1. A Methylmercury Budget for San Francisco Bay Donald Yee, San Francisco Estuary Institute

2. Mercury Conceptual Model System is complicated, simplified by single box model Slow response (decades) MeHg matters most (to biota)

3. Methylmercury Conceptual Model Need to track MeHg MeHg <1% of totHg Poor MeHg:totHg correlation Differences from Hg 1 Box Model Methylation & demethylation Potentially rapid (days- months) Sed-water exchange Meth Demeth

4. WWMMBD? What Would the MeHg Mass Budget Do? Synthesize- do Bay data make sense given… –Loading, production, degradation, sed-water exchange, and other processes? Quantitative conceptual model of MeHg –ID key factors for MeHg fate Feasibility/needs of refined model(s) –E.g. temporal & spatial detail What it won’t/can’t do –Identify “hot” spot impacts (1 box) –Predict long term fate (no Hg linkage)

5. MeHg 1 Box Model Adapted from PCB 1 box model –One water compartment –One sediment compartment (10cm mixed layer) –Daily time step –Annually uniform (no seasonality) –Constant uniform mixing –Equilibrium partitioning Simplifications worked for PCBs, PBDEs

6. External Loads (Imports) +Direct atmospheric (wet) deposition 0.1 g/d Area x literature rain MeHg x local rainfall +Delta (Mallard Island) discharge 9.8 g/d Flow x concentration (Region 5 MeHg TMDL) +Local watersheds 4.9 g/d RMP measured watersheds (extrapolated) +Wetlands (upper range estimate) 2.0 g/d Volume x (incoming - outgoing) concentrations +POTWs (16 largest, ~95% discharge) 0.8 g/d Flow x concentration = 17.6 g/d total

7. Internal Load- MeHg Production Function of multiple factors- –Would need complex C & S & Hg model Next best- lab incubation production rates? –Marvin-DiPasquale et al anaerobic incubations Assume portion of sediment layer methylates –Methylating zone in fraction (30%) of sediment

8. Loss Processes Bio-uptake = “export” from Bay 0.13 g/d –Small fish biomass (CDFG) x concentration (RMP) 1-Box Model Losses Volatilization –Air/water partitioning (Lindqvist & Rodhe 1985) Outflow (through Golden Gate) –Tidal mixing (Connolly), assume ocean MeHg ~0 Burial –Fuller sedimentation 0.88cm/yr (~9% of mixed layer)

9. Modeled Processes Degradation –Sediment: Marvin-DiPasquale demethylation rates = 0.083/d (decay) Assume demethylating zone (70% of mixed layer) –Water: Krabbenhoft Petaluma water half life~7 days (0.10/d decay) Benthic flux –In daily resuspension & de/sorption Large uncertainties some parameters –Some have small ~no effect

10. Base Case Run Base case = averaged –initial concentrations (from RMP monitoring) –loading/process parameter values Quick steady state, within ~5% of T 0 –Sediment mass ~ –Water mass lower

11. Base Case Run Mass (inventory) vs daily flux/degrade/produce Water Mass –Net sediment to water exchange, ext load = Degradation>, GG outflow, >> bio-uptake,volatilization Total (Water+Sediment) –Production ~balances degradation >> all other processes * Flux box measurement similar: ~.014 kg/d (Choe et al) Mass in Water0.236kg Ext. Load0.018kg/d Sed to Water*0.021kg/d Water Degrade0.024kg/d GG Outflow0.014kg/d Bio-uptake<0.001kg/d Volatilize<0.001kg/d Mass in Sediment30.8kg Methylate1.82kg/d Sed Degrade1.79kg/d Sed to Water0.021kg/d Burial0.007kg/d

12. Hot &@%$! Model Responds Fast!? Seasonal de/meth rates (winter -30%) ~month response! Yes, but… –Model oversimplifies (mixing, equilibrium) –Processes vary on microscale (e.g. de/methylation) –Still a good order of magnitude tool

13. Parameter Sensitivity ScenarioMass SMass W Base Case30.8 kg0.236 kg Load /330.70.191 Load x331.00.370 WaterDegrade /330.90.317 WaterDegrade x330.60.134 SedDegrade /388.80.556 SedDegrade x310.40.123 Methylate /310.30.123 Methylate x392.00.574

