Authors: Pedro Ivo Bastos De Castro*, Th-Köln, Rui Pedroso, Th-Köln, Sven Lautenbach, Universität Heidelberg, Raul Vicens, Universidade Federal Fluminense
Topics: Applied Geography, Agricultural Geography, Geographic Information Science and Systems
Keywords: Farmland abandonment, Logistic regression, spatial autocorrelation
Session Type: Virtual Poster
Start / End Time: 3:05 PM / 4:20 PM
Room: Virtual 51
Presentation File: Download
Farmland abandonment has been reported in many parts of the world as a household decision driven by environmental, socioeconomic, and institutional factors. Underlined by several national policy interventions and economic shocks, the Brazilian federal state of Rio de Janeiro has been showing an alarming loss of cultivated areas since the end of the 1980s. Nonetheless, local determinants contributing to the farmland abandonment process at a regional level in Rio de Janeiro remain unclear. In this study, we used logistic regression to identify moderating spatial determinants of farmland abandonment in the Norte Fluminense region of Rio de Janeiro, the region most affected by the loss of cultivated areas. We applied Spatial Eigenvector Mapping (SEVM) to account for spatial autocorrelation and assure a reliable estimation of standard errors and p-values in our spatial logistic regression model. We analyzed the spatial determinants of farmland abandonment for the period from 1991 to 2000. Our results indicate that abandonment was more likely to occur in flat areas with lower average precipitation and higher precipitation variability and areas at larger distances to the main roads.