Abstract
Sector sampling is a new and simple approach to sampling objects or borders. This approach would be especially useful for sampling objects in small discrete areas or “polygons” with lots of internal or external edge, but it may be extended to sampling any object regardless of polygon size. Sector plots are wedge-shaped with a fixed sector angle. The probability of object selection is constant and equal to the sector angle in degrees divided by 360°. A unique property of sector sampling is that the point from which the angle originates may be located subjectively when the sector direction is at random. Another advantage over traditional sampling (such as fixed or variable area plots) is that there is no edge effect; that is, there is no altering of selection probabilities of objects close to polygon boundaries. Various approaches are described for deriving polygon means and totals with their associated variances. We review the genesis of sector sampling and develop two new components: sub-sampling using fixed area plots and line sampling using the sector arcs as transects. Sector sampling may be extended to measuring a variety of objects such as trees, shrubs, plants, birds, animal trails and polygon borders.
Highlights
Sector sampling [1,2] was developed, but not confined to sampling objects in small complex polygons with lots of edge
There is no “edge effect” bias in sector plots as the selection probability for an object is the same anywhere in the sector plot, which is independent of proximity to a polygon edge [1]
The probability of object selection under sector sampling is fixed for any given angle α; polygon or sample unit totals, per tree values or numbers of items may be unbiasedly calculated using simple random sampling (SRS) formulae [1] by multiplying sector plot totals by the expansion factor, 360°/α
Summary
Sector sampling [1,2] was developed, but not confined to sampling objects in small complex polygons with lots of edge. There have previously been papers in other disciplines using an angular section of a circle or ellipse such as: sampling the number of myelin or nerve fibers in the traverse sections on nerve fiber bundles [4,5,6]; sampling sunflower heads [7] and crowds of people [8] In these instances, the statistical properties and techniques for calculating means and variances were not examined in any depth. It was not apparently noted that the pivot-point could be established subjectively rather than centrally, or that there was no object selection edge effect bias. Sector sampling may help extend sampling of some novel conditions with considerable edge that are not well sampled using traditional approaches such as points or quadrats [14] or fixed area or variable probability plots or points [15]
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have