Assessment of spatial interactions in farmland abandonment: A case study of Gwangyang City, Jeollanam-do Province, South Korea

Abstract

Globally, countries have experienced substantial increases in farmland abandonment and concomitant consequences, such as reduced agricultural productions, deteriorated local communities, and increased probabilities of wildfire and landslides. For better understanding of the phenomenon, previous studies have focused on finding individual factors leading to farmland abandonment, while the spatial relationships, including influences exchanged among parcels themselves, have been much less examined. In this study, we investigate the factors inducing farmland abandonment in Gwangyang City in Jeollanam-do Province in South Korea, with a specific focus on the spatial interactions. We analyzed the level of spatial autocorrelation of abandoned farmlands using Moran's I and conducted a spatial autoregressive probit regression analysis to assess the factors that explain farmland abandonments, including their spatial dependence and spillover effects. We found that the global Moran's I was 0.42 for locations of abandoned farmland, justifying the use of a spatial model. The results of spatial autoregressive probit model confirmed the spatial interactions among the spatial units and spillover effects, meaning that farmland abandonment can be affected by abandonment of neighboring agricultural parcels, and that an individual parcel simultaneously receives influences of factors within itself and from factors of other parcels, respectively. Our new dispersion index confirmed that disconnection or fragmentation of farmland is associated with farmland abandonments. The finding of this research will help in establishing rural policies, focusing on the community-level cooperation system to prevent future occurrence of the farmland abandonments.