Open Collections

A comparative study of spatial analysis methods for forestry Nelder experiments

University of British Columbia

The Nelder (1962) series of systematic spacing wheel designs define compact and spatially explicit layouts for forestry spacing experiments. One difficulty in the analysis of Nelder experiments is that the compact arrangements of trees may result in significant correlations among neighbouring variable values when these are geographically ordered. When this is the case, classical analysis methods, based on random sampling models and ordinary least squares estimators, provide inefficient estimates of treatment effects and biased estimates of variance parameters. However, the formal geometric structure of Nelder experiments presents an opportunity to evaluate the properties of alternative analysis methods based on spatial correlated error models. The statistical and practical properties of four spatial analysis methods were evaluated in the context of the interpretation of data from forestry Nelder (1962) experiments. The spatial analysis methods considered were based on either regionalized variables or nearest neighbour models and restricted maximum likelihood estimators. The validity and relative efficiency of the spatial analysis methods were assessed under different magnitudes of spatial autocorrelation through a simulation study. The same spatial analysis methods were also applied to data from a Nelder experiment in the University of British Columbia Malcolm Knapp Research Forest in Haney, British Columbia. The spatial analysis methods based on regionalized variables models were valid and more efficient than the classical analysis method when data were strongly spatially correlated. The spatial analysis methods based on nearest neighbour models were not valid and were less efficient than the classical analysis method. Application of the spatial analysis methods required more intensive estimation procedures.

Thesis/Dissertation

application/pdf

2009-08-04

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/11615

Forestry

University of British Columbia

UBCV

DSpace