Conference Agenda

In recent years, solar-induced chlorophyll fluorescence (SIF) showed great potential for monitoring terrestrial photosynthesis, but existing SIF products retrieved from atmospheric sensors typically feature coarse spatial resolutions from several to hundreds of kilometers. Recently, the Chinese Terrestrial Ecosystem Carbon Inventory Satellite, known as Goumang satellite, launched in August 2022, carries an unique SIF Imaging Spectrometer (SIFIS), the first spaceborne sensor specifically designed for retrieving terrestrial SIF globally. SIFIS provides a substantially improved spatial resolution (along track: 370 m; across track: 800 m), representing an important advance in SIF remote sensing capabilities. Benefiting from its high spectral resolution (0.24 nm) in the spectral range of 664-786 nm, SIFIS SIF can be accurately retrieved using a data-driven approach. In addition, the SIFIS SIF showed excellent spatial and temporal agreement with airborne AisaIBIS data and independent datasets. We also show first results from the recent launched Fengyun 3H satellite which carries on a GHG instrument with potential SIF retrievals at 2km*2km.

Solar-induced chlorophyll fluorescence (SIF) provides a direct optical window into photosynthetic functioning and has emerged as a powerful proxy for gross primary productivity (GPP) from space. Yet most existing satellite SIF products are derived from polar-orbiting missions, which provide only instantaneous snapshots and therefore cannot resolve the dynamic, sub-daily evolution of photosynthesis. Although the OCO-3 mission offers observations at multiple local times, its within-day sampling remains sparse, limiting its ability to characterize full diurnal trajectories. As a result, an hourly-scale SIF product that can routinely capture diurnal photosynthetic dynamics is still lacking. The Tropospheric Emissions: Monitoring of Pollution (TEMPO) mission provides a unique opportunity to fill this gap. TEMPO is a geostationary hyperspectral spectrometer over North America with hourly revisit frequency, sub-nanometer spectral resolution across the visible wavelengths, and kilometer-scale spatial resolution, enabling dense temporal sampling under nearly stable viewing geometry. Here we develop and apply a data-driven retrieval framework to derive hourly red SIF from TEMPO radiance observations across North America. The resulting product delivers spatially continuous fields of red SIF at hourly resolution, allowing systematic characterization of diurnal variability across ecosystems. Our results reveal pronounced ecosystem-dependent diurnal patterns in red SIF, including consistent afternoon depression, indicative of short-term photosynthetic downregulation. Comparisons with flux-tower GPP demonstrate that TEMPO-derived red SIF closely tracks sub-daily photosynthetic dynamics and improves the representation of diurnal variability relative to polar-orbiting SIF products. These findings highlight TEMPO’s potential to enable routine, continent-scale monitoring of photosynthetic regulation and stress through hourly red SIF observations.

In 2026, the ESA FLEX mission will provide global measurements of sun-induced fluorescence (SIF), offering insights into plant photosynthetic activity as well as plant health and stress. In this context, data from existing space-borne hyperspectral missions can play a complementary role by helping with the preparation of commissioning activities, contributing to the validation of FLEX mission products, and, more generally, supporting the analysis and interpretation of FLEX data. This contribution presents the DESIS and EnMAP hyperspectral missions, emphasizing the complementary aspects that can support the FLEX scientific objectives.

The DESIS mission is a collaboration between the German Aerospace Center (DLR) and the U.S. company Teledyne Brown Engineering. DESIS is an imaging spectrometer with 30 m spatial resolution and a swath width of 30 km, operating in the VNIR range (400 – 1000 nm). It is located on the International Space Station (ISS), which allows observations of targets in the latitude range from 55° North to 52° South at varying local times. With a spectral full width half maximum (FWHM) of 3.5 nm, DESIS has a relatively high spectral resolution compared to similar missions. This has allowed the use of DESIS data for SIF retrieval, as presented in another contribution at this workshop (see Buffat et al.).

