Abstract
We consider an adaptive cluster sampling procedure for assessment of contaminant risk to ground water from soil at Owode-Onirin, Lagos, Nigeria. This method provides a more efficient way of sampling sparse but highly clustered population, and exploits the advantage of spatial clustered population to yield more efficient estimates of the concentrations of heavy metals pollutants of ground water from soil. A model-assisted estimation approach that incorporates models into the design-based estimation frame work was considered.