Intersatellite temperature bias: Elimination through statistical calibration

Global cloud imagery (GCI) is constructed from multiple satellite platforms that simultaneously monitor the earth. The GCI overcomes sampling limitations that are inherent to measurements from an individual platform, and provides a continuous and high-resolution description of the global convective pattern. However, it must reconcile inconsistencies in the measurements from different platforms. Escaping operational stages of error detection is a spurious brightening (cooling of brightness temperature), which appears sporadically in the composited imagery and must be removed a posteriori. The spurious brightening is shown to follow from a bias between measurements from polar-orbiting platforms and those from geostationary platforms. The bias is related to the zenith angle dependence of geostationary measurements, which enables its efficient removal. GCI is then composited from satellites in which the zenith angle-dependent bias has been removed a priori. The corrected imagery is shown to be virtually free of the systematic error, leaving a more accurate representation of the global convective pattern.