay are connected to those outcrops. Hence, the possible source of U could be

459

the effluent of Warrior Coal Field, as the coal contains a substantial amount of U as an impurity

460

(Zielinski et al., 2007).

461

Overall there was a decrease in flux of Mn, Hg, Fe, and Zn in the bay from spring to fall

462

precipitation event. However, their concentrations in sediment (19500.00±707.11 mg/Kg;

463

110.00±0 mg/Kg; 30.00±1.41 mg/Kg; 0.03±0.00 mg/Kg) were very high, and their high flux in

464

bay following spring precipitation event may be due to the prevalence of higher reducing

465

conditions (lower ORP values) in the bay. Warrior Coal Field and the chemical industrial plants

466

in the watershed are likely the sources of Hg contamination in Weeks Bay. The concentration of

467

Hg was high following the spring precipitation event compared to the fall precipitation event

468

possibly because of dilution following the summer and fall precipitation events. Previous studies

469

have reported the presence of high concentrations of Mn, Zn, and Cu in the shallow regions of

470

the Gulf of Mexico and suggested their release from reductive dissolution of bottom sediments

471

via decomposing organic matter (Joung and Shiller, 2016; Slowey and Hood, 1971). Another

472

possible source of Mn and Zn could be fertilizer such as basic slag containing Ca, Mg, Mn, and

473

Zn commonly used in the agricultural areas of Southeastern United States. Also, application of

474

ammonium fertilizers can lead to nitrification and soil acidification that accelerates leaching of

475

cations adsorbed to soil.

476

Both NH4+ and NO2- fluxes were negatively related to NO3- flux (p >0.05, r= -0.31 and

477

p<0.05, r= -0.21) in the bay. The weak inverse relationship of both NH4+ and NO2- against NO3-

478

could be due to the active nitrification/denitrification process in the bay. The high flux of NO2-, 21

479

high amount of DO, and low flux of NO3- following spring precipitation event compared to their

480

concentration following the summer and fall precipitation events could indicate high anaerobic

481

denitrification process following spring precipitation and discharge (Kim et al., 2010).

482

High DOC was observed following the spring precipitation event compared to summer

483

and fall precipitation events possibly because of higher residence time of the river effluents in

484

Weeks Bay and vice versa. This pattern suggests that the input of DOC is hydrologically

485

controlled, that is, the flow path is shifted from deep, low-DOC soil horizons to shallow, high-

486

DOC horizons following spring with low amount of precipitation to fall with high amount of

487

precipitation events (Dalzell et al., 2007; Hu et al., 2016; Zhou et al., 2016). Thus our study

488

shows that the DOC flux into Weeks Bay and then to NGoM through Mobile Bay increases with

489

discharge.

490

Stolpe et al. (2010) reported the presence of U-CDOM colloids, Mn in Fe-rich colloids

491

and Cu-CDOM complexes in the Pearl, Mississippi, and Atchafalaya Rivers. A positive

492

correlation was observed between the fluxes of Fe and DOC in the bay following all precipitation

493

events (p<0.05, r=0.86) indicating their co-occurrence as Fe-DOM colloids. According to

494

Novotnik et al. (2018) the mobility of U is also controlled by its oxidation state and the presence

495

of complexing agents like inorganic carbon, PO43-, and DOM. We observed a significant positive

496

relationship (p<0.05, r=0.73) between the fluxes of U and PO43-, however we did also observe a

497

week but significant correlation between DOC and U (p<0.05, r=0.33) in our dataset, but did not

498

observe any correlation between Cu and DOC in Weeks Bay.

499

4.2 Variation DOM in response to precipitation

22

500

The variation of both absorption and fluorescence indices from low spring precipitation

501

through the moderate summer to high fall precipitation revealed the input of less humified,

502

simpler, low molecular weight, microbially reprocessed or protein-like DOM into the bay

503

following the precipitation events. The SUVA254 values had minimum change as a function of

504

precipitation and discharge, however the ANOVA analysis confirmed that the variability of

505

SUVA254 between and within each precipitation event were significantly different (p<0.05). The

506

difference in mean SUVA254 values following the three seasonal precipitation events (spring,

507

summer and fall) indicates DOM export from multiple sources. That is high aromatic and high

508

molecular weight DOM input from adjoining herbaceous wetlands, woody wetlands, and

509

evergreen forests and low aromatic DOM from the croplands, pasture, and urbanized areas (Lu et

510

al., 2013). Meanwhile, high value of SUVA254 following the event of summer-1 precipitation

511

could be due to higher input of soil-derived DOM along with other terrestrially derived DOM,

512

both with high aromatic character, into the bay from dry riparian areas and surrounding woody

513

herbaceous wetlands. The higher and fluctuating SUVA254 values during the events of

514

precipitation in summer and fall could also be due to the input of humic DOM from the coal

515

effluents from warrior coal fields. The release of humic substances from coal is been widely

516

reported (De Souza and Bragança, 2018; Valero et al., 2014). The SR values, proxy for DOM

517

molecular weight, were also high following the summer-1, summer-2 and fall precipitation

518

events compared to spring precipitation event indicating high influx of low molecular weight or

519

simpler microbial humic-like or protein-like DOM from the croplands, pasture, and urbanized

520

areas into the bay following the high amount of precipitation during the summer and fall seasons.

