A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä
Monitoring changes in boreal peatland vegetation after restoration with optical satellite imagery (2024)
Isoaho, A., Elo, M., Marttila, H., Rana, P., Lensu, A., & Räsänen, A. (2024). Monitoring changes in boreal peatland vegetation after restoration with optical satellite imagery. Science of the Total Environment, 957, Article 177697. https://doi.org/10.1016/j.scitotenv.2024.177697
JYU-tekijät tai -toimittajat
Julkaisun tiedot
Julkaisun kaikki tekijät tai toimittajat: Isoaho, Aleksi; Elo, Merja; Marttila, Hannu; Rana, Parvez; Lensu, Anssi; Räsänen, Aleksi
Lehti tai sarja: Science of the Total Environment
ISSN: 0048-9697
eISSN: 1879-1026
Julkaisuvuosi: 2024
Ilmestymispäivä: 26.11.2024
Volyymi: 957
Artikkelinumero: 177697
Kustantaja: Elsevier
Julkaisumaa: Alankomaat
Julkaisun kieli: englanti
DOI: https://doi.org/10.1016/j.scitotenv.2024.177697
Julkaisun avoin saatavuus: Avoimesti saatavilla
Julkaisukanavan avoin saatavuus: Osittain avoin julkaisukanava
Julkaisu on rinnakkaistallennettu (JYX): https://jyx.jyu.fi/handle/123456789/98818
Julkaisu on rinnakkaistallennettu: https://urn.fi/URN:NBN:fi:oulu-202411276937
Tiivistelmä
Restoration can initiate a succession of plant communities towards those of pristine peatlands. Field inventory-based vegetation monitoring is labour-intensive and not feasible for every restored site. While remote sensing has been used to monitor hydrological changes in peatlands, it has been less used to monitor post-restoration changes in vegetation composition. We utilised vegetation inventories from Finnish peatland monitoring network containing 10-year before-after-control-impact monitoring data from 150 peatland sites, representing three peatland types (spruce mire forests, pine mire forests, open mires), and optical observations from Landsat 5–9 and Sentinel-2 satellites. We employed non-metric multidimensional scaling (NMDS) to produce floristic gradients, representing wetness and productivity, from the vegetation data. We constructed random forest regression models with NMDS dimensions, i.e. floristic gradients, as response variables and satellite imagery variables as the predictors. Our results show that the floristic gradients in different peatland types should be monitored with different satellite imagery variables. However, midsummer NIR and red band consistently explain variation in the gradients in all peatland types. Our results indicate that the gradients and the post-restoration changes in them can be modelled with reasonable accuracy in open mires and sparsely treed pine mire forests but not in densely treed spruce mire forests. We suggest that optical satellite imagery can serve as a proxy for assessing the post-restoration vegetation changes in peatlands with little or no trees.
YSO-asiasanat: suot; turvemaat; ojitus; kasvillisuus; ennallistaminen; satelliittikuvat; satelliittikuvaus; koneoppiminen
Vapaat asiasanat: remote sensing; machine learning; mires; land use; drainage
Liittyvät organisaatiot
OKM-raportointi: Kyllä
Alustava JUFO-taso: 2
