Abstract
Based on UN Comtrade data (2004 to 2024), this paper constructs a directed weighted global corn trade network using complex network analysis to examine its spatiotemporal evolution, structural resilience, and recovery strategies. Findings reveal a multipolar trade pattern: exports shifted from the U.S. dominance toward competition among Brazil, France, South Africa, etc., while imports diversified from traditional Asian countries to include Mexico, Spain, Egypt, etc. The network became tighter and more efficient, with increasing density, clustering, average degree, and shorter average path length, showing small-world properties. The U.S., France, and South Africa exhibited strong intermediary control, Europe maintained influence advantages, Turkey and Thailand’s hub roles were enhanced. Resilience tests show improved resistance to random failures but high vulnerability to targeted attacks—core node failure causes sharp efficiency drops. Edge attacks are less disruptive due to route redundancy. Simulations show strong self-recovery under random shocks, yet deterministic hub attacks cause cliff-like collapse, requiring export diversification. This study provides a basis for trade evolution, risk identification, and supply chain resilience.
Download PDF
Full Article
Resilience Assessment and Recovery Strategy of Global Corn Trade Network
Jingyi Wang, Shitong Guo, and Jinzhuo Wu *
Based on UN Comtrade data (2004 to 2024), this paper constructs a directed weighted global corn trade network using complex network analysis to examine its spatiotemporal evolution, structural resilience, and recovery strategies. Findings reveal a multipolar trade pattern: exports shifted from the U.S. dominance toward competition among Brazil, France, South Africa, etc., while imports diversified from traditional Asian countries to include Mexico, Spain, Egypt, etc. The network became tighter and more efficient, with increasing density, clustering, average degree, and shorter average path length, showing small-world properties. The U.S., France, and South Africa exhibited strong intermediary control, Europe maintained influence advantages, Turkey and Thailand’s hub roles were enhanced. Resilience tests show improved resistance to random failures but high vulnerability to targeted attacks—core node failure causes sharp efficiency drops. Edge attacks are less disruptive due to route redundancy. Simulations show strong self-recovery under random shocks, yet deterministic hub attacks cause cliff-like collapse, requiring export diversification. This study provides a basis for trade evolution, risk identification, and supply chain resilience.
DOI: 10.15376/biores.21.4.12104-12120
Keywords: Corn; Complex network; Spatiotemporal evolution; Resilience
Contact information: School of Civil Engineering and Transportation, Northeast Forestry University, Harbin 150040, China; *Corresponding author: [email protected]
INTRODUCTION
Corn is a globally vital crop for food, feed, and industry, and one of the most actively traded agricultural commodities. In the past two decades, globalization and regional economic integration have driven sustained growth in global corn production, consumption, and trade. Increasing cross-border exchanges have made international corn flows more complex and interconnected. Meanwhile, global food security has faced growing uncertainty due to natural disasters, geopolitical conflicts, trade sanctions, and other external shocks. As an important globally traded agricultural commodity, corn is vulnerable to disruptions affecting major producing, exporting, and importing countries. Therefore, understanding the structural evolution of the global corn trade system and evaluating its ability to withstand external shocks are important for global and regional food security (Sartori and Schiavo 2015).
Advances in systems science and big data have established complex network analysis as an effective framework for examining the systemic structure of transnational commodity trade (Newman 2003; Garlaschelli and Loffredo 2005). In this framework, countries or regions can be represented as nodes, while trade relationships are represented as directed and weighted edges according to the direction and volume of commodity flows (Garlaschelli and Loffredo 2009). Previous studies have mainly examined the structural characteristics and evolution of international trade networks, identifying features such as heterogeneous connectivity, small-world properties, and core-periphery structures (Watts and Strogatz 1998; Barabási and Albert 1999; Serrano and Boguñá 2003; Tzekina et al. 2008). These structural characteristics indicate that international agricultural trade is not composed of independent bilateral relationships but rather forms an interconnected system in which disruptions to important nodes or trade relationships may propagate through the network.
