Skip to contents

Model and computation

Consult the detailed technical document (PDF) as needed for model definitions, priors, sampler updates, and prediction. The articles below introduce the models through worked examples.

Fixed-area plot sampling

Begin with the shared simulated forest. The magnitude article estimates overall tree density; the composition article estimates species shares conditional on known totals. Combine magnitude and composition to estimate DBH distributions, then species-by-DBH stand tables. All five articles use the same downloadable population.

A simulated forest for plot-sampling examples

Prepare the shared forest, simulate tree counts, and calculate known population quantities.

Magnitude models for fixed-area plots

Estimate forest and stand tree density and compare with truth and a direct estimate.

Species composition from fixed-area plots

Estimate species proportions conditional on known total tree counts.

DBH distributions from fixed-area plots

Estimate forest and stand DBH distributions by combining magnitude with one ordered composition variable.

Species-by-DBH distributions from fixed-area plots

Combine magnitude and nested composition to estimate species-by-DBH stand tables.

Group, spatial, and space-time effects

These articles extend the full model. Stratum effects describe differences among groups of stands, first through intercepts and then through canopy-height slopes. Spatial effects describe continuous variation using distance between locations. A further example adds effects to DBH allocation within species. The space-time article combines repeated plot measurements across years. Each uses its own simulated population in the same forest and provides the data needed to run its code.

Stratum effects on tree density and species composition

Add stratum effects, examine combined intercepts, and summarize forest and stand predictions.

Canopy-height relationships that vary among strata

Fit stratum-specific intercepts and slopes in magnitude and species allocation, then compare combined coefficients and predicted relationships with truth.

Spatial effects on tree density and species composition

Fit spatial magnitude and composition effects, map combined intercepts, and compare forest and stand predictions with truth.

Random effects in within-species DBH distributions

Add a shared spatial factor to DBH allocation within species, examine its loadings, and compare forest and stand estimates with truth.

Space-time effects with repeated plot measurements

Fit an inventory with partial remeasurement, estimate spatial and temporal dependence, and summarize forest distributions by year.

Inventory studies

Estimating Harvard Forest stand tables from sampled plots

Apply the models to real forest inventories. The Harvard Forest study compares stratum and spatial effects against a mapped census, then assesses estimation error and interval coverage across 1,000 sampling layouts.