A document by Paul Poli. Click on the document to view its contents.

We investigated various aspects of the in-orbit performance of SEVIRI on Meteosat-10 (launch: 05 Jul 2012) and -11 (launch: 15 Jul 2015) with images, where Mercury or Venus appeared in a corner. These objects are of similar or smaller size than the instantaneous field of view, and therefore they are well suited for checks of geometric requirements. From comparing the position of Venus in different channels we conclude that the North-South distance between the two focal planes is shorter than the nominal value by 0.66 km at SSP (Sub-Satellite Point) with Meteosat-10 and longer by 1.44 km at SSP with Meteosat-11. The tilt of the detector array against the equator is less than 0.0037° for SEVIRI on Metosat-10. The sampling with narrow channels is 3.0016 km, with a one-sigma uncertainty of 30 cm at sub-satellite point. The tests we carried out to check the geometric performance of the instrument confirmed that SEVIRI is compliant with the requirements. The Point Spread Function as determined from the image of a planet agrees well with the expectations based on its combination with the finite impulse response. Finally we determined the stability of the calibration coefficients from the counts obtained on the planetary targets and found the reproducibility of the measurements of planetary fluxes similar to those of vicarious calibration targets. Hence planets are a promising alternative to established methods of in-flight characterisation and validation of imagers.

Tropical cyclones (TCs) are the most destructive weather systems that form over the tropical oceans, with 90 storms forming globally every year. The timely detection and tracking of TCs are important for advanced warning to the affected regions. As these storms form over the open oceans far from the continents, remote sensing plays a crucial role in detecting them. Here we present an automated TC detection from satellite images based on a novel deep learning technique. In this study, we propose a multi-staged deep learning framework for the detection of TCs, including, (i) a detector - Mask Region-Convolutional Neural Network (R-CNN), (ii) a wind speed filter, and (iii) a classifier - CNN. The hyperparameters of the entire pipeline is optimized to showcase the best performance using Bayesian optimization. Results indicate that the proposed approach yields high precision (97.10%), specificity (97.59%), and accuracy (86.55%) for test images.

Global-scale changes in water vapor and responses to surface temperature variability since 1979 are evaluated across a range of satellite and ground-based observations, a reanalysis (ERA5) and coupled and atmosphere-only CMIP6 climate model simulations. Global-mean column integrated water vapor increased by 1\%/decade during 1988-2014 in observations and atmosphere-only simulations but coupled simulations overestimate trends because internal climate variability suppressed observed warming in this period. Decreases in low-altitude tropical water vapor in ERA5 and ground-based observations before around 1993 are inconsistent with simulations and increased column integrated water vapor in a satellite dataset since 1987. AIRS satellite data does not capture the increased tropospheric water vapor since 2002 in other satellite, reanalysis and model products. However, global water vapor responses to interannual temperature variability is consistent across datasets with increases of $\sim$4-5\% per K near the surface and 10-15\%/K at 300 hPa. Global water vapor responses are explained by thermodynamic amplification of upper tropospheric temperature changes and the Clausius Clapeyron temperature dependence of saturation vapor pressure that are dominated by the tropical ocean responses. Upper tropospheric moistening is larger in climate model simulations with greater upper tropospheric warming.