Preparing for the Copernicus Hyperspectral Imaging Mission for the Environment (CHIME)
Marco Celesti
The Copernicus Hyperspectral Imaging Mission for the Environment (CHIME), developed by the European Space Agency (ESA), stands as a cutting-edge initiative that promises to revolutionize our understanding of the Earth’s environment. This advanced mission harnesses the power of hyperspectral imaging, a technology that captures a vast range of wavelengths, to unveil unprecedented insights into our planet’s natural processes, ecosystem health, and environmental changes.
CHIME’s primary objective is to provide scientists, researchers, and policymakers with comprehensive and highly accurate data for monitoring and assessing environmental dynamics. The mission’s hyperspectral sensors are capable of detecting an extensive range of electromagnetic wavelengths, enabling it to capture detailed spectral signatures of Earth’s surface materials and vegetation. This level of precision is poised to unlock valuable information about the health of forests, vegetation types, soil composition, and water bodies.
One of the key features of CHIME is its ability to observe the environment across various temporal and spatial scales. This flexibility allows scientists to monitor gradual changes and sudden events, such as forest fires, pollution incidents, and natural disasters, with an unprecedented level of accuracy. The data gathered by CHIME will contribute to enhancing our understanding of climate change, land use patterns, and ecosystem dynamics, ultimately aiding in the development of effective environmental management strategies.
In addition to its scientific significance, CHIME holds great potential for applications in agriculture, forestry, urban planning, and disaster management. The information derived from CHIME’s hyperspectral data can lead to more informed decision-making, resource allocation, and sustainable practices.
As ESA continues to develop and refine the CHIME mission, anticipation grows within the global scientific community. The potential to unveil new insights into the Earth’s environment and provide data-driven solutions to pressing environmental challenges showcases CHIME’s significance in shaping our understanding of the planet we call home. With its state-of-the-art technology and unparalleled capabilities, CHIME holds the promise of becoming a cornerstone in Earth observation, enabling us to safeguard and preserve our environment for generations to come.
Discovering the world’s aquatic ecosystems through spaceborne spectroscopy: status and prospects
Nima Pahlevan, Astrid Bracher, Maycira Costa, Richard Lucas
With the planned launch of several imaging spectrometers by the end of the decade, the aquatic remote sensing community and its user base are ushering in a new era where scientific studies and applications will leverage dramatically increased information content expected to decode novel biological/biogeochemical discoveries. Over the past few years, the Italian space agency’s PRISMA and German Aerospace Center’s EnMAP spectrometers have offered opportunities to prove the potential utility of this enhanced spectral capability relative to existing multispectral missions over fresh and coastal waters. NASA’s recently launched Earth surface Mineral dust source InvesTigation (EMIT) imaging spectrometer aboard the International Space Station (ISS) also serves additional over-water hyperspectral imagery in rarely studied semi-arid regions of the world. The highly anticipated launch of the Plankton, Aerosol, Cloud, and ocean Ecosystem (PACE) mission, which will be the first operational hyperspectral ocean color mission, will empower scientists and satellite practitioners to measure the global oceans, coastal zones, and large inland waters from a hyperspectral imager and two multi-angular polarimeters. Toward the end of the decade, the enthusiasm will be fulfilled by the launch of the Geostationary Littoral Imaging and Monitoring Radiometer (GLIMR), Copernicus Hyperspectral Imaging Mission for the Environment (CHIME), and Surface Biology Geology (SBG) missions. This informational session invites presentations and discussions centered around the current, upcoming, planned, and proof-of-concept missions with hyperspectral capability enabling advancements in regional and/or global aquatic science and applications in the years to come.
Exploring the synergies between imaging spectrometer missions and between multispectral data and imaging spectrometer data for advancing applications
Uta Heiden
Imaging spectrometers have advanced our understanding of the dynamic processes of the Earth ecosystems by quantifying geochemical, biochemical and biophysical properties of the surface. They collect data in many consecutive spectral bands allowing the quantification of the photon-matter interactions with very high spectral detail and accuracy.
Recent years have seen a dramatic increase in the number of orbital imaging spectrometers, with missions by various agencies including the Italian PRISMA mission, the Germany/US-American DESIS mission, the German EnMAP mission, the US mission EMIT, the Japanese HISUI and the Chinese GaoFeng-5 mission. More will follow in the near future, such as NASA’s SBG and ESA’s CHIME mission that anticipate global coverage at regular intervals. Additionally, frequent and regular data from the Sentinels are available and can strengthen applications due to various data fusion approaches.
