Elsevier Editorial System(tm) for Science of the Total Environment Manuscript Draft Manuscript Number: Title: Seasonal and annual trends in forage fish mercury concentrations, San Francisco Bay Article Type: Full Length Article Keywords: Mercury; Seasonal; Interannual; Trend; Forage fish; Clevelandia ios; Atherinops affinis; Menidia audens Corresponding Author: Mr. Ben K Greenfield, MS Corresponding Author's Institution: University of California - Berkeley First Author: Ben K Greenfield, MS Order of Authors: Ben K Greenfield, MS; Aroon R Melwani, MS; Rachel M Allen; Darell G Slotton, PhD; Shaun M Ayers; Katherine H Harrold; Katherine Ridolfi, MS; Andrew Jahn, PhD; Jill L Grenier, PhD; Mark B Sandheinrich, PhD Abstract: San Francisco Bay is contaminated by mercury (Hg) due to historic and ongoing sources, and has elevated Hg concentrations throughout the aquatic food web. We monitored Hg in forage fish to indicate seasonal and interannual variation and trends. Interannual variation and long-term trends were determined by monitoring Hg bioaccumulation during September - November, for topsmelt (Atherinops affinis) and Mississippi silverside (Menidia audens) at six sites, over six years (2005 to 2010). Seasonal variation was characterized for arrow goby (Clevelandia ios) at one site, topsmelt at six sites, and Mississippi silverside at nine sites. Arrow goby exhibited a consistent seasonal pattern from 2008 to 2010, with lowest concentrations observed in late spring, and highest concentrations in late summer or early fall. In contrast, topsmelt concentrations tended to peak in late winter or early spring and silverside seasonal fluctuations varied among sites. The seasonal patterns may relate to seasonal shifts in net MeHg production in the contrasting habitats of the species. Topsmelt exhibited an increase in Alviso Slough from 2005 to 2010, possibly related to recent hypoxia in that site. Otherwise, directional trends for Hg in forage fish were not observed. For topsmelt and silverside, the variability explained by year was relatively low compared to sampling station, suggesting that interannual variation is not a strong influence on Hg concentrations. Although fish Hg has shown long-term declines in some ecosystems around the world, San Francisco Bay forage fish did not decline over the six-year monitoring period examined. Suggested Reviewers: Bruce A. Monson PhD Minnesota Pollution Control Agency Bruce.Monson@state.mn.us Expertise in collection, analysis, and interpretation of mercury monitoring data in fish and other matrices. Representative publication: Monson BA. Trend Reversal of Mercury Concentrations in Piscivorous Fish from Minnesota Lakes: 1982−2006. Environmental Science & Technology 2009; 43: 1750-5. Matthew Chumchal PhD

Assistant Professor, Biology, Texas Christian University m.m.chumchal@tcu.edu Expertise in interpreting bioaccumulation patterns of Hg in forage fish, including influence of biological variables. Representative publication: Chumchal MM, Drenner RW, Fry B, Hambright KD, Newland LW. Habitat-specific differences in mercury concentration in a top predator from a shallow lake. Transactions Of The American Fisheries Society 2008; 137: 195-208. R.A. (Drew) Bodaly PhD Project Leader, Penobscot River Mercury Study dbodaly@penobscotmercurystudy.org Expertise in analysis of mercury in fish, including assessment of temporal trends. Representative publications: Bodaly RA, Fudge RJ. Uptake of mercury by fish in an experimental boreal reservoir. Archives of Environmental Contamination and Toxicology 1999; 37: 103-9. Bodaly RA, St. Louis VL, Paterson MJ, Fudge RJP, Hall BD, Rosenberg DM, et al. Bioaccumulation of mercury in the aquatic food chain in newly flooded areas. In: Sigel A, Sigel H, editors. Metal ions in biological systems, Vol. 34: Mercury and its effects on environment and biology. Marcel Dekker Inc., New York, 1997, pp. 259–87 Linda M Campbell PhD Associate Professor & Senior Research Fellow, Saint Mary's University LM.Campbell@smu.ca Temporal dynamics and ecology of Hg bioaccumulation. Representative publication: Zhang L, Campbell LM, Johnson TB. Seasonal variation in mercury and food web biomagnification in Lake Ontario, Canada. Environmental Pollution 2012; 161: 178-84. Opposed Reviewers:

Environmental Health Sciences Division

School of Public Health

University of California, Berkeley

50 University Hall #7360

Berkeley, CA 94720-7360

Tel: 1 510 507 2365

E-mail: bengreenfield@berkeley.edu

Mae Sexauer Gustin, Professor

Editor – Science of the Total Environment

Department of Environmental and Resource Science

Fleischmann Agriculture, Room # 126 Mail Stop 370

University of Nevada-Reno

1664 N. Virginia Street

Reno, NV 89557

Email: mgustin@cabnr.unr.edu

August 10, 2012

Dear Dr. Gustin:

Enclosed for your consideration is a manuscript entitled “Seasonal and annual trends in

forage fish mercury concentrations, San Francisco Bay,” by Ben K. Greenfield and nine

coauthors. We submit this manuscript for possible publication in Science of the Total

Environment.

The study reports on temporal trends in forage fish mercury concentrations in nearshore

regions of San Francisco Bay. The primary contribution of the study is that it contrasts temporal

trends in multiple habitats, fish species, and time scales within the same water body. There was a

surprising lack of consistency among species and sampling locations at both seasonal and

interannual time scales. These results suggest that local rather than regional conditions strongly

influence mercury variation at the time scales examined. The discussion interprets how

cooccuring fish species with different habits may exhibit contrasting bioaccumulation trends.

This manuscript is not being considered for publication elsewhere, and reports entirely

new findings. The analysis of interannual trends reinterprets data previously published in

Greenfield and Jahn (Environmental Pollution 2010; 158: 2716-24) on fish concentrations from

2005 to 2007, in combination with previously unpublished data from 2008 to 2010. The three

analyses of seasonal trends use previously unpublished data only. Potential reviewers include

Bruce Monson, Matthew Chumchal, R.A. (Drew) Bodaly, and Linda Campbell, and any others

you might consider appropriate.

Thank you for considering this manuscript.

Sincerely,

Ben Greenfield

Corresponding Author

Cover Letter

Local site conditions and habitats affect fish mercury temporal variation

Inconsistent seasonal and interannual variation among species and sites

Arrow goby, a mudflat burrower, had elevated mercury in summer

Mercury increased in the winter for the marine pelagic topsmelt

Silverside mercury increased from 2005 to 2010 at a hypoxia affected site

*Highlights (for review)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

1

1

Seasonal and annual trends in forage fish mercury concentrations, San Francisco Bay 2

3

Ben K. Greenfielda,b, Aroon R. Melwania,1, Rachel M. Allen a,2, Darell G. Slottonc, Shaun 4

M. Ayersc, Katherine H. Harrolda,3, Katherine Ridolfia,4, Andrew Jahnd, J. Letitia Greniera,5, 5

Mark B. Sandheinriche 6

7

8

a. San Francisco Estuary Institute, 4911 Central Ave., Richmond, CA 94804 9

b. Corresponding author. Environmental Health Sciences Division, School of Public 10 Health, University of California-Berkeley, 50 University Hall #7360, Berkeley, CA 94720-11 7360 Tel: 1 510 507 2365 E-mail: bengreenfield@berkeley.edu 12

c. Department of Environmental Science and Policy, University of California-Davis, One 13 Shields Avenue, Davis, CA 95616, USA dgslotton@ucdavis.edu 14

d. 1000 Riverside Drive, Ukiah, CA 95482 andyjahn@mac.com 15

e. River Studies Center, University of Wisconsin-La Crosse, 1725 State Street, La Crosse, 16 Wisconsin 54601, USA sandhein.mark@uwlax.edu 17

1. Current address: Department of Biological Sciences, Macquarie University, NSW 2109, 18 Australia aroon.melwani@gmail.com 19

*ManuscriptClick here to download Manuscript: Final text.docx Click here to view linked References

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

2

2. Current address: Civil and Environmental Engineering, University of California-20 Berkeley, CA, USA rmallen86@gmail.com 21

