Effects on vegetation due to geo-energy technologies, an Earth Observation approach: Salar de Atacama study case

New technologies have been developed to accomplish the Sustainable Development Goals, where Geoenergy is key for the transition. In recent years, some studies have suggested that changes in underground conditions could have a secondary impact on the local environment, for example in Carbon Capture and Sequestration leakage of CO2 can impact the health of surrounding vegetation; on Lithium Mining, fluctuations in the water level may affect the health or dynamics of the vegetation.

We propose a scalable Earth-observation-based framework for the spatio-temporal analysis of vegetation dynamics over time. Combining multi-sensor satellite data: Landsat constellation and Sentinel-2 imagery. We complemented these results with meteorological data from ECMWF Reanalysis v5 (ERA5); and in-situ water level datasets.  The proposed approach integrates ecological interpretation with methodological robustness, enabling the systematic assessment of vegetation presence, persistence, and variability over time. We performed a validation analysis using PlanetScope imagery and in-situ meteorological data.

The workflow combines NDVI, and phenological metrics to generate a spatio-temporal vegetation dynamic characterization. The resulting products are designed to be readily integrated with complementary datasets, such as meteorological records and water-level observations, enabling their incorporation into impact assessments and environmental decision-making process.

Salar de Atacama (SdA), one of the world’s largest salt flats used for lithium production, hosts a variety of wetlands that are exposed to both extreme climatic conditions and increasing anthropogenic pressures. Understanding vegetation variability in this context is crucial for establishing robust environmental baselines and supporting long-term monitoring strategies.

Two areas of the salt flat have been selected to perform the analysis: i) surrounding area to Laguna Tebenquiche in the northern area (LTB); and ii) Vegas de Silao and Palolao on the South-East Border (SEB).

Both sites showed the peak NDVI season between late August to early November. Rainfall is concentrated between January and March, with high-magnitude events, confirming the dependence of vegetation on the availability of underground water.

By contrast, temperature exhibits a smooth, symmetric seasonal cycle, suggesting that thermal variability is secondary compared to hydrological pulses, but it should be considered in the snow melting process in the higher part of the basin.

The relative stability of LTB stations suggests shallow groundwater access, stable geomorphological settings, or vegetation assemblages adapted to predictable seasonal forcing. In contrast, the strong interannual variability observed at SEB indicates the vegetation activation is governed by episodic moisture availability.

NDVI peaks lag water-level increases, indicating a delayed vegetation response consistent with subsurface water availability rather than direct rainfall forcing. This lag is visible across sensors and persists across multiple years, reinforcing its ecological significance.

Water-level dynamics in the study area are influenced by multiple interacting processes, including lithium brine extraction rates, aquifer recharge from the upper basin, and direct recharge from rainfall. This study only evaluates the relevance of the methodology for vegetation assessment; it does not estimate the contribution of brine extraction to water levels.

The proposed methodology, when applied to SdA, demonstrates that the methodology effectively captures spatial and temporal patterns of vegetation response in a geoenergy context.