. A comparison of two estimates of standard error for a ratio-of-means estimator for a mapped-plot sample design in Southeast Alaska. Forest surveys Alaska, Southeast Statistical methods; Forests and forestry Alaska, Southeast Statistical methods; Estimation theory; Bootstrap (Statistics)Error analysis (Mathematics); Forest surveys; Forests and forestry; Estimation theory; Bootstrap (Statistics); Error analysis (Mathematics). u C 0) o "° 0) O) eg c g â¡ Ketchikan â Stikine â Chatham. r i I 500 1,000 3,000 Number of bootstrap iterations Figure 6âDifferences betw
. A comparison of two estimates of standard error for a ratio-of-means estimator for a mapped-plot sample design in Southeast Alaska. Forest surveys Alaska, Southeast Statistical methods; Forests and forestry Alaska, Southeast Statistical methods; Estimation theory; Bootstrap (Statistics)Error analysis (Mathematics); Forest surveys; Forests and forestry; Estimation theory; Bootstrap (Statistics); Error analysis (Mathematics). u C 0) o "° 0) O) eg c g â¡ Ketchikan â Stikine â Chatham. r i I 500 1,000 3,000 Number of bootstrap iterations Figure 6âDifferences between bootstrap and traditional estimates of standard errors as a percentage of the bootstrap estimate, by number of bootstrap iterations, for estimates of gross cubic-foot growth on productive forest-land area within unreserved national forest lands, south- east Alaska. Discussion For the response variables listed, there is little difference between the two estimates. The largest percentage differences between error estimates are in the Ketchikan unit. The forest resource of the Ketchikan unit is more variable than in the other two units; it is ecologically transitional (approaching the southern extent of the Sitka spruce/western hemlock zone), and logging has been more intensive and extensive resulting in greater variety of age and size classes. Greater variability between bootstrap and traditional estimates may indicate the number of iterations was low for those estimates. Relative rapidity of data processing is a significant client service consideration. Rapid responses are important when providing clients with inventory results. Although bootstrap estimates can be generated for any of the population estimates FIA produces, the computational resources needed to generate those estimates for each cell in all the output tables typical for FIA reporting exceed those needed for ROM estimates. Use of the ROM estimator reduces production time dramatically. Acknowledgments Kev
Size: 1377px × 1815px
Photo credit: © The Book Worm / Alamy / Afripics
License: Licensed
Model Released: No
Keywords: ., bookcol, booksubjectforestsandforestry, booksubjectforestsurveys
