The precision of species proportion by area when estimated by angle counts and yield tables

Abstract

In Europe yield tables are frequently used to get information on future total crops as well as on current and future increments per hectare and thus the database for forest valuations, decisions on allowable cut and other management decisions. These yield tables usually claim to describe potential growth and yield per hectare of even-aged pure species stands. As a prerequisite to the use of these yield tables in mixed species, even-aged stands, the area to which every species has to be related needs to be known. The way to calculate this area is to split up the area of the mixed stand into areas each of which is occupied by a 'fully stocked' pure stand of one species. Stocking is measured in terms of observed basal area versus basal area given by the yield tables. Thus the species proportion by area is calculated from the ratios between the observed and the respective yield table basal areas. In order to obtain the necessary sample size for a given precision of the species proportion by area, its variance is calculated from Gauss's law of error propagation with the species proportion by basal area and the ratio between the respective basal areas as they are given by an appropriate yield table as independent variables. Estimates of these variables indicate that 3 to 12 angle counts per stand (basal area factor BAF = 4 m ha~) are sufficient for a standard error of ±10 per cent for the species proportion by area, and 9 to 50 angle counts for one of ±5 per cent.