Modelling past and future impacts of droughts on tree mortality and carbon storage in Norway spruce stands in Germany

• Published in: Ecological modelling 2025-02, Vol.501, p.110987, Article 110987
• Main Authors: Anders, Tim, Hetzer, Jessica, Knapp, Nikolai, Forrest, Matthew, Langan, Liam, Tölle, Merja Helena, Wellbrock, Nicole, Hickler, Thomas
• Format: Article
• Language: English
• Subjects: Central Europe; Climate change; Drought; Dynamic vegetation models; Picea abies; Tree mortality
• ISSN: 0304-3800
• DOI: 10.1016/j.ecolmodel.2024.110987

•Annual weather variables (e.g. summer climatic water balance anomalies) in combination with variables that describe the general climate at a site (e.g. long-term temperature) can explain historical variations and the recent increase in Norway spruce tree mortality since 2018.•Elevated spruce tree mortality due to drought events significantly impacts spruce forest biomass stock and harvested wood amounts.•Future drought-related mortality is predicted to strongly reduce spruce forest biomass stocks and harvest until the end of the century, impacting forest ecosystem services.•Large uncertainty in predictions ranging from periodically recurring drought impacts similar to those observed after 2018 to catastrophic dieback affecting over more than half of the forest area of Germany.•Norway spruce is considered a very high risk species for forest plantations in Germany. Between 2018 and 2020, Central Europe experienced severe droughts. Leading to unprecedented increase in Norway spruce (Picea abies) tree mortality, which has not been reproduced in process-based forest or dynamic vegetation models (DVMs). In this study, we developed logistic regression models for drought-related Norway spruce mortality using climate and weather anomalies and mortality data from the forest monitoring plot network of the German crown condition survey. We integrated these models into the DVM LPJ-GUESS. Under historical conditions (1998–2020), our models reproduce observed temporal and spatial mortality patterns. Future simulations (2021–2070) under the RCP2.6 and RCP8.5 climate scenarios show periodic increases in Norway spruce mortality. Even though the drought-mortality models reproduce past dynamics similarly well, they do not agree on the timing and magnitude of future drought-related mortality events. Including drought mortality in the DVM reveals substantial reductions in aboveground biomass in 2070 (e.g. -18 % in RCP2.6 and -36 % in RCP8.5 (mean across all simulations)), compared to baseline simulations without drought mortality. According to the model, drought-related reductions of potential harvest across Germany could accumulate to 310 million Mg of C (RCP2.6) and 447 million Mg of C (RCP 8.5) in the period from 2021 to 2070. Our study underscores the severe risk of large-scale future Norway spruce forest diebacks in Germany. Predictions of the magnitude and timing of such dieback events are, however, still highly uncertain. Nevertheless, such events should be considered i