Corrections for cluster-plot slop

Harry T. Valentine, Mark J. Ducey, Jeffrey H. Gove, Adrian Lanz, David L.R. Affleck

Research output: Contribution to journalArticlepeer-review

Abstract

Cluster-plot designs, including the design used by the Forest Inventory and Analysis program of the USDA Forest Service (FIA), are attended by a complicated boundary slopover problem. Slopover occurs where inclusion zones of objects of interest cross the boundary of the area of interest. The dispersed nature of inclusion zones that arise from the use of cluster plots precludes the use of most of the slopover-correction methods that apply to solitary plots. One exception is the walkthrough method, which corrects for slopover bias in radially symmetric cluster-plot designs. In this article, we provide a modification of the walkthrough method, the "walkabout method," which is applicable to some asymmetric cluster-plot designs, such as the one used by FIA. We also present two general correction methods, the "vectorwalk method" and the "reflection method," both of which are applicable to cluster-plot designs with satellite subplots arranged in a regular or irregular pattern. The reflection method has two sets of protocols, one that applies to radially symmetric designs and to asymmetric designs with random orientation; the other set applies to asymmetric designs with fixed orientation. Both the vectorwalk and reflection methods incorporate the walkthrough method on a subplot-by-subplot basis. All four methods correct for slopover bias for straight or curved boundaries, and where work outside the tract is impossible.

Original languageEnglish
Pages (from-to)55-66
Number of pages12
JournalForest Science
Volume52
Issue number1
StatePublished - Feb 2006

Keywords

  • Boundary overlap
  • Edge effect
  • Monte Carlo integration
  • Reflection method
  • Slopover
  • Vectorwalk method
  • Walkabout method
  • Walkthrough method

Fingerprint

Dive into the research topics of 'Corrections for cluster-plot slop'. Together they form a unique fingerprint.

Cite this