Model-based prediction and ascription of deforestation risk within commodity sourcing domains: Improving traceability in the palm oil supply chain

Abstract

Palm oil accounts for approximately 50% of global vegetable oil production, and trends in consumption have driven large-scale expansion of oil palm (Elaeis guineensis) plantations in Southeast Asia. This expansion has led to deforestation and other socio-environmental concerns that challenge consumer goods companies to meet no deforestation and sustainability commitments. In support of these commitments and supply chain traceability, we seek to improve on the current industry standard sourcing model for ascribing social and environmental risks to particular actors. Using passive geolocational traceability data (n = 3,355,437 cellular pings) and machine learning models, we evaluate the industry standard sourcing model, and we predict with high accuracy the undisclosed sourcing domains for 1,570 Indonesian and Malaysian palm oil mills on the Universal Mill List (as of November 2021). In combination with the World Wide Fund for Nature - Netherlands Forest Fore-sight, we use our predicted sourcing domains to provide an illustrative example of the proportional allocation of future deforestation, carbon loss, and biodiversity risk to relevant actors, permitting targeted outreach, contract negotiation, and mitigation of large-scale resource degradation. This example depends on a subset of model predictions in the absence of disclosed traceability data. The utilization of additional predictions or disclosed traceability data would influence and improve the results.