Sensitivity of deep convective cloud climatology to inconsistency of spectral response functions among various generations of satellite sensors

Building a homogenous and reliable cloud climatology requires merging observations collected over decades with generations of satellites. Each generation introduces improved instrumentation, causing heterogeneity in terms of spatial, radiometric, and/or spectral resolution. Even if a specific spectral channel is preserved on different satellites (heritage bands), detailed characteristic of a channel may still be different among individual devices. Specifically, each sensor will reveal a different sensitivity to radiation of given vale length, a feature quantitatively defined by the sensor’s Spectral Response Function (SRF).

In this study we evaluate how the differences in SRFs impact the homogeneity of global cloud climatology. We focus on specific type of clouds (deep convective clouds, DCCs), and consider all generations of the Meteosat satellites that have been in service since 1977. Over that time Meteosats featured the three following imagers: MVIRI (Meteosat Visible Infra-Red Imager; first generation of Meteosat), SEVIRI (Spinning Enhanced Visible Infra-Red Imager; second generation), and FCI (Flexible Combined Imager; third generation).

In order to meet the study goal we simulate radiances for MVIRI, SEVIRI and FCI by convolving sensors’ SRF with a high-spectral-resolution data collected globally with the Infrared Atmospheric Sounding Interferometer (IASI) onboard MetOp satellite. Two infrared heritage channels of Meteosat are examined: the infrared window channel (IR, ~11 mm) and the water vapor absorption channel (WV, ~6.5 mm). Both are essential for DCC detection.

Based on simulated radiances we detect deep convective clouds examining brightness temperature difference between WV and IR channels. Whenever the difference is greater than 0K, an IASI pixel is flagged ‘DCC’. Then a DCC frequency is computed globally, and individually for MVIRI, SEVIRI and FCI. Statistics is developed for three summer months of 2013 (June-August).

The results demonstrate how much of a difference in DCC cloud amount between MVIRI, SEVIRI, and FCI may be attributed to a different spectral sensitivity in IR/WV for that sensors. More generally, we address the question how sensitive is a (DCC) cloud climatology to spectral evolution of clouds observing instrumentation.

This research was funded by the National Science Centre of Poland. Grant no. UMO-2020/39/B/ST10/00850.