Quantifying Tree-species Specific Responses to the Extreme 2022 Drought in Germany

In recent decades, forests are increasingly suffering from so-called hotter droughts. This is because rising temperatures increase the atmospheric water demand under drought, thereby amplifying plant-water consumption. As a consequence, soil-water potentials are reaching more extreme values which eventually may result in xylem cavitation and possibly tree die-back. For instance, the extreme 2018 drought resulted in extraordinary die-back frequencies in numerous tree species across Central Europe such as European beech, Norway spruce, and Scots pine. Since forests render an important agent in terms of climate change mitigation, increasing forests' climate-change resilience via forest conversion is essential to preserve their integrity and consequently ecosystem services related to carbon sequestration.

 

In this context, better knowledge of species-specific drought responses is highly valuable since it allows for determining critical drought thresholds beyond which the functional integrity of trees is at threat. Such knowledge may serve to parameterize dynamic vegetation models, which then can be deployed to project tree-species-specific performance under various climate change scenarios. Eventually, this may guide forest managers to select more climate-resilient tree species portfolios in terms of forest conversion.

 

Yet, species-specific drought responses are typically derived from plot-based physiological monitoring networks. While this approach provides valuable and highly precise data on tree physiology and it suffers from a relatively low replication. To overcome this low-spatial replication, large-scale assessments using satellite-based remote sensing appear a promising research avenue. For instance, the recently released European forest condition monitor (EFCM) provides information on forest canopy conditions in near real-time, which was shown to successfully quantify drought impacts under previous droughts (Buras et al., 2020, 2021). However, the EFCM currently does not allow for species-specific assessments given a lack of corresponding data, in particular tree-species classifications.

 

To overcome this research gap and to deepen our knowledge of species-specific drought responses, we here present a machine-learning-based tree-species classification using Moderate Resolution Imaging Spectroradiometer (MODIS) of Germany which is subsequently used to stratify EFCM data to quantify the species-specific response of most abundant tree-species (in particular beech, oak, spruce, and pine) to the extreme 2022 drought. Preliminary results indicated a successful calibration-validation of the tree-species classification, with the corresponding F1-scores in the order of 0.55 – 0.7 and true-skill statistics in the order of 0.6 – 0.78, indicating average to good performance. Once applied to stratify the EFCM data we provide a well-replicated large-scale assessment of tree-species-specific drought response, which will improve our understanding of species-specific climate-change resilience. Corresponding information can then be used further to parameterize dynamic vegetation models, which eventually can be deployed to obtain projections of tree performance under various climate-change scenarios.

 

Buras A, Rammig A and Zang C S 2020 Quantifying impacts of the 2018 drought on European ecosystems in comparison to 2003 Biogeosciences 17 1655–72

Buras A, Rammig A and Zang C S 2021 The European Forest Condition Monitor: Using Remotely Sensed Forest Greenness to Identify Hot Spots of Forest Decline Frontiers in Plant Science 12 2355