Selenium in North San Francisco Bay: Conceptual Model and Recommendations for Numerical Model Development

Technical Memoranda 4 and 5

Prepared by Tetra Tech

Sujoy Roy, Limin Chen, Bill Mills, and Tom Grieb

Presentation to

Technical Review Committee

September 16, 2008

Objectives

Explain important selenium-related processes, and lay out broad areas of agreement in the scientific literature.

Summarize spatial and temporal trends in selenium data, with a focus on concentrations in bivalves, waterfowl and fish, so that they can be compared against toxicological and health-based guidelines.

Highlight data gaps and uncertainties of relevance to the TMDL.

Guide the development of a numerical model that is proposed to be used to link selenium sources quantified in TM-2 to biota.

Technical Support Documents

TM-2: Source Analysis

TM-4: Conceptual

Model

TM-5: Recommendation

for Numerical Model

TM-6: Model Application and

Results

Overview of Presentation

Summary of processes in • Water column• Sediments• Phytoplankton/bacteria• Fish and birds

Recent data from NSFB

Recommendations for numerical model development

Selenium Cycling in North San Francisco Bay

Water Column Selenium

Selenium (+VI) (selenate): Present in very oxidizing environments, and does not adsorb strongly to particulates.

Selenium (+IV) (selenite): This form of selenium is common in oxygenated estuarine waters and can be taken up microbes and algae more readily than selenate.

Selenium-II (selenide): Selenides can form through the uptake of oxidized selenium by plankton or microbes, where the selenium is biologically reduced and incorporated into organic compounds.

In addition, particulate selenium can exist in one of these forms or as elemental selenium (Se-0).

Relative importance of loadings from different sources Total

(kg/yr)Dissolved

(kg/yr)Particulate

(kg/yr) Uncertainty

Sources:

Atmospheric deposition

17.8-163.5 13.7-78.1 1.4-85.4 High

Local tributaries 354-1511 (354-834 best estimate)

- 118.21 High

Municipal and industrial wastewater

250 - - Low

Refineries 538 - - Low

Input from Delta 1,110-11,752(mean: 3,962)

814-9,736(Mean: 3,354)

151-1,5092

(mean: 698)Moderate

Sacramento River at Freeport

670-2,693 (mean: 1,577) for 1991-2007

Moderate

San Joaquin River at Vernalis

760-7,2705 (mean: 2,972) for 1994-2007

838-4,711 (mean: 2,289) for 1991-2007

Moderate

Sediment 293 18.24 275 Moderate-High

South Bay 106 High

Sinks:

Outflow 4,5003 3,7503 7503 Moderate-High

Sediment Dredging 82.5 82.5 Moderate

San Joaquin River at Vernalis

San Joaquin River at Vernalis

Date

1/1/84 1/1/86 1/1/88 1/1/90 1/1/92 1/1/94 1/1/96 1/1/98 1/1/00 1/1/02 1/1/04 1/1/06 1/1/08

Sel

eni

um ( g

/L)

0.0

2.0

4.0

6.0

8.0

10.0

Total Se (SWAMP) Dissolved Se (Cutter and Cutter, 2004)

Se

in p

art

icu

late

( g

/g)

0.0

0.5

1.0

1.5

2.0

2.5

Salinity0 5 10 15 20 25 30 35

Dis

solv

ed

Se

( g

/L)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Sep 86Nov 97Oct 98Nov 99

Particulate and Dissolved Se: Low Flow (Source: Cutter Research Group papers)

No significant change over 1986-1999 in particulate concentrations

Key Findings-Water and Sediments

Evaluation performed using data from the mid-1980s to 2001, with additional water and sediment data from the RMP

Although variable, field data show minimal change in particulate selenium concentrations patterns (expressed as mg/g) despite changes in dissolved concentrations in the late 1990’s.

The role of particulates in the uptake of selenium implies that the modeling of the freshwater residence time and uptake in the bay is critical to understanding potential risk.

From the sediment coring performed to date, there is limited knowledge of the historical signal of selenium in the bay, especially conditions that were prevalent prior to large scale irrigation in the San Joaquin Valley.

Bioaccumulation

Definition: Bioaccumulation is a general term for the accumulation of substances in an organism or part of an organism.

For selenium, little direct uptake from the dissolved phase for most species, besides algae and bacteria

Uptake occurs through particulates and higher particulate selenium concentrations should result in greater bioaccumulation

In Lab Tests, Algal Uptake Response Non-Linear to Ambient Concentrations

With some exceptions, over a wide range of concentrations, algal concentrations are relatively similar.Baines and Fisher, 2001

Resident Bivalve Data (USGS)

Bivalve Data from Carquinez Strait

Year

1976 1985 1995 1999

Se

in T

issu

e (

g/g

dry

wei

ght)

0

2

4

6

8

10

12

14

16

Mussel

Corbicula

Potamocorbula

Refinery cleanup in 1998

Bivalve Data By Location

Sac R SJR

Zooplankton Data

Year

80 85 90 95 00 05 10

Mus

cle

sele

nium

con

cent

ratio

n (

g/g)