521

The BIX values were always less than one during all four trips indicating the dominance

522

of organic matter from external sources or allochthonous input. Increases in both BIX (10.26 %)

23

523

and FI (2.03 %) indices along with fluctuating HIX values as discharge increases also indicates

524

addition of fresh, microbially derived, DOM into the bay through the river discharge in response

525

to precipitation (Fig. 5). The sources for this pool of fresh microbial organic matter could be

526

human-modified lands such as croplands, pasture, and urbanized areas. Previous studies have

527

reported that agricultural and urban areas export DOM with enhanced signatures from microbial

528

sources (Lu et al., 2013; Sankar et al., 2018).

529

Concomitant increase protein-like (C2: U=118, p<0.05 and C4: U=118, p<0.05) and soil-

530

derived components (C3: U=104, p<0.05) and relatively constant input of terrestrial humic-like

531

(C1: U=72, p>0.05) component from the spring low precipitation through moderate to fall high

532

precipitation indicate continuous input of protein-like and soil derived DOM to the bay from

533

agricultural and urbanized areas following precipitation (Fig. 7). The reason for the low

534

variability of the component C1 during all field campaigns could be due to the continuous supply

535

of terrestrial humic-like DOM from the woody herbaceous wetlands present all along the banks

536

of the bay (Sankar et al., 2018). Furthermore, the component C1 has been found in multiple

537

studies as a background fluorescence in fluvial waters due to its refractory nature (Sankar et al.,

538

2018; Singh et al., 2017; 2010). The largest variation of soil derived DOM component (C3)

539

following summer-1 precipitation (IQR= 0.04) could be because of its large influx due to

540

flushing from within and above the dry soil horizons during the initial period of summer

541

precipitation.

542

Based on the variation of both optical indices and DOM composition, it is quite evident

543

that precipitation in the watershed has mobilized a considerable amount of allochthonous fresh

544

microbially reprocessed DOM into the bay from the Weeks Bay Watershed. Additionally, there

545

is a possibility of allochthonous microbially reprocessed DOM input into Weeks Bay through the 24

546

Mobile Bay from Mobile Bay Watershed especially during the events of high tides, however

547

more data is required from the Mobile Bay to ascertain this claim.

548

4.3 Relationship of salinity and dissolved constituents

549

The trace elements and nutrients showed a conservative mixing behavior in the Weeks Bay.

550

Increasing trends in the fluxes of As, Cu, U, NO3-, PO4-, and NO2- with increasing salinity was

551

observed because the field data collection during fall, summer-1, and summer-2 precipitation

552

events coinciding with high tides. Decreasing trends in the fluxes of Fe, Mn, Zn, Hg, and NH4+

553

with increasing salinity indicates their release from bay bottom owing to the highly reducing

554

conditions especially during spring precipitation events. The precipitation induced freshwater

555

input caused an increase in fluxes of As, Cu, U, NO3-, PO4- and NO2-. On the other hand,

556

prevalence of higher redox conditions in the bay bottom during dry weather conditions increased

557

the Fe, Mn, Zn, Hg concentrations along with NH4+ by dissolution of their oxidized forms. The

558

low inverse relationship of Hg and salinity suggests dilutionary effects during the events of

559

precipitation or atmospheric deposition and settling in the bay. The inverse relationship of DOC

560

and TDN with salinity supports the shifting of hydrology from spring low to fall high

561

precipitation events (SI Fig. 4 & 5).

562

4.4 DOM and trace element relationships

563

The samples collected following fall precipitation event were positioned all along the

564

positive side of the PC1 axis indicating protein-like or microbial humic-like DOM input into the

565

bay from the croplands, pasture, and urbanized areas of the watershed. On the contrary, both of

566

the samples collected following summer-1 and summer-2 precipitation events were located on

567

the negative side of the PC2 axis indicating more of microbially reprocessed and soil derived

25

568

DOM input into the bay following summer precipitation (Fig. 8). While all samples collected

569

following the spring precipitation event were in the negative side of the PC1 axis indicating less

570

microbial humic-like or protein-like and more terrestrial humic-like DOM input. The PCA plot

571

clearly shows a strong connection between the fresh, microbially or anthropogenically derived,

572

simpler DOM components (C2 and C4), the trace elements (As, Cu, U), and the nutrients (PO43-

573

and NO3-). Their mutual relationship clearly indicates that As, Cu, U and inorganic N and P

574

nutrients were flushed into the bay along with microbially derived DOM components (C2 and

575

C4) from the adjoining agricultural and urbanized areas (Fig 8 and SI. Fig.6). Furthermore, the

576

positive relationship of pH, with components C2, C4 along with As, Cu, U and PO43- and, NO3-

577

indicates that they entered the bay as the result of water discharge through the rivers following

578

the high amount of precipitation during both fall and summer seasons. Hg shows no correlation

579

with any of the PARAFAC components, nutrients, trace elements, or pH indicating its

580

introduction into the bay through atmospheric deposition or dilutionary effect due to its low

581

concentration. The strong positive relationship of Hg, DOC, and ORP could be due to the

582

formation of Hg-DOM complexes in the bay water under reducing conditions (Gu et al., 2011;

583

Harris et al., 2012; Ravichandran, 2004). A strong positive correlation of Fe and DOC indicates

584

the existence of Fe-DOM colloids in the bay (Stolpe et al., 2010). Similarly, the bootstrap

585

resampling technique also confirmed increase in flux of As, Cu and U with seasonal

586

precipitation, that is from spring to fall in the bay (Fig. 9). On the other hand, the flux of Hg, Mn,

587

and Fe to the bay was high following low amount of precipitation during the spring implying the

588

effect of dilution or the presence of low redox conditions (Table 3). We are also not ruling out

589

the fact that such change in trends of elemental flux in the bay could also be the result of the

26

590

change in redox conditions due to groundwater surface water interactions in the bay. However

591

we do not have associated groundwater data to prove such claim.

