Improving the Precision of Forestry Experimentation with Covariance Analysis

I.C. Zobi (1) , J.P. Pascal (2) , P. Couteron (3) , K.B. Kouadio (1)

1 INP-HB - Institut National Polytechnique Félix Houphouët-Boigny [Yamoussoukro]
2 LBBE - Laboratoire de Biométrie et Biologie Evolutive - UMR 5558
3 IFP - Institut Français de Pondichéry

This study aims to provide a tool of study to forester and forest survey researchers. It deals with the effect of a method of thinning on tropical forest dynamic. In this method, elimination of trees species of no commercial value was done in order to increase the growth rate of the whole group of commercial ones. Specifically, this study shows how a statistical method which is covariance analysis can be used to solve an inadequacy of experimental design. This inadequacy results from a difference between the initial total basal area of the treated quadrants and that of the untreated (standard) quadrants. The gain of precision is 1.054. It provided the following estimates of the increase rate in basal area of the commercial species: 0.73 m2 ha-1 year-1 for the treated quadrants and 0.92 m2 ha-1 year-1 for the untreated ones. When the initial total basal area of the quadrants is not considered like an auxiliary variable, the values of the basal area of the commercial species are: 0.53 m2 ha-1 year-1 for the treated quadrants and 1.05 m2 ha-1 year-1 for the untreated ones.

Mots clés: Forestry experimentation; Covariance analysis

Domaines: Sylviculture, foresterie

I.C. Zobi, J.P. Pascal, P. Couteron, K.B. Kouadio. Improving the Precision of Forestry Experimentation with Covariance Analysis. Journal of Applied Sciences, 2008, 8 (5), pp.866-870. ⟨10.3923/jas.2008.866.870⟩. ⟨hal-02949581⟩