Abstract
Tropical Deep Convective Clouds (DCC) are the brightest, nearly isotropic, tropopause-level earth invariant targets or solar diffusers, making them suitable as pseudo invariant calibration sites (PICS). DCC are observed by both low earth orbit and geostationary (GEO) satellites, since they are found over all tropical domains. DCC calibration is a statistical technique that collectively analyzes all identified DCC over a geographical domain by a simple IR threshold. After the pixel level application of a DCC Bidirectional Reflectance Distribution Function (BRDF), the DCC monthly mode statistic can accurately verify the temporal stability of the sensor response. The temporal standard error is equal to or better than other desert and polar ice PICS. Another advantage of DCC is that they are spectrally flat in the visible and near IR, since they are above the tropopause where there is little water vapor absorption. The most difficult aspect of DCC absolute calibration is the calibration transfer from a reference instrument. The GSICS community has decided on Aqua-MODIS 0.65µm as the calibration reference. As the calibration of future VIIRS or CLARREO hyper-spectral instruments improves, the superior calibration can be transferred back in time using Simultaneous Nadir Overpasses (SNO)s. Since DCC calibration is a statistical method that relies on ~100K DCC identified pixels, the feasibility of pixel-level matching between reference and target sensor is virtually impossible. It also preferable to characterize the DCC with a reference instrument so that the absolute calibration can be inferred even when there is no simultaneous GEO and MODIS data available.
If the frequency histogram of the GEO and MODIS pixel level radiances are very similar and obtained over the same domain and time range with the same DCC algorithm, then the GEO and MODIS mode radiance should be similar. This assumption will be tested by transferring the Aqua-MODIS reference calibration using ray-matched coincident radiance pairs over DCC. Sensitivity to other factors such as GEO and MODIS threshold temperature and pixel resolution differences will also be examined in obtaining the histogram mode DCC radiance.
Deriving a Geostationary Visible Sensor Calibration Reference using DCC Targets Tied to the Aqua-MODIS Band 1 Calibration
Tropical Deep Convective Clouds (DCC) are the brightest, nearly isotropic, tropopause-level earth invariant targets or solar diffusers, making them suitable as pseudo invariant calibration sites (PICS). DCC are observed by both low earth orbit and geostationary (GEO) satellites, since they are found over all tropical domains. DCC calibration is a statistical technique that collectively analyzes all identified DCC over a geographical domain by a simple IR threshold. After the pixel level application of a DCC Bidirectional Reflectance Distribution Function (BRDF), the DCC monthly mode statistic can accurately verify the temporal stability of the sensor response. The temporal standard error is equal to or better than other desert and polar ice PICS. Another advantage of DCC is that they are spectrally flat in the visible and near IR, since they are above the tropopause where there is little water vapor absorption. The most difficult aspect of DCC absolute calibration is the calibration transfer from a reference instrument. The GSICS community has decided on Aqua-MODIS 0.65µm as the calibration reference. As the calibration of future VIIRS or CLARREO hyper-spectral instruments improves, the superior calibration can be transferred back in time using Simultaneous Nadir Overpasses (SNO)s. Since DCC calibration is a statistical method that relies on ~100K DCC identified pixels, the feasibility of pixel-level matching between reference and target sensor is virtually impossible. It also preferable to characterize the DCC with a reference instrument so that the absolute calibration can be inferred even when there is no simultaneous GEO and MODIS data available.
If the frequency histogram of the GEO and MODIS pixel level radiances are very similar and obtained over the same domain and time range with the same DCC algorithm, then the GEO and MODIS mode radiance should be similar. This assumption will be tested by transferring the Aqua-MODIS reference calibration using ray-matched coincident radiance pairs over DCC. Sensitivity to other factors such as GEO and MODIS threshold temperature and pixel resolution differences will also be examined in obtaining the histogram mode DCC radiance.