592

Only a few studies have evaluated the source of trace elements in the coastal areas of

593

NGoM (Lafabrie et al., 2011; Telfeyan et al., 2017; USEPA, 1999). In the present work, we

594

correlated DOM sources and its chemical properties with trace elements, which was found to be

595

effective in the identification of the source of trace elements. Our results suggest that

596

precipitation in the watershed and the subsequent water discharge through both the rivers into the

597

bay affected the quality and chemistry of the Weeks Bay water. In addition to freshwater

598

discharge from the Fish and Magnolia Rivers, Weeks Bay receives tidal flows from Mobile Bay

599

as well (Passeri et al., 2016). The amount of Mobile Bay water incursion to Weeks Bay or the

600

concentration of water quality parameters in the Mobile Bay was not investigated in this study.

601

However, trends of TDS and salinity in the Weeks Bay observed in this study were similar on all

602

four field campaigns, which can be attributed to tidal effects during the time of field work. The

603

DO concentration of an aqueous solution reduces in response to increase in salinity and

604

temperature (Lay et al., 2010), hence, the low DO content of the bay water following the events

605

of summer and fall precipitation were coinciding with both high salinity and temperature, and

606

high DO concentrations were coinciding with both low salinity and temperature following spring

607

with low amount of precipitation. Freshwater input into Mobile Bay happens mainly through the

608

Tombigbee River, the Alabama River, and their tributaries; which drains a much larger

609

watershed. We acknowledge that the incursion of water from Mobile Bay, which includes

610

freshwater from the Mobile Bay Watershed and seawater, to Weeks Bay may have an effect on

611

the water quality of the Weeks Bay especially during high tides. Collection of field data from the

27

612

Mobile Bay and adjoining coastal waters, and hydrodynamic modeling will help in quantifying

613

the effect of flow of Mobile Bay waters to Weeks Bay.

614

5. Conclusion

615

The trace element, nutrient, and DOM amount in the Weeks Bay was very much

616

dependent on precipitation in the watershed and subsequent water discharge into the bay through

617

the two blackwater rivers, Fish and Magnolia. Statistical analyses of water quality parameters

618

following spring low-fall high precipitation events indicated that the concentration of trace

619

elements As, Cu, and U, and the nutrients NO3- and PO43- were higher in the Weeks Bay

620

following fall high-precipitation events. Moreover, the availability of microbial humic-like or

621

protein-like DOM components increased in the bay along with soil-derived, humic component

622

following precipitation during summer and fall seasons. This clearly indicates that following the

623

fall precipitation, the trace elements As, Cu, and U are exported to the bay along with

624

microbially processed DOM from various anthropogenic sources present in the watershed such

625

as agricultural areas or industries. Further, the spectroscopic properties, BIX (< 1 during the

626

study period), FI (gradually increasing) and SUVA254 (decreasing) along with increasing

627

microbial humic-like, protein-like and soil-derived humic DOM components following higher

628

amount of summer and fall precipitation indicated that the trace metals were mobilized to the bay

629

along with allochthonous DOM. Results also revealed that the trace element and DOM fluxes in

630

Weeks Bay was largely dependent on hydrologic characteristics (precipitation and dishcarge)

631

and land use land cover. Future work may include exploring the inclursion of contaminants into

632

Weeks bay through Mobile Bay and its drainage system during high tides.

633 634

Increasing precipitation events are projected due to global climate change, which can affect DOM-trace element mobility in the coastal waterbodies. This study highlights the 28

635

importance of DOM composition in determining sources of trace elements in the watershed.

636

Increasing precipitation events are projected due to global climate change, which can affect

637

DOM-trace element mobility in the coastal waterbodies in the southern states and worldwide.

638

This study highlights the importance of DOM composition in determining sources of trace

639

elements in the watershed. Trace metals analysis can be time consuming and expensive., while

640

DOM fluorometric analysis is comparatively inexpensive. In this article we have demonstrated

641

the use of DOM composition as PARAFAC components to source track contaminant trace

642

element sources. This method can be utilized to identify the contaminant sources in other regions

643

around the globe.

644

Acknowledgement

645

The research was supported by the faculty start-up grant to Dr. Padmanava Dash. The authors are

646

thankful to Ms. Devon Flickinger and Mr. Rusch Ragland of Department of Geosciences,

647

Mississippi State University and Dr. Scott Phipps of the Weeks Bay National Estuarine Research

648

Reserve for their help during field campaigns.

649

References

650

Apeti, D.A., Lauenstein, G.G., Evans, D.W., 2012. Recent status of total mercury and methyl

651

mercury in the coastal waters of the northern Gulf of Mexico using oysters and sediments

652

from NOAA’s mussel watch program. Mar. Pollut. Bull. 64, 2399–2408.

653

doi:10.1016/j.marpolbul.2012.08.006

654

Bandyopadhyay, S., Ghosh, K., Varadachari, C., 2014. Multimicronutrient slow-release fertilizer

655

of zinc, iron, manganese, and copper. Int. J. Chem. Eng. 2014, 1–7.

656

doi:10.1155/2014/327153 29

657

Benedetti, M.F., Van Riemsdijk, W.H., Koopal, L.K., Kinniburgh, D.G., Gooddy, D.C., Milne,

658

C.J., 1996. Metal ion binding by natural organic matter: From the model to the field.

