18 October 2024
09 October 2024
About
EUMETSAT has been entrusted by the European Commission (EC) to extend the operational atmospheric portfolio of the Sentinel-3 (S3) mission, for the benefit of Near Real Time (NRT) atmospheric and marine users, operational services, EUMETSAT Member States and the Copernicus programme. Based on this mandate, EUMETSAT leads the development of an Aerosol Layer Height (ALH) product using Top Of the Atmosphere (TOA) Near-InfraRed (NIR) spectral radiances, including the (di)oxygen A band (O2-A) absorption acquired by the Ocean and Land Colour Instruments (OLCI) on-board both S3 A and B satellites. O2-A band measurements are widely used to retrieve the height of scattering layers in the Earth’s atmosphere, such as clouds. This capability is exploited in the recently developed Sentinel-3 OLCI Cloud Top Pressure (CTP) but sensitivity studies show that these measurements also carry information on aerosol layer height.
In the current project, a prototype ALH algorithm, OLCOAH (OLCI O2-A ALH) is developed, based on the S5P/TROPOMI ALH algorithm chain principles. The S5P ALH algorithm, developed by KNMI for the European Space Agency (ESA) and Copernicus, relies on O2-A band measurements and is being further developed for the Sentinel-4 and Sentinel-5 missions. The heritage and experience gained by KNMI are being applied to OLCI to develop a similar ALH product for Sentinel-3, with necessary adjustments to account for the different measurement and instrument types, notably the known spectral smile properties, broad spectral channels, and fine spatial resolution. The algorithm includes a state-of-the-art AI (neural network-based) radiative transport framework and a fast optimal-estimation retrieval scheme, enabling near-real time (NRT) retrievals with proper accuracy estimates and user-friendly quality indicators and statistics.
The OLCOAH Sentinel-3 OLCI ALH product is being developed alongside a new auxiliary Sentinel-3 optical pre-calculated Land Surface Reflectivity (LSR) product. Together with the recent Sentinel-3 OLCI CTP product, this will extend the atmospheric portfolio of the Sentinel-3 mission. Furthermore, preliminary constraints on aerosol height may help to improve the estimation of water reflectance in the blue spectrum. This improvement is currently being studied in the framework of the OLCI Ocean Colour – Standard Atmospheric Correction (OC-SAC) processor development, led by EUMETSAT.
Figure 1 provides an example image of a smoke plume from the devastating ‘Camp Fires’ in 2018 over the Pacific, showing the height of the plume and a few thin clouds overlying it. The ALH in this figure was derived from S5P/TROPOMI O2-A band measurements.
Objectives
- Develop a state-of-the-art ALH prototype algorithm using S3 A&B OLCI O2-A band measurements, incorporating the latest insights from the S5P, S4, and S5 ALH algorithm experiences. The algorithm should be based on a neural network forward model and an optimal estimation retrieval algorithm.
- Anticipate the needs for future Near Real Time (NRT) processing and fast computing constraints.
- Define user requirements for the prototype S3 OLCI ALH (OLCOAH) algorithm, including cloud masking, smile correction, L1B spatial resolution, surface albedo fitting, etc.
- Validate the OLCOAH algorithm and determine the accuracy of the ALH using existing aerosol layer height and vertical distribution data. Examples of such data include TROPOMI ALH data, MISR stereoscopic data, CALIOP active sensing data (weighted extinction), and ground-based EARLINET data.
- Optionally, validate the co-developed LSR data using the OLCOAH.
Overview
The project consists of several work packages performed simultaneously. During the initial science review period, the OLCI L1B requirements and spatial resolution are determined, a time-dependent smile correction is implemented and a suitable OLCI measurement set is selected, including both Sentinel-3 A and B missions. Next, the OLCOAH prototype algorithm will be developed, based on the existing TROPOMI ALH. A neural network will be trained to simulate global OLCI-specific reflectance measurements, and initial validation using collocated TROPOMI ALH measurements will be conducted to test and optimize the OLCOAH algorithm. Finally, a final prototype algorithm will be created following current TROPOMI and OLCI conventions and extensively validated using the selected OLCI A&B data set and the reference data mentioned above. Optionally, the project can be extended to include validation of the newly developed auxiliary LSR surface product in the OLCOAH algorithm, initially using the existing Sentinel-5 Precursor2 Directional LER surface reflectivity climatology.
