Addendum: The supplementary documents referenced in this manuscript are posted on the Web site https://www.fia.fs.fed.us/library/sampling/index.php
For purposes of sampling and estimation, Forest Inventory and Analysis (FIA) subdivides the total land area of the United States into mutually exclusive populations and subpopulations. Populations are usually defined by county boundaries or by public ownerships that may or may not cross county boundaries (e.g., national forests). In cases where the sample size for individual counties is insufficient, groups of counties may be combined into a super-county to form a single population with adequate sample size. Based on user request, counties occasionally are split into subpopulations to accommodate enumerated (known) acreages supplied by public agencies (e.g., National Forest System and The Bureau of Land Management). This is done to ensure that FIA totals match the county-level acreages reported by the requesting agencies. Each FIA population and subpopulation has a known number of plots and a known area of land, obtained from the U.S. Census Bureau, from which population estimates are derived. Each is sampled and processed as a separate entity, so estimates of grand totals and their variances for groups of populations and subpopulations are additive. For example, State-level estimates are obtained by totaling the estimates from all populations and subpopulations bounded by the State.