Acquiring imaging spectroscopy data across the globe demands a significant coordination effort for acquisition planning, harmonized data quality activities and open data policy. Because of the breadth and diversity of imaging spectroscopy, careful consideration of how the data is managed and the processing and distribution systems are designed will be crucial to enabling its use and preserving the consistency of combined datasets.
This session shall include presentations exploring the synergies between multispectral sensors and imaging spectrometers as well as contributions that combine the different available imaging spectrometers for improving the density of such measurements, for larger spatial coverage of analysed areas, etc. for the development of novel or improved higher-level application products. We would like to see different data fusion concepts and discuss it with a larger community.
Results from the EMIT imaging spectroscopy mission on the International Space Station
Robert O. Green and David R. Thompson
The Earth Surface Mineral Dust Source Investigation (EMIT) is a NASA mission to operate a Visible Shortwave Infrared (VSWIR) imaging spectrometer onboard the International Space Station (ISS). EMIT’s primary aim is to improve our understanding of the mineral dust cycle, a critical part of the Earth System. Mineral dust is an important atmospheric aerosol in the climate system. It erodes from arid land surfaces and is emitted up into the atmosphere where it can disperse over wide areas. Once airborne, it directly modifies atmospheric radiative forcing by absorbing and scattering solar radiation, a process that is critically dependent on the mineral composition of the dust itself. EMIT aims to map the surface mineralogy of these dust-forming areas so that Earth System Models can better predict the composition of mineral dust entering the atmosphere and the resulting impacts on both local and planetary radiative forcing. The target mask for EMIT includes the Earth’s arid regions and their adjacent lands, ISS timing margin, and calibration sites.
Since its launch and installation in July 2022, EMIT has delivered thousands of radiance, reflectance, and mineralogy products to NASA archives in support of this objective. It has also proven useful for a range of other objectives, including the measurement of methane and carbon dioxide greenhouse gas point sources and the mapping of snow properties. EMIT’s imaging spectroscopy measurements cover a large fraction of Earth’s land area, including a wide range of diverse surface cover types and regions. EMIT data and science data processing code are publicly available for all researchers via NASA archives and github respectively.
This session will discuss early results from the EMIT mission, including imaging spectrometer performance and calibration, reflectance retrievals and validations, methane point sources, and other early science results.
Status and applications of the PRISMA mission at the turn of 5 years in orbit
Giorgio Licciardi, Ettore Lopinto, Patrizia Sacco, Maria Elena Cianfanelli, Maria Daraio
PRISMA, the first ASI hyperspectral mission is close to 5 years in orbit, during which it has seen an increasing exploitation of the data among the scientific / applicative community. This special section will present an updated status of the mission, of the CALVAL activities conducted by the system & scientific teams and of the range of innovative applications enabled by the PRISMA technology. A special insight on two themes deemed particularly relevant for spaceborne imaging spectroscopy exploitation will drive the session to the final panel discussion in which the engagement of participants in insightful discussions will allow the sharing of information and experiences. The special session will include 1) Mission Status; 2)Cal/Val #1 (system activities), 3)Cal/Val #2 (scientific activities), 4) Review of Applications, 5) Insight on thematic area #1, 6) Insight on thematic area #2 and 7) Short panel discussion.
EnMAP’s first two years in orbit- current status and recent activities
Nicole Pinnel, Saskia Förster, Anke Schickling and Sabine Chabrillat
EnMAP (Environmental Mapping and Analysis Program) is a high-resolution imaging spectroscopy mission designed to monitor and characterise the Earth’s environment by providing accurate information on the state and evolution of terrestrial and aquatic ecosystems. EnMAP is equipped with a prism-based dual spectrometer capable of making observations in the spectral range between 418.2 nm and 2445.5 nm with 224 bands and high radiometric and spectral accuracy and stability.
EnMAP was launched into a Sun-synchronous orbit on 1 April 2022 on a SpaceX Falcon 9. The mission successfully completed the commissioning phase at the end of October 2022 and the data has since been freely available to the user community after registration at https://planning.enmap.org/.