With the development of temporal network analysis, increasing attention has been paid to the dynamic evolution and risk characteristics of international trade networks. Existing studies have developed various approaches for measuring temporal centrality and identifying critical nodes (Rocha and Masuda 2014; Taylor et al. 2017; Bi et al. 2021). In agricultural trade, Aparicio and Pinilla (2019) analyzed the structural characteristics of grain trade and the evolution of import and export relationships, while Wang et al. (2018) constructed agricultural trade networks for different commodities and found that highly connected countries often act as important hubs. Gutierrez et al. (2021) further demonstrated the pronounced core-periphery structure of the global wheat trade network (Fair et al. 2017). More recently, Wu et al. (2026) examined the evolution and risk propagation of the global major wooden furniture trade network and showed that targeted disruptions to core trading nodes can substantially reduce network efficiency and connectivity. These studies provide an important basis for understanding the structural evolution and vulnerability of global commodity trade networks (Niu et al. 2024; Zuo et al. 2024).
However, existing research has focused primarily on network structure and its temporal evolution, while relatively limited attention has been paid to the resilience and recovery capacity of global agricultural trade networks under external disturbances. Structural connectivity and network efficiency do not necessarily imply strong resilience (Gao et al. 2016). Increasing interdependence may improve the efficiency of international commodity circulation, but it may also increase the systemic consequences of disruptions affecting highly influential countries or trade relationships (Li et al. 2020). Therefore, evaluating network resilience is essential for determining whether an increasingly integrated trade system can maintain its essential functions and recover after external shocks (Li et al. 2024).
Network resilience refers not only to a system’s ability to resist and absorb disturbances but also to its capacity to recover after damage (Nan and Sansavini 2017; Tendall et al. 2015; Liu et al. 2022). Resilience assessment commonly evaluates network changes under different disturbance scenarios, such as random failures and targeted attacks on critical nodes or edges (Sharifi 2019; Wei and Xiu 2020). Random disturbances can represent dispersed or unexpected disruptions, whereas targeted attacks can reflect shocks concentrated on structurally important components. Comparing these scenarios can reveal the vulnerability of the trade network and identify critical nodes that may generate disproportionate systemic effects. Notably, the global corn trade network displays an asymmetric resilience profile, demonstrating high robustness against random failures but extreme vulnerability to targeted attacks (Albert et al. 2000). The recovery mechanisms of the global corn trade network following different types of disturbances remain insufficiently explored.
Against this background, this study constructs a directed and weighted global corn trade network based on UN Comtrade data from 2004 to 2024. The study first investigates the spatiotemporal evolution of the network and changes in its structural characteristics and core nodes. It then evaluates network resilience under different node- and edge-attack scenarios by examining changes in network efficiency and structural connectivity. Finally, recovery processes under different recovery strategies are analyzed to assess the network’s ability to restore its functionality following disruption. Through these analyses, this study aims to identify the structural characteristics and critical components of the global corn trade network, clarify its vulnerability to different external shocks, and examine its recovery capacity. The findings are expected to provide a network-based basis for identifying systemic risks and improving the resilience and stability of global corn trade.
METHODS AND DATA SOURCES
Construction of Complex Networks and Analysis Metrics
Based on Watts and Strogatz’s (1998) method for measuring the global extensiveness index of complex networks, this study constructs a global corn trade network, identifies core trading countries through node strength, and analyzes the topological structure characteristics and dynamic evolution of the global corn trade network. The countries (regions) involved in the global corn trade were the nodes in the network, and the trade relations between these countries (regions) were the edges in the network. Taking the trading volume of corn trade between countries (regions) as an index to measure the degree of tightness between network nodes, this paper constructs a corn trade network model G. G = {V, A, W}, where V = {v1 ,v2 ,v3…vn} represents the collection of countries (regions) participating in the world corn trade; A = {aij} represents the set of sides constituting the world food trade network. aij represents the trade relationship between countries i and j. When there is trade between two countries, aij = 1, otherwise, aij = 0; W = {wij} indicates the weight value. The following are the main measurement indicators of the global corn trade network model.
(1) Node strength
Node strength refers to the sum of the weights of all edges starting from the node or pointing to the node, reflecting the total trade volume output or received from the node, that is, the total trade volume of goods or services exported or imported by the country or region to other trading partners. Node strength includes output strength and input strength, which can be calculated as follows:
(2) Average degree
Average degree refers to the average degree of all nodes in the network, which can reflect the average connectivity of nodes in the network, that is, how many other nodes are connected to each node on average. The expression is as follows,
where m is the number of trade relations, and n is the total number of countries (regions).
(3) Network density
Network density is a measure of the ratio of the number of directed connections that actually exist in the network to the maximum number of directed connections that may exist in theory. It describes the degree of tightness between nodes. The expression is as follows,
where D is the network density, L is the actual number of directed edges, and N is the number of nodes in the network.