659

Geochim. Cosmochim. Acta 60, 2503–2513. doi:10.1016/0016-7037(96)00113-5

660

Blanchard, M., Alfredsson, M., Brodholt, J., Wright, K., Catlow, C.R.A., 2007. Arsenic

661

incorporation into FeS2pyrite and its influence on dissolution: A DFT study. Geochim.

662

Cosmochim. Acta 71, 624–630. doi:10.1016/j.gca.2006.09.021

663

Caffrey, J.M., Murrell, M.C., Amacker, K.S., Harper, J.W., Phipps, S., Woodrey, M.S., 2014.

664

Seasonal and Inter-annual Patterns in Primary Production, Respiration, and Net Ecosystem

665

Metabolism in Three Estuaries in the Northeast Gulf of Mexico. Estuaries and Coasts 37,

666

222–241. doi:10.1007/s12237-013-9701-5

667

Camacho, R.A., Martin, J.L., Diaz-Ramirez, J., McAnally, W., Rodriguez, H., Suscy, P., Zhang,

668

S., 2014. Uncertainty analysis of estuarine hydrodynamic models: An evaluation of input

669

data uncertainty in the Weeks Bay estuary, Alabama. Appl. Ocean Res. 47, 138–153.

670

doi:10.1016/j.apor.2014.04.004

671

Dalzell, B.J., Filley, T.R., Harbor, J.M., 2007. The role of hydrology in annual organic carbon

672

loads and terrestrial organic matter export from a midwestern agricultural watershed.

673

Geochim. Cosmochim. Acta 71, 1448–1462. doi:10.1016/j.gca.2006.12.009

674

De Souza, F., Bragança, S.R., 2018. Extraction and characterization of humic acid from coal for

675

the application as dispersant of ceramic powders. J. Mater. Res. Technol. 7, 254–260.

676

doi:10.1016/j.jmrt.2017.08.008

677 678

Duan, S., Bianchi, T.S., Sampere, T.P., 2007. Temporal variability in the composition and abundance of terrestrially-derived dissolved organic matter in the lower Mississippi and 30

679 680

Pearl Rivers. Mar. Chem. 103, 172–184. doi:10.1016/j.marchem.2006.07.003 Fichot, C.G., Benner, R., 2012. The spectral slope coefficient of chromophoric dissolved organic

681

matter (S275-295) as a tracer of terrigenous dissolved organic carbon in river-influenced

682

ocean margins. Limnol. Oceanogr. 57, 1453–1466. doi:10.4319/lo.2012.57.5.1453

683

Gu, B., Bian, Y., Miller, C.L., Dong, W., Jiang, X., Liang, L., 2011. Mercury reduction and

684

complexation by natural organic matter in anoxic environments. Proc. Natl. Acad. Sci. 108,

685

1479–1483. doi:10.1073/pnas.1008747108

686

Hansen, A.M., Kraus, T.E.C., Pellerin, B.A., Fleck, J.A., Downing, B.D., Bergamaschi, B.A.,

687

2016. Optical properties of dissolved organic matter (DOM): Effects of biological and

688

photolytic degradation. Limnol. Oceanogr. 61, 1015–1032. doi:10.1002/lno.10270

689

Harris, R., Pollman, C., Landing, W., Evans, D., Axelrad, D., Hutchinson, D., Morey, S.L.,

690

Rumbold, D., Dukhovskoy, D., Adams, D.H., Vijayaraghavan, K., Holmes, C., Atkinson,

691

R.D., Myers, T., Sunderland, E., 2012. Mercury in the Gulf of Mexico: Sources to

692

receptors. Environ. Res. 119, 42–52. doi:10.1016/j.envres.2012.08.001

693

Hongxia, Y., Jinxu, G., Wei, L., Keyan, T., 2014. Investigating the composition of dissolved

694

organic matter in natural water in rare earth mine using EEM-PARAFAC analysis. Environ.

695

Sci. Process. Impacts 16, 2527–2535. doi:10.1039/C4EM00380B

696

Hu, Y., Lu, Y., Edmonds, J.W., Liu, C., Wang, S., Das, O., Liu, J., Zheng, C., 2016.

697

Hydrological and Land Use Control of Watershed Exports of DOM in a Large Arid River

698

Basin in Northwestern China. J. Geophys. Res. Biogeosciences 121, 466–478.

699

doi:10.1002/2015JG003082

31

700 701 702

Hudak, P.F., 2000. Regional trends in nitrate content of Texas groundwater. J. Hydrol. 228, 37– 47. doi:10.1016/S0022-1694(99)00206-1 Huguet, A., Vacher, L., Relexans, S., Saubusse, S., Froidefond, J.M., Parlanti, E., 2009.

703

Properties of fluorescent dissolved organic matter in the Gironde Estuary. Org. Geochem.

704

40, 706–719. doi:10.1016/j.orggeochem.2009.03.002

705

Jayasumana, C., Fonseka, S., Fernando, A., Jayalath, K., Amarasinghe, M., Siribaddana, S.,

706

Gunatilake, S., Paranagama, P., 2015. Phosphate fertilizer is a main source of arsenic in

707

areas affected with chronic kidney disease of unknown etiology in Sri Lanka. Springerplus

708

4, 1–8. doi:10.1186/s40064-015-0868-z

709

Joung, D., Shiller, A.M., 2014. Dissolved barium behavior in Louisiana Shelf waters affected by

710

the Mississippi/Atchafalaya River mixing zone. Geochim. Cosmochim. Acta 141, 303–313.

711

doi:10.1016/j.gca.2014.06.021

712

Joung, D.J., Shiller, A.M., 2016. Temporal and spatial variations of dissolved and colloidal trace

713

elements in Louisiana Shelf waters. Mar. Chem. 181, 25–43.

