Variance partitioning as a guide for sampling and comparing spatial distributions of Hg and 137Cs in sediments

Distributions of Hg and 137Cs were examined in the sediments of a contaminated reservoir. The data were used to illustrate how the simple technique of variance partitioning can help in the efficient design of statistically sound environmental sampling schemes. Three regions of the reservoir were evaluated: sediment previously exposed during a 5.8-m draw-down of the water column (Zone 1), sediment in the vicinity of an old creek bed (Zone 3) and sediment between the two zones (Zone 2). Variance partitioning indicated that the largest variance components were due to differences between zones (38%) and samples from adjacent sediment cores within a zone (44%). Variation among aliquots (samples taken from the same core) was <1%. This knowledge allowed us to reallocate sampling effort from aliquots to an increase in field sampling without increasing analysis effort. Different release histories, as well as leaching and erosion from the exposed sediments are discussed as possible reasons for observed differences in the spatial distributions of Hg and 137Cs. Hg was found to be equally concentrated in the tow deeper zones (2 and 3) while 137Cs was found to exist in the order of Zone 1