Abstract:
Appropriate allometric models are urgently needed to assess the status and changes in biomass
and carbon of the trees in miombo woodlands occupying large geographical areas in
Mozambique. This study developed two new and interchangeable allometric models for
estimating total above-ground biomass (AGB) of lowland miombo woodlands in the Beira
corridor, central Mozambique, based on stem diameter at breast height (DBH) and stump
diameter (SDI). The Beira corridor study area covers approximately 29,000 km2, of which about
three-quarters is lowland miombo woodland. The SDI-based model is proposed principally for
estimating total AGB (stem, branches, foliage) of harvested trees/shrubs when diameter cannot
be measured at breast height, and thus to reconstruct the former biomass in forests subjected to
logging, or clear-cutting for agriculture. The DBH-based model and SDI-based model were
fitted using data on a destructive sample of 155 trees, which were representative of tree sizes
(diameter and height) and tree species in the Beira corridor area. The following allometric
models were developed: DBH-based model [tDW (kg tree−1) = 0.1754 * (DBH) 2.3238], with
prediction performance, i.e. adjusted R-squared 98–99%; and SDI-based model [tDW (kg
tree−1) = 0.08495 * (SDI) 2.3987], with prediction performance 86–96%. Carbon comprises
50% of biomass. Both the DBH- and SDI-based models can be used for estimating total AGB
of lowland miombo woodlands with a high degree of reliability, based on field inventory within
the Beira corridor region. Evaluation of the mathematical and statistical credibility of these
models, which was carried out on the construction dataset (verification procedure) and
independent dataset (validation procedure), gave satisfactory results. Moreover, when applied on
our data, these models were more appropriate for the Beira corridor than allometric models
found in the literature. However, application of both models should be restricted to the lowland
miombo type in the Beira corridor, not mountain miombo.
