Proyect Number WW sy Battelle 409826 Pacific Northwest Laboratories To R R. 0. Gilbert Subject 1 APR2S “wae Response to Request for Evaluation of Dose Estimates and Future Actions Concerning 10d NIVINOO LON SAOd LNaNNOOd Aq Pamaraoy niweto Clea nup Eniwetok Oe a ind ’ W. J. Bair From <b ep "0Bere 4 [ File/LB April 25, 1978 Date TO Internal Distribution > Tigi” Question 1: Do the recent LLL dose calculations in the draft paper "Assessment of Potential Doses to Populations from the Transuranic Radionuclides of Enitetok Atoll" suggest the current "minimal action level” (40 pCi/g soil) and the mandatory cleanup level (400 pCi/g soil) should be reduced? As you know Bill, I was not a part of the review last August of the Environmental Impact Statement (EIS) for the Eniwetok Cleanup that used the 40 and 400 pCi/g levels. I have not seen the EIS and hence cannot comment on it or previous evaluations of it. However, I do feel that the dose estimates obtained by the above LLL paper for soil concentrations of 400 pCi/gq suggest that additional sampling, statistical analysis, and dose estimation are warranted. It is true, of course, that the LLL dose estimates were obtained under what the authors considered to be conservative assumptions. Hence, the computed doses may be higher than actual conditions will produce. I would like to suggest the following actions: (1) Estimate the dose separately (using LLL's model) for each 1/4 or 1/2 hectare unit on each island. This can be done since each such area has an estimated surface (0-3 cm) soil concentration obtained using AYOLISOdIY 7 NNe¢ assuming various time utilizations in the various areas over an island. This approach is the same as Item 4 in my memo to you dated April 18, 1978 regarding suggested recommendations to DOE. accomplish: 2 months. x NO!LO3N109 1peysav\aj G89 S 54-1900-001 Once these dose estimates for Estimated time to Evaluate whether a "probability” approach to the estimation of doses would be helpful. Let me explain with an example. There is currently great uncertainty concerning the most appropriate value to use for the gut transfer coefficient. The present approach has been to compute doses using various possible values for this parameter and to see how dose estimates are affected. It is generally assumed that there is one "true" gut transfer coefficient (a constant) that should be S$ Pye) $T lb) ly oy AVIM)y Z —_ ‘ON XO9 w30704 the IMP, soil samples, and kriging. an island are in hand, the total dose to an indurdual can be computed used, but due to inadequate data there is uncertainty as to which value is correct. A probability approach would not consider that there is only onetrue value for the transfer coefficient. Rather, the coefficient would be considered to be a random variable with a statistical distribution. (3-71)