3. Current address: Environmental Sciences and Engineering, Gillings School of Global 22 Public Health, University of North Carolina at Chapel Hill, CB #7431, Chapel Hill, NC 23 27599-7431, USA katie.harrold@gmail.com 24

4. Current address: LimnoTech, 1705 DeSales St. NW, Suite 600, Washington, DC 20036, 25 USA kridolfi@limno.com 26

5. Current address: 5923 Marden Lane, Oakland, CA 94611 letitia@letitia.org 27

28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

3

Abstract 29

San Francisco Bay is contaminated by mercury (Hg) due to historic and ongoing sources, 30 and has elevated Hg concentrations throughout the aquatic food web. We monitored 31 Hg in forage fish to indicate seasonal and interannual variation and trends. Interannual 32 variation and long-term trends were determined by monitoring Hg bioaccumulation 33 during September - November, for topsmelt (Atherinops affinis) and Mississippi 34 silverside (Menidia audens) at six sites, over six years (2005 to 2010). Seasonal variation 35 was characterized for arrow goby (Clevelandia ios) at one site, topsmelt at six sites, and 36 Mississippi silverside at nine sites. Arrow goby exhibited a consistent seasonal pattern 37 from 2008 to 2010, with lowest concentrations observed in late spring, and highest 38 concentrations in late summer or early fall. In contrast, topsmelt concentrations tended 39 to peak in late winter or early spring and silverside seasonal fluctuations varied among 40 sites. The seasonal patterns may relate to seasonal shifts in net MeHg production in the 41 contrasting habitats of the species. Topsmelt exhibited an increase in Alviso Slough from 42 2005 to 2010, possibly related to recent hypoxia in that site. Otherwise, directional 43 trends for Hg in forage fish were not observed. For topsmelt and silverside, the 44 variability explained by year was relatively low compared to sampling station, suggesting 45 that interannual variation is not a strong influence on Hg concentrations. Although fish 46 Hg has shown long-term declines in some ecosystems around the world, San Francisco 47 Bay forage fish did not decline over the six-year monitoring period examined. 48

Keywords 49

Mercury; Seasonal; Interannual; Trend; Forage fish; Clevelandia ios; Atherinops affinis; 50

Menidia audens 51

1 Introduction 52

Mercury (Hg) is a globally distributed pollutant that poses serious management 53

concerns because the methylated form (MeHg) bioaccumulates and is toxic to humans 54

and wildlife (U.S. EPA, 1997). Hg pollution varies across multiple time scales, reflecting a 55

range of processes (Wiener et al., 2007). Seasonal variation likely reflects patterns in net 56

Hg methylation in the sediment or water column (Gorski et al., 1999; Heim et al., 2007; 57

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

4

Herrin et al., 1998). Interannual fluctuations (i.e., temporary increases or decreases in 58

specific years) can result from temporary perturbations due to management activities, 59

weather patterns, or other environmental changes (Henery et al., 2010). Long-term 60

trends can result from changes in Hg loading or in processes that affect methylation 61

(Braune et al., 2005; Levinton and Pochron, 2008; Monson, 2009; Munthe et al., 2007). 62

Small fish consumed by piscivorous wildlife (i.e., forage fish) are useful as biosentinels to 63

determine seasonal and interannual variation in MeHg resulting from natural variability 64

or management activities (Wiener et al., 2007). Forage fish have been widely employed 65

in fresh waters to identify changes in Hg exposure due to management perturbations 66

such as reservoir impoundment or lake acidification (e.g., Kelly et al., 1997; Wiener et 67

al., 1990). In addition, sediment and forage fish monitoring has indicated natural 68

seasonal variation in Hg concentrations (e.g., Eagles-Smith and Ackerman, 2009; Fowlie 69

et al., 2008; Gorski et al., 1999; Slotton et al., 1995; Zhang et al., 2012). In freshwater 70

lakes and estuarine wetlands, strong summer increases have been observed, which 71

corresponds to the period of highest risk to embryonic development for some 72

piscivorous bird species (Eagles-Smith and Ackerman, 2009; Gorski et al., 1999). Studies 73

have often been limited in duration to several months, limiting the evaluation of 74

patterns across multiple seasons or years. Comparisons across multiple fish species or 75

multiple sampling locations within large water bodies to determine consistency of 76

seasonal or interannual patterns are also lacking. 77

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

5

San Francisco Bay (the Bay) has elevated mercury (Hg) concentrations in fish and wildlife 78

(Ackerman et al., 2007; Greenfield et al., 2005; Greenfield and Jahn, 2010), with some 79

evidence of sublethal effects to birds and harbor seals (Herring et al., 2010; 80

Schwarzbach et al., 2006). Consequently, there are substantial efforts to control specific 81

Hg sources through the U.S. federally mandated Total Maximum Daily Load (TMDL) 82

program (SFBRWQCB, 2006). An intensive restoration program is underway to restore 83

wetlands surrounding the Bay, including the conversion of thousands of hectares of 84

historic salt ponds and other isolated habitats to wetlands (Goals Project, 1999; Grenier 85

and Davis, 2010). MeHg production can be relatively high in nearshore managed ponds 86

and wetland habitats due to differences in hydrology and redox conditions (Davis et al., 87

2003; Grenier and Davis, 2010; Heim et al., 2007; Miles and Ricca, 2010). Consequently, 88

Hg monitoring in Bay biota is underway to evaluate potential short and long-term 89

changes in MeHg exposure and bioaccumulation due to habitat restoration. 90

We report seasonal and interannual variation in Hg concentrations in San Francisco Bay 91

forage fish. Seasonal variation was characterized over three separate monitoring 92

periods, encompassing three species at sixteen separate locations (Fig. 1). Interannual 93

variation and trends were determined by fall monitoring of two species at six sites over 94

six years. We assessed whether temporal patterns were consistent across sites, 95

sampling events (i.e., date), and species and the relative importance of temporal 96

variation versus spatial variation for explaining Hg concentrations. 97

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

6

2 Methods 98

Temporal Hg trends were evaluated in three San Francisco Bay forage fish species: 99

Mississippi silverside (Menidia audens), topsmelt (Atherinops affinis), and arrow goby 100

(Clevelandia ios). These species were selected based on documented usefulness as 101

biosentinels (Greenfield and Jahn, 2010), and consistent residence at the chosen 102

sampling locations. All sampling was performed by beach seine adjacent to the 103

shoreline. 104

2.1 Sampling dates and site descriptions 105

Four separate sampling efforts were conducted, each focusing on specific time scales 106

and sampling locations. The first effort evaluated seasonal patterns from October 2007 107

to December 2010, at six-week intervals (Table 1). Trends were assessed at four stations 108

within Central San Francisco Bay: Martin Luther King Jr. Regional Shoreline (MLK), 109

Berkeley (BRK), Keller Beach (KEL), and Tiburon (TIB) (Fig. 1). MLK is a subtidal mudflat, 110

abutting a 20 ha wetland complex. MLK is enclosed within San Leandro Bay, a small 111

basin with legacy industrial contamination and outlets to Oakland Harbor and Central 112

San Francisco Bay (Daum et al., 2000). MLK was sampled for topsmelt and arrow goby 113

(Mississippi silverside were only intermittently present). BRK, KEL, and TIB were sampled 114

for topsmelt only. These sites are marine sandy beach sites adjacent to the open waters 115

of Central Bay. Arrow goby (a mudflat inhabitant) and Mississippi silverside (favoring 116

lower salinities) were not present at BRK, KEL, and TIB. Samples were obtained from 117

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

7

these three sites as part of long-term sampling performed by US Fish and Wildlife 118

Service (Stockton, CA). It was not possible to collect topsmelt at all sampling events, due 119

to their heterogeneous distribution in the Estuary throughout the year. For example, 120

sample collection at TIB began in August, 2008. 121

The second sampling effort entailed seasonal sampling at four seasonal sampling 122

periods: October 2010, January 2011, May 2011, and July 2011. This effort covered four 123

Bay locations: Mallard Slough, Alviso Slough (both in Lower South Bay), Eden Landing 124

