Files
Abstract
This dissertation comprises three integrated studies that examine the relationship between oil palm expansion and filovirus spillover risk in sub-Saharan Africa, combining remote sensing, spatial econometrics, and structural economic modeling.Chapter 2 presents a novel methodology for generating annual estimates of oil palm plantation establishment using MODIS satellite imagery. The study builds upon existing Landsat-based methods by addressing temporal data gaps through a machine-learning classification framework that utilizes XGBoost and multiple vegetation indices. Although the model does not estimate land cover independently, it successfully detects bare-soil signals indicative of plantation establishment across 17 African countries from 2000 to 2020. Validation against the Descals et al. (2024) dataset shows high temporal and spatial consistency, capturing post-2013 expansion in both smallholder and industrial plantations.
Chapter 3 quantifies the epidemiological effects of oil palm expansion using a panel of 10,674 grid cells. Spatial regressions demonstrate that industrial plantations significantly increase the probability of filovirus spillovers. In contrast, smallholder plantations reduce risk at low to moderate densities but lose this protective effect when overly clustered. These findings underscore the importance of plantation structure and scale in shaping disease ecology, suggesting that land-use configuration is a critical determinant of zoonotic emergence.
Chapter 4 incorporates these empirical insights into a spatial-dynamic bioeconomic model. The model captures household-level land allocation decisions under uncertainty and external health risks. Simulation results indicate that private land-use decisions diverge from the social optimum due to unpriced spillover externalities. A uniform tax of US$15.30 per metric ton of crude palm oil is shown to reduce palm oil output by only 1.5% while internalizing the health risk externality and generating US$86.70 million in public revenue and US$112 million in annual health benefits. These results highlight the efficiency of modest fiscal instruments in realigning incentives toward socially desirable outcomes.
Collectively, the three chapters present a data-driven policy framework to mitigate zoonotic disease spillover risk while promoting agricultural growth, thereby demonstrating the feasibility of aligning public health and environmental goals through targeted economic interventions.