Progress towards the assimilation of all‐sky infrared radiances: an evaluation of cloud effects

Abstract

As a step toward the assimilation of cloud‐affected infrared radiances in multi‐layer cloud conditions, this study evaluates cloud effects on model first‐guess simulations (background) and observations using the Infrared Atmospheric Sounding Interferometer (IASI) radiances. It is found from an extensive statistical analysis that over oceans the magnitude of observation‐minus‐background departures (O–B) – even in the most cloud‐sensitive window channels – is typically less than 10 K for 85% of all‐sky IASI data. A parameter has been developed to express the magnitude of the cloud effect based upon observed and simulated cloudy radiances. It is shown that the variations in the standard deviation (SD) of O–B departures can be described (and thus predicted) by this cloud effect parameter – such that the probability density function (PDF) of O–B normalized with predicted O–B SD exhibits a near‐Gaussian form. It is argued that the predicted cloud effect can be used in an assimilation context to define cloud‐dependent quality controls and aid observation error assignment. Simple linear estimation theory is used to simulate the possible benefits of state‐dependent observation errors according to cloud effect.