(South Bay), and Benicia State Park (San Pablo Bay; Fig. 1). Silverside were captured on 125

all dates and stations, and topsmelt were captured at Alviso Slough and Eden Landing 126

only (Table 1). 127

The third sampling effort focused on seasonal variation in the Napa-Sonoma salt ponds, 128

located north of San Pablo Bay, near the Napa River (Table 1, Fig. 1). Historically part of 129

a large wetland complex, the Napa-Sonoma ponds were originally developed for salt 130

extraction, but are currently managed as waterbird habitat. Mississippi silverside were 131

collected seasonally on December 2009, March 2010, May 2010, and July 2010. 132

Sampling occurred on five ponds (Pond 1, Pond 2, Pond 2A, Pond 4/5, and Pond 9/10) 133

and on one site on the Napa River (Kennedy Park). 134

Finally, annual sampling was performed at six Bay locations from 2005 to 2010: Alviso 135

Slough, Newark Slough, Bair Island, Eden Landing, China Camp, and Benicia State Park 136

(Table 1, Fig. 1). Annual sampling was only performed from September through 137

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

8

November, to determine interannual trends while reducing the influence of 138

confounding seasonal variability. Data collected from 2005 to 2007 were reported 139

previously (Greenfield and Jahn, 2010), and are here reanalyzed in combination with 140

new data from 2008 to 2010. 141

2.2 Sample preparation and analysis 142

For each date and station combination, four to six composites of five to ten individuals 143

per species were targeted. Target total lengths were 25 to 40 mm for arrow goby, 40 to 144

80 mm for silverside, and 60 to 100 mm for topsmelt. Fish samples were frozen on the 145

day of collection in a -20O C laboratory freezer, and analyzed at a later date for total Hg 146

by standard cold vapor atomic absorption (CVAA) spectrophotometry. From 2005 147

through 2007, laboratory analysis was performed at the River Studies Center at the 148

University of Wisconsin – La Crosse, with analytical methods described in Greenfield and 149

Jahn (2010). Laboratory analyses for all samples collected from 2008 to 2011 were 150

performed at the University of California – Davis. Further analytical and QA methods are 151

described in Supplemental Information. All study Hg results are reported on a wet 152

weight basis. 153

2.3 Data analysis 154

Linear models were constructed to evaluate temporal variation at seasonal or annual 155

scales using the linear model function in R version 2.15 (R Development Core Team, 156

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

9

2012). Total Hg was log10 transformed prior to analysis to improve normality and 157

variance homoscedasticity of residuals. Potential input parameters included sampling 158

date (categorical), station, and fish total length. Interaction terms were also evaluated 159

to evaluate potential changes in the relationship between fish length and Hg over time 160

(date*length) or across stations (length*station), as well as whether different stations 161

exhibited different interannual variation or trends (date*station). For the six-week 162

seasonal evaluations (i.e., MLK, BRK, KEL, and TIB stations), each site-species 163

combination was evaluated individually, due to the inconsistent seasonal coverage 164

across sites; a station effect and a date*length interaction were not included. 165

Parameters were retained based on statistical significance of addition to the full model 166

(p < 0.05); the final model contained only significant parameters. 167

The squared semipartial correlation (hereafter, sr2) was determined for each significant 168

parameter. Calculated as the difference in R2 resulting from removing that parameter 169

from the full model, the sr2 indicates the amount of added variation that parameter 170

explains after all other parameters are accounted for. For base terms (i.e., date, station, 171

and length), the sr2 was determined based on the full model, prior to adding any 172

interaction terms. 173

Spatial and temporal patterns were plotted, including concentration averages and 174

confidence intervals (CI). When length was a significant predictor, plotted 175

concentrations were corrected to a standardized length, based on the model length 176

coefficient. To aid in visual comparison of concentration means across sampling events, 177

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

10

error bars (CI) were constructed based on 95% confidence for a pairwise difference 178

between means with unknown mean and unknown standard deviation; i.e., CI = (t0.915, df 179

= n-1) * SD/ , where SD and N are specific to each sampling event (Austin and Hux, 180

2002; Goldstein and Healy, 1995). 181

3 Results 182

All models included date in the final model structure (p < 0.05), indicating temporal 183

variation was present at all time scales examined. Date alone explained much more 184

variation in the individually examined six-week Central Bay sites (squared semipartial 185

correlation [sr2] ranging from 0.57 to 0.93) than the remaining evaluations (sr2 from 0.04 186

to 0.12). A date*station interaction was present in all of the annual and seasonal 187

sampling efforts (sr2 from 0.11 to 0.36), except for the 2011 examination of topsmelt at 188

Alviso Slough and Eden Landing (Table 2). 189

3.1 Seasonal variation in Central Bay, 2007 to 2010 190

191 Arrow goby at MLK and topsmelt at all four sites exhibited strong variation across 192

sampling dates. Date was included, with high sr2, in the models for both species and all 193

stations, as was length for all combinations except topsmelt at TIB and at KEL (Table 2). 194

Length corrected arrow goby sampled at MLK exhibited a clear seasonal pattern from 195

2008 to 2010 (Fig. 2a). In each year, concentrations were lowest in the early spring (i.e., 196

March or April), and then increased and peaked in July (2008) or September (2009 and 197

2010). 198

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

11

Concentrations in topsmelt at all stations exhibited seasonal periodicity that was 199

generally distinct from arrow goby at MLK. Length corrected topsmelt geometric mean 200

concentrations at MLK were elevated in winter (December, 2008, late November, 2009, 201

and January, 2010). Summer concentrations varied among years, with elevated 202

concentrations in July, 2008 and July, 2009 but low concentrations in July, 2010 (Fig. 2b). 203

For the remaining stations, topsmelt concentrations peaked during late winter to early 204

spring of every monitored year (Fig. 2c-e). Annual peak geometric average topsmelt 205

concentrations were highest in February for TIB in 2009; KEL in 2008 and 2009; and BRK 206

(length corrected) in all three years. Similarly, peak concentrations for KEL in 2010 were 207

in March. 208

3.2 Seasonal variation in San Francisco Bay, 2011 209

For both silverside and topsmelt, there was a significant effect of date and station but 210

not length. Station explained a greater proportion of the total variation in fish Hg than 211

date (Table 2). For example, median silverside concentrations at Alviso Slough (0.27 μg 212

g-1) were more than four times those at Benicia State Park (0.06 μg g-1). Silverside 213

exhibited moderately different seasonal patterns among stations (date*station sr2 = 214

0.11). In particular, the lowest concentrations at Mallard Slough were in July 2011, while 215

the lowest concentrations at Benicia State Park occurred in October 2010 (Fig. 3, top 216

panels). Silverside were absent from Alviso Slough in July, but seasonal differences 217

among stations were still present in an analysis excluding Alviso Slough (date*station sr2 218

= 0.22; N = 62). For topsmelt, seasonal variations were consistent between Alviso Slough 219

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

12

and Eden Landing, with concentrations lowest in October and highest in July at both 220

stations (Fig. 3, bottom panels). The increase in median topsmelt concentrations from 221

October to January was consistent with the seasonal pattern found at Central Bay 222

stations. 223

3.3 Seasonal variation in North Bay Ponds, 2010 224

225

Hg in Mississippi silverside from North Bay Ponds was influenced by date, station, and 226

length, but the seasonal pattern was divergent across stations (date*station sr2 = 0.36, 227

Table 2). In particular, Pond 2 Hg exhibited much greater seasonal variation than other 228

stations (Fig. 4), with a dramatic decrease from December, 2009 (median = 0.20 μg g-1) 229

to March (0.14 μg g-1) and May, 2010 (0.045 μg g-1). In contrast, Pond 4/5 and Kennedy 230

Park exhibited increases from December to March (Fig. 4). 231

3.4 Interannual variation 232

For both silverside and topsmelt, there were significant effects of date, station, length, 233

and a date*station interaction (Table 2). For Mississippi silverside, the variability 234

explained by sampling station (sr2 = 0.60) was much greater than that accounted for by 235

date (i.e., year; sr2 = 0.05), fish length (sr2 = 0.05), and a date*station interaction (sr2 = 236

