Open-Earth-Monitor Global Workshop 2026

Increasing forest cover is generally perceived positively. Potentially, more trees can capture more carbon, along with bringing other beneficial ecosystem services if they become sustainable ecosystems. However, this also changes some key properties of the surface of the Earth (e.g. albedo, latent heat flux, etc) that can affect the local climate through biophysical mechanisms, notably by changing the local land surface temperature (LST). This aspect is still largely ignored by policy-makers, in part because they are difficult to model given they depend much on local context which is spatio-temporally variable. As demonstrated by several studies, satellite remote sensing has a role to play here. In this presentation we will illustrate how, within a dedicated OEMC use case, we adapted an established methodology to retrieve the local effect of forest cover change on LST and made the processing pipeline openly available and reproducible. We applied it for the first time on geostationary satellite data with a 15 minute revisit frequency over Europe and Africa to generate estimations of the change of LST along the entire daily cycle. This was done in collaboration with the European Commission Joint Research Centre on their BDAP platform opening the possibility to expand results within their EU Forest Observatory. In parallel work we also explored how the spatial heterogeneity of tree cover could also have an effect on LST and further into cloud formation across Africa.

There is a growing need to understand the inter-relationships between our atmosphere, the biosphere, and what we could call the anthroposphere. These interactions are complex and not always evident, complicating their analysis through a traditional modelling framework. On the other hand, we do have large amounts of climatic and environmental data in the form of spatio-temporal layers, along with socio-economic data expressing human behaviour, which could be combined in a data-driven perspective. Past work has shown that it is possible to summarize the state of the terrestrial biosphere in a few dimensions by applying statistical techniques to Earth system data cubes. Similarly, socio-economic development data can be statistically reduced to some key dimensions.
To investigate correlations between the biosphere, atmosphere and the sociosphere, we developed the initial concept of the planetary health index (PHI). Towards this goal, data from FLUXCOM-X and Eurostat was collected and resampled at quarterly temporal frequency and NUTS3 spatial level. We then developed an interactive dashboard in R Shiny based on interpretable and holistic Canonical Correlation Analysis to visualize spatio-temporal patterns of features and relationships between the 3 spheres. Additional data can be added later on for customisation. This tool was designed and developed for a particular stakeholder, the European Central Bank, who is interested in the physical risk and the impacts that financial institutions have on climate change, biodiversity loss, ecosystem services and their degradation, but also on how these processes also have impacts on financial institutions.

Gross Primary Productivity (GPP) is an essential variable to monitor carbon uptake by terrestrial ecosystems. Mapping GPP is thus of great interest to establish estimates of the global carbon budget, especially when this needs to be done at regional scale such as with the 'REgional Carbon Cycle Assessment and Processes' (RECCAP) initiative coordinated by the Global Carbon Project. While GPP is not directly measurable from space by remote sensing, sun-induced chlorophyll fluorescence (SIF) can, and it has been shown to correlate well with GPP. SIF is a weak signal that can be retrieved from hyperspectral remote sensing and which is emitted by the photosynthetic machinery itself. The downside is that current and past satellite instruments from which SIF can be derived have a very coarse spatial resolution (5 km at best) and are usually very noisy due to the subtle nature of the signal. In this use case, we have thus developed some tools to improve the situation. First, we collaborated to derive a new retrieval method leveraging on machine learning to reduce the noise. Second, we adapted a spatial enhancement methodology enabling users to increase the spatial resolution by combining other remote sensing variables with a simple modelling framework of their choice. This was done on the Copernicus Data Space Ecosystem (CDSE) using the OpenEO framework, allowing users to design their own downscaling approach and applying it lazily to their zones of interest. To compare with in-site GPP estimations from eddy covariance flux towers, we further designed a methodology to characterize the heterogeneity around these towers based on data cubes of Sentinel-2 data.