Ten-Year Assessment of IASI Radiance and Temperature

Marie Bouillon,Marie Doutriaux-Boucher,Dorothée Coppens,Cathy Clerbaux,Thomas August,Elsa Jacquette,Simon Whitburn,Lieven Clarisse,Juliette Hadji-Lazaro,Sarah Safieddine

doi:10.3390/rs12152393

Abstract

The Infrared Atmospheric Sounding Interferometers (IASIs) are three instruments flying on board the Metop satellites, launched in 2006 (IASI-A), 2012 (IASI-B), and 2018 (IASI-C). They measure infrared radiance from the Earth and atmosphere system, from which the atmospheric composition and temperature can be retrieved using dedicated algorithms, forming the Level 2 (L2) product. The operational near real-time processing of IASI data is conducted by the EUropean organisation for the exploitation of METeorological SATellites (EUMETSAT). It has improved over time, but due to IASI’s large data flow, the whole dataset has not yet been reprocessed backwards. A necessary step that must be completed before initiating this reprocessing is to uniformize the IASI radiance record (Level 1C), which has also changed with time due to various instrumental and software modifications. In 2019, EUMETSAT released a reprocessed IASI-A 2007–2017 radiance dataset that is consistent with both the L1C product generated after 2017 and with IASI-B. First, this study aimed to assess the changes in radiance associated with this update by comparing the operational and reprocessed datasets. The differences in the brightness temperature ranged from 0.02 K at 700 cm−1 to 0.1 K at 2200 cm−1. Additionally, two major updates in 2010 and 2013 were seen to have the largest impact. Then, we investigated the effects on the retrieved temperatures due to successive upgrades to the Level 2 processing chain. We compared IASI L2 with ERA5 reanalysis temperatures. We found differences of ~5–10 K at the surface and between 1 and 5 K in the atmosphere. These differences decreased abruptly after the release of the IASI L2 processor version 6 in 2014. These results suggest that it is not recommended to use the IASI inhomogeneous temperature products for trend analysis, both for temperature and trace gas trends.

Highlights

Surface and atmospheric temperatures are both Essential Climate Variables (ECV) that critically contribute to the characterization of Earth’s climate [1]
The differences between the reprocessed and operational radiance values were computed at four selected wavenumbers (700, 1200, 1700, and 2200 cm−1) and for four 1◦ × 1◦ regions with various latitudes/longitudes and characteristics:
If a daily file did not contain any observation of the region with a pixel in the right Field of Regard (FoR), the difference for this day, region, and FoR was set to Not a Number (NaN)

Summary

Introduction

Surface and atmospheric temperatures are both Essential Climate Variables (ECV) that critically contribute to the characterization of Earth’s climate [1]. Recent efforts in the Global Climate Observing System (GCOS) Reference Upper-Air Network (GRUAN) [10] have aimed to standardize quality assurance practices in releasing sondes and processing their data at a few stations Overall, they consist of inhomogeneous records (e.g., from different instruments, calibration, and data processing). Other biases are related to the diversity in instrument characteristics and temperature retrieval algorithms that are based on different assumptions When these datasets are used in climate and Numerical Weather Prediction (NWP) models, the error due to this homogenization of different data records becomes difficult to assess. This emphasizes the importance of using a single instrument with a global spatial coverage and long time series that is homogeneous and consistent for the assessment of climate variables

Objectives

Methods

Results

Discussion

Conclusion