0.11), indicating that the majority of variability in silverside Hg resulted from differences 237

among sampling locations. Station was also most important for topsmelt (sr2 = 0.35), 238

which had somewhat more variability explained by length (sr2 = 0.15) than silverside. As 239

reported previously (Greenfield and Jahn, 2010), concentrations were highest in Lower 240

South Bay stations, and lowest in China Camp and Benicia State Park. 241

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

13

There was an overall lack of directional trend over the six sampling years, with the 242

exception of increasing silverside Hg concentrations over time at Alviso Slough (Fig. 5). 243

Interannual patterns were generally inconsistent across stations for silverside (Fig. 5) 244

and topsmelt (Fig. 6). However, silverside concentrations in 2006 were lower than other 245

years at Alviso Slough, Newark Slough, and China Camp (Fig. 5). Topsmelt exhibited 246

qualitatively similar interannual variation at Newark Slough and Eden Landing (Fig. 6). At 247

Eden Landing, silverside concentrations in 2005 were much higher than other years. 248

Outliers were evident, with silverside samples from China Camp (2006) and Benicia 249

State Park (2008) and a topsmelt sample from Alviso Slough (2009) three to four times 250

the remaining samples. 251

4 Discussion 252

Consistent with previous studies (e.g., Eagles-Smith and Ackerman, 2009; Fowlie et al., 253

2008; Gorski et al., 1999; Slotton et al., 1995; Ward and Neumann, 1999; Zhang et al., 254

2012), we found significant seasonal variation in fish Hg. However, we also observed a 255

striking lack of consistency in temporal patterns across different sites and species, at 256

both seasonal and interannual scales. Species differences included different interannual 257

and seasonal patterns for silverside versus topsmelt in San Francisco Bay and different 258

seasonal patterns for topsmelt versus arrow goby in Central Bay. Site differences were 259

strongest for seasonal variation in silverside from North Bay Ponds, likely related to the 260

hydrologic isolation of these ponds. However, site differences were also apparent in 261

interannual variation in Central Bay sites. Perhaps related to the wide range of habitats, 262

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

14

Hg sources, and Hg methylation rates within a complex estuary such as San Francisco 263

Bay (Davis et al., 2012; Goals Project, 1999; Heim et al., 2007), we observed few 264

consistent temporal patterns of small fish Hg exposure across species and sites. 265

4.1 Seasonal variation 266

267 In the three-year examination of Central Bay sites, strong seasonal variation was evident 268

across the sampling years. Previous examinations of seasonal patterns in biota Hg 269

typically found consistent patterns of seasonal variation across species (Eagles-Smith 270

and Ackerman, 2009; Zhang et al., 2012). In contrast, we found that arrow goby and 271

topsmelt collected at the same site (MLK) exhibited distinct seasonal patterns, with 272

arrow goby Hg highest in late summer and early fall, while topsmelt Hg concentrations 273

were elevated in early spring and variable in the summer. Topsmelt Hg concentrations 274

also tended to increase during late winter and early spring in the other Central Bay sites 275

(2008 to 2010) and Alviso Slough and Eden landing (2010 to 2011), suggesting a 276

consistent pattern for this species across sites. Unlike topsmelt, silverside exhibited 277

notable spatial variation in seasonal patterns; for example, North Bay Pond 2 silverside 278

had a dramatic difference in Hg, not exhibited at other stations. 279

Previously described habitat, home range, and salinity differences among topsmelt, 280

silverside, and arrow goby (Greenfield and Jahn, 2010) could result in differing seasonal 281

and interannual patterns of MeHg exposure among these species. Although the 282

difference in seasonal Hg between topsmelt and arrow goby could relate to differing 283

seasonal growth or dietary shifts, the seasonal patterns were still observed after 284

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

15

adjusting for body size of sampled fishes. We hypothesize that the difference may stem 285

from differing seasonal changes in MeHg concentrations in the contrasting habitats and 286

spatial scales utilized by these species, resulting from spatial differences in net MeHg 287

production. Arrow gobies are sedentary burrow dwellers, indicating a highly localized 288

area, including the demersal food web near their intertidal mudflat burrows (Emmett et 289

al., 1991; Goals Project, 2000; Prasad, 1958). In contrast, topsmelt exhibit onshore and 290

offshore movements with the tides, and thus indicate a spatially broader sampling of 291

the Central Bay bentho-pelagic food web (Greenfield and Jahn, 2010; Visintainer et al., 292

2006). Therefore, in our study the different species indicated different spatial scales as 293

well as different food webs. 294

Available data on seasonal variation in estuarine wetland and mudflat MeHg generally 295

supports the seasonal pattern observed in arrow goby. The mudflats at MLK, and other 296

arrow goby mudflat habitats, are adjacent to emergent tidal wetlands, which often 297

exhibit summer increases in MeHg production and export. Mitchell et al. (2012) 298

observed summer increases in water column concentrations and net export from an 299

estuarine wetland draining the Chesapeake Bay, and Best et al. (2008) found that 300

fragmentation and leaching of two abundant wetland plant species (Spartina foliosa and 301

Salicornia virginica) causes late spring and summer peaks in MeHg release from the 302

China Camp tidal marsh (Fig. 1) to San Francisco Bay. There may also be increased 303

summer MeHg production within the mudflats themselves, associated with elevated 304

temperatures and increasing in situ activity of sulfate reducing bacteria. MeHg 305

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

16

concentrations are seasonally elevated in spring and summer within intertidal mudflat 306

surface sediment and sediment-dwelling polychaetes (Nereis diversicolor) of the Scheldt 307

Estuary, Belgium (Muhaya et al., 1997). 308

Rather than being determined by mudflat MeHg fluctuations, topsmelt Hg is likely 309

influenced by pelagic water column processes and exchange with offshore sediments, 310

since topsmelt appear to spend some time in pelagic and offshore habitats (Greenfield 311

and Jahn, 2010; Visintainer et al., 2006). Consistent with the topsmelt Hg pattern, 312

Conaway et al. (2003) found higher water column MeHg concentrations in February 313

(2000) than July (1999, 2000) in North Bay (i.e., San Pablo Bay to the Sacramento-San 314

Joaquin River Delta), although they found no significant seasonal pattern in Central or 315

South Bay. Luengen and Flegal (2009) also observed a spring increase in Bay water 316

column MeHg concentrations , following a phytoplankton bloom. These limited findings, 317

in combination with our forage fish data, suggest that the timing of MeHg production 318

peaks may differ between the Bay water column (spring peak) and intertidal wetlands 319

(summer peak). 320

Seasonal variability should be considered in sampling design, with long-term trend or 321

spatial studies focusing on a narrow and consistent sampling season (Wiener et al., 322

2007), and evaluations of MeHg risk to piscivorous wildlife focusing on developmental 323

periods for sensitive life stages (Eagles-Smith and Ackerman, 2009; Herring et al., 2010). 324

Changes over time accounted for half to nearly all (57% to 93%) of the variation in 325

forage fish Hg that we sampled every six weeks. Similarly, Eagles-Smith and Ackerman 326

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

17

(2009) found Hg in longjaw mudsucker (Gillichthys mirabilis) and three-spined 327

stickleback (Gasterosteus aculeatus) from managed wetlands to increase 40% during the 328

summer months of 2006. Mercury monitoring in multiple seasons is warranted when 329

there are multiple predator species of concern or uncertainty regarding the most 330

sensitive periods. 331

The contrasting patterns observed among forage fish species and locations in this study 332

indicate a need to understand what prey species and habitats are favored by the 333

piscivorous wildlife of management concern. Seasonal patterns in topsmelt Hg 334

concentrations contrasted with arrow goby and also with longjaw mudsucker and three-335

spined stickleback sampled in South Bay managed ponds (Eagles-Smith and Ackerman, 336

2009), indicating species-specific temporal risk patterns for piscivorous birds. For 337

example, Forster’s tern (Sterna forsteri) predominantly consumes the wetland fish 338

species (Eagles-Smith and Ackerman, 2009), whereas California Least tern (Sterna 339

antillarum browni, a federally endangered bird) largely consumes topsmelt and other 340

