8
Conclusions and Recommendations
DS86 is clearly a much more complete and sophisticated dosimetry system than any of its predecessors, and it calculates the organ doses in the survivors with considerable accuracy. The main component of the dose in organs deep in the body is gamma radiation, and measurement of this component with thermoluminescent techniques yields excellent agreement with the calculations of DS86 to within the accuracy of the measurement technique, possibly about ±10% and in any case within the range of expected uncertainty. However, the agreement might be improved further if the energy response of the TL measurements were reexamined. The most significant of the factors that potentially affect the TL measurements is the increase in the response of the TL with low gamma-ray energy. An estimate of the correction needed to account for the energy response is a decrease in the reported measurement value by an arbitrary 20%.
The neutron component of the dose is small, especially in Nagasaki, and considerably less certain than the gamma-ray component at the time of writing of this report. In 1986, it was apparent that unresolved discrepancies existed between measurements of thermal-neutron activation in cobalt and in europium and calculations of neutron fluence in Hiroshima and Nagasaki. In the revisions proposed since DS86, the discrepancies have essentially been resolved for Nagasaki (mainly because of new and finer group calculations and cross-section improvements) but have tended to become worse for Hiroshima with the addition of 36Cl thermal-neutron activation measurement techniques, possibly amounting to thermal-neutron fluence 1500 m from the epicenter greater by a factor of almost 10 than calculated in DS86.
Determined efforts in the last few years to establish the magnitude of the “neutron problem” and to explain it by examining critically the uncertainties in the measurements themselves have not resulted in definitive conclusions. Issues related to
the background and other uncertainties of some of the samples need to be explored further. And possible modifications of the spectrum issuing from the Hiroshima weapon has not produced a new source that is consistent with the reported measurements of thermal activation at great distances or satisfied the dynamics of the bomb explosion and the sulfur-activation measurements of fast neutrons made close to the bomb soon after its explosion.
This committee and others still working on these problems recognized that measurements of the fast-neutron component were essential to resolve this problem. A new method—measuring 63Ni produced by fast neutrons in copper samples (63Cu (n,p) 63Ni) —was proposed and strongly endorsed by this committee, which urged immediate support for such investigations (NRC 1996). Suitable samples of copper irradiated at the time of the bombing in Hiroshima (August 6, 1945) have been obtained and are being measured in Japan (with 63Ni radioactivity; T1/2 100 y) and in the United States and Germany (by accelerator mass spectrometry). More samples at strategic locations are still urgently being sought in Hiroshima and Nagasaki to expand the measurement program and make it as complete as possible.
At 1000 m, it appears that DS86 is not significantly in disagreement with measurements. Even at greater distances, the discrepancy is probably less than previously reported. However, the data are not certain enough to allow a good estimate of the discrepancy at great distances. Further measurements of 63Ni and 36Cl and improved uncertainty analyses are needed for better definition of the magnitude of the disagreement in the neutron fluence and confirmation of the 32S data for Hiroshima. The new measurements should use pre-established data-quality objectives. Some of the previously reported data, in particular the 36Cl data, need to be reanalyzed to resolve possible errors in the original reported results. Future measurements should include secondary-reference standard reagents and analytical blanks, and when it is appropriate, an isotopic tracer should be processed with each field sample. Preliminary results of the 63Ni fast-neutron measurements suggest that the discrepancy in Hiroshima is smaller (perhaps a factor of 3–5, not a factor of 10, at 1500 m) than suggested by the thermal-neutron activation measurements. Reconciling the relaxation lengths derived from some of these measurements is difficult. Further exploration of the apparent neutron discrepancy will continue with as many additional samples as feasible.
If it turns out that the discrepancy is indeed small and perhaps within the uncertainties of calculation and measurement, the doses to survivors can be regarded as reliable within a given uncertainty range. For the estimation of gamma-ray risk, it has been noted in this report that the neutron component—although small, with reasonable values of RBE in the range of 20–50—could contribute appreciably to the total effect even without any increase in neutron fluence in Hiroshima. A somewhat greater neutron contribution will result if neutron fluence is eventually increased at great distances there. However, it is clear again that although the risk from gamma-rays might, as a consequence, be a little less than previously estimated, it is well within the range of uncertainty known to exist for these estimates (NCRP 1997).
SUMMARY
Although DS86 is a good system for specifying dose to the survivors and for assessing risk, it needs to be updated and revised. Uncertainties have not been fully evaluated and might amount to more than the 25–40% in fractional standard deviations of parameters (Kaul and Egbert 1989). A number of parameters in DS86 have been improved and should be implemented. While the calculated gamma-ray fluences agree well with values measured with thermoluminescence and constitute the main component of the dose to the survivors, more work needs to be done to establish the magnitude of the neutron component and to assess the extent to which the neutron component (small in DS86) affects (lowers) the estimates of gamma-ray risk.
The committee offers the following recommendations regarding the revision of DS86 that is clearly needed and that hopefully will be completed in 2002:
-
The present program of 63Ni measurements should be pursued to completion.
-
All thermal-neutron activation measurements, particularly those with 36Cl and 152Eu, should be reevaluated with regard to uncertainties and systematic errors, especially background (see Chapter 3).
-
Critical efforts to understand the full releases from the Hiroshima bomb by Monte Carlo methods should be continued.
-
Adjoint methods of calculation (i.e., going back from the field situation to the source term) should be pursued to see whether they help solve the neutron problem.
-
Local shielding and local-terrain problems should be resolved.
-
The various parameters of the Hiroshima explosion available for adjustment, including the height of burst and yield, should be reconsidered in the light of all current evidence in order to make the revised system as complete as possible.
-
A complete evaluation of uncertainty in all stages of the revised dosimetry system should be undertaken and become an integral part of the new system.
-
The impact of the neutron contribution on gamma-ray risk estimates in the new system should be determined.