Do Earth Observation Essential Climate Variables (ECVs) provide a consistent picture of land surface dynamics? Insights from a multi-variable analysis over the Mediterranean region

Multi-variable analyses combining several Earth Observation (EO) Essential Climate Variables (ECVs) are increasingly used to investigate biosphere-land surface interactions under climate variability and anthropogenic pressure, including drought dynamics, vegetation responses, irrigation and land management practices, and land-atmosphere exchanges. By integrating optical, thermal, and microwave EO observations, large interdisciplinary initiatives such as the ESA Agricultural Land Abandonment and Climate Change (GLANCE) project aim to characterise coupled biosphere-hydrosphere processes at regional scales. These applications implicitly assume that EO-based ECVs provide a consistent and physically meaningful representation of the Earth system when jointly analysed. However, this assumption has rarely been evaluated explicitly.

Here, we present a systematic multi-annual (approximately two decades) cross-variable consistency analysis of key ESA Climate Change Initiative (CCI) ECVs over the Mediterranean region, conducted in the framework of the GLANCE project. The analysis focuses on soil moisture, precipitation, land surface temperature, vegetation parameters, biomass, and land cover, spanning multiple components of the biosphere-soil-water-atmosphere continuum.

We first assess each ECV by analysing its spatial patterns, seasonal and interannual variability, associated uncertainty, and response to drought events, and by comparing CCI products with external reference datasets. This single-variable assessment reveals substantial differences in some cases between CCI and non-CCI products, with particularly pronounced discrepancies for land cover. Building on this assessment, we investigate cross-variable consistency by jointly analysing the different CCI ECVs, focusing on the spatial correspondence of long-term mean patterns, correlations of temporal anomalies, and joint responses during drought conditions. Precipitation, land surface temperature, vegetation parameters, and soil moisture generally exhibit consistent behaviour, although sometimes with pronounced spatial and/or temporal differences, while biomass and land cover show substantially lower cross-variable consistency.

By explicitly evaluating the consistency of EO-based ECVs across biosphere-relevant variables, this work demonstrates that multi-variable EO analyses cannot assume coherence a priori. The results provide critical guidance for the interpretation and integration of multi-sensor EO datasets in biosphere and land surface studies, and help identify strengths and limitations of individual ECV products for Earth system research.