Decreasing global uncertainty in mapping site & environmental contamination by combining different measurement techniques in a joint (geo)statistical interpretation

Hasan Moudud Md


Huysmans Marijke, (VUB),

SCK•CEN Mentor

Vidmar Tim
+32 14 33 21 10

SCK•CEN Co-mentor

Rogiers Bart
+32 14 33 31 23

Expert group

Crisis Management and Decision Support

PhD started


Short project description

Every country having older nuclear installations or NORM-related industries such as USA, South Korea, France, UK, Slovakia etc., is suffering from problems related to soil contamination (radiological, chemical or mixture of both problems) usually with large volumes of contaminated soil.

For site and environmental decommissioning and remediation, an accurate estimate of the three-dimensional radionuclide contamination and its global uncertainty (measurement and spatial) is important for impact assessment and remediation option evaluation. Geostatistical methods are a standard way to estimate spatial delineation of contamination and form the basis in site and environmental remediation of impact assessment and optimization of remediation. However, uncertainties of estimates in the context of remediation are sometimes very large. Further methodological developments are needed to reduce the uncertainty in estimates of the contaminated volume. Promising routes are a combination of different measurement techniques or campaigns (e.g., different sample volumes in the study of Boden et al., 2013) or adding the third dimension (depth measurements). 


The basic hypothesis of this PhD study is that the combination of different types of data (direct-indirect, destructive-non-destructive, contact-non contact) obtained during different sampling campaigns with different levels of detail and information with stochastic space-time geostatistical methods accounting for different degrees of uncertainty of the different data could lead to reliable estimates of the three-dimensional spatial distribution of contamination and its uncertainty. The main objectives are therefore to develop and improve a measurement technique to measure in-situ the depth distribution of radionuclide contamination, and to develop and apply a methodology to estimate radionuclide contamination and its uncertainty by combining different measurement techniques in a joint (geo)statistical framework. 

An important part is to develop a practical method to measure in-situ the depth distribution of the contamination. Small and shallow drillings (a few meters at most) allow for measuring indirectly contamination at several depths either within the drilling, which with smart shielding and collimators design or by making use of multiple or segmented detectors, may give direction-dependent measurements or on the sampled drilling sample. As most applications of measurements in drilling have limited shielding or measure radiation from all directions, here it is extended to a direction-dependent measurement device or method. It is also investigated how signals from different locations measured simultaneously can be identified. A theoretical analysis to optimize detectors/shielding/collimators precedes a practical demonstration of the methodology at a contaminated site.

In addition, an evaluation is needed of different measurements to supplement direct radionuclide measurements on samples in the laboratory. The main focus is on so-called soft data; data which can be collected in a relative simple and fast way, in situ, but may only give indirect information of the total contamination. Possibilities include in-situ gamma, dose rates, X-ray luminescence, or surface magnetic susceptibility. Evaluation criteria are the easiness of collecting data in a fast way, the signal-to-background ratio, the correlation of the method results with the contamination levels, etc. The applicability of thermoluminescence (TL) and/or optically stimulated luminescence (OSL), which measures the cumulative dose in quartz grains, omnipresent in sandy soils, will be evaluated. This is the first time that this technique will be evaluated to monitor soil contamination. A field campaign will be defined and set up to apply different relevant measurement techniques under realistic conditions. Within this campaign, the newly developed method for in-situ measurement of depth distributions will be tested in combination with a range of state-of-the-art radiological methods combining the expertise of different groups within SCK•CEN and non-radiological methods identified, providing a unique dataset for further analyses and validation of the geo-statistical methods. Potential sites are BR3, BelgoProcess, Kepkensberg (NORM waste site), and others.

The third part is to develop and apply geostatistical algorithms to combine different types of input data obtained during different campaigns to estimate the spatial distribution of the contaminated materials. Crucial points are e.g. how to account for differences in information, accuracy and uncertainty in the different types of data, how to account for different timings of sampling, how to integrate data with different (potentially weighted) spatial supports, and how to use this information as an input for subsequent sampling campaigns in order to reduce global uncertainty etc. The algorithms and the workflow (input of data, data analysis and output of results) will be integrated in a tool applicable at different stages in an environmental remediation project related to sampling, source identification, radiological impact assessment and remediation action evaluation.