Ecosystems
This project concerns the topic Ecosystems and spans the subtopics Collaborative estimation of forest quality parameters and Forest and grassland disaster monitoring with the objective to develop new methods for the respective areas. The forest quality parameters include biomass, tree species and new quality parameters. For biomass, new tools will be developed to map forest state and change from satellite-borne radar with support of airborne laser scanning data. The deliverables are algorithms for correction of synthetic aperture radar (SAR) data for inconsistent weather conditions, decreasing stationary map uncertainties of forest estimates due e.g. fluctuating model uncertainty, wavelength dependence, and improper reference data, and using forest change estimations and auxiliary data to derive forest site-index (SI) maps. For tree species, the methods will be based on a combination of remotely sensed data types: medium-resolution satellite imagery, satellite SAR data, high-resolution images and LiDAR data to derive information both from spectral information structural information. The deliverables are algorithms for tree species identification from satellite images, SAR data, combining satellite images and ALS data, as well as a tree species map for a study area in Sweden. The new quality parameters are related to tree retention elements in the landscape, aiming to characterize and quantify forest structures that are key elements for ecological biodiversity. This objective will be addressed using several high-resolution data sources, such as multispectral satellite imagery and LiDAR data. The deliverables here are algorithms for mapping tree retention elements in the landscape from high-resolution multispectral satellite imagery and LiDAR data. The method for collaborative estimation of forest quality parameters includes the following steps: (1) extraction of features from Remote Sensing (RS) data, (2) linking RS data to reference data on sample plots, (3) training of estimators and classifiers using the reference data, and (4) prediction or supervised classification. For storm damage, the objective is to develop methodology and algorithms and to perform a scientific evaluation of SAR data from two or more satellite sensors for detecting and mapping changes in boreal forests. The approach for mapping storm-felled forest is to use change detection based on backscatter SAR images before and after the changes. Simulated wind-thrown fellings offer a unique in situ dataset with a setting very similar to a natural storm damaged forest. The deliverables here are algorithms for detecting and mapping changes in boreal forests, in particular storm damage, from SAR data from two or more satellite sensors. Two kinds of forest insect damage will be studied: (1) The European spruce bark beetle (Ips typographus [L.]), (2) pine wood nematode (Bursaphelenchus xylophilus). The objectives here are two-fold: (1) Developing the methods for forest insect damage detection at an early stage. (2) Analyzing distribution and spreading patterns of the forest insect damage. The deliverables here are methods for early detection, forest infestation spreading patterns and forecasting models, and prediction maps for large area application. The joint teams currently own 12 ongoing projects (2020 – 2022) with 4 339 700 EUR funding and 4 planned/submitted applications (2021 – 2023) with 1 078 740 EUR that are related to this proposal for Dragon 5 collaboration. For a coherent and comprehensive scientific research, field inventory and other remote sensing data have been collected in 2018 and 2019, and planned more data acquisition in 2020 – 2024. Together with satellite images provided by the Dragon 5 project, the large amount and various datasets will support our studies for forest information extraction.
