MODARIA WG4: Analysis of radioecological data in IAEA TRS to identify key radionuclides and associated parameter values for human and wildlife assessment
IAEA Parameter value compilations WTD: http://www.wildlifetransferdatabase.org
Objectives Using the recent data compilations: • To identify the most important radionuclides, pathways and parameter values – For different source terms – For different exposure situations • Identify data gaps which matter • Provide guidance on need for further data for different source terms • Explore ways to make TRS 472 data more usable by modelling community Consider both human and wildlife
Current focus • Parameter values for ingestion doses Soils, sediments, animals, plants • • CR for human foodchain • CR wo-media for wildlife • Kd values for both • Soil • Sediment – Freshwater – Marine?
Prioritising Data gaps
Wildlife: eg. terrestrial Radionuclide Grasses Lichens & Shrub Annelid Tree Mammal Mollusc Arthopod Bird Reptile Amphibian Arachnid & Herbs Bryophytes Cs Pb Am Sr Cd Pu Ni U Po Ru Mn Th Cl Co Se Sb Ce Eu I Tc Ag Cm Zr Nb Np P S Te n≤ 10 n>10<20 n>20<100 n≥ 100
Human foodchain • Animal products Sheep Element Beef Goat meat Pork Poultry Egg Cow milk Goat milk Sheep milk meat Ag 1 Am 1 1 1 1 2 Ba 2 1 2 1 15 3 1 Be 1 Ca 3 2 1 15 12 St Cd 8 1 2 8 1 1 Ce 1 1 6 1 Cl 1 Co 4 2 2 2 4 1 2 Cr 3 2 1 Cs 58 41 11 22 13 11 288 28 28 Fe 4 1 2 7 St St I 5 1 2 3 4 104 24 7 La 3 Mn 2 1 1 2 3 4 St 1 Mo 1 1 3 7 4 Na 2 1 1 2 7 St 1 Nb 1 1 1 1 1 1 Ni 2 2 1 Np 1 P 1 1 1 St St St Pb 5 2 15 St 1 1 4 2 Pu 5 2 2 n/a 1 Ra 1 11 Ru 3 2 1 1 6 S 3 1 12 St Sb 2 3 Se 1 4 4 12 2 Sr 35 25 8 12 7 9 154 21 4 Te 1 1 1 1 11 1 1 Th 6 3 U 3 2 2 2 3 1 W 7 Y 1 1 Zn 6 6 2 3 4 8 St St Zr 1 1 1 1 6 1
ICRP RAPs CR values Based on data
Approach • Develop a set of criteria to evaluate importance of parameter values Source terms • Magnitude and relative importance of internal dose • Impact of environmental factors on internal dose • estimates related to each parameter value • Analyse data quantity and quality TRS wildlife / some TRS 472 data in spreadsheets • Derived values •
SRS 19 update – parameter value analysis No data for TRS Milk 1.0E+00 Needs work Criteria 1.0E-01 to • Half life Dose (Sv per Bq m-3) understand 1.0E-02 • Magnitude of dose calculations 1.0E-03 • Proportion of dose and • TRS value? 1.0E-04 assumptions 1.0E-05 in SRS update 1.0E-06 and develop Th-229 Th-230 Rb-87 K-40 Re-186m Rb-86 Rb-84 Rb-83 Tc-98 Tl-204 Tc-99 Re-184m Ge-68 Re-184 As-74 Tc-95m Tc-97m Th-234 As-73 Tc-97 Tl-202 Ag-111 Re-187 Ge-71 criteria 1.0E+01 Values included in TRS Milk 1.0E+00 Dose (Sv per Bq m-3) 1.0E-01 1.0E-02 1.0E-03 1.0E-04 1.0E-05
Revision and application of Kd values Aim To harmonise the approach used to develop and extend the range of PDFs available To provide PDFs in a format applicable for modelling for soils and freshwater Collate new data and integrate into current databases Select appropriate statistics to analyse data to develop the PDFs
Soil and freshwater K d databases Starting point: databases created in the frame of TRS 472 Both datasets reported as AM, SD, GM, GSD and min- max values. Probability/cumulative density functions (PDF/CDF) were derived for the freshwater K d dataset. Probability density functions could also be calculated for the soil K d dataset, as it might be a better approach than using best estimate values.
