Food Security and Resilience Implications of the Global Agricultural-Phosphorus Trade Network

With current global population estimates surpassing 9 billion by mid-century, a significant ramp in food production must follow. Most of this yield growth is expected to come from intensification of agriculture in developing... [ view full abstract ]

With current global population estimates surpassing 9 billion by mid-century, a significant ramp in food production must follow. Most of this yield growth is expected to come from intensification of agriculture in developing nations, requiring them to consume synthetic fertilizers at an accelerating rate. Synthetic phosphate fertilizers in particular represent over 70% of the P input to modern agriculture, and are subject to resource criticality, geopolitical conflict, and price volatility. Derived from highly geographically concentrated, non-renewable phosphate rock reserves, P fertilizers dictate some 30-50% of modern agricultural yields, and are a critical material in achieving global food security. Many developing countries – often those with little to no P rock reserves – also lack the P soil accumulation to sustain improved yields, and as such are heavily reliant on continued physical and economic access to these fertilizers via international trade. In attempting to answer questions of security of access to P fertilizers, Network Analysis (NA) offers a number of quantitative methods to pose such systemic questions.

Using principles from NA and Industrial Ecology, this body of work aims to better characterize the key relationships, underlying structure, and systemic vulnerabilities of the global agricultural phosphorus trade network (GAPTN). These goals first require the construction of a weighted, directed network of annual inter-country exchanges of all synthetic P fertilizers and their chemical precursors. In constructing edges, each directed annual commodity exchange is translated to its elemental P weight content, and summed with all other same-direction exchanges between a given pair of countries. Nodes are then simply those countries participating in P trade, importing and/or exporting. In constructing this model, this work aims to (a) investigate time-series features of the GAPTN’s topology over 1994-2014; (b) identify those countries and trade connections most critical to the short- and long-term GAPTN function; and (c) determine which countries are most dependent on the GAPTN’s proper function. We expect these analyses and questions may serve to inform engineering and policy measures that aim to ensure stable access to P nutrients in the developing world, and thereby advance global food security.

Trade partner connections within the GAPTN display mild heterophily, preferentially connecting nodes of a dissimilar number of trade connections. Statistical tests for distribution identification of node degrees suggest a set of heavy-tail distributions, as do the distributions of in-, out-, and total-strength over the given years. Said distributions suggest that a small subset of all the participating countries exchange the vast majority of P resources within the network, leaving it robust to random disruptions yet vulnerable to targeted ones. In conducting a robustness analysis, randomized and targeted removal of edges from the network indicates that some 80% of the GAPTN’s overall activity halts when some 5% of the most critical trade connections are dissolved. In light of these findings, international P management and local-scale closing of nutrient cycles can aid in improving food security while also minimizing the environmental impacts of excessive P fertilizer application.