Quantification of uncertainties of water vapour column retrievals using future instruments
Abstract. This study presents a quantification of uncertainties of total column water vapour retrievals based on simulated near-infrared measurements of upcoming instruments. The concepts of three scheduled spectrometers were taken into account: OLCI (Ocean and Land Color Instrument) on Sentinel-3, METimage on an EPS-SG (EUMETSAT Polar System – Second Generation) satellite and FCI (Flexible Combined Imager) on MTG (Meteosat Third Generation). Optimal estimation theory was used to estimate the error of a hypothetical total water vapour column retrieval for 27 different atmospheric cases. The errors range from 100% in very dry cases to 2% in humid cases with a very high surface albedo. Generally, the absolute uncertainties increase with higher water vapour column content due to H2O-saturation and decrease with a brighter surface albedo. Uncertainties increase with higher aerosol optical thickness, apart from very dark cases. Overall, the METimage channel setting enables the most accurate retrievals. The retrieval using the MTG-FCI build-up has the highest uncertainties apart from very bright cases.
On the one hand, a retrieval using two absorption channels increases the accuracy, in some cases by one order of magnitude, in comparison to a retrieval using just one absorption channel. On the other hand, a retrieval using three absorption channels has no significant advantage over a two-absorption channel retrieval.
Furthermore, the optimal position of the absorption channels was determined using the concept of the "information content". For a single channel retrieval, a channel at 900 or 915 nm has the highest mean information content over all cases. The second absorption channel is ideally weakly correlated with the first one, and therefore positioned at 935 nm, in a region with stronger water vapour absorption.