Adaptive cluster sampling design is used to estimate the abundance the population which is rare, hidden or hard to reach by considering only a small part of the population. In adaptive cluster sampling, an initial sample is selected with a conventional sampling design. Whenever a particular variable of interest satisfies a specified condition for an individual in the sample, units in the neighborhood of that unit are added to the sample. If in turn any of the added units satisfies the condition, still more units are added, and so on.

Sometime a population may occur in environment in which the variable of interest is zero for many of the population units. This is because either there is no object at all or objects are may leave the place for some unknown reason i.e. the population objects has moving pattern.

In this thesis, an environmental sampling design has been proposed by considering the moving pattern of the objects in the population. The probability of availability of an object and the detection probability of each unit has been considered having certain distribution for particular kind of population.

Population total and population mean variance of the estimate have been estimated using simple random sampling as well as the proposed design by considering the selection probability. Comparing these two estimated population mean variance, it has been observed that proposed sampling design produce better precision.