The satellite was realised by OHB System AG and the EnMAP ground segment by the German Aerospace Center (DLR) in Oberpfaffenhofen. The Principal Investigator of EnMAP is the GFZ Potsdam, which is supported in its activities by the EnMAP Science Advisory Group (EnSAG), a body of national and international scientists. The project management of the EnMAP mission lies with the DLR Space Agency.
This informative session invites presentations from the mission consortium to take stock of the mission after two years in orbit and provide updates on its current status, recent developments and activities.
Advances in DESIS data products and applications
Uta Heiden
In 2014, Teledyne Brown Engineering (TBE, USA) and the German Aerospace Center (DLR, Germany) collaborated to build and operate the DESIS instrument. DESIS is installed into the Multi-User System for Earth Sensing (MUSES) platform on the International Space Station (ISS), with TBE providing the infrastructure for operations and data tasking. DLR built the instrument and developed and maintain the Ground Segment of the mission including software for data processing, data calibration and delivery and is responsible for the scientific exploitation of the mission.
Since the launch of the DESIS instrument to the ISS in July 2018, DESIS is collecting imaging spectroscopy data from the Earth’s surface. Today, the DESIS mission is the only currently running imaging spectrometer mission that has a data archive of 5 years. More than 237.000 scenes (status July 2023) of about 30 km x 30 km are available free and open for the science community to derive L3 and higher products. For a multitude of areas, DESIS can collect data more than one time. For the test sites of ESA’s future imaging spectrometer mission CHIME, DESIS provides recurrent observations starting from 2020 until today. This way, the DESIS data archive provides an excellent opportunity for measuring Earth system processes.
Imaging spectrometers have advanced our understanding of the dynamic processes of the Earth ecosystems by quantifying geochemical, biochemical and biophysical properties of the surface. In this way, this session shall present an overview about the mission status, a short update about the DESIS L1 – L2 processors including data calibration and validation, data quality and spatial and temporal matchups with other imaging spectrometer missions. Further, the session shall present advances in the development of imaging spectroscopy techniques to derive novel and reliable application data to (1) identify Earth surface materials such as minerals, urban surface materials and vegetation invasive species; (2) quantify biochemical parameters such as the chlorophyll content of vegetation stands, organic matter content of soils, and water pollutants such as oil spills; (3) identify the background signal shadowing effects in urban areas, soil background for agricultural fields or vice versa with unmixing techniques.
Imaging spectroscopy for climate robust agriculture
Stephanie Delalieux, Stefan Livens
Climate change poses significant challenges to agriculture, affecting crop productivity, water availability, and overall food security. As the agricultural sector strives to adapt to the changing climate, innovative technologies are becoming essential. Imaging spectroscopy offers a promising solution by providing detailed spectral information, enabling climate robust agriculture practices. This special session aims to explore the potential of imaging spectroscopy in addressing climate challenges and fostering sustainable and resilient agricultural systems.
Session Topics:
- Applications of Imaging Spectroscopy to support Climate Robust Agriculture: Presentations on the diverse applications of imaging spectroscopy supporting more resilient agriculture, such as identifying climate-resistant crop varieties, assessing drought stress, identifying early signs of climate-induced stress in crops, thereby allowing timely interventions for adaptation, agroecology, litter detection and quantification, soil management practices.
- Precision Farming with Imaging Spectroscopy: Presentations showcasing how imaging spectroscopy facilitates precision agriculture practices, optimizing resource use and crop performance in a changing climate.
- Data Fusion and Integration: Discussions on integrating imaging spectroscopy data with climate and environmental datasets to enhance accuracy and understanding of agricultural responses to climate change. Also, crop modeling studies driven by imaging spectroscopy features are welcome.
- Machine Learning: Exploring the use of machine learning algorithms to analyze complex spectral data in support of a more robust agriculture.
Hyperspectral remote sensing of vegetation health
Roshanak Darvishzadeh, Miriam Machwitz, Clement Atzberger, Katja Berger
By the year 2050, the world’s population is anticipated to grow to around 9.8 billion. Further, global weather patterns have grown progressively erratic, and the planet is experiencing rising temperatures. Consequently, there has been a notable increase in extreme weather events, altered precipitation patterns and severe droughts that have exacerbated the effects of abiotic and biotic stressors on vegetation health, causing adverse consequences on biodiversity and food security. Hence monitoring vegetation health through their physiological traits that are key to adaptation processes in such stressed environments is crucial. With its massive, detailed spectral information, hyperspectral remote sensing has been demonstrated to be viable in characterising plants’ physiological traits. Yet, its use for assessing, monitoring and analysing the dynamics of these traits under healthy and stressed conditions for developing early detection methods and management strategies has not been widely explored.
