Revealing the distribution and change of abandoned cropland in Ukraine based on dual period change detection method
Выявление распределения и изменений заброшенных сельскохозяйственных угодий в Украине с использованием метода детектирования изменений по двум периодам
2025-02-17
SCID: 54.1/nc5egcs5
Discuss with AI
Ukraineabandoned croplandconflict-affected agriculturedual-period change detectiontime-series NDVI
Figures from the paper
Abstract (AI)
Since the outbreak of the Russia-Ukraine conflict in 2022, Ukraine has experienced different types of abandoned cropland, such as unused and unattended cropland, as a result of war damage, agricultural infrastructure destruction, and refugee outflows. Common methods for detecting abandoned cropland have difficulty effectively identifying and distinguishing these different types. This study proposes a Dual-period Change Detection method to reveal the spatial distribution and changes of different types of abandoned cropland in Ukraine, which can aid in agricultural assessments and international assistance in conflict-affected areas. The method mainly utilizes time-series NDVI data to fit the crop curves corresponding to cropland on a pixel-by-pixel basis, and then establishes discrimination rules for different types of abandoned cropland based on the crop curves, so as to detect unused cropland in the pre-conflict period (2015-2021) as well as unused cropland and unattended cropland in the post-conflict period (2022-2023). Finally, the detection results are validated and accuracy assessed using medium and high resolution spatiotemporal remote sensing imagery interpretation. The results show that the overall accuracy of the abandoned cropland extraction in Ukraine ranges from 83 to 96% during the study period. Before the conflict, the national average unused rate was 1.6%, with the lowest in 2021 and the highest in 2018. In 2022, the unused cropland area was approximately twice the average unused area before the conflict, and it was widely distributed, with the area of unattended cropland reaching 462,000 hectares, mainly in the eastern part of Ukraine. In 2023, compared to 2022, the unused cropland area decreased by 67.8%, while unattended cropland increased by 116.7%. Both types of abandoned cropland exhibited spatial clustering, with major clusters identified in the Crimea region, Kherson Oblast, Zaporizhzhia Oblast, and Donetsk Oblast.
Key Findings
1
A Dual-period Change Detection method uses pixel-level time-series NDVI crop-curve fitting to distinguish abandoned cropland types before and after the conflict.
2
Abandoned cropland showed spatial clustering, with major concentrations in Crimea, Kherson, Zaporizhzhia, and Donetsk oblasts.
3
Abandoned-cropland extraction achieved overall accuracies ranging from 83% to 96% based on medium- and high-resolution remote-sensing interpretation.
4
Before the conflict, Ukraine’s national average unused-cropland rate was 1.6%; in 2022, unused cropland approximately doubled relative to the pre-conflict average.
5
From 2022 to 2023, unused cropland decreased by 67.8%, while unattended cropland increased by 116.7%; 2022 unattended cropland covered 462,000 hectares.
6
The method detects pre-conflict unused cropland during 2015–2021 and separates post-conflict unused and unattended cropland during 2022–2023.
Research Object
Abandoned cropland in Ukraine, including unused and unattended cropland during the pre-conflict (2015–2021) and post-conflict (2022–2023) periods
Research Subject
The spatial distribution, temporal changes, and spatial clustering of different types of abandoned cropland
Publication Details
Publication Date
2025-02-17
Journal
Scientific Reports
Publisher
Nature Portfolio
ISSN
2045-2322
Cited by
12
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest