nuclearsafety.gc.ca nuclearsafety.gc.ca 1 Thompson et al. (2005) - PowerPoint PPT Presentation

Canadian Nuclear Commission canadienne Safety Commission de sûreté nucléaire Canadian Benthic Data Set Canadian Benthic Data Set S teve Mihok S teve Mihok EMRAS II Meeting, Vienna, Austria January 27, 2010 EMRAS II Meeting, Vienna, Austria January 27, 2010 nuclearsafety.gc.ca nuclearsafety.gc.ca 1

Thompson et al. (2005) Thompson et al. (2005) 12 contaminants, As, Cr, Cu, Mo, Ni, Pb, Se, U, V, Pb-210, Po-210, Ra-226 (N=1,020 - 2,269) Uranium mining regions with co-located benthos sampling & organic depositional sediments 132 Ontario & Saskatchewan sites 190 genera and/or species Criteria/methods follow Persaud et al. (1992) as used for Ontario LEL / SEL guidelines (As, Cr, Cu, Ni, Pb) 90 th percentile SSLC for each taxon 5 th percentile LEL, 95 th percentile SEL calculated No curve fitting, no dose calculations EMRAS II Meeting, Vienna, Austria January 2010

Original Methods Original Methods “Weighted” percentile as in Persaud et al. (1992), also calculated “closest observation” percentile (SPSS) Weighted value typically higher Uranium LEL 104 vs 32, SEL 5,874 vs 3,410 ug/g Six data selection criteria Minimum of 10 sites per taxon, lost considerable data Concentration range 2 orders of magnitude (V, Cr x) Spatial range (35 reference sites, 97 contaminated) Mainly benthic species (81% defined as infaunal) Minimum of 20 SSLCs for LEL/SEL calculation Data mean ~30 for SSLCs, LEL/SEL range N=28-59 EMRAS II Meeting, Vienna, Austria January 2010

Interpretation Interpretation LEL/SEL values highly sensitive to exact percentile calculation method, large effect on Uranium values “Good” predictions of “no impacts” with weighted LELs, “Poor” predictions for many SELs (21 sites) Exploratory data analysis with multiple contaminant indices or data subsets was not informative (As & Ni associated with uranium deposits in Saskatchewan) Radionuclide thresholds differ from ERICA model predictions of potential impacts, Why? Hardly any data on contaminant concentrations in benthos vs sediments, just a few taxa sampled EMRAS II Meeting, Vienna, Austria January 2010

Benthos - Radiation vs Metals? ? Benthos - Radiation vs Metals? ? % of Community Control Umine site ? Sediments LEL range Benthos equivalent at Bq / g dw Thompson et al. 2005 10 µGy/h (ERICA) Ra-226 0.1 – 0.6 0.6 Pb-210 0.5 - 0.9 80 Po-210 0.6 – 0.8 600 EMRAS II Meeting, Vienna, Austria January 2010

Example of Uranium SSLC’s Example of Uranium SSLC’s URANIUM URANIUM Ont+Sask + Bancroft N=56 Taxa Intuitive Weighted LEL (ug/g) 32.0 104.4 SEL (ug/g) 3410.0 5874.1 Mean SSLC (ug/g) 693.5 1064.9 SD SSLC 1319.8 1683.2 Median sample size for SSLC 21 Mean sample size for SSLC 32 EMRAS II Meeting, Vienna, Austria January 2010

Discussion: Dose Calculations Discussion: Dose Calculations Calculations, % Dose Data Issues Notes Approach I – E % U-238 Nat uran 49% 11 - 0 96% Abundance DCF Th-234 =U-238? 0 - 26 X = Parent? External U-234 Nat uran 49% 12 - 0 96% Abundance DCF Th-230 = Ra-226? 10 - 0 X = Daughter? Ra-226 Measured 20 -1 99% Rn-222 @30%? Ra 24 - 59 X Retention % CRITICAL Pb-210 = % Ra-226 0 - 1 74% Site-specific? Bi-210 = Pb-210 0 -13 X Po-210 = % Ra-226 23 - 0 70% Site-specific? CRITICAL U-235 (?) Nat uran 2% n/a 96% Abundance DCF I-E example Greer Lake 2004 EMRAS II Meeting, Vienna, Austria January 2010