714

doi:10.1016/j.marchem.2016.03.003

715

Keul, N., Morse, J.W., Wanninkhof, R., Gledhill, D.K., Bianchi, T.S., 2010. Carbonate

716

chemistry dynamics of surface waters in the Northern Gulf of Mexico. Aquat. Geochemistry

717

16, 337–351. doi:10.1007/s10498-010-9091-2

718

Kim, S.W., Miyahara, M., Fushinobu, S., Wakagi, T., Shoun, H., 2010. Nitrous oxide emission

719

from nitrifying activated sludge dependent on denitrification by ammonia-oxidizing

720

bacteria. Bioresour. Technol. 101, 3958–3963. doi:10.1016/j.biortech.2010.01.030

32

721

Kothawala, D.N.., Murphy, K.R.., Stedmon, C.A.., Weyhenmeyer, G.A.., Tranvik, L.J.., 2013.

722

Inner filter correction of dissolved organic matter fluorescence. Limnol. Oceanogr. Methods

723

11, 616–630. doi:10.4319/lom.2013.11.616

724

Kothawala, D.N., Stedmon, C.A., Müller, R.A., Weyhenmeyer, G.A., Köhler, S.J., Tranvik, L.J.,

725

2014. Controls of dissolved organic matter quality: Evidence from a large-scale boreal lake

726

survey. Glob. Chang. Biol. 20, 1101–1114. doi:10.1111/gcb.12488

727

Lafabrie, C., Major, K.M., Major, C.S., Cebrián, J., 2011. Arsenic and mercury bioaccumulation

728

in the aquatic plant, Vallisneria neotropicalis. Chemosphere 82, 1393–1400.

729

doi:10.1016/j.chemosphere.2010.11.070

730

Lay, W.C.L., Liu, Y., Fane, A.G., 2010. Impacts of salinity on the performance of high retention

731

membrane bioreactors for water reclamation: A review. Water Res. 44, 21–40.

732

doi:10.1016/j.watres.2009.09.026

733

Le Pape, P., Blanchard, M., Brest, J., Boulliard, J.C., Ikogou, M., Stetten, L., Wang, S., Landrot,

734

G., Morin, G., 2017. Arsenic Incorporation in Pyrite at Ambient Temperature at Both

735

Tetrahedral S-Iand Octahedral FeIISites: Evidence from EXAFS-DFT Analysis. Environ.

736

Sci. Technol. 51, 150–158. doi:10.1021/acs.est.6b03502

737

Lewis, M., Chancy, C., 2008. A summary of total mercury concentrations in flora and fauna near

738

common contaminant sources in the Gulf of Mexico. Chemosphere 70, 2016–2024.

739

doi:10.1016/j.chemosphere.2007.09.020

740

Lewis, M.A., Russell, M.J., 2015. Contaminant profiles for surface water, sediment, flora and

741

fauna associated with the mangrove fringe along middle and lower eastern Tampa Bay.

742

Mar. Pollut. Bull. 95, 273–282. doi:10.1016/j.marpolbul.2015.04.001 33

743

Lin, H., Chen, M., Zeng, J., Li, Q., Jia, R., Sun, X., Zheng, M., Qiu, Y., 2016. Size

744

characteristics of chromophoric dissolved organic matter in the Chukchi Sea. J. Geophys.

745

Res. Ocean. 121, 6403–6417. doi:10.1002/2016JC011771

746

Lin, T.Y., Wei, C.C., Huang, C.W., Chang, C.H., Hsu, F.L., Liao, V.H.C., 2016. Both

747

Phosphorus Fertilizers and Indigenous Bacteria Enhance Arsenic Release into Groundwater

748

in Arsenic-Contaminated Aquifers. J. Agric. Food Chem. 64, 2214–2222.

749

doi:10.1021/acs.jafc.6b00253

750

Lu, Y., Bauer, J.E., Canuel, E.A., Yamashita, Y., Chambers, R.M., Jaffé, R., 2013.

751

Photochemical and microbial alteration of dissolved organic matter in temperate headwater

752

streams associated with different land use. J. Geophys. Res. Biogeosciences 118, 566–580.

753

doi:10.1002/jgrg.20048

754

Miller, M.., Phipps, S.W., Major, C.S., Major, K.M., 2011. Effects of Environmental Variation

755

and Non-point Source (NPS) Nutrient Pollution on Aquatic Plant Communities in Weeks

756

Bay National Estuarine Research Reserve (WBNERR), AL. Estuaries and Coasts (2011) 34,

757

1182–1193. doi:10.1007/s12237-011-9422-6

758

Molina, M., Aburto, F., Calderón, R., Cazanga, M., Escudey, M., 2009. Trace element

759

composition of selected fertilizers used in Chile: Phosphorus fertilizers as a source of long-

760

term soil contamination. Soil Sediment Contam. 18, 497–511.

761

doi:10.1080/15320380902962320

762

Perry, H., Isphording, W., Trigg, C., Riedel, R., 2015. Heavy metals in red crabs, Chaceon

763

quinquedens, from the Gulf of Mexico. Mar. Pollut. Bull. 101, 845–851.

764

doi:10.1016/j.marpolbul.2015.11.020 34

765 766 767 768 769

Ravichandran, M., 2004. Interactions between mercury and dissolved organic matter - A review. Chemosphere 55, 319–331. doi:10.1016/j.chemosphere.2003.11.011 Reuter, J.H., Perdue, E.M., 1977. Importance of heavy metal - organic matter interactions in natural waters. Geochim. Cosmochim. Acta 41, 325–334. Ruecker, A., Uzun, H., Karanfil, T., Tsui, M.T.K., Chow, A.T., 2017. Disinfection byproduct

770

precursor dynamics and water treatability during an extreme flooding event in a coastal

771

blackwater river in southeastern United States. Chemosphere 188, 90–98.

