Geostatistical modelling of the spatial distribution of sewage pollution in coastal sediments

The spatial correlation of a series of sedimentary fecal stanols and sterol s data obtained from 39 surface sediment sampling stations in the eastern w aters of Hong Kong was analysed by geostatistical techniques. Coprostanol [ 5 beta] (5 beta(H)-holestan-3 beta-ol), cholestanol [5 alpha] (5 alpha(H)-c olestan-3 beta-ol) and cholesterol abundantly found in human feces were use d as molecular markers for sewage pollution monitoring in the coastal sedim ents. The [5 beta/(5 beta + 5 alpha)] ratio in the study area was found to range from 0.04 to 0.61 ((x) over bar = 0.27) and that of [5 beta/(5 alpha + cholesterol)] ranged from 0.02 to 0.73 ((x) over bar = 0.19). Semi-variog rams of the two ratios were found to fit isotropic spherical models: for [5 beta/(5 beta + 5 alpha)]: c(0) = 4 x 10(-3), c = 6.5 x 10(-3) a = 12000, D -c = -0.024 and R = 0.91; for 5 beta/(5 alpha + cholesterol): c(0) = 7.5 x 10(-3), c = 8.5 x 10(-3), a = 25000, D-c = -0.061 and R = 0.67. Because of the more precise covariance model produced, the former stanol/sterol ratio was adopted as the sewage pollution marker for the kriging of contour map f or sewage pollution. Three major sources of sewage pollution in the eastern waters of Hong Kong were highlighted, they were the densely populated Vict oria Harbour, the sludge disposal site and the Inner Mirs Bay. A detailed d ispersion pattern of sewage pollution around the sludge disposal site was s tudied. Our results demonstrate the usefulness of kriging in providing a ho listic picture on the extent and spatial correlation of sewage pollution in coastal marine environment. (C) 1999 Elsevier Science Ltd. All rights rese rved.