Atherinopsidae (Elliott, 2005; Elliott et al., 2007). Arrow goby appears to be a 341

component of the diet of mudflat feeding birds, including greater yellowleg (Tringa 342

melanoleuca) and long-billed dowitcher (Limnodromus scolopaceus) (Goals Project, 343

2000; Reeder, 1951). 344

The unique MeHg bioaccumulation pattern in silverside from North Bay Pond 2 indicates 345

that site-specific biogeochemical conditions can develop in these isolated ponds. Pond 346

2, like several of the North Bay Ponds, is fairly hydrologically isolated, lacking major 347

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

18

inflows or outflows, and exhibiting muted tidal action. Silverside in Pond 2 exhibited the 348

strongest temporal fluctuation among all sites and species examined, with a more than 349

fourfold concentration decline between December, 2009 and May, 2010. Possible 350

mechanisms to explain this variation include seasonal differences in net MeHg 351

production or variations in trophic status causing biodilution. 352

Beyond this study, seasonal patterns in MeHg bioaccumulation vary by region, species 353

and habitat. Wiener et al. (2007) suggest that in freshwater forage fish, Hg 354

concentrations and mass tend to be highest during the summer months. Three-spined 355

stickleback and longjaw mudsucker in Lower South Bay managed wetlands exhibited 356

distinct seasonal patterns from arrow goby and topsmelt in our study, with highest 357

concentrations in late May of 2006 (Eagles-Smith and Ackerman, 2009). Slotton et al. 358

(2007) found site and species differences in seasonal patterns. In bi-monthly sampling of 359

11 North Bay and Delta sites from November 2005 to November 2006, silverside Hg 360

concentrations peaked in February, May, July, or November depending on sampling 361

location. Seasonal variation in small fish Hg is likely driven by seasonal patterns in water 362

MeHg concentrations, which also vary among study systems. For example, maximum 363

MeHg concentrations in Florida Everglades wetlands and Venice Lagoon mudflats peak 364

in midsummer (Bloom et al., 2004; Hurley et al., 1998). Lavaca Bay, Texas exhibited 365

highest sediment to water column MeHg fluxes in late winter (Gill et al., 1999), and in 366

Davis Creek Reservoir, California, and Devil’s Lake, Wisconsin, fall destratification 367

resulted in substantial food web MeHg uptake (Herrin et al., 1998; Slotton et al., 1995). 368

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

19

4.2 Interannual variation 369

Interannual variation, though significant, explained limited variation in fish Hg compared 370

to previously documented (Greenfield and Jahn, 2010) spatial differences among 371

sampling stations. Both silverside and topsmelt exhibited interannual variation in Hg 372

concentrations that differed among sites, suggesting an influence of local conditions on 373

temporal Hg patterns. 374

Site specific differences in Hg patterns may be associated with the hydrologic 375

management of former salt ponds adjacent to South Bay sites including Alviso Slough, 376

Bair Island, and Newark Slough (Miles and Ricca, 2010). These ponds are currently 377

managed for wildlife as part of a long-term restoration project. This project could 378

potentially alter biotic MeHg exposure due to the high sediment MeHg concentrations 379

in some of the salt ponds and the potential for hydrologic modification, including cycles 380

of wetting and drying, to increase MeHg production or release (Grenier and Davis, 381

2010). In Eden Landing, silverside were elevated in 2005 compared to other years, which 382

could also be associated with pond management activities. Miles and Ricca (2010) 383

observed a decline in surface sediment MeHg in Eden Landing Salt Ponds associated 384

with restoration, from winter (i.e., early) 2005 to winter 2006. In this restoration, water 385

exchange between salt ponds and Old Alameda Creek (the location of the Eden Landing 386

small fish sampling site) was increased via flood control gates (South Bay Salt Pond 387

Restoration Project, 2006). The 2005 elevated silverside concentrations at the Eden 388

Landing site could have resulted from short term exposure to MeHg released from the 389

Eden Landing pond complex, as they transitioned from hydrologically isolated ponds 390

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

20

having elevated sediment MeHg production to ponds with a greater hydrological 391

connection to the South Bay. Monitoring of Alviso Slough is underway to determine 392

whether restoration of greater connectivity between tidal channels and managed ponds 393

will cause temporal changes in forage fish Hg (D.G. Slotton, UC Davis, Pers. comm.). 394

At Alviso Slough, decreases in fall dissolved oxygen and consequent increases in MeHg 395

production may have caused the increases in silverside Hg from 2005 to 2010, and 396

topsmelt high outlier results in 2009 and 2010. The time period of increasing fish Hg 397

corresponds with reported fish kills in the Guadalupe River draining to Alviso Slough in 398

2008, 2009, and 2010 (City of San Jose et al., 2011). The fish kills occurred around the 399

time we collected fish, and are believed to be related to low dissolved oxygen levels in 400

the river channel (R. Schlipf, SFRWQCB, pers. comm.). Also during this time period, 401

compliance monitoring of outfall water from one of the salt ponds that drains to Alviso 402

Slough (Pond A7) indicated an annual increasing rate of hypoxic water between 2005 403

and 2009, which pond managers attribute to periodic algae blooms within the pond. 404

Continuous dissolved oxygen (DO) measurements were obtained and the 10th percentile 405

of DO concentration was reported on a weekly basis to indicate potential for hypoxia 406

(United States Fish and Wildlife Service and United States Geological Survey, 2012). The 407

average result between July 19 and October 3 was 3.4, 4.4, 1.9, 0.6, and 1.0 mg L-1 DO 408

over the five year duration. Hypoxic conditions are often associated with increased 409

MeHg production in aquatic systems, due to activity of sulfate reducing bacteria 410

(Gilmour et al., 1992; Watras et al., 1995). We hypothesize that the increasing silverside 411

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

21

Hg and topsmelt outliers in 2009 and 2010 at Alviso Slough may be associated with 412

increased incidence of hypoxia in the Guadalupe River and Alviso Slough. Hypoxic 413

conditions would also explain the observed fish kills during this same time period, and 414

suggest the need for synthesis of local dissolved oxygen monitoring data. 415

Consistent long term directional trends for Hg in forage fish were not apparent over the 416

six-year study duration. With the exception of Alviso Slough, most stations did not 417

exhibit increases or decreases over time, and spatial differences were greater than 418

temporal variation. Fish Hg has exhibited long-term declines in many ecosystems 419

(Levinton and Pochron, 2008; Munthe et al., 2007), but this declining trend is 420

inconsistent among water bodies and may be reversing in recent years (Monson, 2009). 421

In San Francisco Bay, PCBs and legacy pesticides have declined but Hg concentrations 422

have exhibited no consistent upward or downward trends in Bay biota (Greenfield et al., 423

2005; Gunther et al., 1999; Stephenson et al., 1995). This is likely due to extensive 424

mixing of historic Hg throughout the system and continued atmospheric and fluvial 425

inputs, resulting in generally stable concentrations of total Hg in sediment and rates of 426

change on the order of decades to centuries (Davis et al., 2012; Yee et al., 2011). 427

5 Acknowledgements 428

Field sampling assistance was provided by J.L. Grenier, A. Robinson, S.B. Shonkoff, S. 429

Slater, T. Grenier, M. Delaney, C. Austin, R. Looker, M. Odaya, B. Stanford, M. Lent, G. 430

Nelson, R. Wilder, E. Volkman, P. Hrody, and staff of the USFWS-Stockton, with funding 431

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

22

from the Interagency Ecological Program. We thank California Department of Parks and 432

Recreation, Department of Fish and Game, East Bay Regional Parks District, and U.S. Fish 433

and Wildlife Service for site access and sampling guidance. J.A. Davis and J. Griswold 434

constructively reviewed drafts of the manuscript. This study was funded by the Regional 435

Monitoring Program for Water Quality in San Francisco Bay (SFEI Publication ###), with 436

additional support by a US EPA STAR Graduate Fellowship to BG. 437

6 References 438

Ackerman JT, Eagles-Smith CA, Takekawa JY, Demers SA, Adelsbach TL, Bluso JD, et al. 439

Mercury concentrations and space use of pre-breeding American avocets and 440

black-necked stilts in San Francisco Bay. Science of the Total Environment 2007; 441

