Remote sensing the radionuclide contaminated Belarusian landscape : predicting ¹³⁷Cs content in 'Pinus sylvestris'

• Main Author: Goldsmith, Paul Christian
• Published: Kingston University 2007
• Subjects: 550; Earth systems and environmental sciences : Geography and environmental studies
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.485444

The 1986 Chernobyl accident resulted in distribution of Radiocaesium ([sup]137Cs) throughout much of the northern hemisphere. [sup]137Cs still persists in the environment, particularly in Belarus where up to 80% of the radioactive fallout occurred. Monitoring [sup]137Cs across space and through time is essential to remediate contamination and remote sensing, particularly imaging spectrometry has potential to provide information at appropriate spatial and temporal scales. This thesis explores that potential by assessing whether imaging spectrometry can be used to monitor [sup]137Cs content of vegetation. The research has three main components; field radiometry, laboratory experimentation and spaceborne hyperspectral imagery (EO-1 Hyperion). Field radiometry collected spectral and biochemical data from Pinus sylvestris growing on the Belarusian landscape contaminated with varying levels of [sup]137Cs, and investigated the links between foliar biochemistry, [sup]137Cs specific activity and spectral reflectance. Significant differences exist between spectra of Pinus sylvestris contaminated with different levels of [sup]137Cs as well as key foliar biochemicals of chlorophyll, nitrogen, cellulose, lignin and water. Partial correlation identified which parts of the electromagnetic spectrum [sup]137Cs contamination had impacts upon in relation to particular biochemicals. Prediction of leaf/needle biochemicals from spectra was possible using regression techniques, but predicting [sup]137Cs specific activity from biochemicals was not possible at a statistically significant level. Laboratory experimentation results supported fieldwork findings and exhibited significant differences in spectral response between contaminated and non-contaminated trees, also highlighting non-linearity in uptake and response of Pinus sylvestris to [sup]137Cs (with subsequent remote sensing implications). Spectra extracted from hyperspectral imagery also show significant difference in their spectra despite assumptions and constraints in interpretation of this imagery. The thesis concludes by suggesting that presently, the use of imaging spectrometry to monitor [sup]137Cs specific activity of vegetation (0.07 - 39.9 Bq/g) is not feasible; mainly due to immature understanding of [sup]137Cs impacts on plant biochemicals and the links between foliar biochemistry and [sup]137Cs content. Research limitations present opportunities for further research which may allow future monitoring of [sup]137Cs levels using imaging spectrometry.