Methodology pieces from the PERSEUS multi-model forest carbon intercomparison. Each note is a short methods finding from the five-state pilot or the CONUS extension plan, suitable for citing alongside the main framework paper.
The hybrid yield form fit per cell across 48 states (CONUS reserve 11,607 to 13,701 Tg C; experimental, un-anchored), plus an IPCC-style harvested wood products layer: net forest+products change turns standing-only -43%/ -68% (harvest/intensive) into -1,155/-3,145 Tg C.
A Chapman-Richards-with-decline hybrid (best CV across 48 states), sawtimber/pulpwood/residue product allocation with product-resolved CONUS biomass yield (standing sawtimber share 29% to 35% as stands mature), and a FIADB vs TreeMap area-expansion comparison.
The empirical peak-decline yield curves applied to TreeMap 2022 pixels and grown 100 years under four management scenarios. The no-harvest reserve reproduces the established CONUS figure (10,002 to 14,813 Tg C, +48%); conservation +5%, harvest -43%, intensive -68%. Served as model yc_treemap_spatial_v1 across all four management buckets.
Oregon's statewide GCBM landed: 5 of 6 states now cluster tightly between libcbm/GCBM ratios 0.74 and 0.80 under canonical B1.1 v6 — gaps of 20.5 to 25.6 percent (WA, MN, IN, ME, OR), with GA the lone outlier at 1.05 consistent with its warm-donor placeholder. Replacing uniform-FT with FIA EXPNS expansion factors flips the picture: MN reaches parity, four states overshoot, ME drops further below. Inventory stratification, not engine differences, is the dominant component of cross-engine uncertainty.