Recurrent climate-driven disturbances impact on the health of European forests that reacted with increased tree dieback and mortality over the course of the last four decades.

Deep learning (DL) models are gaining popularity in forest variable prediction using Earth observation (EO) images. However, in practical forest inventories, reference datasets are often represented by plot- or stand-level measurements, while high-quality representative wall-to-wall reference data for end-to-end training of DL models are rarely available.

Our objective was to develop a method for the assessment of forest area and structural variables for cases in which the availability of representative ground reference data is poor and these data are not collected from the whole area of interest.

In this study, we introduce an improved semisupervised deep learning approach, and demonstrate its suitability for modeling the relationship between forest structural parameters and satellite remote sensing imagery and producing forest maps.

Sustainable forest management requires up-to-date and reliable information, not only on forest structure, but also on its primary productivity and carbon stocks, and on changes. Information is increasingly demanded for forest management purposes, to prove responsible asset management to the shareholders, and to follow the European policies and treaties.

ICP Forests has monitored the European forests since the late 1980s. This Brief describes temporal and spatial developments in tree defoliation, a widely adopted indicator of forest health and vitality. Crown defoliation is defined as the loss of foliage on a tree compared to a fully foliated reference tree.

The European Union is today, in October 2020, a very different place from where it was in October 2015, when a call for proposals was launched that resulted in the funding of the DataBio project (among many others). We are now a Union that is planning its recovery from the most severe health emergency in our history, an emergency that has affected our economy very negatively.

A method to detect abrupt land cover changes using hierarchical clustering of multi-temporal satellite imagery was developed. The Autochange method outputs the pre-change land cover class, the change magnitude, and the change type.

We compared the performance of Sentinel-2 and Landsat 8 data for forest variable prediction in the boreal forest of Southern Finland. We defined twelve modelling setups to train multivariable prediction models with either multilayer perceptron (MLP) or regression tree models with the brute force forward selection method.