Vicarious Radiometric Calibration of the Hyperspectral Imaging Microsatellites SPARK-01 and -02 over Dunhuang, China

Two wide-swath hyperspectral imaging microsatellites, SPARK-01 and -02, were launched on 22 December 2016. Radiometric calibration coefficients were determined for these two satellites via a calibration experiment performed from the end of February to the beginning of March 2017 at the high-altitude, homogenous Dunhuang calibration site in the Gobi Desert in China. In-situ measurements, including ground reflectance, direct transmittance, diffuse-to-global irradiance ratio, and radiosonde vertical profile, were acquired. A unique relative calibration procedure was developed using actual satellite images. This procedure included dark current computation and non-uniform correction processes. The former was computed by averaging multiple lines of long strip imagery acquired over open oceans during nighttime, while the latter was computed using images acquired after the adjustment of the satellite yaw angle to 90°. This technique was shown to be suitable for large-swath satellite image relative calibration. After relative calibration, reflectance, irradiance, and improved irradiance-based methods were used to conduct absolute radiometric calibrations in order to predict the top-of-atmosphere (TOA) radiance. The SPARK-01 and -02 satellites passed over the calibration site on 7 March and 28 February 2017, during which time fair and non-ideal weather occurred, respectively. Thus, the SPARK-01 calibration coefficient was derived using reflectance- and irradiance-based methods, while that of SPARK -02 was derived using reflectance- and improved irradiance-based methods. The sources of calibration uncertainty, which include aerosol-type assumptions, transmittance measurements, water vapor content retrieval, spectral wavelength shift and satellite image misregistration, were explored in detail for different calibration methods. Using the reflectance and irradiance-based methods, the total uncertainty for SPARK-01 was estimated to be 4.7% and 4.1%, respectively, in the <1000 nm spectral range. For SPARK-02, total uncertainties of 8.1% and of 5.9% were estimated using the reflectance- and improved irradiance-based methods, respectively. The calibration methods were also verified using MODIS images, which confirmed that the calibration accuracies were within the expected range. These in-situ measurements, analyses, and results provide a basis for in-orbit radiometric calibration of the SPARK-01 and -02 satellites. These experiments strongly support the use of diffuse-to-global ratio measurements in in-situ vicarious calibration experiments and the addition of spectrally continuous measurements for direct transmittance, which is important for hyperspectral satellite sensors.