Quantifying Patterns and Impacts of Land-Use Change within a Major Amazonian Agricultural Frontier

Noble, Ciar Dorje (2025) Quantifying Patterns and Impacts of Land-Use Change within a Major Amazonian Agricultural Frontier. Doctoral thesis, University of East Anglia.

PDF Restricted to Repository staff only until 30 September 2927. Request a copy

Abstract

Deforestation and forest fragmentation are among the greatest threats to tropical forest biodiversity, yet their large-scale impacts remain difficult to quantify. This thesis bridges the gap between localised ecological field data and regional land-use changes to predict biodiversity loss within the >113,000 km2 Portal da Amazˆonia (PdA; northern Mato Grosso, Brazil), one of the most extensively deforested agricultural frontiers in Amazonia. Across five data chapters, I apply diverse methodologies to provide novel insights into the impacts of anthropogenic activities on biodiversity and landscape structure. First, I evaluate the impact of imperfect detection on species richness and beta-diversity estimates within fragmented landscapes, revealing that while observed species counts underestimate impacts of reductions in forest patch size, Hierarchical Multi-Species Occupancy Models (HMSOMs) can largely overcome these biases. I then apply HMSOMs to fragmented dung beetle assemblages, identifying links between species’ forest habitat specificity and fragmentation sensitivity, and a critical threshold in patch size beyond which specialists rapidly decline. Next, I examine associations between fragmentation and fires, finding that small forest patches (<100 ha) made disproportionately large contributions to annual forest fire cover throughout the PdA, likely due to the increased influence of desiccating edge-effects. I also show that land tenure shapes spatial patterns of forest disturbance in the PdA, with smallholder activities disproportionately linked to deforestation, fragmentation, and fire, while largeholders retain greater potential to preserve continuous forest tracts. Finally, I attempt to predict the community composition of seven vertebrate and invertebrate taxa across >50,000 km2, revealing that fragmentation amplified regional biodiversity loss resulting from deforestation, as forest specialist vertebrate richness declined faster than forest cover over 36 years. Together, these findings highlight the need for policies that integrate species-specific responses, land-use planning and fire management at regional scales, with emphasis on retention of large forest patches and restoration of landscape connectivity.

Item Type:	Thesis (Doctoral)
Faculty \ School:	Faculty of Science > School of Environmental Sciences
Depositing User:	Chris White
Date Deposited:	23 Dec 2025 08:59
Last Modified:	23 Dec 2025 08:59
URI:	https://ueaeprints.uea.ac.uk/id/eprint/101486
DOI:

Actions (login required)

View Item