Assessing the benefits of improved spatiotemporal resolution of current geostationary imagers for surface solar irradiance retrievals based on the S2VSR campaign

The current advanced geostationary imagers including the GOES-R ABI and MTG FCI instruments offer significant improvements in terms of spatio-temporal resolution compared to previous instruments, featuring pixel sizes for solar channels down to 500x500m², and scan frequencies up to 1 per min. While these capabilities enable us to better resolve small-scale variability in clouds and radiation, our understanding of the practical benefits for monitoring cloud development and retrieving surface solar irradiance remains limited. One key reason is the limited representativity of many ground-based remote sensing observations serving as potential reference, which are point-like in nature. In contrast, satellite-derived quantities correspond to extended spatial domains.

To improve our knowledge about the small-scale structure and variability of clouds and its influence on solar radiation, the Small-Scale Variability of Solar Radiation (S2VSR) campaign was conducted at the ARM Southern Great Plains (SGP) site in summer 2023. A unique sensor network consisting of 60 autonomous pyranometer stations developed at the Leibniz Institute for Tropospheric Research was deployed at the SGP site for a 12-week period. Stations were distributed across a 6x6 km2 domain centered around the ARM SGP Central Facility. Together with operational ARM measurements including cloud profiling and a stereo-photogrammetric 4D reconstruction of clouds, this campaign offers an unprecedented dataset for studying cloud-induced small-scale variability in solar irradiance, resolving fluctuations down to the second- and decameter-scale.

In the present contribution, a preliminary analysis of the benefits of 500m-resolution retrievals based on the GOES-R ABI imager using the S2VSR data will be given. Specifically, the deviation of satellite retrievals of surface solar irradiance from single-site measurements caused by the limited representativity will be quantified. An estimate of the instantaneous retrieval uncertainty will be given for different cloud situations. Also, the effects of navigation accuracy and the impact of two different parallax correction strategies will be quantified.