Schotte et al. (2026) Comparing absolute and standardized drought indices for modelling tree mortality of spruce, beech, pine, and oak based on the Crown Condition Survey in Germany

Identification

• Journal: Forest Ecology and Management
• Year: 2026
• Date: 2026-03-27
• Authors: Laura-Marie Schotte, Angelika Olesch, Sebastian Fuchs, Luca Smekal, Bernd Ahrends
• DOI: 10.1016/j.foreco.2026.123741

Research Groups

Northwest German Forest Research Institute (NW-FVA), Department of Environmental Control

Short Summary

This study evaluated the association of absolute aridity and standardized drought indices with tree mortality for four major tree species in Germany (1990–2022). It found that standardized drought indices, particularly the Standardized Precipitation Evapotranspiration Index (SPEI), better explained mortality for Norway spruce, European beech, and Scots pine, with effects increasing over longer aggregation periods of up to five years.

Objective

• To evaluate the extent to which absolute aridity (Climatic Water Balance, CWB; Site Water Balance, SWB) and standardized drought indices (Standardized Precipitation Evapotranspiration Index, SPEI; CWB anomalies) are associated with tree mortality for Norway spruce, European beech, Scots pine, and oak in Germany.
• Hypothesis 1: Absolute aridity and standardized drought indices differ in their ability to explain tree mortality, with standardized indices providing a better representation of mortality responses due to their characterization of relative deviations from long-term climatic means.
• Hypothesis 2: The strength of the relationship between drought conditions and tree mortality depends on the aggregation period, with longer drought aggregation periods exerting a stronger influence.
• Hypothesis 3: Drought-mortality relationships vary among tree species.

Study Configuration

• Spatial Scale: Germany, based on a 16 km × 16 km grid with 481 plots.
• Temporal Scale: 1990–2022 (32 years).

Methodology and Data

• Models used: Correlation analyses (point-biserial correlations), Generalized Additive Models (GAMs) with a binomial error distribution and logit link function.
• Data sources:
  • Tree mortality data: German Crown Condition Survey (WZE) from 1990 to 2022, covering 16,473 individual trees across 481 plots.
  • Climate data: Daily climate data regionalized from German Weather Service (DWD) stations, aggregated to monthly means/sums for temperature, precipitation, and potential evapotranspiration (PET, calculated using FAO standard).
  • Soil data: Second National Forest Soil Inventory (NFSI II) for Available Water Capacity (AWC) of the soil, used for Site Water Balance (SWB) calculation on a subset of 413 plots.
  • Drought and aridity indices: Climatic Water Balance (CWB), Site Water Balance (SWB), CWB anomalies, and Standardized Precipitation Evapotranspiration Index (SPEI) calculated for aggregation periods of 3, 12, 24, 36, and 60 months.

Main Results

• Tree mortality exhibited distinct species-specific responses and spatial patterns, with Norway spruce showing the highest mortality rates (up to 10% in 2020), significantly higher than European beech, Scots pine, and oak (0–1%).
• For Norway spruce, European beech, and Scots pine, standardized drought indices (SPEI and CWB anomalies) showed stronger associations with mortality than absolute aridity measures (CWB, SWB).
• The strength of the relationship between drought and mortality generally increased with longer aggregation periods, with the 60-month SPEI (SPEI60) providing the best model performance for spruce, beech, and pine. This indicates that cumulative drought effects over several years are more critical.
• Generalized Additive Models (GAMs) showed high statistical power (AUC > 0.8) for all species and SPEI aggregations in distinguishing between surviving and dead trees.
• For Norway spruce, SPEI60 models closely reflected observed mortality rates, achieving the highest performance (AUC = 0.916).
• For European beech, SPEI models partially captured mortality trends, with SPEI60 showing the best performance (AUC = 0.941), despite generally low observed mortality rates.
• For Scots pine, SPEI models captured general temporal trends, with SPEI60 providing the best fit (AUC = 0.834).
• For oak, SPEI-based models did not adequately explain mortality dynamics, showing the lowest model performance and no clear superior fit across aggregation periods (highest AUC = 0.856 for SPEI3). Oak mortality was more strongly associated with absolute SWB and temperature-related variables than with standardized drought indices.

Contributions

• The study demonstrates that standardized drought indices, particularly the SPEI, are robust indicators for predicting drought-related tree mortality across dominant tree species in Germany, outperforming absolute aridity measures for spruce, beech, and pine.
• It highlights the critical importance of considering longer drought aggregation periods (up to five years) to capture cumulative and persistent drought stress effects on tree mortality, especially in the context of increasing multi-year drought events.
• The research provides species-specific insights into drought-mortality relationships, which are essential for improving mortality predictions and informing climate-adapted tree species selection in forest management.
• It suggests that decision support systems for forest management could benefit from incorporating standardized, species-specific drought indices to enhance the accuracy of future mortality predictions.

Funding

This study was part of the joint research project MultiRiskSuit, funded by the German Federal Ministry of Food and Agriculture (BMEL) and the Federal Ministry for the Environment, Nature Conservation, Nuclear Safety and Consumer Protection (BMUV) under grant agreement No. 2220WK41A4 through the Agency for Renewable Resources (FNR) within the Forest Climate Fund (WKF) program.

Citation

@article{Schotte2026Comparing,
  author = {Schotte, Laura-Marie and Olesch, Angelika and Fuchs, Sebastian and Smekal, Luca and Ahrends, Bernd},
  title = {Comparing absolute and standardized drought indices for modelling tree mortality of spruce, beech, pine, and oak based on the Crown Condition Survey in Germany},
  journal = {Forest Ecology and Management},
  year = {2026},
  doi = {10.1016/j.foreco.2026.123741},
  url = {https://doi.org/10.1016/j.foreco.2026.123741}
}