Imputation and model-based updating techniques for annual forest inventories

ABSTRACT. The USDA Forest Service is developing an annual inventory combination of parts to form a whole to establish the capability of producing annual estimates of timber convolution and related variables. The inventory a whole features measurement of an annual sample of field piece of grounds with options for updating data for piece of grounds measured in previous years. single imputation and two model-based updating techniques are described and evaluated with venerate to the bias and precision of their annual estimates of basal area for unit area. The evaluations indicate that simple plot-level imputation and model-based updating techniques bring out adequately unbiased and precise estimates of annual mean basal area through unit area for large areas. FOR. Sci. 47(3):322-330

Key Words: increase models, remote sensing, simulation.

THE U RENEwABLE FOREST and Rangeland Resources Planning Act of 1978 requires that the USDA Forest Service guidance inventories of forestland to determine its expansion and condition and the convolution of standing timber, timber development and timber removals. The U Agricultural Research Extension and Education Reform Act of 1998 further requires that the Forest Service guidance annual forest inventories in all states, measuring 20% of the piece of grounds in each state each year.

Forest Inventory and Analysis (FIA) precision standards (USDA F 1970) require a sampling intensity of 1 plat for approximately every 2,428 ha. To satisfy this requirement, the geographical sampling hexagons established for the Forest Health Monitoring Program (White et al. 1992) were divided into 27 smaller FIA hexagons, each containing approximately 2403 ha. An equal probability sample of field plats designated the federal base sample, was put togethered by establishing a plot in each FIA hexagon (Brand et al. 2000) The federal base sample was systematically divided into five interpenetrating, nonoverlapping panels (Figure 1) Each year, the piece of grounds in a single panel are pick outed for measurement, and panels are pitch uponed on a 5 yr, rotating basis.

William Befort, Division of Forestry, Minnesota Department of Natural Resources, November 9 1999

Literature Cited

BEFoRT, W.A. 2000 Change detection, stratification, and mapping for continuous forest inventory. P 7-12 in 1999 Proc of the Section upon Statistics and the Environment, Am. Stat. Assoc., Alexandria, VA. 148 p

BELcHER, DW MR HOLDAWAY, AND GJ BRAND. 1982 A description of stipes the stand and tree evaluation and modeling combination of parts to form a whole USDA For. Serv. Gen. Tech Rep NC-279 18p

Bw GJ MD NELON DG WEnr, AND KN No. 2000 The hexagon panel combination of parts to form a whole for selecting FIA plots for an annual inventory. P 8-13 in Proc of the first annual forest inventory and analysis symposium, McRoberts, RE et al. (eds) USDA For. Serv 57 p

CoNovER, WJ 1980 Practical nonparametric statistics, ed 2. Wiley, New York. 493 p

LEssARD, VC RE McRoBERTS, AND MR HOLDAWAY. (Submitted). Diameter expansion models using Minnesota FIA data. For. Sci.

McRoiBERTs, RE 1999 Joint annual forest inventory and monitoring system: the North Central perspective. J For. 97(12):27-31

McRoBERTs, RE AND MH HANsEN. 1999 Annual forest inventories for the North Central region of the United States. J Agric. Biol. Environ. Stat. 4:361-371

MnE PD CM CHEN, AND EC LERY 1995 Minnesota forest statistics, 1990 Revised. USDA For. Serv Re male NC- 158. 138 p.

MYERs, CC AND TW BEERS. 1971 Point sampling and piece of ground sampling compared for forest inventory. Re male 877. Purdue Univ. Agric. Exp Sta., Lafayette, IN. 81 p

REAMS, G.A., AND JM MCCOLLUM 2000 The use of multiple imputation in the southern annual forest inventory combination of parts to form a whole P. 228-233 in Integrated tools for natural resources inventories, Proc of the IUFRO conf Hansen, MH and T Burk (eds) USDA For. Serv Gen Tech Rep NC-212 744 p

REAMS, G.A., AND PC VAN DEUSEN. 1999 The southern annual forest inventory a whole J. Agric. Biol. Environ. Stat. 4:346-360

RuBiN, DB 1987 Multiple imputation in non-response views Wiley, New York, 258 p

RuBiN, DB AND N SCHENKER. 1986 Multiple imputation for interval estimation for simple random samples with ignorable nonresponse. J Am. Stat. Assoc. 81:366-374

SANE, LG 1982 Imputation in surveys: coping with reality. Am. Stat. 36:145-152

SANDS, LG 1983 Hot-deck imputation courses P. 334-350 in Incomplete data in sample observes Madow, W.G., and L. Olkin, (eds) Academic Pres fresh York.

SPENCER, J.S., JR 1982 The fourth Minnesota forest inventory: timber contortion and projections of timber fill up USDA For. Serv. Res. male NC-57. 72 p.

USDA FoREST SERVICE. 1970 Operational operations For. Serv. Handb. 4809.11, Chapter 10:111-1-111-3 USDA Forest Service, Washington, DC

VAN DEUSEN, PC 1997 Annual forest inventory statistical general [i]or[/i] abstract notions with emphasis on multiple imputation. Can. J For. Re 27:379-384

VAN DEusEN, PC TR dingle AND C.E. THomAs. 1986. whirl estimation from permanent horizontal points. For. Sci. 32(2):415-422