0

10

20

30

40

50

60

Refinery Cleanup

White Sturgeon Muscle Tissue Data

Splittail and Sturgeon Data Compared

Concentrations in Diving Ducks

1980 1990 2000 2010

Se

len

ium

co

nce

ntr

atio

ns

in s

urf

sco

ter

(mu

scle

, g

/g)

0

10

20

30

40

50

60

70

Refinery cleanup

Bioconcentration/Bioaccumulation Factors

Media Concentration Bioconcentration/Bioaccumulation Factors

Water Column 0.10 μg/l (low flow)0.12 μg/l (high flow)

Seston 0.73 μg/g (low flow)0.49 μg/g (high flow)

7,300 L/kg (low flow)4,800 L/kg (high flow)

Plankton 3 μg/g 25,000 – 100,000 L/kg

Sediment 0.25 μg/g 2,200 L/kg

Bivalve P. amurensis 11 μg/g 100,000 L/kg3 – 4 times plankton concentrations

White Sturgeon (liver) 24.1 μg/g 219, 000 L/kg2 - 3 times P. amurensis concentrations

Zooplankton 4.5 μg/g (low flow)1.9 μg/g (high flow)

17, 200 - 40, 900 L/kg0.6 – 1.5 times plankton concentrations

Splittail (liver) 11.4 μg/g 103,600 L/kg2 - 6 times zooplankton concentrations

Key Findings-Selenium in Biota

In addition to the load changes, there have been recent changes in the ecology of the bay, primarily an increase in phytoplankton productivity to levels seen prior to the P. amurensis invasion.

There is no information now on what the P. amurensis populations are, and whether a decline in its numbers may reduce the levels of bioaccumulation in the future.

Laboratory data show the significant effect of algal species on selenium uptake, although this information is not collected in the field.

Although 1999-2005 particulate, bivalve, and white sturgeon selenium data do not show decreases from prior periods, the very small amount of bird muscle data does indicate a decrease

Dissolved Particulate Prey Items Predators

Se

len

ium

in D

iffe

ren

t C

om

pa

rtm

en

tsK

ey

Fa

cto

rs f

or

TM

DL

A

nal

ysis

Speciation data needed for accurate understanding of uptake.

Speciation data shows a decline in selenite concentrations after refinery cleanup in mid-1998.

No speciation data available after 1999.

The bioavailability of different particulate forms varies, with algal selenium being more bioavailable than other forms.

Particulate selenium data available for select years, most recently for 1999.

In recent years algal concentrations in NSFB have increased.

Bivalve uptake of selenium is primarily through particulates.

Different bivalve species have different uptake rates.

Bivalve data are not available after 1999. Prior data show stable concentrations in the late 1990s.

Limited recent data on predator species.

Most data show similar results over the late 1990s, although a small amount of diving duck muscle tisse data suggest a small decrease.

Uncertainty in Predicting Bioaccumulation

Selenate, Selenite, and Organic Selenide

Bacteria, Phytoplankton

Mineral Particles

Organic Detritus

Water Column Species (e.g., zooplankton)

Benthic Species (e.g., bivalves)

Splittail, Striped bass

White Sturgeon

Diving Ducks (Greater and Lesser Scaup,

Surf Scoter)

Numerical Model Development

Why needed? An approach to link point and non-point sources to numeric targets adopted in the TMDL

Numeric targets could be in the form of water column and/or tissue concentrations and have not yet been proposed for the NSFB TMDL

Specific scenarios to be modeled will be developed later in the TMDL process

Modeling Approaches (1): Presser and Luoma (2006) Title: Forecasting Selenium Discharges to the San

Francisco Bay-Delta Estuary: Ecological Effects of a Proposed San Luis Drain Extension

All loads enter at the head of the estuary, concentration gradient from head of the estuary to Golden Gate a function of salinity

Bioaccumulation in different trophic levels using DYMBAM and regressions from data

Calibrated to field data from the bay

Scenarios tested include baseline and different levels of loads from San Joaquin Valley, different runoff scenarios; and different bioavailability and assimilative efficiency

Modeling Approaches (2): Meseck and Cutter (2006)

Title: Evaluating the Biogeochemical Cycle of Selenium in San Francisco Bay through Modeling

Focused on North bay; uses a 1-D model from Rio Vista on Sacramento River to the Golden Gate. Developed using ECoS modeling tool

Representation of selenium biogeochemistry and conversion between different forms, calibrated to Cutter group data collected in the 1980s and 1990s

No representation of bioaccumulation processes

Recommendation in TM-5

Use existing peer-reviewed modeling information to the extent feasible

Employ the Meseck and Cutter (2006) approach for selenium transport, and the Presser and Luoma (2006) approach for bioaccumulation

Calibrate to detailed water chemistry and biological data from ~1999

Validate for water chemistry to RMP data in 2001 and 2005

Predictions for other years with a given hydrology