Putting Abandoned Farmlands in the Legend of Land Use and Land Cover Maps of the Brazilian Tropical Savanna

Включение заброшенных сельскохозяйственных угодий в легенду карт землепользования и земельного покрова тропической саванны Бразилии
Ivo Augusto Lopes Magalhães, Edson Eyji Sano, Édson Luís Bolfe, Gustavo Bayma-Silva
2025-12-27

Brazilian CerradoFully Connected Neural NetworkSentinel-2 multispectral imageryfarmland abandonmentland use and land cover mapping
Farmland abandonment is becoming a growing land use challenge in the Brazilian Cerrado, yet its extent, spatial distribution, and underlying drivers remain poorly understood. This study addresses the following question: Can deep learning methods reliably identify abandoned farmlands in tropical savanna environments using multispectral satellite images? To answer this question, we used a Fully Connected Neural Network (FCNN) classifier to map abandoned farmlands in the municipality of Buritizeiro, Minas Gerais State, Brazil, using Sentinel-2 images acquired in 2018 and 2022. Seven land use and land cover (LULC) classes were mapped using visible and near-infrared bands, spectral indices, spectral mixture components, and principal components as input parameters for the CNN. The LULC map for 2022 achieved high classification performance (overall accuracy = 94.7%; Kappa coefficient = 0.93). Agricultural areas classified in 2018 as annual croplands, cultivated pastures, eucalyptus plantations, or harvested eucalyptus that transitioned to grasslands or shrublands in 2022 were considered abandoned. Based on this definition, we identified 13,147 hectares of abandoned land in 2022, representing 4.7% of the municipality’s agricultural area in 2018. Most abandoned areas corresponded to eucalyptus plantations established for charcoal production. This study provides the first deep learning-based assessment of farmland abandonment in the Cerrado. Our findings demonstrated the potential of FCNN classifiers for detecting abandoned farmlands in this biome and provide important contribution for public policies focused on ecological restoration, carbon sequestration, and sustainable agricultural planning.
1
A Fully Connected Neural Network mapped seven land-use and land-cover classes in Brazil’s Cerrado using Sentinel-2 imagery from 2018 and 2022.
2
Most abandoned areas originated from eucalyptus plantations established for charcoal production and later transitioned to grasslands or shrublands.
3
The 2022 land-use and land-cover map achieved 94.7% overall accuracy and a Kappa coefficient of 0.93.
4
The study identified 13,147 hectares of abandoned farmland in 2022, equivalent to 4.7% of the municipality’s agricultural area in 2018.
5
The study provides the first deep-learning-based assessment of farmland abandonment in the Cerrado and supports restoration, carbon sequestration, and sustainable agricultural planning.

Abandoned farmlands in the Brazilian Cerrado, mapped in the municipality of Buritizeiro using Sentinel-2 imagery

The spatial extent, distribution, and detection of farmland abandonment, including transitions from agricultural land uses to grasslands or shrublands

Publication Details
Publication Date
2025-12-27
Journal
Publisher
ISSN
Access Type
Author Information
Authors
Ivo Augusto Lopes Magalhães
Edson Eyji Sano
Édson Luís Bolfe
Gustavo Bayma-Silva
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%