Greer Lake 2004 “PSL2” Chironomid Greer Lake 2004 “PSL2” Chironomid Measured U-238 / Th-230 at equal activity Ra, Pb, Po at 2x U / Th values, Radon estimated @30% Alpha RBE 40, Amiro (1997) DCFs [Tailings present in lake] Rn-222 = Pb-210 30% Ra Rn-222 = Bi-210 (=Pb- 30% Ra 210) Po-210 Pb-210 (=Pb-210) Ra-226 Bi-210 (=Pb- 210) Po-210 (=Pb-210) Ra-226 U-238 Th-230 U-238 Th-230 Th-234 (=U- U-234 Th-234 (=U- U-234 238) 238) Internal Dose 2,538 uGy/h External Dose 36 uGy/h ? EMRAS II Meeting, Vienna, Austria January 2010

Reference Lakes - Not in equilibrium Reference Lakes - Not in equilibrium Schmoo Lake 2002 Fulton Lake 2004 Po/Pb = 2x U-238, 10x Th-230, Po/Pb = 4x U-238 / Th-230 6x Ra-226 2x Ra-226 Bi-210 (=Pb- 210) Bi-210 (=Pb- Rn-222 = Po-210 (=Pb- 210) Pb-210 Pb-210 Po-210 30% Ra 210) (=Pb-210) Rn-222 = Ra-226 30% Ra Th-230 Ra-226 U-238 U-234 Th-234 (=U- U-238 Th-230 Th-234 (=U- 238) 238) U-234 Internal Dose 12.1 uGy/h Internal Dose 2.6 uGy/h (alpha RBE of 40) EMRAS II Meeting, Vienna, Austria January 2010

Decisions (May need a Radon DCF? ) Decisions (May need a Radon DCF? ) PSL2 Internal Calc ERICA Tool – Internal + External? RBE = 10, DCFs for insect larvae Alpha RBE 40, Amiro DCFs, generic benthos & bivalve mollusc ~ same? U-238 = Natural Uranium SA? 49% nat uranium (DCFs vary ±) Th-234 Not included = daughter or parent - Decision? U-234 Included in U-238 49% nat uranium Th-230 = daughter = daughter or parent - Decision? Ra-226 = Measured = Measured or parent Rn-222 @30% Ra (~ vert bone) ? % retained ? Decision critical Pb-210 Measured or = radon ~30% missing, calculated @~2x? Bi-210 = Pb-210 = Pb-210, not critical to dose? Po-210 Measured or = radon ~30% missing, calculated @~2x? U-235 Included in U-238 2% nat uranium EMRAS II Meeting, Vienna, Austria January 2010

Discussion: Improving SSLCs ? Discussion: Improving SSLCs ? Curve-fitting to obtain SSLCs Chironomus - Arsenic (N=97) Even this large Frequency Cumulative % data set is not 20 120% a good fit to log 18 100% 16 distribution? 14 80% Frequency 12 90 th percentile 10 60% is in the flat 8 40% part of CDF 6 and may 4 20% 2 remain as 0 0% poorly defined? 7 3 1 5 9 3 7 1 5 9 e . . . . . . . . . . r 0 0 0 0 0 1 1 2 2 2 o - - M log Sediment concentration EMRAS II Meeting, Vienna, Austria January 2010

Statistical Interpretation Statistical Interpretation Curve fitting to obtain LEL/SEL thresholds (5%) will also be affected by small sample sizes Bootstrapping for confidence intervals, what to do about outliers (e.g. very high Uranium in Link Lakes) Sensitivity Analysis (sites, SSLCs, taxonomic groups) Multivariate analyses for interactions among metals and radionuclides, “fingerprints” for severe effects? Prospects for more data Objectives to be defined Path forward on data / analysis … J. Garnier LaPlace EMRAS II Meeting, Vienna, Austria January 2010

A question from WG 4 on Tuesday A question from WG 4 on Tuesday The Biota modeling group will be estimating dose to fish and benthic invertebrates at Beaverlodge for the next mid-term EMRAS meeting. Is there any interest in modeling population effects, e.g. for lake whitefish and/or fingernail clams ( Pisidium spp.)? EMRAS II Meeting, Vienna, Austria January 2010