While the data analysis and sampling design methods that can be applied depend strongly on the situation and specific goals of initial nuclear site characterization, the overall strategy often takes the form of the generic workflow illustrated in this flowchart.
The starting point we consider here is the request for initial nuclear site characterization to a radiological characterization team. Such a request can come from different kinds of actors, and can come with different amounts of detail. Following this request, a clear list of all objectives and identification of the constraints is absolutely required, and might require some iterations with the applicant to agree on the goals and priorities:
Decommissioning is a multi-disciplinary operation and the involvement of specialized staff performing the next stages of the decommissioning project can be highly beneficial. Technical feasibilities/constraints in the next decommissioning stages might strongly influence the initial characterization program. Effective communication and a common basis of understanding are essential. Extensive compartmentalizing might result in misinterpretation, non-optimal solutions and wrong decisions.
The highest-priority objective should be tackled first in most cases, and the cycle along the different objectives is started.
If feasible, data sets containing large amounts of data below detection limit should be avoided for a proper statistical analysis and whenever possible tackled during the strategy development. However, this is not always possible; for example due to low threshold values, the limitations of measurement techniques, but also due to unclear initial objectives or potentially changing objectives or thresholds during the characterisation process. Depending on the case, it might be very wise to use advanced statistical methods for dealing with samples below detection limit (Kim, Hornibrook, & Yim (2020), Pérot (2020)).
All prior information that is available and relevant for the investigated case should be gathered as a first step (historical records, mappings, incidents, etc.). If some radiological data would already be available, a first analysis to check if the objective is achieved is probably very useful, even if the results come with lots of uncertainty. In D&D, such prior information is nearly always available. We are working on historical installations and/or sites that have been shut down, or are going to be. Therefore, there is always a history of the exploitation phase, with available data, so this initial data-gathering step is of vital importance.
The data analysis succeeding the data collection consists, in general, of the following steps: Pre-processing, exploratory data analysis, the actual data analysis, and potentially a post-processing step. If the objective is not achieved, a sampling design should be proposed using the most appropriate method(s) given all prior information and the data analysis result. Following the design, the corresponding characterization campaign should be performed. Additional characterization can reveal unexpected issues, and often revisiting the gathering of prior information is then useful. After the additional characterization, the updated dataset is again analysed, and this iterative procedure is continued until the objective is finally reached. The entire process can then be repeated to tackle the remaining objectives. Once all objectives have been achieved, the initial characterization study should be reported in a transparent way, making clear what has been measured, which results were obtained from the data analysis, and how large the corresponding uncertainty is. The different steps are more extensively discussed one by one below.
Performing a radiological characterization program in two or more stages/phases can be efficient and effective to tackle areas with higher uncertainties. Unfortunately, this might not always possible due to planning constraints.