Ratio estimation to improve estimates of top height from suboptimal samples in forest inventory plots in Québec

Top height is one of the most useful variables provided by sample plot measurements conducted during forest inventory programs. Combined with tree age, it allows the determination of site index, a common measure of forest site productivity. Unfortunately, there are cases where the number of site trees is suboptimal; this potential problem can lead to imprecise estimates of top height and site index. We approached this problem using ratio estimation techniques and incomplete data sets from forest inventory plots in Québec. We tested different combinations of suboptimal sampling of site trees to estimate top height based on their diameters, using only 2 or 3 site trees per 0.04-ha plot, instead of 4. The results show that suboptimal sampling of site trees should not be conducted randomly, and that some combinations consistently performed better than others. Also, some combinations never outperformed others and should be avoided when estimating top height of some important tree species in Québec using suboptimal sampling of site trees. Finally, for the best combinations, we concluded that the ratio estimation technique generally provided precise and unbiased estimates of top height. Key words: top height, site index, site trees, ratio estimation, forest inventory