| 10 Jun 2026
EarthCARE MSI Level-1 Data: First In-Orbit Calibration and Validation
Abstract. The successful launch of the EarthCARE satellite on 28 May 2024, with its unique combination of active and passive instruments, enables new insights into advanced cloud-aerosol-radiation interactions in the atmosphere. The Multi-Spectral Imager (MSI) onboard the EarthCARE satellite provides horizontal observations of reflected visible, near-infrared, and short-wave infrared (VNS) radiation and emitted thermal infrared (TIR) radiation across a 150-km swath. Following 1.5 years of calibration and validation, the MSI instrument is performing very well.
This paper focuses on the main findings of the validation and calibration of the MSI Level-1 data since the launch. One major improvement in the MSI Level-1 processing was the introduction of band-specific pixel geolocation and inter-band coregistration corrections. These improvements have been shown to achieve the necessary instrument performance, successfully meeting the specifications for geolocation accuracy of ±500 meters and coregistration accuracy of ±0.3 pixels.
The radiometric accuracy of MSI was assessed against the Flexible Combined Imager (FCI) onboard the Meteosat Third Generation (MTG) satellite using collocated cloud scenes. The results showed systematic offsets in MSI Level-1 VNS radiances compared to the corresponding FCI spectral bands. These findings led to the conclusion to perform an ad-hoc vicarious calibration of the MSI VNS radiances against FCI. The initial verification results for the improved and reprocessed MSI Level-1 BA dataset confirm the successful mitigation of the observed offsets between MSI and FCI with relative biases below 1 %.
In addition, the VNS reflectances exhibited unexplained across-track variations that could not be attributed to the radiance measurements. Monitoring of the spectral solar irradiance, measured by MSI's onboard diffusers, revealed across-track artifacts and day-to-day variations exceeding expected solar variability. Therefore, the measured spectral solar irradiance has been replaced by theoretical values in the MSI Level-1 product. All of these updates led to a significantly improved radiometric and geometric stability.
In conclusion, I believe that this manuscript should be accepted after minor revisions and the addition of several clarifications.
This manuscript describes the calibration and correction procedures for the Multi-Spectral Imager (MSI) onboard the EarthCARE satellite, which was launched in May 2024. The paper is broadly divided into two major topics: geometric correction and radiometric calibration. The manuscript provides a clear explanation of the EarthCARE mission objectives and a thorough description of the MSI calibration process. The correction and calibration methodologies are presented in a logical and systematic manner, and the technical content appears accurate. Therefore, this paper can be regarded as an important reference for future MSI science applications.
Overall, I do not find any major issues with the manuscript. However, I have identified several points that should be clarified before acceptance. These are minor comments rather than critical flaws, but addressing them would further improve the quality and completeness of the paper. I encourage the authors to revise the manuscript accordingly and incorporate the following comments into the final version.
1, Line 33
The manuscript states that “only MSI provides radiance and brightness temperature measurements accurately co-located with the footprint of the active instruments with consistent viewing geometry.” However, it would be appropriate to mention that the CloudSat project provides the MODIS-AUX dataset, in which Aqua/MODIS observations are collocated with CPR footprints.
2, Line 81
Can PICS-based calibration only be applied to targets with moderate reflectance levels? Is it possible that radiometric calibration cannot be adequately performed for very bright targets? Please clarify the upper reflectance range for which this method is considered applicable (e.g., up to approximately what reflectance percentage).
3, Line 273
In the figure presented here, the solar irradiance has been normalized to a Sun–Earth distance of 1 AU. Is the MSI Level-1C radiance product also normalized to 1 AU, or does it represent absolute radiance corresponding to the actual Sun–Earth distance at the time of observation?
4, Table 3
The radiance simulations consider only three cloud-top pressure (CTP) values: 200, 500, and 800 hPa. These correspond approximately to cloud altitudes ranging from 10 km to 2 km. For spectral bands that are sensitive to water vapor absorption, these three cases may not adequately represent the full range of atmospheric conditions. It would be helpful to discuss the radiance differences between the 200 hPa and 800 hPa cases and demonstrate that these three simulations are sufficient for the intended analysis.
5, Table 4
The Spectral Band Adjustment Factors (SBAFs) appear to be characterized not only by their median (MED) values but also by their standard deviations (STD). If available, I suggest adding the STD values to Table 4 to provide a better indication of the variability and uncertainty associated with the SBAFs.
6, Line 310
What level of radiance variability between MSI and FCI is expected due to anisotropic cloud reflectance? Please provide an estimate of the magnitude of this effect (e.g., whether it can be constrained within a certain percentage).
7, Line 439
For the alternative SWIR calibration approach, fixing the cloud effective radius may not be an ideal assumption. Can the error introduced by prescribing a fixed cloud effective radius be quantified in advance? A discussion of the expected uncertainty associated with this assumption would strengthen the analysis.
That's all of my comments