14. WDMMBD? What Did the MeHg Mass Budget Do? Did Bay data make sense? –Base case near starting state- near “right” Baywide? –Non-unique solution (e.g. offsetting errors?) Feasibility/needs of refined model(s) –1 box driven by steady state/equilibrium –Basis for more detailed model? Much higher data needs Key factors affecting MeHg fate –External loads have small/medium effect –Very sensitive to de/methylation rates

15. Management Strategy – Dr. Evil Acquire $1 Million Option A- Control Methylation: Sterilize the Bay (thermonuclear device) Option B- Control Demethylation: Equip sharks w/ UV lasers to photodemethylate

16. Management Strategy -RMP Option C- RMP Mercury Strategy: –Where biota affected (food web entry) –ID disproportionate (high leverage) pathways –ID intervention opportunities –IF strategy finds locations where critical pathways (e.g. de/methylation) may be acted on THEN act (e.g.holding ponds, aeration, dredging, nutrient reductions, etc) –Monitor & model management effectiveness “adaptive management” (Unfortunately likely > $1 million)

17. Acknowledgements Too many to list… “If I have seen further it is by standing on ye shoulders of Giants” – Sir Isaac Newton

19. Atmospheric (Wet) Deposition No local data –RMP MDN station only measured totHg Literature rainfall MeHg (avg 0.11 ng/L) … –Watras & Bloom (1989 Olympic Penins. WA 0.15ng/L) –Risch et al (2001-2003 Indiana, 0.06ng/L) –St Louis et al (1995, ELA area, 0.05ng/L) –Mason et al (1997, Still Pond, MD, HgT x %MeHg avg = 0.04ng/L) x Local annual precipitation (0.45m/y) = 0.10 g/d deposition Baywide

20. Discharges from… Delta (SWRCB Region 5) –Flow weighted avg concentration x mean annual discharge = 4.7g/d in Hg TMDL –Revised to w/ later data Local watersheds –Extrapolate w/ SIMPLE Model (modeling mine + urban + non-urban areas) Local MeHg data, extrapolated to Bay area (3.6 g/d) Local Hg data x MeHg%, extrapolated to Bay area (6.2 g/d) –Use average of above 4.9g/d

21. Discharges from… Wetlands –Wetland Goals est. 40k acres wetland (1.6e8 m 2 ), assume 0.3m overlying water every day –Petaluma marsh extrapolation ~50% water particulate settles -1.2g/d ebb tide dissolved conc ~2.5x flood tide (max 5x at Petaluma) +3.2g/d = net 2g/d load to Bay –USACE Hamilton AAF leaching assumptions 0.8%/d of net production = 4.0g/d load –Stephenson et al showed net import and export different events for Suisun Marsh May be difficult to refine net load

22. Discharges from… POTWs –Annual mean conc x discharge for 16 largest plants (loads for each plant calculated then summed) = 0.79g/d Conc range 0.04-1.3ng/L (mean ~0.42ng/L) Discharge 14-165e9 L/y (sum ~2.15e9g/d ~95% of discharge volume)

23. Bio-uptake “Loss” Phytoplankton? –Cloern 2002-2004 productivity ~210gC/m 2 y –Hammerschmidt MeHg 0.5ng/g ww =5ng/g dw –LakeMichMassBal algal MeHg = 30 ppb dw –C → CH 2 O, geomean MeHg 12ng/g –= 19.5g/d MeHg into phytoplankton? Phytoplankton rapid turnover (growth~0.3/d?), reversible “loss” from water/sed pools, loss estimate probably too high Small fish? –Slater (CDFG, IEP) young of year pelagic fish est. 0.01-0.25g/m 3 (Suisun lowest, Central highest, mostly anchovies) mean ~0.17g/m 3 ww biomass –RMP anchovy Hg 0.049µg/g ww = 0.13g/day MeHg into fish biomass (<1% of phyto?) –Expect less (short term) cycling than algae, “irreversible” net loss by incorporation into higher trophic levels