772

doi:10.1016/j.chemosphere.2017.08.122

773

Sankar, M.S., Dash, P., Singh, S., Lu, Y.H., Mercer, A.E., Chen, S., 2018. Effect of photo-

774

biodegradation and biodegradation on the biogeochemical cycling of dissolved organic

775

matter across diverse surface water bodies. J. Environ. Sci. (China) 1–18.

776

doi:10.1016/j.jes.2018.06.021

777

Shank, G.C., Evans, A., Yamashita, Y., Jaffé, R., 2011. Solar radiation-enhanced dissolution of

778

particulate organic matter from coastal marine sediments. Limnol. Oceanogr. 56, 577–588.

779

doi:10.4319/lo.2011.56.2.0577

780

Signes-Pastor, A., Burló, F., Mitra, K., Carbonell-Barrachina, A.A., 2007. Arsenic

781

biogeochemistry as affected by phosphorus fertilizer addition, redox potential and pH in a

782

west Bengal (India) soil. Geoderma 137, 504–510. doi:10.1016/j.geoderma.2006.10.012

783

Singh, S., D’Sa, E.J., Swenson, E.M., 2010. Chromophoric dissolved organic matter (CDOM)

784

variability in Barataria Basin using excitation-emission matrix (EEM) fluorescence and

785

parallel factor analysis (PARAFAC). Sci. Total Environ. 408, 3211–3222.

786

doi:10.1016/j.scitotenv.2010.03.044 35

787

Singh, S., Dash, P., Silwal, S., Feng, G., Adeli, A., Moorhead, R.J., 2017. Influence of land use

788

and land cover on the spatial variability of dissolved organic matter in multiple aquatic

789

environments. Environ. Sci. Pollut. Res. doi:10.1007/s11356-017-8917-5

790

Slowey, F.J., Hood, D.W., 1971. Copper, manganese and zinc concentrations in Gulf of Mexico

791

waters. Geochim. Cosmochim. Acta 35, 121–138. doi:https://doi.org/10.1016/0016-

792

7037(71)90052-4

793

Spencer, R.G.M., Aiken, G.R., Dornblaser, M.M., Butler, K.D., Holmes, R.M., Fiske, G., Mann,

794

P.J., Stubbins, A., 2013. Chromophoric dissolved organic matter export from U.S. rivers.

795

Geophys. Res. Lett. 40, 1575–1579. doi:10.1002/grl.50357

796

Stedmon, C. a., Bro, R., 2008. OCEANOGRAPHY : METHODS Characterizing dissolved

797

organic matter fluorescence with parallel factor analysis : a tutorial. Limnol. Oceanogr.

798

Methods 6, 572–579. doi:10.4319/lom.2008.6.572

799

Stedmon, C. a., Markager, S., 2005. Tracing the production and degradation of autochthonous

800

fractions of dissolved organic matter using fluorescence analysis. Limnol. Oceanogr. 50,

801

1415–1426. doi:10.4319/lo.2005.50.5.1415

802

Stolpe, B., Guo, L., Shiller, A.M., Hassellov, M., 2010. Size and composition of colloidal

803

organic matter and trace elements in the Mississippi River, Pearl River and the northern

804

Gulf of Mexico, as characterized by flow field-flow fractionation. Mar. Chem. 118, 119–

805

128. doi:10.1016/j.marchem.2009.11.007

806

Telfeyan, K., Breaux, A., Kim, J., Cable, J.E., Kolker, A.S., Grimm, D.A., Johannesson, K.H.,

807

2017. Arsenic, vanadium, iron, and manganese biogeochemistry in a deltaic wetland,

808

southern Louisiana, USA. Mar. Chem. 192, 32–48. doi:10.1016/j.marchem.2017.03.010 36

809

Telfeyan, K., Breaux, A., Kim, J., Kolker, A.S., Cable, J.E., Johannesson, K.H., 2018. Cycling of

810

oxyanion-forming trace elements in groundwaters from a freshwater deltaic marsh. Estuar.

811

Coast. Shelf Sci. 204, 236–263. doi:10.1016/j.ecss.2018.02.024

812

Tribovillard, N., Riboulleau, A., Lyons, T., Baudin, F., 2004. Enhanced trapping of molybdenum

813

by sulfurized marine organic matter of marine origin in Mesozoic limestones and shales.

814

Chem. Geol. doi:10.1016/j.chemgeo.2004.08.011

815 816 817 818

USEPA, 1999. The Ecological Condition of Estuaries in the Gulf of Mexico. EPA 620-R-98-004 80. USEPA, 1983. Sample preservation. pp.xv-xx. In Methods for Chemical Analysis of Water and Wastes, EPA-600/4-79-020. U.S.E.P.A., Cincinnati, Ohio, USA

819

USGS, 1999. Collection of Water Samples. TWRI Book 9, Chapter A4.

820

Valero, N., Gómez, L., Pantoja, M., Ramírez, R., 2014. Production of humic substances through

821

coal-solubilizing bacteria. Brazilian J. Microbiol. 45, 911–918. doi:10.1590/S1517-

822

83822014000300021

823

Weng, L., Temminghoff, E.J.M., Lofts, S., Tipping, E., Van Riemsdijk, W.H., 2002.

