One purpose of this paper is to report on a multistage sampling scheme used in a Michigan peat resources study. A second purpose is to show how variations of this method can be applied in a sequential manner to adjust for incoming data or changing budgetary parameters.

The sampling plan for this study is a multistage sampling scheme utilizing clustering and multiple stratification. The State of Michigan has been stratified into soil-geomorphic provinces in order to group together organic soils formed on similar associated land surfaces. Within each province the unit of interest is the county or portion of the county. Each county or portion is then sampled. Each township within a county is stratified into four levels of organic soil deposit size, based on dot counts of soil survey maps. The townships form stratified clusters, and within each township a stratified sample of sections is chosen. Within each section systematic transect samples of the organic soils are obtained. The sampling plan should result in good point and interval estimates of organic soil characteristics by section, township, county, province, and total state evaluation.

The advantages and disadvantages of this sampling method are discussed in detail. Its advantages include: simplicity to implement in the field; ability to change with different geographic, personnel, or budgetary restrictions; and ability to be modified in a sequential manner as data are collected. The major disadvantage is the resulting relatively complex estimating equations. An attempt is made to simplify the explanation of these equations.

The utilization of this sampling scheme in conjunction with soil survey maps gives a very efficient method of estimating the total amount and characteristics of peat over a large geographic area.