384: 452-66. 442

Austin PC, Hux JE. A brief note on overlapping confidence intervals. Journal of Vascular 443

Surgery 2002; 36: 194-5. 444

Best EPH, Hintelmann H, Dimock B, Bednar AJ. Natural cycles and transfer of mercury 445

through pacific coastal marsh vegetation dominated by Spartina foliosa and 446

Salicornia virginica. Estuaries and Coasts 2008; 31: 1072-88. 447

Bloom NS, Moretto LM, Scopece P, Ugo P. Seasonal cycling of mercury and monomethyl 448

mercury in the Venice Lagoon (Italy). Marine Chemistry 2004; 91: 85-99. 449

Braune BM, Outridge PM, Fisk AT, Muir DCG, Helm PA, Hobbs K, et al. Persistent organic 450

pollutants and mercury in marine biota of the Canadian Arctic: An overview of 451

spatial and temporal trends. Science of the Total Environment 2005; 351: 4-56. 452

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

23

City of San Jose, Santa Clara Valley Water District, EOA Inc. Follow-up evaluation of 453

stormwater pump station discharges in San Jose, 2011, pp. 41 454

Conaway CH, Squire S, Mason RP, Flegal AR. Mercury speciation in the San Francisco Bay 455

estuary. Marine Chemistry 2003; 80: 199-225. 456

Daum T, Lowe S, Toia R, Bartow G, Fairey R, Anderson J, et al. Sediment contamination 457

in San Leandro Bay, CA. San Francisco Estuary Institute, San Francisco Bay 458

Regional Water Quality Control Board, California Department of Fish and Game, 459

and Port of Oakland, Richmond, CA, 2000, pp. 52. 460

http://www.sfei.org/node/2103 461

Davis JA, Looker RE, Yee D, Marvin-Dipasquale M, Grenier JL, Austin CM, et al. Reducing 462

methylmercury accumulation in the food webs of San Francisco Bay and its local 463

watershed. Environmental Research 2012; In review. 464

Davis JA, Yee D, Collins JN, Schwarzbach SE, Luoma SN. Potential for increased mercury 465

accumulation in the Estuary food web. San Francisco Estuary and Watershed 466

Science 2003; 1(1): Article 4. 467

Eagles-Smith CA, Ackerman JT. Rapid changes in small fish mercury concentrations in 468

estuarine wetlands: implications for wildlife risk and monitoring programs. 469

Environmental Science and Technology 2009; 43: 8658-64. 470

Elliott ML. Diet, prey, and foraging habits of the California least tern (Sterna antillarum 471

browni). San Francisco State University, 2005, pp. 89 472

Elliott ML, Hurt R, Sydeman WJ. Breeding biology and status of the California least tern 473

Sterna antillarum browni at Alameda Point, San Francisco Bay, California. 474

Waterbirds 2007; 30: 317-25. 475

Emmett RL, Hinton SA, Stone SL, Monaco ME. Distribution and abundance of fishes and 476

invertebrates in west coast estuaries, Volume II: Species life history summaries. 477

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

24

ELMR Report No. 8 NOAA/NOS Strategic Environmental Assessments Division, 478

Rockville, MD, 1991, pp. 329 479

Fowlie AR, Hodson PV, Hickey MBC. Spatial and seasonal patterns of mercury 480

concentrations in fish from the St. Lawrence River at Cornwall, Ontario: 481

Implications for monitoring. Journal of Great Lakes Research 2008; 34: 72-85. 482

Gill GA, Bloom NS, Cappellino S, Driscoll CT, Dobbs C, McShea L, et al. Sediment−water 483

fluxes of mercury in Lavaca Bay, Texas. Environmental Science and Technology 484

1999; 33: 663-9. 485

Gilmour CC, Henry EA, Mitchell R. Sulfate stimulation of mercury methylation in fresh-486

water sediments. Environmental Science and Technology 1992; 26: 2281-7. 487

Goals Project. Baylands Ecosystem Habitat Goals. A report of habitat recommendations 488

prepared by the San Francisco Bay Area Wetlands Ecosystem Goals Project. U.S. 489

Environmental Protection Agency, San Francisco, Calif./S.F. Bay Regional Water 490

Quality Control Board, Oakland, CA. , 1999, pp. 328. 491

http://www.sfei.org/sfbaygoals/ 492

Goals Project. Baylands Ecosystem Species and Community Profiles: Life histories and 493

environmental requirements of key plants, fish and wildlife. Oakland, CA: 494

Prepared by the San Francisco Bay Area Wetlands Ecosystem Goals Project. P.R. 495

Olofson, editor. San Francisco Bay Regional Water Quality Control Board, 2000. 496

http://www.abag.ca.gov/bayarea/sfep/reports.html 497

Goldstein H, Healy MJR. The graphical presentation of a collection of means. Journal of 498

the Royal Statistical Society Series A-Statistics in Society 1995; 158: 175-7. 499

Gorski PR, Lathrop RC, Hill SD, Herrin RT. Temporal mercury dynamics and diet 500

composition on the mimic shiner. Transactions of the American Fisheries Society 501

1999; 128: 701-12. 502

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

25

Greenfield BK, Davis JA, Fairey R, Roberts C, Crane D, Ichikawa G. Seasonal, interannual, 503

and long-term variation in sport fish contamination, San Francisco Bay. Science 504

of the Total Environment 2005; 336: 25-43. 505

Greenfield BK, Jahn A. Mercury in San Francisco Bay forage fish. Environmental Pollution 506

2010; 158: 2716-24. 507

Grenier JL, Davis JA. Water quality in South San Francisco Bay, California: current 508

condition and potential issues for the South Bay Salt Pond Restoration Project. 509

Reviews of Environmental Contamination and Toxicology 2010; 206: 115-47. 510

Gunther AJ, Davis JA, Hardin DD, Gold J, Bell D, Crick JR, et al. Long-term 511

bioaccumulation monitoring with transplanted bivalves in the San Francisco 512

Estuary. Marine Pollution Bulletin 1999; 38: 170-81. 513

Heim WA, Coale KH, Stephenson M, Choe K-Y, Gill GA, Foe C. Spatial and habitat-based 514

variations in total and methyl mercury concentrations in surficial sediments in 515

the San Francisco Bay-Delta. Environmental Science and Technology 2007; 41: 516

3501-7. 517

Henery RE, Sommer TR, Goldman CR. Growth and methylmercury accumulation in 518

juvenile Chinook salmon in the Sacramento River and Its Floodplain, the Yolo 519

Bypass. Transactions of the American Fisheries Society 2010; 139: 550-63. 520

Herrin RT, Lathrop RC, Gorski PR, Andren AW. Hypolimnetic methylmercury and its 521

uptake by plankton during fall destratification: a key entry point of mercury into 522

lake food chains. Limnology and Oceanography 1998; 43: 1476-86. 523

Herring G, Ackerman JT, Eagles-Smith CA. Embryo malposition as a potential mechanism 524

for mercury-induced hatching failure in bird eggs. Environmental toxicology and 525

chemistry / SETAC 2010; 29: 1788-94. 526

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

26

Hurley J, Krabbenhoft D, Cleckner L, Olson M, Aiken G, Rawlik P. System controls on the 527

aqueous distribution of mercury in the northern Florida Everglades. 528

Biogeochemistry 1998; 40: 293-311. 529

Kelly CA, Rudd JWM, Bodaly RA, Roulet NP, St. Louis VL, Heyes A, et al. Increases in 530

fluxes of greenhouse gases and methyl mercury following flooding of an 531

experimental reservoir. Environmental Science and Technology 1997; 31: 1334-532

44. 533

Levinton JS, Pochron ST. Temporal and geographic trends in mercury concentrations in 534

muscle tissue in five species of Hudson River, USA, fishes. Environmental 535

Toxicology and Chemistry 2008; 27: 1691-7. 536

Luengen A, Flegal AR. Role of phytoplankton in mercury cycling in the San Francisco Bay 537

estuary. Limnology and Oceanography 2009; 54: 23-40. 538

Miles AK, Ricca MA. Temporal and spatial distributions of sediment mercury at salt pond 539

wetland restoration sites, San Francisco Bay, CA, USA. The Science of the total 540

environment 2010; 408: 1154-65. 541

Mitchell CPJ, Jordan TE, Heyes A, Gilmour CC. Tidal exchange of total mercury and 542

methylmercury between a salt marsh and a Chesapeake Bay sub-estuary. 543

Biogeochemistry 2012; Published online. 544

Monson BA. Trend Reversal of Mercury Concentrations in Piscivorous Fish from 545

Minnesota Lakes: 1982−2006. Environmental Science & Technology 2009; 43: 546

1750-5. 547

Muhaya BBM, Leermakers M, Baeyens W. Total mercury and methylmercury in 548

sediments and in the polychaete Nereis diversicolor at Groot Buitenschoor 549

(Scheldt estuary, Belgium). Water Air and Soil Pollution 1997; 94: 109-23. 550

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

27

Munthe J, Bodaly RA, Branfireun BA, Driscoll CT, Gilmour CC, Harris R, et al. Recovery of 551

mercury-contaminated fisheries. Ambio 2007; 36: 33-44. 552

Prasad RR. Notes on the habitat and habits of Clevelandia ios (Jordan and Gilbert). 553

