Eliades et al. (2026) Forests in a semi-arid climate die with a memory: satellite signals predict forest mortality years after drought

Identification

• Journal: Journal of Forestry Research
• Year: 2026
• Date: 2026-03-03
• Authors: Filippos Eliades, Dimitrios Sarris, Felix Bachofer, Silas Michaelides, Chris Danezis, Diofantos Hadjimitsis
• DOI: 10.1007/s11676-026-02016-z

Research Groups

• Remote Sensing and GeoEnvironment Lab, Department of Civil Engineering and Geomatics, Cyprus University of Technology, Limassol, Cyprus
• Eratosthenes Centre of Excellence, Limassol, Cyprus
• KES Research Centre, Nicosia, Cyprus
• KES College, Nicosia, Cyprus
• German Aerospace Center (DLR), Earth Observation Center (EOC), Wessling, Germany

Short Summary

This study investigates the relationship between satellite-derived vegetation indicators and meteorological drought indices to understand tree mortality mechanisms in semi-arid Cypriot forests, revealing that severe drought conditions trigger mortality and that vegetation response is linked to multi-year climate memory effects, with indicator effectiveness varying by species and post-mortality stage.

Objective

• Investigate the relationship between remotely sensed indicators of forest decline and meteorological drought indicators across seasonal, monthly, and multi-year timescales.
• Examine the relationship between the 2008 and 2016 forest mortality events in Cyprus and drought conditions.
• Identify the most effective drought and remote sensing vegetation indicators for explaining forest decline and mortality in conifers and evergreen broadleaves.
• Determine the temporal scales at which these indicators are most effective.

Study Configuration

• Spatial Scale: Woodlands of Cyprus, specifically three sites: Akamas (10,805 m²), Machairas (19,179 m²), and Stavrovouni (75,055 m²). The study focused on Juniperus phoenicea, Pinus brutia, and Quercus alnifolia populations at the xeric limits of their distribution.
• Temporal Scale: Data collection and analysis spanned from 1990 to 2020 (30-year record), with analyses conducted at monthly, seasonal, annual, and multi-year scales (including 2–10 year running means and 10-year moving windows).

Methodology and Data

• Models used:
  • Remote Sensing Indicators: Normalized Difference Moisture Index (NDMI), Enhanced Vegetation Index (EVI), Gross Primary Productivity (GPP), Leaf Area Index (LAI), Normalized Burn Ratio (NBR), Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Soil-Adjusted Vegetation Index (SAVI).
  • Drought Indicators: Standardized Precipitation Index (SPI), Standardized Precipitation-Evapotranspiration Index (SPEI), Palmer Drought Severity Index (PDSI).
  • Statistical Methods: Cook’s Distance, Kalman smoothing filter, Spearman’s correlation, running mean correlation analysis, moving-window correlation analysis.
• Data sources:
  • Satellite: Landsat 5, 7, 8, and 9 images (Surface Reflectance products) retrieved via Google Earth Engine (GEE).
  • Observation: Daily temperature and precipitation records from three meteorological stations in Cyprus (Akamas/Smiges, Machairas/Kionia, Stavrovouni/Kornos) from the Cyprus Department of Meteorology, supplemented by data from Polis Chrysochous (NOAA’s Global Historical Climatology Network - GHCN).
  • Reanalysis: ERA5 climate reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF) for cross-validation.

Main Results

• Very severe drought conditions (SPI-12 and SPEI-12 < -2, or PDSI-12 < -5) were found to exceed the capacity of Juniperus phoenicea, Pinus brutia, and Quercus alnifolia to sustain healthy stands at their xeric distribution limits in Cyprus, aligning with observed tree mortality events in 2008 and 2016.
• For conifers (J. phoenicea, P. brutia), NBR and NDMI were the most effective remote sensing indicators for explaining vegetation response to drought before and immediately after mortality, with NBR correlating strongly with SPI/SPEI for J. phoenicea and PDSI for P. brutia.
• After several years post-mortality, NDVI became a better descriptor of decadal vegetation response to climate for conifers, likely due to its effectiveness in capturing severe defoliation.
• For the evergreen broadleaf Q. alnifolia, remote sensing indicators struggled to capture vegetation dynamics in dense stands before mortality; however, post-mortality, NDVI provided the highest accuracy, attributed to faster and more severe defoliation making changes detectable.
• The annual vegetation response to drought-induced mortality was related to climate averaged over 4 to 7 past years (including the monitoring year), depending on the species, suggesting a 'memory' effect potentially linked to cumulative depletion and slow recharge of deeper moisture pools.

Contributions

• This study presents the first remote-sensing-based assessment documenting robust multi-year associations between canopy condition and prior climate in semi-arid forests, providing independent evidence for a storage-mediated 'memory' effect.
• It offers a novel and comprehensive analysis combining remote sensing and meteorological drought indicators at long-term scales to characterize tree mortality in Mediterranean ecosystems.
• The research highlights the critical importance of distinguishing between gymnosperm and angiosperm vegetation dynamics when assessing drought impacts, as well as the varying detection capabilities of different satellite remote sensing indicators under severe drought stress.
• The findings provide key insights for regional drought assessment and long-term ecosystem monitoring, informing the development of early warning systems and species-specific drought mitigation strategies for forest management and conservation planning.

Funding

• "EXCELSIOR": ERATOSTHENES: Excellence Research Centre for Earth Surveillance and Space-Based Monitoring of the Environment H2020 Widespread Teaming project (Grant Agreement No 857510).
• Government of the Republic of Cyprus through the Directorate General for the European Programmes, Coordination and Development.
• Cyprus University of Technology.
• Open access funding provided by the Cyprus Libraries Consortium (CLC).

Citation

@article{Eliades2026Forests,
  author = {Eliades, Filippos and Sarris, Dimitrios and Bachofer, Felix and Michaelides, Silas and Danezis, Chris and Hadjimitsis, Diofantos},
  title = {Forests in a semi-arid climate die with a memory: satellite signals predict forest mortality years after drought},
  journal = {Journal of Forestry Research},
  year = {2026},
  doi = {10.1007/s11676-026-02016-z},
  url = {https://doi.org/10.1007/s11676-026-02016-z}
}