Use of vector analysis to determine the effects of harvest residues on early growth of second-rotation Sitka spruce

Vector analyses have been used to examine the effects of harvest residues o n the growth of Sitka spruce planted 18 months after clearfelling a first r otation of the same species. Four treatments were imposed at the time of pl anting: retention of all residues; retention of fine twigs and foliage; pla cement of coarse branch material onto plots; and removal of all harvest res idues. Sample trees were harvested each year for five years after planting and the biomass and nutrient contents of whole crowns were quantified. Harvest residues had a large effect on mean tree height (p < 0.001). Coarse and fine components produced similar increases when compared to plots with no residues although the response to fine material occurred earlier. After five years, residues had increased mean tree height by 40%. Residues also increased the mean dry weight per tree although the response to coarse mate rial was not significant until the age of five years, at which time trees i n residues had 166% more biomass (p< 0.001). Tree growth was greatest where both forms of residue were retained on site. Interpretation of vector nomograms, foliage concentrations and concentratio n ratios indicated that the growth response was unlikely to he related to c hanges in nutrient supply for the first two years of treatment. Early treat ment effects may have been mediated through changes in weed competition or microclimate. After three years, treatment effects were likely to have been related to altered nutrient supply, particularly those for nitrogen and po tassium. According to vector analyses, potassium was not limiting tree grow th in whole-tree harvested plots although foliage levels were below critica l levels for young Sitka spruce and potassium : nitrogen ratios indicated a nutrient imbalance in these plots. Such an imbalance may arise from the ra pid loss of potassium from decomposing harvest residues accompanied by a mo re gradual release of nitrogen that may relate more closely to the nutrient demand of newly planted trees. The contention that vector analyses allow a standardized comparison between treatments without confounding with other factors has to be questioned. In the present study, different interpretations resulted from analyses based on comparing samples of different age-classes or from different canopy posi tions. The method has also been shown to be very sensitive to the choice of critical test statistic used in assessing each vector component. We conclu de that vector analyses can provide a useful visualization of treatment eff ects on tree growth and nutrition, Its relevance to the assessment of susta inable forest management is limited by the need to have treatment compariso ns available and interpretations should be corroborated by other evidence b efore firm conclusions can be drawn. (C) 1999 Elsevier Science B.V. All rig hts reserved.