Recent Dynamics of Forest Canopy Cover Loss in Germany

Abstract

Forests cover about one third of Germany’s land area. They are an important economic factor and fulfil several other ecosystem services at the same time. Furthermore, forests play a major role as carbon sink and are therefore essential for Germany to achieve national climate goals. According to the latest national forest inventory (NFI), conducted in 2021 and 2022, German forests turned into a net carbon source due to exceptional losses. These losses are mainly attributed to a severe drought situation between 2018 and 2020 and beyond. Compound and cascading effects of droughts, heatwaves and insect infestations led to large-scale disturbances, often resulting in salvage and sanitary logging. In essence, the decadal sampling of the NFI in Germany does not keep pace with the dynamics on the ground. In response to this new challenge, several policies such as the national forest strategy 2050 promote the development of digital technologies with the potential to monitor forests and to provide updates regularly and at higher temporal resolution in order to complement the in-situ measurements from the NFI. One key technology is remote sensing because of its capabilities to cover large areas synchronously on a regular basis. While Earth Observation (EO) is well-implemented in the monitoring of forests in the tropics, its potential is far from being deployed in operational monitoring in Germany. However, several EO-based forest monitoring tools have been developed over the past years. While there are some similarities, there are also differences between the products. To date, none of these tools is able to address all requirements of the annual crown condition assessments used to assess forest condition in Germany based on field samples. The information the EO-based tools provide is often complementary to the field-based assessments as they reveal additional spatio-temporal details and characterize canopy condition from a top-view perspective rather than the ground-view perspective. Recently, we provided a first assessment of forest canopy cover losses (FCCL) in Germany between January 2018 and April 2021 (Thonfeld et al., 2022). This was the first EO-based national study of this kind. It is different from all later tools in that it quantifies loss areas, i.e. forest areas that were either cleared or that were covered by larger patches of standing dead trees. Here, we provide an update until including September 2024 based on fully reprocessed and extended time series. We revised the approach to include Landsat-9 and used all available Sentinel-2A/B and Landsat-8/9 images with less than 80 % cloud cover. We masked out unusable data due to clouds, cloud shadows, snow and other artefacts and computed the Tasseled Cap-based Disturbance Index (DI), an index that normalizes across all forest pixels and hence reduces seasonality. We filtered the resulting time series with a moving median filter and used the 10th percentile to compute monthly composites. These were subtracted from the median full year 2017 DI. We applied a fixed threshold and counted how often it was exceeded. After exceeding the threshold six times in a row, an anomaly was considered a FCCL. The FCCL time series covers September 2017 through September 2024 at monthly resolution. Further, we overlaid our FCCL results with a tree species map and computed statistics for the four major tree species in Germany, namely spruce, pine, oak and beech. Our results show that at the beginning of the observation period several larger storm events caused losses distributed over large parts of Germany with a focus in Northern Germany. As the impacts of the 2018-2020 drought became effective, spruce-dominated areas became the hotspots of canopy cover loss. Accordingly, three major regions of FCCL emerged. Although spruce forests are particularly affected by recent-past FCCL, virtually all species show serious signs of stress, often resulting in dieback and associated logging. Hence, we found canopy cover losses across all major tree species. Considering the whole time series, most severe dynamics were recorded in central Germany, covering a region from the far West to the far East of Germany. In total, about 850,000 ha were affected by FCCL, corresponding to 8.4 % of the forest area. In some administrative districts of Germany about 45 % of the total forest area was cleared. Only a small share of FCCL can be attributed to land use changes for road construction, settlement expansion, or open-pit mining. The FCCL information is of great importance for forest authorities and forest managers. It can be used to specify the need of seedlings and to plan replanting with species that are better adapted to future climate. With our contribution we demonstrate how spatially explicit information about forest disturbances can be provided in high temporal resolution to support the national forest monitoring.