(4) Average clustering coefficient
The average clustering coefficient is the average of the clustering coefficients of all nodes, which is used to measure the clustering degree of nodes in the network. It reflects the average probability of trade between trading partners in the trade network. The greater the value, the higher the degree of aggregation of trading partners. The average clustering is computed as follows,
where Ti is the actual number of edges between the neighbors of node i, and ki is the number of trading partners.
(5) Average path length
Average path length reflects the efficiency of resource and information transmission between nodes in the trade network. Because the edge weight in this study represents the corn trade volume between countries or regions, a larger weight indicates stronger trade connectivity rather than a longer network distance. Therefore, to apply shortest-path analysis to the weighted trade network, the trade volume is transformed into an effective distance using its reciprocal. Specifically, for a directed trade relationship from node i to node j, the effective distance is defined as follows,
where wij represents the corn trade volume from country or region i to country or region j, and dij represents the corresponding effective trade distance. Thus, larger trade volumes correspond to shorter effective distances, indicating stronger and more accessible trade connections. The shortest path between two nodes is determined by minimizing the sum of effective distances along all feasible directed paths connecting them.
Based on this transformation, the average path length (L) is calculated as follows:
where dij represents the shortest path length between node i and node j.
(6) Betweenness centrality
The betweenness centrality measures the frequency of a node in the shortest path of all node pairs. Specifically, it indicates the extent to which a node plays an intermediary role in connecting other node pairs. Nodes with higher betweenness centrality usually have stronger control and influence in the network. The expression is as follows,
(7) Eigenvector centrality
Eigenvector centrality is an index to measure the importance of nodes. It considers not only the number of connections of nodes, but also the importance of their neighbors. This centrality measurement method can better reflect the influence of nodes in the network, because the nodes connected with high importance nodes will also be considered more important. The expression is as follows:
Assessment of Trade Network Resilience
In this study, the complex network analysis method was used to establish a multi-scenario simulation framework to quantify the structural resilience of the global corn trade system. Through constructing two modes of random attack and deliberate attack, the stability and adaptability of the global corn trade network under external pressure are systematically evaluated. Taking the network efficiency, the proportion of largest connected subgraphs, the proportion of isolated nodes, the average clustering coefficient, and the average degree as resilience assessment indexes, the impact mechanism of random attack and deliberate attack on the corn supply network was evaluated, and the scene model of interruption interference was established to simulate the dynamic response characteristics of corn supply network under interruption interference. In random attack, nodes are randomly selected to attack (i.e., remove nodes), which simulates the situation that the network suffers from aimless and random damage, and reflects the robustness of the network to random failures. In deliberate attacks, key network hubs identified by degree centrality are removed sequentially to simulate the propagation of targeted shocks, such as geopolitical conflicts and trade sanctions. Through calculating the change rule of the resilience assessment index, this paper deeply analyzes the cascading collapse effects that may be caused by the failure of the core node, so as to reveal the potential vulnerable links and resilience characteristics of the trade network. The following are the main resilience assessment indicators.
(1) Overall network efficiency
The overall network efficiency is the average value of the reciprocal of the shortest path between all node pairs in the network. It measures the efficiency of information transmission in the network and reflects the efficiency of communication between all node pairs in the network. A higher overall efficiency results in more efficient information transmission in the network, that is, the difficulty of information transmission between nodes. The expression is as follows,
where dij represents the shortest path length between node i and node j.
(2) Largest connected subgraph
The largest connected subgraph refers to the connected component with the largest number of nodes in an undirected graph. Specifically, if a subgraph formed by a subset V’ of the vertex set V of G satisfies the condition that there exists a path between any two vertices, and there is no connected subgraph encompassing more vertices, then this subgraph is called the largest connected subgraph. In directed graphs, the concept of the largest connected subgraph can be extended to the largest strongly connected subgraph, which means that there exists a bidirectionally reachable path between any two nodes in the subgraph. The relative size of the largest connected subgraph is often used to measure the invulnerability or connectivity of a network, and its expression is as follows,
where S represents the number of nodes in the largest connected subgraph, and S0 denotes the initial number of nodes in the network.
Resilience Recovery Strategy of Global Corn Trade Network
The evaluation of network structure resilience mainly simulates the resistance of the network to external interference and the resilience of the network to the initial state after the disturbance stops. Dynamic interrupt simulation continuously disturbs the network in each time period according to the rules through programming. Due to the correlation between trade supply and demand, the interference will have a ripple effect and affect the operation efficiency of the network. After the disturbance, the failed nodes and edges are restored according to the rules, and the network performance changes dynamically in this process (Sterbenz et al. 2010). Generally speaking, the external disturbance of the trade network is mostly from the randomness and deliberate attack.