The proposed session is planning to bring together researchers working in the field of hyperspectral remote sensing and vegetation health and stress to discuss relevant and innovative achievements as well as the existing shortcomings. The session will consist of studies in different terrestrial ecosystems, such as croplands and forests and at various levels (leaf, canopy and landscape) using field, airborne and newly launched satellite spectrometers as well as the upcoming satellites (e.g. CHIME) capabilities.
The session will include 4-5 presentations which will then be followed by a round table discussion.
Hyperspectral Remote Sensing of Forest Traits
Roshanak Darvishzadeh, Andrew Skidmore, Martin Schlerf
Forests serve as the primary hub for converting carbon dioxide into carbon and oxygen globally. The European Union’s forest strategy necessitates continuous and up-to-date evaluations of the state of European forests, building upon the groundwork laid in the Mapping and Assessment of Ecosystems and their Services (MAES). Accurate mapping and prediction of forest structural and functional traits play a crucial role in evaluating forest states and guiding the protection and responsible management of the valuable natural resources found within forests. The increasing focus in forest management in Europe on uneven-aged and mixed forest stands and single-tree precision forestry (Fassnacht et al. 2023) could lead to a rising demand for hyperspectral and very high-resolution images.
Hyperspectral remote sensing with a large number of continuous narrow spectral bands has been demonstrated to be effective in characterizing several physiological traits in forest ecosystems. The advancement of hyperspectral sensor technology, expansion of leaf and canopy radiative transfer models and developments of new machine learning methods have opened opportunities for monitoring unique physiological traits which were hard to study. While studies have successfully estimated various species traits using airborne and field spectrometry data, estimation of these variables is still not widely explored using the new generation of hyperspectral satellites.
This session seeks to gather researchers who are investigating forest structural and functional traits using hyperspectral remote sensing. The primary goal is to foster discussions on the latest innovative breakthroughs in this area while also addressing existing limitations or shortcomings that need to be overcome. The session will consist of studies at different levels, from leaf to canopy and landscape level, using laboratory and airborne and satellite spectrometers as well as forthcoming satellites like CHIME and their capabilities.
This session will include 4-5 presentations and will be followed by a round table discussion.
Advancements in field and laboratory measurements of vegetation spectra
Miina Rautiainen and Lucie Homolová
The increasing availability of hyperspectral remote sensing data allows complex analysis of terrestrial vegetation ecosystems. However, knowledge of the variability and spectral behaviour of key ecosystem components (individual species, leaves, woody structures, understorey vegetation) is still essential for physically-based vegetation reflectance models and interpretation of remote sensing data.Non-imaging and imaging spectroradiometers can be used under laboratory and field conditions to capture spectral properties of vegetation. This session invites oral presentations on recent advances invegetation spectroscopy, covering topics such as 1) laboratory and field measurements of vegetation spectra, including their elements such as leaves, needles, woody structures, 2) creation and application of spectral libraries for physically based modelling.
Assessment of advancing water quality monitoring with hyperspectral satellite imagery and explainable machine learning
Patricia Urrego, Ana B. Ruescas and Katalin Blix
Water quality monitoring is of utmost importance for the sustainable management and protection of our natural water resources. Traditional remote sensing methods of water quality estimation have been effective but often limited in scope and resolution. In recent years, there has been a growing interest in leveraging advanced technologies to enhance the accuracy and efficiency of water quality monitoring. Hyperspectral imaging is an example of a cutting-edge technology, which allows for the collection of high resolution spectral data and provides valuable insights into water quality parameters, including chlorophyll levels, suspended sediment concentrations and dissolved organic matter. Hyperspectral images could also be an asset for developing new products and parameters for water science.
