Lystematic sampling of two-dimensional surfaces and related problems

Abstract

The paper is largely concerned with twenty-one possible methods of sampling a plane area, with points as sampling units, for the purpose of estimating a portion of this area having certain defined characteristics. These methods result from a combination, two at a time, of random, stratified and systematic sampling in two perpendicular directions, with or without alignment of the sampled points. Eleven of these methods involve systematic sampling in one or both directions. The estimate of the proportion of the area of interest is simply the proportion of points in the total area, falling within the area having the defined characteristics. For each method the variance function is derived. Fourteen different types of space covariance functions are involved in these variance functions.

Keywords:

• plane area
• satellite photographs
• pixels
• random sampling
• stratified sampling
• aligned
• conditional probability
• space covariance functions
• topographic variances
• line transects