Random disturbance may randomly select a certain proportion of nodes to damage them, such as natural disasters, epidemics, and other non-human and unpredictable impacts. Deterministic disturbances involve purposeful attacks on certain nodes, such as geopolitical conflicts and trade frictions, which may lead to failure of some nodes.
The relevant rules are as follows: (1) Random attack: nodes or edges are randomly removed. Recovery follows first-failure-first recovery, so this scenario represents random recovery. (2) Deliberate node attack: nodes are removed in descending order of degree centrality, out-degree, in-degree, or betweenness centrality. Nodes with higher centrality fail earlier. Because recovery follows first-failure-first recovery, these nodes are also restored earlier. This scenario therefore represents high-centrality-node-first recovery. (3) Deliberate edge attack: edges are removed in descending order of trade volume. Edges with larger trade volumes fail earlier and are restored earlier under first-failure-first recovery. This scenario therefore represents key-corridor-first recovery. (4) Disturbance scale: when nodes are disturbed, the number of failures is about 30% of all nodes, approximately 60 nodes. When edges are disturbed, the specified number of failures is 175. The economic interpretation is that core exporters, key transit nodes, and major trade corridors are usually prioritized in real-world recovery.
The simulations therefore compare random recovery with high-centrality-node-first recovery and key-corridor-first recovery. In practice, node removal represents the temporary loss of a country’s effective participation in corn trade rather than its physical disappearance. This may result from geopolitical conflicts, trade restrictions, or insufficient port capacity.
Data Sources and Processing
Based on the United Nations Comtrade Database (https://comtradeplus.un.org/), this paper selects corn (HS 1005) from 2004 to 2024 as the research object, uses physical units of mass (tonnes) to quantify the trade flow, collects and organizes the data, and constructs a global corn trade network encompassing 220 countries (regions), forming a dynamic trade matrix. By setting three characteristic time sections in 2004, 2014, and 2024, the study reveals the evolution law, resilience assessment, and recovery strategy of the global corn trade network.
RESULTS AND DISCUSSION
Changes of Core Nodes of the Global Corn Trade Network
Table 1 summarizes the shifting roles of major nations in the global corn trade. The decline of U.S. export dominance and the rise of Brazil are driven by a combination of factors. The U.S.–China trade friction prompted China to diversify its corn imports, while Brazil expanded corn production through double-cropping and improved logistics. The Black Sea conflict disrupted Ukraine’s exports, further accelerating Brazil’s rise. On the import side, Mexico, Spain, Egypt, and China increased their purchases due to growing feed demand and food security concerns.
These shifts reflect not only market competition but also geopolitical realignment and national strategies to reduce dependence on single suppliers. Changes in corn import patterns may also reflect variations in livestock-feed demand, population, income, and dietary patterns; however, these demand-side factors are not directly examined in the present network model.
Table 1. Top 10 Nodes in Node Strength (Unit: 107t)
Evolution Characteristics of the Global Corn Trade Network
(1) Evolution of network scale
The trajectory of the global corn trade network reveals a decisive shift toward systemic integration. As shown in Fig. 1, the number of participating nations increased from 195 to 220, yet the growth in trade linkages was even more pronounced, rising from 1,378 to 2,225. This gap between node growth and link accumulation points to a marked increase in network density, reflecting more frequent and complex exchanges among trading partners. The data indicate that global corn trade has evolved from isolated regional interactions into a tightly woven, interdependent structure. This intensified connectivity reflects broader shifts, including trade policy liberalization, logistics optimization, and the global imperative to strengthen food security through diversified supply sources.