However, the wealth of data generated by hyperspectral imaging poses challenges in the analysis and interpretation of results. This is where explainable machine learning comes into play. Machine learning algorithms have demonstrated remarkable capabilities in extracting patterns and correlations from complex datasets, but their lack of transparency often leaves scientists and decision-makers hesitant to fully trust their results. Explainable machine learning methods seek to bridge this gap by providing clear, interpretable explanations for the predictions and decisions made by these algorithms. In the context of water quality monitoring, XAI can offer crucial insights into the relationships between hyperspectral data – in synergy with multispectral satellite data and in-situ measurements – and various water quality indicators.
Validation of L2A products: content and format for a global joint effort to validate atmospheric correction products
Raquel de los Reyes, Martin Bachmann, Maximilian Brell, Bringfried Pflug, Andreas Hueni
Atmospheric correction processors implemented in the ground segment of hyperspectral EO missions have become a standard to deliver final products with compensated atmospheric effects, ready to be used in the analysis of bio- and geophysical variables. In order to ensure the quality of derived products, networks like AERONET are used for the validation of the atmospheric parameters retrieved during the atmospheric correction to characterize the atmosphere and final correction. Around 400 stations are typically operational around the globe to provide measurements of Earth atmosphere properties like aerosols, water vapor (WV), etc. Already these derived parameters are included in intercomparison experiments like the ACIX / CMIX series.
But as identified in the outcomes of projects like the Copernicus Cal/Val Study (CCVS) and other surveys, the direct validation of the BOA surface reflectance is not as advanced. The resources are limited but gaining momentum with CEOS, ESA and EU supported initiatives like HyperNETs or consolidated CalVal TOA (Top-Of-Atmosphere) networks like RadCalNet, providing BOA in-situ measurements. However, these permanently instrumented sites don’t offer the diversity of surface types and atmospheric conditions necessary to widely validate the BOA products. Therefore, there is the need for supplement these validation data by in-situ measurements from ad-hoc campaigns. Such campaigns often provide reference data for a wider variety of surface types but they are not foreseen for longer time periods. They are also constrained by weather prediction and limited by funding budgets, so no often and long-distance campaigns can be often organized. In addition, adverse atmospheric conditions, if not seasonal, are difficult to predict and therefore to plan, so faster response is expected from geographical local teams.
These campaigns are often planned for overpasses of a specific mission, so there is always a RS + in-situ measurements set and the TOA data from some of these RS missions are often available to the scientific community. There, any atmospheric correction processor, able to process the TOA data from a particular mission, could test its performance if the corresponding in-situ measurements were available.
For these reasons, a global joint effort sharing the data from these ad-hoc campaigns would be very helpful for the validation of AC processors, especially if a common data format and content is agreed.
This session intends to bring together developers of atmospheric correction processors, experts involved in related Cal/Val infrastructures and field campaigns, as well as and researchers interested in validating the quality of the hyperspectral BOA reflectance data.
Topic include, but are not limited, to the harmonization of campaign planning, measurement and upscaling protocols and exchange formats, the definition of mandatory information needed for the validation of the BOA products.
Emulation for imaging spectroscopy applications
Jorge Vicent Severa, Jouni Susiluoto, Jochem Verrelst
Recent years have seen a large increase in the development and availability of hyperspectral data from satellite missions, such as ENMAP, PRISMA, CHIME, SBG, FLEX, and DESIS. These missions provide unprecedented insights into the Earth’s surface, but they also pose new challenges for processing and analyzing the data. One of the challenges is the computational cost of running radiative transfer models (RTMs). RTMs are used to simulate the interaction of light with the Earth’s surface, and they are essential for accurately interpreting hyperspectral data. However, RTMs can be very computationally expensive, especially for pixel-wise calculations. While Look-Up Tables (LUTs) interpolation is often used to speed up the calculation of operational data processing applications, LUTs are also challenged by the large data volume needed to meet accuracy requirements.
Emulation is a technique that can be used to overcome the computational cost and accuracy limitations of LUT interpolation. Emulation is a statistical approach that uses a small number of RTM simulations to build a predictive model that can be used to calculate the spectral outputs generated by an RTM. Emulation has several advantages over LUT interpolation. First, it is much faster for pixel-wise calculations. Second, it is more accurate than interpolation for LUTs of moderate size, especially for input bio/geophysical conditions that are not well-represented in the LUT. Third, it is more flexible, as it can be used to simulate a wider range of input conditions.
