Mercury Contamination in Central Valley Sport Fish: Highlights from 2006

1
Mercury Contamination in Central Valley Sport
Fish Highlights from 2006

• Aroon R. Melwani, S.N. Bezalel, J.A. Hunt, J.L.
  Grenier,
• J.A. Davis (SFEI)
• Gary Ichikawa and Billy Jakl (CDFG)
• Wes Heim and Autumn Bonnema (MLML)
• Margy Gassel (OEHHA)

Presentation at the Fish Mercury Project Annual
Meeting, Sacramento, CA. June 5th, 2007
2
Outline of Presentation

• Goals of FMP sport fish monitoring in 2006
• Brief review of 2005 sport fish mercury data
• FMP Sampling in 2006
• Data Analysis to relate concentrations between
  species, and to predict mercury normalized for
  size
• Results from 2006 sampling
• characterizing mercury in sport fish
• spatial trends in mercury

3
Goals of FMP Sport fish Monitoring

• Goal 1 Characterize mercury to assess health
  risks of consuming contaminated fish
• Goal 2 Characterize spatial and temporal trends
  in mercury concentration

4
FMP 2005

• High Hg on Cosumnes, Feather, and American
• Increase in mercury closer to the Delta on San
  Joaquin R.
• Mixed message on Sacramento R.
• Low concentrations in the Central Delta

5
Sport fish Sampling in 2006

• 2006 was the second year of sampling
• 1400 fish, 30 species, 47 sites
• Coordinated sampling with the CVRWQCB

6

• 2006 Locations
• Sacramento River watershed
• 45 of 47 sites were for advisory development
• 2 restoration sites targeted
• No index or intensive sites were sampled for
  sport fish

7
Field Sampling

• May to early December 2006
• California Department of Fish and Game
• Electrofisher boat and fyke nets
• Clean handling techniques
• Length measurements

8
Mercury Analysis at MLML

• Total mercury
• Fillet (muscle tissue)
• Wet weight
• Individual analyses (not composites)
• Passed preliminary QA (van Buuren 2006)
• Including lab inter-comparison study
• (3 of samples n 40)

9
Concentration Categories

• Mercury concentrations are presented in four
  categories

• Allows for consistent comparison between species

10
Length mercury Relationships

• Examined relationship using linear model when
  data were appropriate
• For other species, size limits applied, based on
  peak in length-frequency distribution
• Narrow size limits according to EPA guidance
• Size of smallest fish should be no less than 75
  of the largest

11
General Linear Mixed Model

• Advantages over ANCOVA model comparison,
  increase inference space
• Evaluation of model types was recommended by the
  Peer Review Panel after the first year of the
  Project
• Site random effect
• Region fixed effect
• Length and length2 covariates
• First order interaction terms (e.g., sitelength)
• sites assigned to one of ten regions in
  watershed (e.g., San Joaquin River, Sacramento
  River, Central Delta, etc)

12
Model Application

• Largemouth bass and channel catfish data were
  appropriate
• Compared a suite of a priori models
• Evaluate best model using Akaiki Information
  Criterion (AIC)
• Selected model used to calculate mean
  concentration at a standard length

13
Largemouth Bass
size limits
14
Channel Catfish
size limits
15
Common Carp
16
Redear Sunfish
size limits
17
Largemouth Bass is most frequently the highest
species

• More green and yellow sites in the higher
  watershed
• More orange and red sites close to Delta

18
Redear Sunfish and Rainbow Trout are most
frequently the lowest species

• All sites higher in the watershed are green
• Higher proportion of yellow sites closer to the
  Delta

19

• Largemouth
• Bass
• Widespread in the watershed
• Higher mercury in north Delta and lower Sac.
• Lower concentrations on higher Sacramento R.

20

• Channel Catfish
• Similar spatial pattern to bass
• Catfish less widely distributed
• High concentration on Cosumnes R.
• Lower concentrations in northern portion of the
  watershed

21
Common Carp

• High concentration on eastern and western
  drainages
• Higher close to Delta
• Low concentration on mainstem of Sacramento R.

22
Sacramento Pikeminnow

• No clear pattern
• Limited distribution
• Concentrations of
• 0.4 ppm in both the lower and upper watershed
• All sites 0.1 0.5 ppm

23

• Redear Sunfish
• Much lower concentrations compared to other
  species
• Spatial difference in mercury more subtle for
  redear sunfish

24
Model Results

• FINAL MODEL included length2mercury (slope)
• LMB significant effect of site and region
• CCAT significant effect of region
• After 3rd year of sampling, the larger dataset
  will provide most rigorous examination of models
  for these species
• Complex model required to detect spatial patterns
  
• Significant variables may included region, site,
  length, and species

25
N
S
Largemouth Bass (2006 sites) Lengthmercury by
site
26

• Largemouth Bass
• Spatial Patterns
• Mercury mean CI estimated for standard length
  of 350 mm
• American Cosumnes gtgt others regions
• Central Delta ltlt Sac

27

• Channel Catfish
• Spatial Patterns
• Mercury mean CI estimated for standard length
  of 425 mm
• Western drainages
• ltlt Eastern drainages
• and Sacramento R.

28
Common Carp Spatial Pattern
29
Redear sunfish Spatial Patterns

• Mercury mean CI
• Using length limits

• Size limits applied, ngt5
• Sites from N to S on Y-axis

30
FMP Progress Summary

• Goal 1 Characterize mercury concentrations in
  fish
• Mercury contamination of sport fish was
  significant
• Concentrations varied by species and over space
• Average concentrations were most often in the
  moderate yellow category (0.1 0.5 ppm)
• Largemouth bass and channel catfish were high in
  mercury
• Redear sunfish and rainbow trout were low in
  mercury

31
FMP Progress Summary

• Goal 2 Characterize spatial patterns
• Mercury higher in the lower Sacramento and North
  Delta lower concentrations in the upper
  Sacramento
• Pattern stronger in large piscivorous fish
• Tributary loading of MeHg on lower Sacramento
  River from historic sources (i.e., Sierra Nevada,
  Coast Range)
• 3-fold increase in MeHg water concentrations
  significantly related to 350-mm largemouth bass
  (Chris Foes data)

32
FMP Progress Summary

• Goal 2 Characterize temporal patterns
• Unable to do so with most locations sampled in
  2005 not having been sampled in 2006
• Few overlapping largemouth bass sites sampled in
  2006 were somewhat higher than 2005
• Rigorous treatment of inter-annual variation with
  3 years of data (next year)

33
(No Transcript)