An early warning system based on machine learning detects huge forest loss in Ukraine during the war

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F60460709%3A41330%2F25%3A103609" target="_blank" >RIV/60460709:41330/25:103609 - isvavai.cz</a>

Result on the web

<a href="https://doi.org/10.1016/j.gecco.2025.e03427" target="_blank" >https://doi.org/10.1016/j.gecco.2025.e03427</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.gecco.2025.e03427" target="_blank" >10.1016/j.gecco.2025.e03427</a>

Result language

angličtina

Original language name

An early warning system based on machine learning detects huge forest loss in Ukraine during the war

Original language description

The Ukrainian-Russian historical relationship has culminated in the latest Ukraine conflict, a multifaceted crisis transcending Ukraine's borders. This protracted war has triggered heavy environmental consequences in addition to its huge human toll. Among them, the loss of forests stands out, especially in Ukraine's agriculturally rich regions, where those remaining forested areas provide fundamental ecosystem processes. Wars have been well-documented for their adverse environmental impact, including the deliberate destruction of vital natural resources like forests, agriculture, and water supplies. In conflict zones, remote sensing combined with Artificial Intelligence is an indispensable tool for monitoring forest loss since this technology offers distance and secure data acquisition, enabling the identification and quantification of forest cover changes almost in real time. We employed Random Forest, a supervised machine learning classification algorithm, in conjunction with high-quality satellite imagery, to quantify the forest loss in Ukraine during the war, between 2022 and 2023. We found that forest loss in Ukraine was 807.56 km2 and 771.81 km2 in 2022 and 2023, respectively. We have now evidence that, in 2022-2023, most of the regions affected by the conflict show a high increase in forest loss compared to 2021 (Donets'k 2-year loss: 180.25 km2; Kharkiv: 181.38 km2; Kherson: 214.14 km2; Kyiv: 268.37 km2; Luhans'k: 195.4 km2), whereas in other areas not directly involved in the conflict we did not find any significant losses. This represents evident proof that the forest loss we detected in 2022-2023 within Ukrainian regions affected by the war (65.8 % of the whole country's forest loss) may be directly related to the conflict. Detecting these ecological impacts with an early warning system based on AI is vital for safeguarding ecosystems in conflict zones and underscores the urgency of environmental preservation amidst armed conflicts.

Czech name

—

Czech description

—

Type

J_imp - Article in a specialist periodical, which is included in the Web of Science database

CEP classification

—

OECD FORD branch

10619 - Biodiversity conservation

Project

—

Continuities

S - Specificky vyzkum na vysokych skolach

Publication year

2025

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Name of the periodical

GLOBAL ECOLOGY AND CONSERVATION

ISSN

2351-9894

e-ISSN

2351-9894

Volume of the periodical

58

Issue of the periodical within the volume

APR 2025

Country of publishing house

NL - THE KINGDOM OF THE NETHERLANDS

Number of pages

11

Pages from-to

—

UT code for WoS article

001416457800001

EID of the result in the Scopus database

2-s2.0-85215128713