Fig. 1. Evolution of global corn trade scale during the period of 2004 to 2024
The global corn trade network evolution maps are delineated using Gephi (0.9.2) network visualization software based on corn trade flow data exceeding 10,000 tons each year (Fig. 2). The spatiotemporal evolution of the global corn trade network reveals a fundamental transition from a localized, unipolar structure into a highly integrated, polycentric global system. In 2004, the network exhibited a distinct core-periphery architecture heavily dominated by the U.S. as the primary export hub, while import demands were concentrated among a few traditional Asian nations. By 2014, this rigid structure began to decentralize as South American producers, particularly Brazil and Argentina, substantially expanded their market shares, establishing new regional trade clusters and driving up overall network density. By 2024, this multipolar paradigm consolidated further. Driven by accelerating import demands across developing economies and the rising prominence of emerging suppliers from Eastern Europe and the Black Sea basin, trade routes have become substantially denser, more intricate, and widely distributed. On a systemic level, these shifting topological patterns illustrate deep market integration and a structural reorganization. This evolution effectively reduces single-node dependencies by embedding operational redundancy, thereby enhancing the overall resilience and substitution capacity of the global agricultural supply chain.
Fig. 2. Evolution of the global corn trade network during the period of 2004 to 2024
(2) Evolution of network structural characteristics
Table 2 details the structural evolution of the global corn trade network over the past two decades, indicating a clear trajectory toward systemic integration and heightened efficiency. Over this period, the network exhibited consistent growth in connectivity and compactness. Network density rose approximately 30%, shifting from 0.034 to 0.045, which signals an expanding involvement of peripheral trading nations. Concurrently, the average degree experienced a threefold increase, surging from 9.415 to 31.934. This sharp rise points to a substantial expansion of active trading partnerships worldwide.
The topological changes also highlight robust small-world properties within the global supply chain. The average path length decreased from 2.876 in 2004 to 2.524 in 2024, indicating improved network connectivity and trade transmission efficiency. However, this decline may also reflect increasing dependence on key trading hubs. Therefore, the reduction in path length should be interpreted together with changes in node centrality. Meanwhile, the average clustering coefficient remained consistently high, moving from 0.372 to 0.408, which underlines the stable cohesion of localized, triadic trading groups. Although the slight drop in this coefficient between 2014 and 2024 reflects the emergence of long-range cross-regional linkages, the overall slowing growth rates suggest that the network density is nearing structural saturation. Future network evolution is therefore expected to shift from quantitative expansion toward qualitative optimization driven by coordination among core nations.
Table 2. Evolution of Measurement Indicators of Global Corn Trade Network
(3) Network centrality analysis
Tables 3 and 4 track the shifting landscape of country-specific roles within the global corn trade network, evaluated through betweenness and eigenvector centralities. Betweenness centrality metrics reveal that traditional hub nations, specifically the U.S., France, and South Africa, consistently maintained their leading intermediary control over the two-decade period.
Table 3. Measurement of Global Corn Trade Network in Terms of Betweenness Centrality
Although the betweenness centrality of the U.S. declined over time, it remained among the highest in the network, indicating its continued importance as a key intermediary. This suggests that the increasing efficiency of the network may coexist with persistent dependence on major hubs, which could amplify the impact of disruptions to critical nodes. Concurrently, the network witnessed the structural rise of emerging economies, as seen when Turkey climbed significantly to secure the fourth position, highlighting its growing function as a vital crossroads connecting Europe and Asia. Other nations, including Thailand, Serbia, Brazil, Spain, and China entered the top 10 intermediary centers for the first time, underscoring a broader expansion of Asian and emerging market influence.
In terms of eigenvector centrality, which reflects a node’s influence based on the importance of its trading partners, European countries maintained a dominant presence despite internal structural realignments. France consistently held the top rank, while the Netherlands achieved the fastest growth in influence to secure the second position. Concurrently, nations, such as Poland and Hungary, advanced into the top ten, indicating the rising integration and status of Eastern Europe within the trade system. Although the relative influence of traditional Western European hubs like Germany experienced a slight decline, the gradual inclusion of the U.S. into the core eigenvector circle further illustrates a transition toward a more polycentric network. Collectively, these dynamic shifts demonstrate that while traditional powers still anchor the network, the rise of emerging hubs has fostered a more diversified and distributed structure, effectively mitigating single-country dependencies.
Table 4. Measurement of Global Corn Trade Network in Terms of Eigenvector Centrality
Evolution and Analysis of Structural Resilience of Global Corn Trade Network
(1) Resilience assessment of global corn trade network under node attack
The structural resilience of the global corn trade network under random and deliberate node attacks is evaluated using four indicators: network efficiency, the proportion of the largest strongly connected subgraph (LSCC), the proportion of isolated nodes, and the average clustering coefficient (Fig. 3). Baseline comparisons across the characteristic cross-sections indicate that initial network efficiency, connectivity, and local clustering have enhanced over the two-decade period, reflecting a highly integrated trading system. However, the response of the network to node removal varies substantially between the two attack modes, demonstrating a clear asymmetric resilience profile.
