Integrating Field and Remote Sensing Approaches to Quantify Forest Structure and Biomass in the Miombo Ecoregion.

Loading...
Thumbnail Image

Journal Title

Journal ISSN

Volume Title

Publisher

Namibia University of Science and Technology.

Abstract

Miombo ecoregion spans more than 3.8 million km² across eleven countries in eastern and southern Africa, forming the continent’s largest tropical dry forest system. The vegetation structure (VS) is shaped by strong rainfall gradients, marked seasonality, and recurrent fires, resulting in a heterogeneous mosaic of tall trees, multi-stemmed shrubs, and dense understorey layers. This structural complexity creates challenges for interpreting above-ground biomass (AGB) patterns and for accurately monitoring VS using global remote-sensing products. This thesis aimed to elucidate the influence of environmental gradients and VS on AGB partitioning and LiDAR canopy height estimation across the Miombo ecoregion using integrated field and remotely sensed data. I explored: (i) the influence of rainfall, disturbance, topography, and soils on biomass allocation across tree size classes and the contribution of large trees; (ii) the contribution of understorey vegetation (stem diameter < 5 cm) to total biomass and species diversity under contrasting rainfall conditions; (iii) the influence of canopy structural attributes on the subcanopy layer (vegetation height < 3 m) and associated ecological properties; and (iv) the extent to which structural and environmental gradients affect the accuracy of spaceborne LiDAR canopy height estimates. Rainfall emerged as the dominant driver of AGB, increasing from 17 to 133 Mg ha⁻¹ along a gradient of 345 to above 2000 mm mean annual precipitation. However, the biomass share of the largest 1% of stems indicated a slight decreasing trend along the gradient. AGB accumulation in Miombo, therefore, results from broad cohorts of medium and large stems rather than a few larger individuals as seen in humid forests. Elevation and fire reduced total AGB but did not diminish biomass stored in large trees, underscoring their resistance to disturbance. Incorporating understorey into AGB assessments revealed a substantial and previously overlooked carbon pool in dry woodlands. Small trees, saplings, and shrubs contributed up to 28.2% of total AGB in Namibia, compared with only 2.5% in wet DRC. Understorey composition shifted along the rainfall gradient, from shrub-dominated in dry sites to sapling-dominated in wet forests, illustrating how rainfall, fire, and canopy closure shape biomass partitioning and recruitment. Terrestrial laser scanning provided fine-scale insight into subcanopy patterns. Dense lower canopies promoted subcanopy density and biomass, whereas higher gap fraction reduced subcanopy density and xviii exerted non-linear effects on AGB. Greater variability in canopy cover reduced both species richness and biomass, highlighting the importance of canopy continuity in woodland. These structural gradients also conditioned spaceborne LiDAR performance. The Global Ecosystem Dynamics Investigation (GEDI) systematically underestimated canopy height (mean bias -2.79 m), with stronger underestimation in dry woodlands and slight overestimation in wetforests. Regionally calibrated Random Forest models integrating GEDI metrics, climatic and structural variables improved height prediction (explained variance: 64%; bias: –0.27 m), outperforming existing global products. This thesis demonstrates that Miombo biomass is distributed across broad demographic cohorts, that small-diameter vegetation constitutes an important yet overlooked carbon component, and that VS governs the performance of spaceborne LiDAR. These findings establish new empirical foundations for carbon accounting, forest monitoring, and remote-sensing applications in the region.

Description

Thesis/Dissertation submitted in fulfilment/ of the requirements for the Doctorate degree in PhD in Natural Resource Management and Ecology at the Namibia University of Science and Technology.

Citation

Mwamba, H. D. (2026). Integrating Field and Remote Sensing Approaches to Quantify Forest Structure and Biomass in the Miombo Ecoregion [Unpublished doctoral dissertation]. Namibia University of Science and Technology.

Endorsement

Review

Supplemented By

Referenced By