Structural Property Estimation from Density Samples for Western Woods

To improve the reliability of structural property information on a continuing basis for engineering applications of today’s wood, a sampling system is needed which incorporates modern statistical techniques, yet in the interest of economy produces data that may be compatibly combined with existing property information. Since it appears reasonable to assume that the regression of mechanical properties on specific gravity is independent of within tree location, the existing FPL strength-specific gravity data can be used to establish such regressions. By utilizing a statistical technique called double sampling, one of the variables, in this case, the more readily obtained specific gravity of standing trees, can be sampled with greater precision to produce a more reliable prediction of average strength for today’s wood. This is the methodology of Phase IV of the Western Wood Density Survey which makes it possible to evaluate all species on a reasonably equitable and continuing basis by means of intensive specific gravity sampling of standing trees. The representative quality of this sampling compared with exsiting data is revealed by the fact that the survey encompassed all commercial forest land in all 12 western states based on field plots located at the intersections of a 6.9 mile grid. For example, 9,000 Douglas-fir trees were sampled as compared with the sample size of 300 trees reported for Bulletin 479. Current estimates of average strength properties were made by progressively combining geographical “areas” (smallest area for which reliable species volume estimates are available) into “unit areas” (at least one percent of the total species volume in the United States, and at least 20 tree specific gravity samples) and into “unit area combinations” (smallest areas for which reliable volume and average property estimates can be made, i.e., for groups of “unit areas”, for species, or for combinations of species, depending on practical marketing considerations). Unit areas are combined by computing volume weighted averages of “unit area” average properties. Guidelines are drawn for combining species into any desired groupings, with appropriate description of the corresponding heterogeneous property distributions.

You must be logged in to download any documents. Please login (login accounts are free) or learn how to Become a Member