824

Complexation with dissolved organic matter and solubility control of heavy metals in a

825

sandy soil. Environ. Sci. Technol. 36, 4804–4810. doi:10.1021/es0200084

826

Yamashita, Y., Jaffé, R., 2008. Characterizing the interactions between trace metals and

827

dissolved organic matter using excitation-emission matrix and parallel factor analysis.

828

Environ. Sci. Technol. 42, 7374–9. doi:10.1021/es801357h

829

Zhang, Y., Liu, X., Osburn, C.L., Wang, M., Qin, B., Zhou, Y., 2013. Photobleaching Response

37

830

of Different Sources of Chromophoric Dissolved Organic Matter Exposed to Natural Solar

831

Radiation Using Absorption and Excitation-Emission Matrix Spectra. PLoS One 8.

832

doi:10.1371/journal.pone.0077515

833

Zhou, Z., Stolpe, B., Guo, L., Shiller, A.M., 2016. Colloidal size spectra, composition and

834

estuarine mixing behavior of DOM in river and estuarine waters of the northern Gulf of

835

Mexico. Geochim. Cosmochim. Acta 181, 1–17. doi:10.1016/j.gca.2016.02.032

836

Zielinski, R.A., Foster, A.L., Meeker, G.P., Brownfield, I.K., 2007. Mode of occurrence of

837

arsenic in feed coal and its derivative fly ash, Black Warrior Basin, Alabama. Fuel 86, 560–

838

572. doi:10.1016/j.fuel.2006.07.033

839

List of figures

840

Fig. 1 Land use and land cover map of the study area with sampling points indicated

841

Fig. 2 Precipitation in the Weeks Bay watershed and discharge through the Fish and Magnolia

842

Rivers

843

Fig. 3 Variation in the fluxes of trace elements to Weeks Bay following spring (n=12), summer-1

844

(n=12), summer-2 (n=13) and fall (n=12) sampling events. ‘n’ denotes number of samples.

845

Fig. 4 Variation in fluxes of nutrients to Weeks Bay following spring (n=12), summer-1 (n=12),

846

summer-2 (n=13) and fall (n=12) sampling events. ‘n’ denotes number of samples.

847

Fig. 5 Variation of absorption (SUVA254, SR) and fluorescence indices (HIX, FI, BIX) in Weeks

848

Bay water following spring (n=12), summer-1 (n=12), summer-2 (ML, n=13) and fall (n=12)

849

sampling events. ‘n’ denotes number of samples.

38

850

Fig. 6 Excitation (blue curve) and emission (red curve) loadings for four dissolved organic

851

matter components derived using a parallel factor (PARAFAC) model that was validated using

852

split-half analysis as well as random initialization. Contour plots corresponding to the

853

components identified in the PARAFAC model are shown for each of the components.

854

Fig. 7 Variation of parallel factor (PARAFAC) components in Weeks Bay water following

855

spring (n=11), summer-1 (n=12), summer-2 (n=12) and fall (n=11) sampling events. ‘n’ denotes

856

number of samples.

857

Fig. 8 Principal component analysis (PCA) biplot for the Weeks Bay water chemistry. Samples

858

collected following spring (L, n=11), summer-1 (n=12), summer-2 (n=11) and fall (n=11)

859

sampling events were indicated by distinct symbols. ‘n’ denotes number of samples.

860

Fig. 9 Bootstrap resampling results of the discharged normalized concertation of trace elements

861

in Weeks Bay water following spring (n=1000), summber-1 (n=1000), summer-2 (n=1000) and

862

fall (n=1000) sampling events. ‘n’ denotes number of bootstrap means.

863

List of Tables

864

Table 1 The amount of land use and land cover area in km2 and in percentages of Weeks Bay

865

watershed

866

Table 2 The average amount of precipitation in the Weeks Bay watershed and discharge through

867

the Fish and Magnolia Rivers during five days prior to the sampling events during spring,

868

summer-1, summer-2 and fall field campaigns

869

Table 3 Distribution of water quality parameters for Weeks Bay waters collected during four

870

field campaigns

39

871

Table 4 Parallel factor (PARAFAC) analysis components information and their identification

872

Table 5 Bootstrap 2.5 %, 50% and 97.5 % quantile values of trace elements following spring,

873

summer-1, summer-2 and fall water sampling events.

40

Tables:

Table 1 Land Classes Open Water Developed, Open Space Developed, Low Intensity Developed, Medium Intensity Developed, High Intensity Barren Land Deciduous Forest Evergreen Forest Mixed Forest Shrub/Scrub Herbaceuous Hay/Pasture Cultivated Crops Woody Wetlands Emergent Herbaceuous Wetlands

Area (Km2) 4.49 22.02 7.89 2.75 0.56 2.20 0.31 22.14 0.64 8.79 13.29 49.02 64.86 37.88 4.79

Area (%) 1.86 9.11 3.27 1.14 0.23 0.91 0.13 9.16 0.27 3.64 5.50 20.29 26.84 15.68 1.98

Table 2 Name of the Date of sampling sampling events

Spring Summer-1 Summer-2 Fall

05/09/2016 06/19/2016 06/26/2016 09/30/2016

Average amount of precipitation five days prior to the sampling (mm) 0.66 3.51 0.15 2.03

Average discharge through Fish River five days prior to sampling (m3/s)

5.51± 4.93 2.19± 0.03 1.70± 0.03 1.93± 0.10

Average discharge through Magnolia River five days prior to sampling (m3/s)

0.78± 0.06 0.77± 0.22 0.60± 0.02 1.80± 2.48

Table 3 Water Quality Parameter pH

ORP (mV)

