Provider: ingentaconnect
Database: ingentaconnect
Content: application/x-research-info-systems
TY - ABST
AU - Williams, M. S.
AU - Schreuder, H. T.
TI - Documentation and Evaluation of Growth and Other Estimators for the Fully Mapped Design Used by FIA: A Simulation Study
JO - Forest Science
PY - 1995-08-01T00:00:00///
VL - 41
IS - 2
SP - 26
EP - 45
KW - Classical variance estimator
KW - area estimation
KW - length of boundaries
N2 - For the proposed mapped design of the USDA Forest Service, Forest Survey and Inventory (FIA) units, in a simulation study for a realistic mapped population, Grosenbaugh's ingrowth estimator Î_{1} and Van Deusen's survivor growth estimator ΔS_{2} are best. If additivity of estimates is desired, ΔY_{1} should be used for total growth estimation. If additivity is not required, a plausible assumption as FIA users become more statistically literate, Roesch's estimator ΔY_{4}, which is 12% more efficient than ΔY_{1} should be used if survivor ongrowth trees can be cored to determine their first time period diameters. If not, Grosenbaugh's nonadditive estimator ΔY_{2} should be used which is 8% more efficient than ΔY_{1}. The classical variance estimator yields unbiased estimates of the variance of the parameter estimates as expected. Area estimation, change in area estimation, percentage of plots in more than one condition, means per tree, and changes in means per tree over time can be estimated reliably if sample size is large enough and conditions or changes in conditions are not too infrequent. The length of boundaries between conditions are consistently underestimated. For. Sci. Monogr. 31:26-45.
UR - http://www.ingentaconnect.com/content/saf/fs/1995/00000041/A031s031/art00004
ER -