Proceedings of the National Institute of Sciences of India. Part B, Biological 554

sciences 1958; 24: 314-24. 555

R Development Core Team. R: A language and environment for statistical computing. R 556

Foundation for Statistical Computing, Vienna, Austria, 2012. http://www.R-557

project.org/ 558

Reeder WG. Stomach analysis of a group of shorebirds. Condor 1951; 53: 43-5. 559

Schwarzbach SE, Albertson JD, Thomas CM. Effects of predation, flooding, and 560

contamination on reproductive success of California Clapper Rails (Rallus 561

longirostris obsoletus) in San Francisco Bay. Auk 2006; 123: 45-60. 562

SFBRWQCB. Mercury in San Francisco Bay Total Maximum Daily Load (TMDL) Proposed 563

Basin Plan Amendment and Staff Report for Revised Total Maximum Daily Load 564

(TMDL) and Proposed Mercury Water Quality Objectives. California Regional 565

Water Quality Control Board San Francisco Bay Region, Oakland, CA, 2006, pp. 566

116. http://www.swrcb.ca.gov/rwqcb2/TMDL/sfbaymercurytmdl.htm 567

Slotton DG, Ayers SM, Weyand RD. CBDA biosentinel mercury monitoring program 568

second year draft data report. University of California, Davis, CA, 2007, pp. 92. 569

http://www.sfei.org/cmr/fishmercury/DocumentsPage.htm 570

Slotton DG, Reuter JE, Goldman CR. Mercury uptake patterns of biota in a seasonally 571

anoxic northern California Reservoir. Water Air and Soil Pollution 1995; 80: 841-572

50. 573

South Bay Salt Pond Restoration Project. Pond Breach Dates. 2006. Accessed August 10, 574

2012. http://www.southbayrestoration.org/monitoring/breachdates.html 575

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

28

Stephenson MD, Martin M, Tjeerdema RS. Long-term trends in DDT, polychlorinated 576

biphenyls, and chlordane in California mussels. Archives of Environmental 577

Contamination and Toxicology 1995; 28: 443-50. 578

U.S. EPA. Mercury Study Report To Congress. U.S. Environmental Protection Agency, 579

1997. Available from http://www.epa.gov/mercury/report.htm 580

United States Fish and Wildlife Service and United States Geological Survey. South Bay 581

Salt Pond Restoration Project 2011 Annual Self-Monitoring Report. Prepared for 582

San Francisco Bay Regional Water Quality Control Board, 2012, pp. 169. 583

http://www.southbayrestoration.org/monitoring/ 584

Visintainer TA, Bollens SM, Simenstad C. Community composition and diet of fishes as a 585

function of tidal channel geomorphology. Marine Ecology Progress Series 2006; 586

321: 227-43. 587

Ward SM, Neumann RM. Seasonal variations in concentrations of mercury in axial 588

muscle tissue of largemouth bass. North American Journal of Fisheries 589

Management 1999; 19: 89-96. 590

Watras CJ, Bloom NS, Morrison KA, Claas SA, Gilmour CC, Craig SR. Methylmercury 591

production in the anoxic hypolimnion of a dimictic seepage lake. Water, Air, and 592

Soil Pollution 1995; 80: 735-45. 593

Wiener JG, Bodaly RA, Brown SS, Lucotte M, Newman MC, Porcella DB, et al. Monitoring 594

and evaluating trends in methylmercury accumulation in aquatic biota. In: Harris 595

R, Krabbenhoft DP, Mason R, Murray MW, Reash R, Saltman T, editors. 596

Ecosystem Responses to Mercury Contamination: Indicators of Change. CRC 597

Press, Pensacola, FL, 2007, pp. 87-122 598

Wiener JG, Fitzgerald WF, Watras CJ, Rada RG. Partitioning and bioavailability of 599

mercury in an experimentally acidified Wisconsin lake. Environmental Toxicology 600

and Chemistry 1990; 9: 909-18. 601

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

29

Yee D, McKee LJ, Oram JJ. A regional mass balance of methylmercury in San Francisco 602

Bay, California, USA. Environmental toxicology and chemistry 2011; 30: 88-96. 603

Zhang L, Campbell LM, Johnson TB. Seasonal variation in mercury and food web 604

biomagnification in Lake Ontario, Canada. Environmental Pollution 2012; 161: 605

178-84. 606

607

608

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

30

Table 1. Summary of sampling efforts. 609 610

Frequency Species Stations (Abbreviation) Dates

6 Week Arrow goby Martin Luther King Jr. Regional Shoreline (MLK) Nov, 2007 – Nov, 2010

6 Week Topsmelt Martin Luther King Jr. Regional Shoreline (MLK) Nov, 2007 – Nov, 2010

6 Week Topsmelt Tiburon (TIB) Aug, 2008 – Oct, 2010

6 Week Topsmelt Berkeley (BRK) Oct, 2007 – Dec, 2010

6 Week Topsmelt Keller Beach (KEL) Oct, 2007 – Dec, 2010

Seasonal Mississippi silverside

Alviso Slough, Eden Landing, Artesian Slough, and Benicia State Park (SF Bay)

Oct, 2010 – July, 2011

Seasonal Topsmelt Alviso Slough and Eden Landing (SF Bay) Oct, 2010 – July, 2011

Seasonal Mississippi silverside

Pond 1, Pond 2, Pond 2A, Pond 4/5, Pond 9/10, and Napa River at Kennedy Park (Napa-Sonoma Ponds)

Dec, 2009 – July, 2010

Annual Mississippi silverside

Alviso Slough, Newark Slough, Bair Island, Eden Landing, China Camp, and Benicia State Park (SF Bay)

Oct, 2005 – Nov, 2010

Annual Topsmelt Alviso Slough, Newark Slough, Bair Island, Eden Landing, China Camp, and Benicia State Park (SF Bay)

Nov, 2005 – Oct, 2010

611 612

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49

31

Table 2. Linear model results for each sampling effort. N = number composite samples included in analysis. Model structure includes 613

all significant parameters and interactions included in model. sr2 = squared semipartial correlation. NS = not significant (p > 0.05). 614

615

Sampling effort N Dates a

Stations Model structure Model

R2

Date

sr2

Length

sr2

Station

sr2

Date*Station

sr2

MLK arrow goby 6 week 95 23 1 Date + Length 0.75 0.65 0.02

MLK topsmelt 6 week 101 21 1 Date + Length 0.60 0.57 0.12

TIB topsmelt 6 week 34 9 1 Date 0.93 0.93 NS

BRK topsmelt 6 week 57 15 1 Date + Length 0.89 0.71 0.04

KEL topsmelt 6 week 58 17 1 Date 0.72 0.72 NS

SF Bay silverside seasonal 78 4 4 Date*Station 0.82 0.08 NS b 0.56 0.11

SF Bay topsmelt seasonal 44 4 2 Date + Station 0.77 0.12 NS b 0.65 NS

Napa-Sonoma marsh silverside seasonal

131 4 6 Date*Station + Length 0.83 0.10 0.16 0.18 0.36

SF Bay silverside annual 115 6 6 Date*Station + Length 0.76 0.05 0.05 0.60 0.11

SF Bay topsmelt annual 130 6 6 Date*Station + Length 0.78 0.04 0.15 0.36 0.16

616

a. Only dates with N 2 included. b. 0.05 < p < 0.08617

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

32

Figure captions 618

Figure 1. Study sampling locations. Note: SF Bay Seasonal/Interannual indicates sites 619

that were monitored annually, a subset of which were also monitored seasonally (Table 620