Under targeted deliberate attacks, the network exhibits high vulnerability to the removal of core hubs, specifically major exporters such as the U.S. and Brazil. Network efficiency and the LSCC proportion undergo a precipitous decline at low attack thresholds, confirming that the trade network possesses a prominent scale-free property with a heavy reliance on central hubs. Concurrently, the growth rate of isolated nodes accelerates over the studied period, as peripheral importing nations rapidly lose connection pathways during the initial phases of targeted disruption. This trend highlights the intensifying polarization of the core-periphery structure, wherein disruptions at major export centers pose significant risks to the supply security of vulnerable edge nodes. Additionally, although the network forms denser local trade clusters over time, targeted attacks quickly dismantle these cohesive groups, causing a sharp initial drop in the average clustering coefficient. Conversely, the network exhibits strong robustness against random failures, as all evaluated indicators deteriorate gradually and linearly even under high disruption intensities.
Fig. 3. Variations of network efficiency with attack intensity from 2004 to 2024 (values are averages over 100 simulations)
(2) Resilience assessment of global corn trade network under edge attack
Figure 4 illustrates the topological variations of the global corn trade network under random and deliberate edge attacks, simulating the obstruction of specific trade corridors. Systemic response is evaluated via network efficiency, the largest connected subgraph (LSCC) proportion, average clustering coefficient, and average degree.
Under deliberate attacks, network efficiency exhibits non-linear degradation, initially decreasing gradually and decelerating rapidly after exceeding a 40% attack proportion to retain only 30% of its baseline value at 80% disruption. Conversely, under random failures, the network retains over 70% efficiency even at 60% link removal, indicating that targeted disruptions of critical corridors severely impair transmission capacity. The LSCC proportion exhibits a distinct threshold effect under both modes. The breakdown threshold for deliberate attacks occurs at approximately 45%, noticeably earlier than under random failures, demonstrating that targeted edge removal accelerates fragmentation. For the average clustering coefficient, deliberate attacks cause a steeper decline than random disruptions, yet the reduction remains moderate, retaining 55% of its baseline value at 50% attack intensity, which implies that localized clustering structures buffer link losses. Meanwhile, the average degree decreases identically and linearly in both modes. Although an increased initial average degree over time reflects dense trade integration, it implies greater absolute connectivity losses under identical attack intensities.
Collectively, while the network exhibits redundancy and substitution capabilities against route disruptions, the lower collapse threshold under targeted edge attacks underscores the structural importance of key trade corridors.
Fig. 4. Changes of network eigenvalues under edge attack mode from 2004 to 2024 (values are averages over 100 simulations)
Recovery Strategy of Global Corn Trade Network
The dynamic responses of the global corn trade network efficiency to various node-based disturbances and subsequent recovery processes in 2024 are presented in Fig. 5. The network efficiency declines steadily under all scenarios, reaching a minimum at the end of the disturbance phase before gradually recovering. The impact of random node removal is noticeably less severe than that of targeted attacks on core nodes.
Fig. 5. Changes of network efficiency to random and deliberate node attacks in 2024 (values are averages over 100 simulations)
Fig. 6. Changes of network efficiency to random and deliberate edge attacks in 2024 (values are averages over 100 simulations)
When key hubs are identified and removed based on degree centrality or betweenness centrality, the network efficiency experiences a rapid and substantial decrease. This indicates that the network is highly sensitive to the failure of central trading partners, as their removal disrupts information flow and results in systemic fragmentation. In contrast, the more stable recovery trajectory following random disturbances highlights the resilience of the network to unpredictable, non-human impacts. These findings indicate that the stability of primary export hubs is the decisive factor for maintaining the structural integrity of the global corn trade network. The recovery of trade network connectivity does not imply an immediate recovery of physical corn supply. Planting, crop growth, harvesting, storage, and transportation require additional time, creating an inherent delay between network recovery and actual supply recovery. Moreover, corn quality may deteriorate during prolonged storage or transportation due to moisture, pests, and fungal contamination. Therefore, the restoration of trade connections does not necessarily indicate an immediate restoration of effective corn supply.