TDS (mg/L)

Salinity (PSU)

DO (mg/L)

Temperature (°C)

Spring

Summer-1

Summer-2

Fall

Avg:8.06±0.31 Avg:8.20±0.34 Avg=8.69±0.23 Avg=8.78±0.20 IQR=0.23 IQR=0.48 IQR=0.26 IQR=0.13 Max=8.9 Max=8.57 Max=8.95 Max=9 Min =7.71 Min=7.59 Min=8.13 Min=8.21 Avg=-58.24±44.08 Avg=-17.63±27.81 Avg=-31.06±14.90 Avg=-79.78±47.62 IQR=78.65 IQR=42.80 IQR=22.15 IQR=14.70 Max=-6.10 Max=32.60 Max=-3.90 Max= -41.4 Min =-119.30 Min =-57.0 Min =-57.40 Min = -224.8 Avg=1984.25±2069.54 Avg=7125.92±3318.26 Avg=5684.17±2486.38 Avg=10973.50±3464.25 IQR=4097 IQR= 4729.75 IQR= 3009.25 IQR= 2524.75 Max=4261 Max= 11790 Max= 9611 Max= 17120 Min =7 Min =742 Min =86 Min = 3063 Avg=2.20±2.30 Avg=8.30±4.08 Avg=6.47±2.95 Avg=13.25±4.46 IQR=4.56 IQR=5.90 IQR=3.63 IQR=3.31 Max=4.75 Max= 14.17 Max=11.32 Max=21.43 Min =0.01 Min = 0.74 Min =0.08 Min =3.32 Avg=5.80±4.36 Avg=1.62±0.12 Avg=1.54±0.18 Avg=0.38±0.02 IQR=8.09 IQR=0.10 IQR=0.15 IQR=0.02 Max=12.6 Max=1.81 Max=1.79 Max=0.41 Min =1.59 Min =1.39 Min =1.07 Min =0.35 Avg=27.00±3.66 Avg=29.27±0.68 Avg=31.47±0.54 Avg=26.20±0.72 IQR=6.95 IQR=1.12 IQR=0.68 IQR=1.07 Max=30.83 Max=30.14 Max=32.06 Max=27.55 Min =23.43 Min =28.26 Min =30.33 Min =25.17

Table 4 Components (Ex/Em) unit nm

Description and reference

C1

(<250-330/448)

Terrestrial humic-like component, common in freshwater environment C1 Kothawala et al., 2014 C Coble 1996 C1 Singh et al., 2017 C1 Yamashita et al.,2010 C2 Sankar et al., 2018

C2

(<250-310/388)

Microbial Humic like, Anthropogenic origin, Protein like, common in agriculture dominated watershed, Prevalent in waste water, Microbially reprocessed, like marine humic-like C4 Yamashita et al.,2010 C2 Kothawala et al., 2014 C2 Singh et al., 2017 C3 Kowalczuk et al., 2010 C3 Hosen et al., 2014

C3

(270-400/504)

Humic-like, Terrestrial origin, high molecular weight and aromatic organic compounds, Soil derived, reduced terrestrial, common in wide range of freshwater environments such as wetlands and rangelands, humic acid-type C3 Kothawala et al., 2014 C4 Kowalczuk et al., 2010 C3 Singh et al., 2017 C5 Yamashita et al.,2010 C2 Hosen et al., 2014 C2 Lu et al., 2013

C4

(280/334)

Tryptophan-like, Protein like, autochthonous production; autochthonous algal and microbially derived proteins; Protein like indicative of recent production, also commonly found in croplands, wastewater, industrial, and livestock wastes C8 Fellman et al., 2009 C4 Singh et al., 2017 T Coble, 1996 C6 Kowalczuk et al., 2010 C7 Yamashita et al.,2010 C4 Hosen et al., 2014 C4 Singh et al., 2014

Table 5 Sampling Session Bootstrap Quantiles As Mn U Cu Zn Hg Fe

Flux during Spring (mgs-1) 2.5 % 50 % 97.5 % 20.41 20.76 242.55 265.78 1.50 1.62 5.45 5.72 7.33 9.05 0.38 0.42 1279.96 1429.23

21.10 288.54 1.72 6.00 11.62 0.45 1600.28

Flux during Summer-1 (mgs-1) 2.5 % 50 % 97.5 % 12.26 137.14 0.99 4.35 4.39 0.08 369.5742

13.07 156.38 1.17 5.18 7.59 0.09 411.07

13.83 175.91 1.36 6.72 13.74 0.10 459.99

Flux during Summer-2 (mgs-1) 2.5 % 50 % 97.5 %

Flux during Fall (mgs-1) 2.5 % 50 % 97.5 %

12.67 105.61 0.64 3.56 1.93 0.02 211.66

40.96 43.46 147.81 155.98 2.26 2.41 52.79 71.70 3.09 3.43 0.11 0.21 406.25 442.37

13.49 110.57 0.69 3.77 2.12 0.02 219.19

14.19 117.14 0.74 3.94 2.37 0.03 227.38

46.39 164.18 2.60 96.11 3.84 0.40 483.63

Figures:

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

C1

C2

C3

C4

Fig. 6

Fig. 7

F

ig. 8

Fig. 9

Highlights: •

Dissolved organic matter compositions were used for source-tracking trace metals.

•

Flux of As, Cu, U, PO4, and NO3 correlated with protein-like and soil derived DOM.

•

The trace element and DOM mobility was controlled by precipitation and discharge.

•

Multivariate statistics revealed anthropogenic sources for the trace metals.

We do not have any conflicts of interests.