Urban Sprawl Undermines Agricultural Supply Chain Resilience: Amplifying Role of Resource Misallocation and Its Threshold Effects

Abstract

The acceleration of urbanization and its resultant expansion of urban sprawl (US) have significantly reshaped agricultural supply chain resilience (ASCR). In this study, we focused on the effects of US and the induction of imbalances in resource allocation. To accomplish this goal, we analyzed the data using a Geographically Weighted Random Forest (GWRF) with SHAP values, a moderated effects model, and Bayesian Kernel Machine Regression (BKMR). These analyses enabled us to evaluate how US undermines ASCR via resource misallocation (RM), with an emphasis on the nonlinear dynamics and spatial heterogeneity of RM. Our results indicate that an apparent nonlinear threshold effect exists between US and ASCR. When US remains below a critical threshold value of 0.33, moderate expansion enhances supply chain resilience. However, once this limit is surpassed, the negative impact of US intensifies exponentially. In addition, we found evidence of spatial disparities, with eastern coastal regions exhibiting more pronounced vulnerabilities, while specific megacity clusters displayed a positive response. RM functions as an amplifier, under conditions where capital distortion (CD) and land distortion (LAD) substantially magnify the marginal adverse effects of US at lower levels. In contrast, labor distortion (LD) demonstrates relatively stable effects, though these effects were spatially differentiated. Collectively, these distortions operate through a cyclical mechanism of “market inefficiency—spatial crowding—labor mismatch,” the net effect of which systematically erodes resilience. Compared with previous studies that mainly focus on linear mean effects or single-factor misallocations, in this study we employed a combined GWRF–SHAP and BKMR framework to systematically reveal the threshold-dependent spatially heterogeneous effects of US and the nonlinear amplification mechanisms of multidimensional RM. With this approach, quantifiable evidence for enhancing agricultural supply chain resilience management in developing countries can be achieved. Accordingly, policy recommendations are proposed to underscore the urgency of establishing urban growth boundary controls and corrective mechanisms for land misallocation to sever the amplification loop of this process.