The temporal dynamics of global corn trade network efficiency under various edge-based disturbances and subsequent recovery processes in 2024 are depicted in Fig. 6. As trade connections are systematically removed, network efficiency declines gradually, reaching its minimum at the end of the disturbance phase before initiating a steady recovery. In comparison to node-based attacks, edge-based disturbances result in a more moderate and linear reduction in network performance. This suggests that the network maintains a degree of structural redundancy, allowing it to preserve core functionality through alternative trade routes despite the removal of individual edges. The observed recovery trajectories indicate that the network is capable of effectively restoring its performance levels once external disruptions are ceased. These findings emphasize that the presence of diverse and redundant trade channels is essential for ensuring the overall robustness of the global corn trade network.
In addition to trade connectivity, the resilience of corn supply can be affected by upstream agricultural inputs, environmental conditions, and transportation constraints. For example, fertilizer availability, soil conditions, maritime shipping, and port capacity may influence corn production and delivery. These factors are not explicitly incorporated into the present trade network model and should be considered in future assessments of global corn supply resilience.
CONCLUSIONS
- The global corn trade network has evolved into a more integrated, polycentric system characterized by small-world properties. However, this progress is accompanied by a deepening core-periphery polarization, where peripheral countries remain highly dependent on a limited number of central export hubs.
- While traditional powers such as the U.S. maintain significant intermediary control, their relative dominance has diminished. The emergence of new hubs, such as Turkey, Thailand, and Brazil, alongside the rising influence of Eastern European nations, indicates a shift toward a more diversified and distributed trade structure.
- The network exhibits high robustness against random failures, benefiting from increased connectivity and local clustering. Conversely, it shows high vulnerability to deliberate attacks on core nodes. The failure of these central hubs can trigger systemic collapse, reflecting the inherent risks of a concentrated trade structure.
- Simulations show that the network is more sensitive to node loss than to edge disruption. Recovery follows first-failure-first recovery, so prioritizing core nodes and key corridors shortens recovery time but may reinforce hub dependence. Enhancing resilience therefore requires diversified export sources and import channels, protection of critical nodes and corridors, and redundant routes.
ACKNOWLEDGMENTS
The authors would like to acknowledge the support of the Joint Project of the Natural Science Foundation of Heilongjiang Province (LH2021G001).
REFERENCES CITED
Albert, R., Jeong, H., and Barabási, A.-L. (2000). “Error and attack tolerance of complex networks,” Nature 406(6794), 378-382. https://doi.org/10.1038/35019019
Aparicio, G., and Pinilla, V. (2019). “International trade in cereals and the collapse of the first wave of globalization,” Journal of Global History 14(1), 44-67. https://doi.org/10.1017/S1740022818000347
Barabási, A.-L., and Albert, R. (1999). “Emergence of scaling in random networks,” Science 286(5439), 509-512. https://doi.org/10.1126/science.286.5439.509
Bi, J., Jin, J., Qu, C., Zhan, X., Wang, G., and Yan, G. (2021). “Temporal gravity model for important node identification in temporal networks,” Chaos, Solitons and Fractals 147, article 110934. https://doi.org/10.1016/j.chaos.2021.110934
Fair, K. R., Bauch, C. T., and Anand, M. (2017). “Dynamics of the global wheat trade network and resilience to shocks,” Scientific Reports 7(1), article 7177. https://doi.org/10.1038/s41598-017-07202-y
Gao, J., Barzel, B., and Barabási, A.-L. (2016). “Universal resilience patterns in complex networks,” Nature 530(7590), 307-312. https://doi.org/10.1038/nature16948
Garlaschelli, D., and Loffredo, M. I. (2005). “Structure and evolution of the world trade network,” Physica A: Statistical Mechanics & Its Applications 355(1), 138-144. https://doi.org/10.1016/j.physa.2005.02.075
Garlaschelli, D., and Loffredo, M. I. (2009). “Generalized Bose-Fermi statistics and structural correlations in weighted networks,” Physical Review Letters 102(3), Article ID 038701. https://doi.org/10.1103/PhysRevLett.102.038701
