IASI temperature and water vapor retrievals – error assessment and validation

N Pougatchev,X Calbet,T Hultberg,O Oduleye,P Schlüssel,K St Germain,G Bingham,T August,B Stiller

doi:10.5194/acp-9-6453-2009

Abstract

Abstract. The METOP-A satellite Infrared Atmospheric Sounding Interferometer (IASI) Level 2 products comprise retrievals of vertical profiles of temperature and water vapor. The error covariance matrices and biases of the most recent version (4.3.1) of the L2 data were assessed, and the assessment was validated using radiosonde data for reference. The radiosonde data set includes dedicated and synoptic time launches at the Lindenberg station in Germany. For optimal validation, the linear statistical Validation Assessment Model (VAM) was used. The VAM uses radiosonde profiles as input and provides optimal estimate of the nominal IASI retrieval by utilizing IASI averaging kernels and statistical characteristics of the ensembles of the reference radiosondes. For temperatures above 900 mb and water retrievals above 700 mb, level expected and assessed errors are in good agreement. Below those levels, noticeable excess in assessed error is observed, possibly due to inaccurate surface parameters and undetected clouds/haze.

Highlights

Atmospheric sounders, i.e., systems that remotely measure atmospheric thermodynamic parameters and constituents, are important sources of data for numerous practical and scientific applications such as Numeric Weather Prediction (NWP) and climate studies
Radiosonde launches for the validation campaign and Infrared Atmospheric Sounding Interferometer (IASI) individual Field Of View (FOV) retrievals are schematically presented in Fig. 1
The assessed and expected total retrieval errors are in good agreement above the 900 mb level and are significantly smaller than the temperature variance, which means that the IASI temperature measurements are very informative

Summary

Introduction

Atmospheric sounders, i.e., systems that remotely measure atmospheric thermodynamic parameters and constituents, are important sources of data for numerous practical and scientific applications such as Numeric Weather Prediction (NWP) and climate studies. Rodgers and Connor (2003) demonstrated that even when two different systems perform the measurements on the same state of the atmosphere, a sensible comparison cannot be reduced to a straightforward, point-by-point analysis of differences; proper statistical methods should be used instead to reconcile what we will call characteristic difference error They developed an approach that has been applied to validation of the MIPAS ozone and MOPITT carbon monoxide satellite measurements (Barret et al, 2003). The best estimate of the true atmospheric state and corresponding nominal satellite measurement are provided by the linear statistical Validation Assessment Model (VAM) For this particular study, the VAM uses correlative radiosonde profiles as input and returns the optimal estimate of the nominal IASI retrieval by utilizing IASI averaging kernels and statistical characteristics of the ensembles of the reference radiosondes. Bold lower case symbols denote column vectors; upper case bold typeface denotes matrices, and regular italicized typeface is reserved for scalars

Error assessment and validation technique

Basic relations

Estimation of non-coincidence errors and retrieval noise

Data description

Error assessment

Conclusions