Spatio-Temporal Vegetation Dynamics and Climate Drivers in Cyprus: A Multi-Decadal Analysis Using Remote Sensing and Causal Inference

Kurzfassung

The Mediterranean region is highly vulnerable to climate change, with increasing drought frequency, declining precipitation, and rising temperatures posing major threats to ecosystems, water resources, and agriculture. Cyprus, located in the eastern Mediterranean faces particular pressure, holding the highest Water Exploitation Index among all EU member states. Its unique ecosystems, including maquis and phrygana vegetation, are increasingly threatened by changing climatic conditions, while socioeconomic consequences of drought, including water scarcity and reduced agricultural productivity, underscore the urgent need for improved understanding of vegetation-climate interactions. This study investigates spatio-temporal patterns of vegetation index anomalies across Cyprus and their climatic drivers over a 25-year period from March 2000 to December 2025. Vegetation dynamics are analyzed using the MODIS MOD13Q1 V6.1 product (250 m, monthly composites), with the analysis stratified by stable land-cover and land-use (LULC) classes derived from the ESA Climate Change Initiative (CCI) land cover dataset (available for the years 2000-2022). Trend analysis is performed to characterize inter- and intra-annual variability of vegetation indices across different LULC types. To identify causal climate drivers and potential time-lag effects, the PCMCI causal inference framework is applied using TerraClimate data, incorporating variables including precipitation, temperature, soil moisture, vapor pressure deficit, and surface solar radiation. The analysis is conducted across 34 hydrological subbasins, with causal graphs aggregated seasonally to reveal dominant land-atmosphere interactions. Results reveal marked seasonal and spatial heterogeneity in vegetation responses to climate variability. Analysis of the significant drought year 2007-2008 illustrates the utility of the approach: negative Enhanced Vegetation Index (EVI) anomalies were already detectable in winter months, with the most pronounced negative deviations occurring in March, primarily in the districts of Nicosia and Larnaca and the Mesaoria valley. Causal analysis identifies water availability (expressed through soil moisture and atmospheric demand) as the dominant driver of vegetation dynamics across nearly all seasons and subbasins, consistent with findings from comparable Mediterranean studies. Seasonal causal graphs reveal a strong negative coupling between solar radiation and EVI in winter, while precipitation exerts a strong positive influence on soil moisture in all seasons except summer, where this relationship reverses. Soil moisture, in turn, positively influences EVI particularly in spring and summer, with time lags of up to three months captured in the analysis. These findings contribute to a deeper understanding of vegetation-hydrology-climate interactions in Cyprus and provide a robust methodological framework applicable to other water-scarce Mediterranean environments, ultimately supporting evidence-based land and water management under ongoing climate change.