Gutierrez, E., Adenso-Diaz, B., and Lozano, S. (2021). “Analysis and vulnerability of the international wheat trade network,” Food Security 13(1), 113-128. https://doi.org/10.1007/s12571-020-01117-9
Li, Y., Zobel, C. W., Seref, O., and Chatfield, D. (2020). “Network characteristics and supply chain resilience under conditions of risk propagation,” International Journal of Production Economics 223, article 107529. https://doi.org/10.1016/j.ijpe.2019.107529
Li, Z., Zhuoming, R., Zhao, Z., and Weng, W. (2024). “Topological perturbations on resilience of the world trade competition network,” Humanities and Social Sciences Communications 11(1), 1-9. https://doi.org/10.1057/s41599-024-03045-x
Liu, X., Li, D., Ma, M., Szymanski, B. K., Stanley, H. E., and Gao, J. (2022). “Network resilience,” Physics Reports 971, 1-108. https://doi.org/10.1016/j.physrep.2022.04.002
Nan, C., and Sansavini G. (2017). “A quantitative method for assessing resilience of interdependent infrastructures,” Reliability Engineering & System Safety 157,35-53. https://doi.org/10.1016/j.ress.2016.08.013
Newman, M. E. J. (2003). “The structure and function of complex networks,” SIAM Review 45(2), 167-256. https://doi.org/10.1137/S003614450342480
Niu, X. Y., Yu, Z. Y., Chen, H. Y., and Chen, W. (2024). “Change and resilience of the energy trade backbone network in the Belt and Road region,” Resources Science 46(11), 2137-2149. https://doi.org/10.18402/resci.2024.11.04
Rocha, L. E. C., and Masuda, N. (2014). “Random walk centrality for temporal networks,” New Journal of Physics 16(6), Article ID 063023. https://doi.org/10.1088/1367-2630/16/6/063023
Sartori, M., and Schiavo, S. (2015). “Connected we stand: A network perspective on trade and global food security,” Food Policy 57, 114-127. https://doi.org/10.1016/j.foodpol.2015.10.004
Serrano, M. A., and Boguñá, M. (2003). “Topology of the World Trade Web,” Physical Review E 68(2), Article ID 015101. https://doi.org/10.1103/PhysRevE.68.015101
Sharifi, A. (2019). “Resilient urban forms: A review of literature on streets and street networks,” Building and Environment 147, 171-187. https://doi.org/10.1016/j.buildenv.2018.09.040
Sterbenz, J. P. G., Hutchison, D., Çetinkaya, E. K., Jabbar, A., Rohrer, J. P., Schöller, M., and Smith, P. (2010). “Resilience and survivability in communication networks: Strategies, principles, and survey of disciplines,” Computer Networks 54(8), 1245-1265. https://doi.org/10.1016/j.comnet.2010.03.005
Taylor, D., Myers, S. A., Clauset, A., Porter, M. A., and Mucha, P. J. (2017). “Eigenvector-based centrality measures for temporal networks,” Society for Industrial and Applied Mathematics 15(1), 537-574. https://doi.org/10.1137/16M1066142
Tendall, D. M., Joerin, J., Kopainsky, B., Edwards, P., Shreck, A., Le, Q. B., and Six, J. (2015). “Food system resilience: Defining the concept,” Global Food Security 6, 17-23. https://doi.org/10.1016/j.gfs.2015.08.001
Tzekina, I., Danthi, K., and Rockmore, D.N. (2008). “Evolution of community structure in the world trade web,” The European Physical Journal B 63(4),541-545. https://doi.org/10.1140/epjb/e2008-00181-2
Wang, X., Qiang, W., Niu, S. W., Liu, A. M., Cheng, S. K., and Li, Z. (2018). “Analysis of the global agricultural trade network and its evolution,” Journal of Natural Resources 33(6), 940-953. https://doi.org/10.31497/zrzyxb.20180403
Watts, D. J., and Strogatz, S. H. (1998). “Collective dynamics of small-world networks,” Nature 393, 440-442. https://doi.org/10.1038/30918
Wei, Y., and Xiu, C. (2020). “Study on the concept and analytical framework of city network resilience,” Progress in Geography 39(3), 488-502. https://doi.org/10.18306/dlkxjz.2020.03.013
Wu, J., Guo, S., Wang, J., and Ruffino, R. (2026). “Evolution and risk propagation mechanisms of the global major wooden furniture trade network,” Forest Products Journal 76(1), 110-123. https://doi.org/10.13073/FPJ-D-25-00053
Zuo, Z. L., Cheng, J. H., Guo, H. X., and Zhan, C. (2024). “Resilience evolution and evaluation of the trade network structure in the lithium industry chain,” China Population, Resources and Environment 34(02), 155-166. https://doi.org/10.12062/cpre.20230911
Article submitted: July 2, 2026; Peer review completed: September 12, 2026; Revised version received: October 14, 2026; Accepted: October 18, 2026; Published: October 30, 2026.
DOI: 10.15376/biores.21.4.12104-12120