General scientific consensus is growing that natural resources play an important role in violent conflicts. Both scarcity and abundance contribute to violent conflict, depending on the socio-economic and political conditions at play. Therefore, the criticality of natural resources might better describe the link to violent conflict than scarcity or abundance. Yet, research struggles to unravel the precise socio-environmental mechanisms that bring about natural resource conflicts. Consequently, interventions in natural resource management have shown limited effectiveness towards conflict prevention and peacebuilding.

The overarching aim of this thesis is to advance the understanding of mechanisms underlying violent natural resource conflicts towards effective prevention. This thesis focuses on violent natural resource conflicts within states since 1989, because with the end of the Cold War the character of armed conflicts turned increasingly intrastate. I pose the following questions: (1) What constitutes critical natural resources? (2) Are natural resources and violent conflicts related quantitatively, after considering their complex socio-economic and political contexts? (3) What are the core socio-environmental causal pathways linking natural resources and violent conflicts? (4) What opportunities are available to prevent violent conflict from resource use? A multi-method approach was applied, including systematic literature reviews, open-coding of text data, predictive modelling with logistic regression and machine learning techniques, causal network mapping, and a network analysis of the causal map.

The first paper of this thesis shows that existing descriptions of critical natural resources overemphasize their economic importance at the expense of ecosystem support and socio-cultural functions of natural resources; and that non-renewable resources are overrepresented compared to renewables. Therefore, it proposes a new, holistic, definition of criticality for natural resources, based on a hierarchy of human needs instead of only economic value. The second and third papers confirm the importance of natural resource variables in numerically predicting conflict, though their effects are often mediated by the intervening socio-economic variables. More specifically, renewables such as water and food are important predictors of conflict, while non-renewables are less important than prior research suggested. Of all socio-environmental interactions identified, food production interacts most strongly with its economic and demographic context. The fourth paper identified key structural points on the causal paths towards natural resource conflicts, which potentially function towards effective prevention. To reverse vicious circles of conflict to virtuous circles of peace, my analysis found that it is necessary to improve state and natural resource management institutions, balance resource availability and demand, limit population movements, and ensure adequate livelihoods. My analysis further found that resource exploitation and the revenues generated thereof, as well as environmental conservation and curbing environmental degradation can stabilize peaceful situations.