DBH thresholds systematically bias biomass estimates and stand structural analyses in African savannas

Abstract

30 in tropical dry forests, where heterogeneous stand structures and abundant small-diameter 31 woody stems can likewise influence biomass estimation and ecological interpretation.32 Threshold selection is therefore not a marginal methodological detail but a major source of 33 systematic bias.Small-diameter stems dominate stem counts in savannas 10 and can 34 cumulatively contribute substantial biomass, particularly in degraded or recovering systems 35 where juvenile cohorts prevail 11 .Excluding these stems can underestimate biomass and 36 obscure regeneration dynamics, whereas including very small individuals may increase stem 37 density and hinder comparability across studies 12 .Moreover, DBH-related uncertainty can 38 exceed that associated with the choice of allometric model, and measurement errors near 39 threshold boundaries may propagate into 10-20% uncertainty in carbon estimates 13 .This 40 pattern is not confined to savannas: in northern boreal peatland forests, excluding small-41 diameter stems has been shown to bias growth-trend estimates in a consistent, positive 42 direction, because the omitted biomass is concentrated in young cohorts 14 .43 These issues are especially consequential in Africa, where savanna inventories inform GHG 44 reporting, Reducing Emissions from Deforestation and Forest Degradation (REDD+), 45 initiatives, the African Forest Landscape Restoration Initiative (AFR100), and biodiversity 46 assessments 15 .While shrub and coppice biomass may be quantified through whole-plot 47 destructive harvesting, thereby avoiding individual-shoot measurements, most regional and 48 national assessments depend on DBH-based inventories and allometric equations.Yet, DBH 49 thresholds are rarely standardized or explicitly justified.Here, we argue that DBH threshold 50 selection is an under-recognized methodological choice that systematically shapes savanna 51 biomass and diversity estimates, and we propose an evaluative framework to identify, quantify, 52 and transparently report this hidden source of uncertainty, providing a foundation for future 53 threshold-adjustment methods.54 2. How DBH Thresholds Shape Reported Savanna Structure 55 Minimum (DBH) thresholds applied in African savanna inventories vary widely across regions, 56 ecological contexts, and study objectives (Table 1).In Sudan savanna, reported thresholds range 57 from ≥5 cm to ≥10 cm 16 , with some large-scale surveys omitting explicit DBH criteria 58 altogether despite reporting broad ranges of stem density.Similar variability is evident across 59 the African savanna biome, where thresholds span from very low cut-offs (≤2 cm) in Sudanian 60 systems to ≥15 cm in some miombo woodlands.Semi-arid and Sahelian savannas commonly 61 apply intermediate thresholds (≥5-6 cm), while ≥10 cm is frequently used in production-62 oriented or management-focused inventories.Although such variation may reflect differences