The state-of-the-art in emulation for imaging spectroscopy is rapidly evolving. Recent advances have focused on improving the accuracy and flexibility of emulators, as well as on developing new methods for training and evaluating emulators. One of the most promising recent advances is the use of deep learning for emulation. Deep learning is a type of machine learning that can learn complex relationships from data. This makes it well-suited for the task of emulating RTMs, which can be very complex.
In this session, we will present the state-of-the-art of emulation in the field of imaging spectroscopy. We will share the latest advances in emulation methods, and we will discuss the challenges and opportunities of using emulation for hyperspectral data analysis.
Thermal Infrared (TIR) remote sensing special session
Jennifer Adams, Agnieszka Soszynska, Elnaz Neinavaz
Thermal InfraRed (8-14 µm TIR) remote sensing provides a unique way to obtain parameters, such as Land Surface Temperature and Emissivity (LST&LSE), which play a critical role in understanding Earth systems and processes. LST&LSE derived from thermal remote sensing data is valuable for numerous applications, including landscape characterisation, analysis and change; quantification of energy fluxes and energy balance (e.g., biosphere, hydrosphere, cryosphere and the atmosphere); detection of thermal anomalies (e.g., urban heat island); extreme events detection by natural (e.g., volcanic eruptions) or man-made (e.g., gas flares, burning oil spills, forest fire) causes; and climate applications.
The need to more accurately monitor LST&LSE combined with advancements in technology and upcoming new satellites missions with unprecedented spatial and temporal resolutions are opening many new opportunities. Notably, growing advancements and needs for hyperspectral TIR data will open new possibilities not only in application areas such as plant stress, geology, geothermal systems and landscape classification, but also in algorithm development, mission preparation and Calibration/Validation activities.
This special session, organised by the EARSeL Thermal Remote Sensing SIG, aims to showcase cutting-edge advancements in the TIR remote sensing including the new methods, data processing procedures, new data acquisition and solutions available, that will be crucial to ensure the adoption of TIR remote sensing data by end-users.
Hyperspectral imaging of chlorophyll fluorescence across scales
PART I: retrieval and modeling trends | PART II: sampling strategies and interpretation
Uwe Rascher, Shari Van Wittenberghe, Bastian Siegmann, Juliane Bendig, MªPilar Cendrero-Mateo
The combination of hyperspectral real surface reflectance and effective chlorophyll fluorescence can be used to assess plant health, productivity, and response to environmental stress. To this end, the upcoming 8th Earth Explorer mission FLuorescence EXplorer (FLEX) is dedicated to investigating chlorophyll fluorescence from space. Retrieving solar induced fluorescence (SIF) from hyperspectral data is a complex process that involves modeling the fluorescence emitted by vegetation from measured solar irradiance and reflected radiance. Atmospheric effects, instrumental noise, and vegetation variability are the main factors encountered in SIF retrieval. A number of different SIF retrieval methods are available, including spectral fitting, physical modeling, and statistical methods. The choice of SIF retrieval method depends on several factors, such as the type of remote sensing data available and the desired accuracy. SIF trends are driven by first-order absorbed photosynthetically active radiation (APAR) and second-order excess light energy dissipated as non-photochemical quenching (NPQ). The relative contributions of APAR and NPQ to SIF trends can vary with plant species, growth stage, and environmental conditions. There are a number of different modeling approaches that have been used to model SIF dynamics; some of the most common approaches include physical, statistical, and hybrid models. The development of SIF models is often limited in terms of spatial and temporal data coverage. This can make it difficult to model long-term trends. In addition, the SIF signal can be affected by a variety of environmental and vegetation structural factors, such as cloud cover, atmospheric conditions, and plant leaf area index and/or fractional cover. This makes it difficult to isolate the effects of environmental stresses on SIF trends. In order to validate SIF modeling trends and/or advance fluorescence reflectance-based products (i.e., APAR and NPQ), an appropriate sampling strategy should be defined. However, depending on the specific application, it is important to ensure that the samples are representative of the target area. Interpretation of reflectance-based products and SIF data can be challenging as it requires a good understanding of the underlying physics and biology. However, with careful analysis, this technique can provide valuable insights into plant health and status.
This session will discuss state of the art studies facing the above described SIF related challenges. It will be divided into two parts [PART I] retrieval and modeling trends and [PART II] sampling strategies and interpretation. Each part will consist of 4-5 presentations.