EnMAP is a German satellite operated by DLR. EnMAP observes from a Sun-synchronous orbit (11:00 local time descending node) with a 27-day repeat cycle, which allows re-observing a target approximately every four days by varying the off-nadir tilt angle. The instrument consists of two spectrometers: a VNIR spectrometer operating in the range 420–1000 nm and a SWIR spectrometer operating in the range 900–2450 nm. The average spectral FWHM of EnMAP is 6–11 nm in the VNIR and 7–11.5 nm in the SWIR. The spatial resolution is, like in DESIS, 30 m with a swath width of 30 km.
In terms of calibration, EnMAP is equipped with several on-board calibration devices, which allow very precise monitoring of instrument performance, while DESIS relies on vicarious calibration and shows higher spectral variability. Both missions use similar processing chains that provide products at three different levels: L1B (at-sensor top-of-atmosphere radiance), L1C (orthorectified and geolocated top-of-atmosphere radiances), and L2A (orthorectified and geolocated bottom-of-atmosphere reflectances). These data can be accessed via the DLR EOWEB portal (https://eoweb.dlr.de/) and are also available for direct download from the EOC-Geoservice (https://geoservice.dlr.de/) after user registration for each mission. The EnMAP mission catalogue dates back to April 2022, while the DESIS catalogue extends back to October 2018.

In this contribution, we discuss characteristics of EnMAP and DESIS data that are relevant for FLEX, including acquisition strategies, spectral coverage, radiometric performance, spatial resolution, geolocation accuracy, and data access. Potential use cases are presented, such as the inclusion of information derived from the EnMAP SWIR range, identification of suitable validation areas, as well as mission intercomparison activities. The contribution highlights how EnMAP and DESIS data can complement FLEX observations by offering independent hyperspectral radiance and reflectance measurements with higher spatial resolution.

With the FLEX launch approaching, downscaled SIF datasets are essential to bridge coarse-resolution observations and the sub-kilometer applications envisioned for the mission. A key feature of FLEX is its tandem flight with Sentinel-3 (S3), allowing S3 observations to provide the ancillary atmospheric and vegetation information required for sub-kilometer SIF applications. Building on this tandem-mission concept and addressing the current scale gap, we developed the first global downscaled SIF (S3-SIF₇₄₃) dataset with a spatiotemporal resolution of 300 m and 4 days.

Our methodology integrates TROPOspheric Monitoring Instrument in the 743–758 nm retrieval window (TROPOSIF₇₄₃) with S3 OLCI radiances and S3-derived vegetation traits through a Random Forest regression framework implemented on Google Earth Engine (GEE). This approach uses S3's 300 m spectral capabilities to downscale coarse SIF observations, consistent with the FLEX–S3 observing strategy. Model training over Europe achieved robust performance against TROPOSIF₇₄₃ reference data (R² = 0.86, RMSE = 0.14 mWm⁻² sr⁻¹ nm⁻¹), and the approach was subsequently extended globally. Validation against ground-based tower observations across four sites spanning 2017–2021 confirmed that S3-SIF₇₄₃ effectively reproduces seasonal dynamics across diverse ecosystems, with correlation coefficients ranging from R = 0.40 to 0.70 (all P < 0.001). Pixelwise comparisons against TROPOSIF₇₄₃ in 2020 demonstrated strong spatial consistency, particularly in temperate agricultural regions, with the highest R² values in croplands (R² = 0.70) and deciduous forests (R² = 0.67–0.75). Global mapping revealed coherent patterns of photosynthetic activity, with peak values in tropical rainforests and major agricultural zones. Importantly, S3-SIF₇₄₃ reduces retrieval noise relative to TROPOSIF₇₄₃ and provides unprecedented insights into sub-kilometer spatial heterogeneity.

By utilizing S3’s spectral information and GEE’s scalable processing capabilities, S3-SIF743 provides sub-kilometer SIF information that is directly relevant to FLEX spatial scales and science objectives. Our work offers a practical framework for FLEX-related studies, while ensuring temporal continuity before and during the FLEX mission lifetime. As such, this work contributes a complementary, operational pathway that bridges current multi-mission SIF capabilities with the high spectral precision of FLEX, supporting the mission’s long-term goal of advancing quantitative monitoring of terrestrial photosynthesis.

