Spatially predicting ecosystem service patterns in boreal drained peatlands forests using multisource satellite data
Year: 2025
Type: article
Abstract: • Predicted single and summed ecosystem services (ESs) using satellite data. • Model performance varies among single ESs and summed-ESs. • Incorporating auxiliary predictors improves model accuracy. • Closed canopy areas hold higher ESs values than partial or open canopy areas. Boreal drained-peatland forests provide diverse, interlinked ecosystem services (ESs), critical for informed decision-making in forest management. We mapped five ESs: bilberry yield, visual amenity, biodiversity conservation, carbon storage, and timber production using Landsat 8–9, Sentinel-2, and PlanetScope data. By combining these five ESs variables, we calculated a summed-ESs variable to capture overall ESs in drained peatland forests. Our objectives included assessing the influence of sensor resolution, auxiliary data, and the feasibility of scaling ESs predictions across varying canopy covers (closed, partial, and open). Using spectral bands and indices, we applied random forest regression, achieving explained variances (R 2 ) of 13–75 % for single ESs and 58–67 % for summed ESs. Sensor performance varied, with Landsat (R 2 22–69 %), Sentinel-2 (R 2 25–75 %), and PlanetScope (R 2 13–65 %). Incorporating auxiliary variables from seven-year-old LiDAR data improved model R 2 value by 1–24 %. We successfully scaled ESs predictions to map spatial distributions across the study area, with high ESs value in closed-canopy areas. These findings demonstrate satellite imagery’s effectiveness for spatial ESs prediction, supporting sustainable drained-peatland forest management.
Institution: Natural Resources Institute Finland
Language: English
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
SDGs: Life on land, Climate action
Open Access status: gold
Estimated APC paid: $3,550
Award: G3426158051
OpenAlex ID: W4409595851