Complex mosaic landscapes: novel methodologies to understand dynamics of cocoa farming in Ghana

Thesis

Complex mosaic landscapes are some of the most prolific and important in the tropics. Often consisting of a mixture of small scale agriculture, natural vegetation and settlements, these areas support the livelihoods of hundreds of thousands of people and are the backbone of some globally important crops, such as cocoa. These landscapes are inherently complex due to the presence and interaction of numerous social and environmental factors across multiple spatial and temporal scales. Given the importance and abundance of these landscapes, disproportionately little is known about the patterns and processes which sustain or disrupt them.

This thesis aims to advance the understanding of complex mosaic landscapes by using novel methodologies to capture and analyse the patterns and processes of these systems. Cocoa farming landscapes surrounding Kakum National Park in southern Ghana are used as a case study. Data were acquired in three distinct regions surrounding the park, namely the area surrounding the villages of Aboabo , Ahomaho and Kwameamoabeng.

This thesis first presents detailed methods for undertaking Unmanned Aerial Vehicle surveys using commercially available equipment with RGB cameras. Next a detailed account of how to processes this imagery using commercially available softwares is outlined. The first data chapter of this thesis examines using machine learning approaches to classify land use within complex mosaic landscapes only using RGB drone imagery. These first chapters highlight the significant insights generated by this thesis towards capturing the patterns of complex mosaic landscapes in the tropics. The later part of the thesis focuses on the dynamics and processes which govern complex mosaic landscapes by framing them as social-ecological systems. The second data chapter provides a detailed summary of land use in the area of study while also quantifying the impact of humans on the environment by calculating and comparing the Human Appropriated Net Primary Productivity between the three study regions. The final data chapter presents a study of the dynamics between humans and the environment during a climatic shock perturbation. Using combination of social, environmental and spatial data the resilience of households is assessed. Indicators of resilience are then compared at a regional scale to understand some of the multi-scale dynamics of resilience to the climate shock. The thesis is then concluded and future research directions are identified.

Authors: Moore, C

• DOI: 10.5287/ora-xrpbp6gar
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: cocoa; UAV; Ghana; El Nino; image classification; Human Appropriated Net Primary Productivity; tropical forest agriculture; social-ecological systems; drone; smallholder agriculture
• Subjects: Drone aircraft in remote sensing; Climate change adaptation; Cocoa; Remote-sensing images; Community-supported agriculture