Solar-induced chlorophyll fluorescence (SIF) is a re-emitted signal arising directly from the photosynthetic process that can currently be retrieved from instruments such as TROPOMI on-board of Sentinel-5P. We expect this signal to have some potential in diagnosing vegetation stress, and in particular before this stress is reflected by decreases in greenness indicators. To do so, we would need to disentangle the physiological component, i.e. fluorescence efficiency (ΦF), from other confounding factors present in the SIF signal. Satellite SIF observations (SIF_obs) are strongly modulated by illumination conditions. The canopy structure also determines the fraction (f_esc) of fluorescence that escapes the canopy to the sensor. This is further compounded by the spatial heterogeneity of vegetation elements within the nadir observation footprint. Methods are available to remedy these effects, but they are typically applied only after multiple instantaneous SIF observations have been forced into a convenient common grid. We hypothesize that such harmonization should be done prior to gridding, thereby ensuring these are done over the correct spatio-temporal supports. This would enable reducing uncertainties due to mismatches in space and time.

We therefore propose deriving SIF efficiency (ΦF) from TROPOMI by normalizing SIF_obs with radiation and canopy features prior to spatial gridding. We normalize SIF_obs by photosynthetically active radiation (PAR) and near-infrared reflectance of vegetation (NIRv), where NIRv serves as a proxy for vegetation greenness and canopy structure. We explore ΦF over Germany using multi-source NIRv and PAR datasets and TROPOMI SIF from three retrieval products. We assemble downward shortwave radiation as a proxy for PAR from multiple sources spanning different spatio-temporal resolutions (e.g. MSG, ERA5, TROPOMI). NIRv is derived from other sources (e.g. MODIS, Sentinel-3, Sentinel-2) and aggregated to the TROPOMI footprint, as well as from TROPOMI native TOA reflectance products.

To evaluate whether the ΦF derived from our different combinations of SIF retrievals and alternative PAR–NIRv combinations are sensitive to vegetation stress, we analyse them before and during compound hot and dry (CHD) events. We use the Dheed dataset, an ERA5-based CHD event database at 0.1° spatial resolution. The temporal consistency and sensitivity of ΦF to CHD events is assessed across biomes, with emphasis on homogeneous rainfed croplands and forests.

Capturing the sub-daily variability of terrestrial ecosystem productivity is one of the next frontiers for satellite remote sensing. The changes in the environmental drivers of an ecosystem can often be more radical within the arc of a day than over several months. Plant stress can notably kick in during the afternoon, leading to an afternoon depression of photosynthesis that would ideally be detectable from fluorescence yield measured from space. FLEX will open a new era for fluorescence measurement from space. Being the first mission specifically dedicated to this task, it will offer a comprehensive measurement of the fluorescence signal along with the confounding factors complicating its relationship to photosynthesis. By reaching a spatial resolution of 300m we will largely avoid the complexity of dealing with landscape-level structural heterogeneity. However, its morning overpass and its low revisit frequency remain a major limitation to progress towards capturing diurnal dynamics, or the diel scale as it is otherwise known. How can we capitalize on the advantage of FLEX to progress towards the goal of characterizing terrestrial ecosystem diel productivity? The present contribution discusses a roadmap to this end, being therefore directly in-line with the workshop main’s theme of “Combining multi-scale, multi-mission EO data for closing the temporal and spatial gap” and the workshop’s objective to “Plan future activities related to fluorescence and merged data sets”. It involves combining data from various missions (Meteosat Second Generation, Sentinels 2, 3, and 5P), along with downscaling and data fusion tools being developed in Horizon Europe projects Open Earth Monitor (OEMC) and NextGenCarbon, to attempt extending FLEX level 3 products of fluorescence yield to diel frequency. Furthermore, we also elaborate on the potential opportunity to retrieve SIF directly from geostationary satellites such as Sentinel 4 in order to support this goal.