Deconstructing Adsorption Variability: the prediction of spatial uncertainty in pollutant movement from contaminated land

The re-use of land contaminated by pollutants from manufacturing or other industrial processes is currently high on the political agenda. Consequently, there is a growing need for good quality assessment of the risk (both to groundwater and directly to humans residing/working on such land) from historical/current pollution. This multi-disciplinary project seeks to understand and quantify the influence of factors controlling adsorption of pollutants into soils; adsorption (not absorption) means that the pollutant gets stuck in the soil and ceases to move and so no longer poses a threat to groundwater although it may still be a threat to people/animals in contact with the soil.

The project is a collaboration between the departments of Geology and Mathematical Sciences at the University of Durham and the department of Civil and Structural Engineering at the University of Sheffield. Soil sampling and chemical and other analyses to build models of how adsorption depends on pollutant and soil properties are being carried out by Durham Geology and Sheffield Engineering The statistical part of the project is the responsibility of Peter Craig who is a statistician working in Durham Mathematical Sciences.

The statistical component involves (i) design of suitable spatial sampling methods for site sampling; (ii) quantifying spatial variability in soil properties within and between sites; (iii) quantifying uncertainty in the adsorption relationships developed in the rest of the project and (iv) pooling these sources of uncertainty for the purposes of site-specific risk assessment. The statistical aspects are being tackled within a Bayesian framework which is well-suited for the ultimate goal of risk assessment.

The project is funded by EPSRC (grants GR/R03129/01 and GR/R02825/01) to a total value of approximate 260000 from January 2001 to May 2004.

Peter Craig
Last modified: Wed Nov 5 15:22:23 GMT 2003