Natural disasters not only pose challenges to global sustainable development but also impact the sustainable growth of regional economies. Existing research has mainly focused on the geographic spillover effects of natural disasters, with less attention to how economic ties between cities transmit and amplify these shocks. This paper utilizes China's geo-coded disaster data and enterprise-level investment records to combine spatial econometric models with improved spatial weight matrices and complex network methods to quantify the spillover effects of natural disasters in geographic and economic networks. The study results indicate that ignoring economic ties between cities leads to an overall economic loss underestimated by about 52.5% to 57.5%. In economic networks, cities with higher centrality or clustering coefficients significantly amplify negative spillover effects; At the same time, as the structure of intercity investment networks becomes increasingly complex, this amplifying effect is further amplified. This article points out that traditional disaster outcome assessments have ignored economic links between cities, and therefore advocates for developing cross-city disaster resilience policies, which should fully incorporate economic ties between cities while considering geographic proximity.

Figure 1: Spatial distribution of natural disaster events in China mainland

Meanwhile, research on complex networks shows that network structure and node status significantly affect the diffusion paths and amplification of external shocks, but the application of this perspective in disaster economics remains limited. Especially against the backdrop of China's unified market construction and rapid development of urban agglomerations, cities have become deeply embedded in cross-regional investment networks, with core cities playing a hub role in the economic network. The impact of natural disasters can spread across regions along economic ties, and may even be further amplified through high-centrality nodes, breaking traditional geographic boundaries. If disaster impact is assessed solely based on geographic proximity, it may both systematically underestimate overall economic losses and affect the rational allocation of disaster prevention resources. Therefore, it is necessary to incorporate intercity economic networks into the analytical framework and re-examine the trans-urban spillover mechanisms of natural disasters and their amplification effects.

Figure 2: Spatial evolution of intercity investment network flows in Chinese mainland

Figure 3: Benchmark empirical results

Second, neglecting economic network connections can systematically underestimate total economic losses. When only the geographic adjacency matrix is used to measure spillover effects, the estimate results are significantly smaller than when economic network connections are also included. The study found that ignoring the intercity investment network as an economic link would underestimate the total economic loss by about 52.5% to 57.5% (Figure 3d). This result directly indicates that economic networks constitute an important channel for the spread of natural disaster risks.

Third, network hub cities and highly clustered cities have an "amplifier" effect. At the node level, a city's position in the economic network determines its impact spillover capacity. Cities with high centrality or clustering coefficients experience significant negative spillover effects when natural disasters occur, exhibiting characteristics of "super-spread nodes" (Figure 4). Among them, the amplification effect of cities with high clustering coefficients is particularly pronounced, indicating that tightly connected economic subnetworks are more likely to form cascading shocks.

Figure 4 Results of the moderating effects of centrality and clustering coefficient

Fourth, as the structure of investment networks becomes increasingly complex, the spillover effects of disasters are intensifying. By dynamically estimating investment networks from different years as spatial weight matrices, the results show that from 2000 to 2017, the negative indirect effects transmitted through the economic network continued to strengthen (Figure 5a). Further analysis shows that the higher the overall structural indicators such as the number of network nodes, number of edges, network density, and clustering coefficient, the greater the spillover intensity of disaster shocks (Figures 5b, 5c). This indicates that deepening economic ties not only improves efficiency but also increases the system's sensitivity to external shocks.

Figure 5 Overall structure of intercity investment networks and results of negative spillover effects

Fifth, the spillover intensity varies among different types of natural disasters, but the overall conclusion is robust. The robustness test shows that high-frequency disasters such as floods and storms have more significant negative spillover effects on the regional economy, while droughts and heatwaves are more likely to have long-term suppressive effects. Regardless of the type of disaster, the additional spillover effects from economic network channels are significant, supporting the robustness of the benchmark results.

The study found that economic networks also constitute important risk transmission channels, and their amplifying effects should not be ignored. Compared to the relatively visible and predictable characteristics of geographic spillover paths, shock transmission in economic networks is often more hidden and may be amplified through supply chain "bull tail effects" or network cascading mechanisms. When cities located at the core of the network with high centrality or clustering experience disasters, their negative externalities are more likely to spread to related areas and even broader regions. As investment links between cities continue to intensify and economic networks become more complex, the systemic risks of disaster shocks also rise.

At the policy level, the study believes that allocating disaster prevention resources solely based on high-risk disaster areas may overlook the systemic importance of core nodes in the economic network. In the Chinese context, disaster-prone areas are concentrated in central and western regions, while economic network hubs are mostly located in core eastern cities. This spatial dislocation may lead to tensions in resource allocation. Therefore, disaster prevention and mitigation policies should focus not only on geographically high-risk areas, but also strengthen prevention and resilience building in core cities within economic networks, while promoting the establishment of cross-city collaborative governance mechanisms. More broadly, as the economic network structures of developing countries become increasingly complex, the impacts of natural disasters are more likely to spread from localized to nationwide. Therefore, economic interconnectedness should be incorporated into systematic assessment frameworks when formulating disaster response strategies.