Soil K d database Soil K d database created in the frame of TRS 472 Excluded data for other materials ( e.g ., sediments; pure soil phases such as clays or Fe-Mn-Al oxides) Around 2900 records for 67 elements, Cs and Sr most values, a few elements with more than 100 entries each ( e.g. , I, U, Co, Sb and Se). Data grouped according to texture/OM and, when possible, soil properties governing their interaction ( cofactors ). Lognormal distribution was assumed (although not tested): GM value defined the K d best estimate (for n > 3), no probability density functions were derived Main aim of this topic within WG4
Update of soil K d database Initially not foreseen, but… Aims of the dataset have (partially) changed. Density functions require large number of entries (n > 10). TRS 472 database: gaps have been identified (either absence of information or low number of entries). More than 50 new documents already critically reviewed and additional data sources (yet) to be checked.
Initial conclusions on soil K d database update Significant increase in the number of elements/records with respect to TRS 472 (especially for RNs such as Am, Eu, Ni, Cs, Sr, U and Co): 4500 entries for 75 elements. Limited number of records for priority radionuclides ( e.g., Ag, lanthanides, actinides). Reconsideration of criteria to accept data: use of types of analogues . • Stable (+ natural/indigenous) isotopes • Actinides / lanthanides • Clays / sediments
Construction of CDFs for K d database Soil K d data fitted to a distribution function, using whole and partial datasets (according to texture/OM; scores; cofactors) and with/without scoring. Use of scores To weight data to obtain CDFs with a lower variability and more representative than those without scoring. Challenge: to agree scores for a minimum number of partial datasets, so the related number of observations is still significant, and valid to be used for any radionuclide Different scores for sorption, desorption and indigenous data
Impact of source data type For Cs n GM GSD Min Max GM GSD 5 Indigenous 38 21878 13 73 650000 24210 19 186 Sorption 335 1023 6.6 4 43445 1517 4.5 129 Desorption 225 692 8.1 26 375000 594 10.3 13
INITIAL CONCLUSIONS ON USE OF PDF/CDF Preliminary lognormal CDFs for Cs, Sr and Am. Soil K d best estimates should not be derived from AM, from GM of dataset or from 50 th percentile of the fitted CDF . 5 th -95 th percentiles reduces data variability (identify outliers ). Use of scores : • introduces expert (personal) judgment in data treatment • Relevance depends on the size and quality of the dataset Challenges: • How to proceed with small datasets (n < 10??) or when statistical descriptors are not available? • Analogues to fill data gaps!
ACTION LIST: SOIL To continue the update of the K d soil database (to be finished Nov 2014!) To start testing the use of analogues in large datasets use of sediments data for clay soils? To explore CDFs sensitivity to scores for small datasets
Freshwater Kd : Suspended vs surface sediments Kd for suspended particulate matter (Kd SPM ) and Kd for surface oxygenated sediments (Kd SOS ) can be linked by considering granulometric size corrections SPM particle C C • For same particles: WaterColum n SOS Kd SPM dissolved dissolved C C WaterColum n SOS • Sediments Particles Sizes > Suspended Particles Sizes: Kd Kd SOS SPM
CDF for Freshwater Kd for SPM and surface sediments From P.Ciffroy Data Base For Cesium 0,25
WHAT IT CAN BE DONE FOR SINGLE VALUE OR SMALL DATASETS (N<3 )? 1. No PDF 2. Exponential distribution (ERICA) 3. Log normal distributions: GM given by the dataset and GSD conservatively estimated from knowledge of GM and GSD couples ( ). Two possibilities 1. The most conservative GSD = GSD max 8.34 2. GSD = F(GM) F(GM) = envelop 𝑯𝑵; 𝑯𝑻𝑬 (Green curve)
Next steps • Define a common structure for freshwater and soils Kds including Kd values and relevant parameters. • Update the references with the new data published or obtained since 2007. • Collate referencees from MODARIA community into the data base following a common template • contributors will be acknowledged • Allowing free access to summarised data.
WG interaction • Queries on parameter values – Need logging systems – Q and A available – requests for pdf • Sustainability of databases – Databases with common QC – Summary data available and regularly updated
Future Plans: Interim meeting – PROPOSED WORKSHOP ON Kd in collaboration with STAR – May week 21 or 22, Location ? – Invited speakers – To discuss improvement and enhancement of Kd databases for soils and freshwater – What to do with small datasets (n < 10??) or when statistical descriptors are not available? – Analogues to fill data gaps! – Sustainability and visibility
Recommend
More recommend
