A. Figueiredofilho et al., NUMBER OF DIAMETERS REQUIRED TO REPRESENT STEM PROFILES USING INTERPOLATED CUBIC-SPLINES, Canadian journal of forest research, 26(7), 1996, pp. 1113-1121
Interpolated cubic splines were used to estimate diameters and partial
or total volumes for 285 trees from Pinus taeda L. plantations in sou
thern Brazil. A range of 4 to 13 interpolated points was tested in an
attempt to determine how many points are necessary to adequately repre
sent stem taper. The various number of combinations with 13 or fewer p
oints for interpolation were compared with a full spline of 16 points,
which according to pertinent literature, estimates volume similarly t
o the traditional volume formulas (Newton, Huber, Smalian). The result
showed that 10 well-placed points are necessary to obtain reliable es
timates of diameter and partial or total volume along the entire stem.
If the top 15% of the stem is ignored, 5 points adequately estimate v
olume, with an average residual of less than 2%. However, when fewer t
han 7 points are used, placement of points is very important to avoid
illogical oscillation inherent in the use of cubic splines.