Page History

...

Technical issues related to the ERICA Tool (on Facilia website)

...

Posted: 28/05/2014
From: 
Zoltán Dezső, Hungary

PDFs for CR values in ERICA Tool?

Question
In the documentations concerning the derivation of CR values used in the ERICA Tools (e.g Hosseini et al., 2008; Beresford et al., 2008) it is written that "where a mean and a standard deviation could be determined from the raw data being used to derive the parameter, a lognormal distribution was applied". However, in my opinion, the means and SDs given are not the geometrical ones but the arithmetical ones.
To our knowledge, the statistical description of lognormal distributions requires the use of geometrical means and GSD. For example, in case of uranium CR of trees in terrestrial environment, in the ERICA Tool the following is given: lognormal(0.006794210843500887,0.014144979928509134,0.0,Infinity), based on 521 data entries. In the IAEA Wildlife Transfer Database, for the same 521 data entries the above data are given as AM and ASD as well as a geometrical mean of 0.002929 kg/kg with geometrical SD of 3.654965.
Our questions are:
1. In case of lognormal distributions the GM gives the most probable value of the distribution and it is usually lower than the AM. Does the ERICA Tools calculate/estimate the geometrical mean based on the AM and ASD given? Furthermore, does it apply as the most probable value?
2. Similarly, in case of Tier 3 calculations, does the ERICA Tools calculate/estimate the GSD based on the AM and ASD given? Does it apply during the probability propagation calculations correctly?

Answer:
The mean and SD given in ERICA are arithmetic (not geometric) and these values are used as the inputs for assumed lognormal distributions. There may be additional information in the 'D-ERICA' report - available from https://wiki.ceh.ac.uk/x/swbbBg
Thorne, 2013 (http://dx.doi.org/10.1088/0952-4746/33/2/N1) discusses AM v's GM application in screening assessments.
See also Avila et al. 2014 (J. Radiol. Prot. 34, 261-262) (http://dx.doi.org/10.1088/0952-4746/34/1/L01) for comment on Thorne (added 26/05/2014)

Answered by Nick Beresford (NERC-CEH), 27/05/2014

...

Posted: 28/05/2014
From: 
Mirjana Cujic, Serbia

...