Extreme weather events are increasing in frequency and intensity, posing a threat to forest ecosystems and eliciting forest-pest outbreaks. In the southern Italian Alps, a dramatic windthrow called Vaia occurred in October 2018, shifting populations of the European spruce bark beetle (Ips typographus) from an endemic to an epidemic phase. Remote-sensing methods are often employed to detect areas affected by disturbances, such as forest-pest outbreaks, over large regions. In this study, a random forest model on the Sentinel-2 images acquired over the south-eastern Alps in 2021 and 2022 was used to detect the outbreak spots. The automatic classification model was tested and validated by exploiting ground data collected through a survey conducted in 2021 and 2022 in both healthy and infested spots, characterized by variable sizes and degrees of infestation. The model correctly identified the forest conditions (healthy or infested) with an overall accuracy of 72% for 2022 and 58% for 2021. These results highlight the possibility of locating I. typographus outbreaks, even in small spots (between 5 and 50 trees) or spots intermixed with healthy trees. The prompt detection of areas with a higher frequency of outbreaks could be a useful tool to integrate field surveys and select forest areas in which to concentrate management operations.

Spruce Bark Beetle Outbreak Prediction through Automatic Classification of Sentinel-2 Imagery

Bozzini, A;Battisti, A;Faccoli, M
2023

Abstract

Extreme weather events are increasing in frequency and intensity, posing a threat to forest ecosystems and eliciting forest-pest outbreaks. In the southern Italian Alps, a dramatic windthrow called Vaia occurred in October 2018, shifting populations of the European spruce bark beetle (Ips typographus) from an endemic to an epidemic phase. Remote-sensing methods are often employed to detect areas affected by disturbances, such as forest-pest outbreaks, over large regions. In this study, a random forest model on the Sentinel-2 images acquired over the south-eastern Alps in 2021 and 2022 was used to detect the outbreak spots. The automatic classification model was tested and validated by exploiting ground data collected through a survey conducted in 2021 and 2022 in both healthy and infested spots, characterized by variable sizes and degrees of infestation. The model correctly identified the forest conditions (healthy or infested) with an overall accuracy of 72% for 2022 and 58% for 2021. These results highlight the possibility of locating I. typographus outbreaks, even in small spots (between 5 and 50 trees) or spots intermixed with healthy trees. The prompt detection of areas with a higher frequency of outbreaks could be a useful tool to integrate field surveys and select forest areas in which to concentrate management operations.
2023
File in questo prodotto:
File Dimensione Formato  
forests-14-01116.pdf

accesso aperto

Descrizione: Spruce Bark Beetle Outbreak Prediction through Automatic Classification of Sentinel-2 Imagery
Tipologia: Published (publisher's version)
Licenza: Creative commons
Dimensione 3.44 MB
Formato Adobe PDF
3.44 MB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11577/3489045
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 12
  • ???jsp.display-item.citation.isi??? 11
  • OpenAlex ND
social impact