Speaker
Description
The spectral dimension in climate applications is crucial for studying the evolution of key climate variables, characterizing relevant driving climate mechanisms, and identifying biases in climate model simulations. Today, this type of studies is made possible by the long time series of spectrally resolved radiances collected from instruments operating since the early 2000s, such as AIRS (2002-present), IASI (2006-present) and CrIS (2011-present). These instruments provide hyperspectral observations in the mid-infrared (MIR) range (667 to 2750 cm⁻¹) of the Earth emission spectrum, capturing essential data on key climate variables such as temperature, water vapor, and clouds. This spectral range also includes absorption lines from a variety of gases (CO₂, CH₄, O₃, N₂O, CO, SO₂, etc.). Additionally, the forthcoming FORUM mission, scheduled for launch in 2027, will offer unique spectrally resolved measurements extending into the far-infrared (FIR) region (100 to 667 cm-1). The long-term sequence of observations from IASI and FORUM will enable further exploration of spectral climate feedback, their representation in climate models across inter-annual and decadal timescales and provide additional constraints for climate model evaluation.
In this work, 12 years (2008-2019) of IASI Metop-A measurements are compared to simulated spectral radiances provided by the EC-Earth GCM (ECE, version 3.3.3) based on the atmospheric and surface fields predicted in all-sky conditions by the model. An innovative strategy is adopted to consider the cloud variability within the large model grid cell (roughly 80-km grid spacing near the equator) and to optimally compare the climate model outputs with the higher spatial resolution (about 15 km of diameter) observations performed by the instrument. The spectral radiances are simulated online every 3 hours by the σ-IASI radiative transfer model, that has been previously embedded in the climate model through the Cloud Feedback Model Intercomparison Project (COSP) module.
Spectral biases in the climate model are identified by comparing the trends of spectral radiances across different wavenumbers. Additionally, the impact of key climate variables on the temporal evolution of observed and simulated radiances is evaluated using precomputed spectral kernels.
