An investigation into the applicability of geostatistical techniques for estimating contamination levels following an accidental release of radioactivity
| Autor: | Neil Higgins, Jackie Carter, Fiona McLaren |
|---|---|
| Rok vydání: | 1997 |
| Předmět: |
Gaussian
Public Health Environmental and Occupational Health Sample (statistics) General Medicine Atmospheric dispersion modeling symbols.namesake Data point Kriging Sample size determination Statistics symbols Environmental science Radiation monitoring Waste Management and Disposal Gaussian network model |
| Zdroj: | Journal of Radiological Protection. 17:73-84 |
| ISSN: | 1361-6498 0952-4746 |
| DOI: | 10.1088/0952-4746/17/2/004 |
| Popis: | In the event of an accidental release of radioactive material into the environment appropriate responses will be aided by the analysis and display of the collected data. This paper considers the suitability of using the geostatistical technique of ordinary kriging to interpolate between environmental measurements. Two factors affecting the accuracy of kriged estimates for unsampled locations are investigated, namely the effects of sample size and of spatial distribution. Sample data were generated by a gaussian atmospheric dispersion model and error estimates were determined by comparison of the kriged estimates at unsampled locations with the gaussian model predictions at these locations. Calculations were made for datasets containing between 20 and 80 points, and it was found that even datasets with low numbers of samples gave encouraging results. For example, 40 data points provided estimates with less than 50% error in 70% of the entire region. Estimation errors were smallest for the region of highest values, and highest on the boundary of the region. These results suggest that ordinary kriging can provide useful information in the assessment of post accident radiation monitoring data. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|