1). 621

Figure 2. Temporal variation in Hg concentrations at six-week collection intervals from 622

late 2007 to 2010. Points are composite sample results and error bars indicate 623

approximate 95% significant differences in means. Note log scale. a. Martin Luther King 624

Jr. Regional Shoreline (MLK) arrow goby, length corrected. b. MLK topsmelt, length 625

corrected. c. Tiburon topsmelt. d. Berkeley topsmelt, length corrected. e. Keller beach 626

topsmelt. 627

Figure 3. Seasonal variation in silverside and topsmelt Hg at four San Francisco Bay 628

monitoring stations. In this and all following plots, points are composite sample results, 629

boxes indicate approximate 95% significant differences in means, and y-axis scales vary 630

among subplots. 631

Figure 4. Seasonal variation in length corrected silverside Hg at Napa-Sonoma ponds. 632

Figure 5. Interannual variation in length corrected silverside Hg at San Francisco Bay 633

stations. 634

Figure 6. Interannual variation in length corrected topsmelt Hg at Bay stations. 635

Figure 1Click here to download high resolution image

Mer

cury

( µ

g g−

1 )

0.02

0.03

0.04

0.05

0.06

0.07

0.080.09

1 O

ct

1 Ja

n 20

08

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

09

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

10

1 A

pr

1 Ju

l

1 O

ct

Figure 2aClick here to download Figure: Fig 2a MLK arrow goby.pdf

Mer

cury

( µ

g g−

1 )

0.03

0.04

0.05

0.06

0.07

0.080.09

0.1

1 O

ct

1 Ja

n 20

08

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

09

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

10

1 A

pr

1 Ju

l

1 O

ct

Figure 2bClick here to download Figure: Fig 2b MLK topsmelt.pdf

Mer

cury

( µ

g g−

1 )

0.04

0.05

0.06

0.07

0.08

0.09

0.11

Oct

1 Ja

n 20

08

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

09

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

10

1 A

pr

1 Ju

l

1 O

ct

Figure 2cClick here to download Figure: Fig 2c TIB topsmelt.pdf

Mer

cury

( µ

g g−

1 )

0.03

0.04

0.05

0.06

0.07

0.08

0.09

1 O

ct

1 Ja

n 20

08

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

09

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

10

1 A

pr

1 Ju

l

1 O

ct

Figure 2dClick here to download Figure: Fig 2d BRK topsmelt.pdf

Mer

cury

( µ

g g−

1 )

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

1 O

ct

1 Ja

n 20

08

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

09

1 A

pr

1 Ju

l

1 O

ct

1 Ja

n 20

10

1 A

pr

1 Ju

l

1 O

ct

Figure 2eClick here to download Figure: Fig 2e KEL topsmelt.pdf

Season

Mer

cury

( µ

g g−

1 )0.

060.

080.

100.

120.

14

Oct 2010 Jan 2011 May 2011 July 2011

Topsmelt, Alviso Sl.

0.04

0.05

0.06

0.07

0.08

Oct 2010 Jan 2011 May 2011 July 2011

Topsmelt, Eden Landing

0.15

0.20

0.25

0.30

0.35

Silverside, Alviso Sl.

0.05

0.10

0.15

Silverside, Eden Landing

0.05

0.10

0.15

0.20

0.25

Silverside, Mallard Sl.

0.04

0.05

0.06

0.07

0.08

Silverside, Benicia S.P.

Season

Mer

cury

( µ

g g−

1 )0.

060.

080.

100.

120.

14

Oct 2010 Jan 2011 May 2011 July 2011

Topsmelt, Alviso Sl.

0.04

0.05

0.06

0.07

0.08

Oct 2010 Jan 2011 May 2011 July 2011

Topsmelt, Eden Landing

0.15

0.20

0.25

0.30

0.35

Silverside, Alviso Sl.

0.05

0.10

0.15

Silverside, Eden Landing

0.05

0.10

0.15

0.20

0.25

Silverside, Mallard Sl.

0.04

0.05

0.06

0.07

0.08

Silverside, Benicia S.P.

Figure 3Click here to download Figure: Fig 3 Seasonal SF.pdf

Season

Mer

cury

( µ

g g−

1 )0.

050.

060.

070.

080.

090.

10

Dec 2009 March 2010 May 2010 July 2010

Napa Marsh Pond 1

0.05

0.10

0.15

0.20

Dec 2009 March 2010 May 2010 July 2010

Napa Marsh Pond 2

0.05

0.06

0.07

0.08

Napa Marsh Pond 2A

0.04

0.05

0.06

0.07

0.08

Napa Marsh Pond 4/5

0.06

0.07

0.08

0.09

0.10

Napa Marsh Pond 9/10

0.06

0.08

0.10

0.12

0.14

Napa River at Kennedy Park

Season

Mer

cury

( µ

g g−

1 )0.

050.

060.

070.

080.

090.

10

Dec 2009 March 2010 May 2010 July 2010

Napa Marsh Pond 1

0.05

0.10

0.15

0.20

Dec 2009 March 2010 May 2010 July 2010

Napa Marsh Pond 2

0.05

0.06

0.07

0.08

Napa Marsh Pond 2A

0.04

0.05

0.06

0.07

0.08

Napa Marsh Pond 4/5

0.06

0.07

0.08

0.09

0.10

Napa Marsh Pond 9/10

0.06

0.08

0.10

0.12

0.14

Napa River at Kennedy Park

Figure 4Click here to download Figure: Fig 4 Seasonal Napa Sonoma marsh.pdf

Year

Mer

cury

( µ

g g−

1 )0.

100.

150.

200.

250.

30

2005 2006 2007 2008 2009 2010

Alviso Sl.

0.05

0.10

0.15

0.20

0.25

0.30

2005 2006 2007 2008 2009 2010

Bair Island

0.05

0.10

0.15

Benicia S.P.

0.05

0.10

0.15

China Camp

0.05

0.10

0.15

0.20

Eden Landing

0.10

0.15

0.20

0.25

Newark Sl.

Year

Mer

cury

( µ

g g−

1 )0.

100.

150.

200.

250.

30

2005 2006 2007 2008 2009 2010

Alviso Sl.

0.05

0.10

0.15

0.20

0.25

0.30

2005 2006 2007 2008 2009 2010

Bair Island

0.05

0.10

0.15

Benicia S.P.

0.05

0.10

0.15

China Camp

0.05

0.10

0.15

0.20

Eden Landing

0.10

0.15

0.20

0.25

Newark Sl.

Figure 5Click here to download Figure: Fig 5 Annual Silverside.pdf

Year

Mer

cury

( µ

g g−

1 )0.

050.

100.

150.

20

2005 2006 2007 2008 2009 2010

Alviso Sl.

0.03

00.

040

0.05

0

2005 2006 2007 2008 2009 2010

Bair Island0.02

00.

030

0.04

0

Benicia S.P.

0.02

50.

035

0.04

5

China Camp0.03

0.05

0.07

Eden Landing

0.03

0.04

0.05

0.06

0.07

Newark Sl.

Year

Mer

cury

( µ

g g−

1 )0.

050.

100.

150.

20

2005 2006 2007 2008 2009 2010

Alviso Sl.

0.03

00.

040

0.05

0

2005 2006 2007 2008 2009 2010

Bair Island0.02

00.

030

0.04

0

Benicia S.P.

0.02

50.

035

0.04

5

China Camp0.03

0.05

0.07

Eden Landing

0.03

0.04

0.05

0.06

0.07

Newark Sl.

Figure 6Click here to download Figure: Fig 6 Annual Topsmelt.pdf

Supplementary Methods DescriptionClick here to download Supplementary Material: Supplemental information.docx