This work focuses on the Outgoing Longwave Radiation (OLR) spectrum, its dependence on geophysical variables (atmospheric temperature, water vapor, clouds), natural variability on different time scales, and changes in the context of external forcing of climate. We demonstrate the need, and explore methods, to use spectrally resolved radiances for more completely understanding the longwave radiative feedbacks and climate sensitivity, for validating General Circulation Model (GCM) simulations, and for the monitoring of climate change.
First, we deploy a tool—Radiative Jacobians (the partial derivatives of OLR to the influencing factors—to quantify the rate at which OLR responds to perturbations in temperature and moisture. It is found that the clear-sky Jacobians are insensitive to formulations of water vapor continuum absorption, and are properly captured by the parameterized radiation code in the Geophysical Fluid Dynamics Laboratory (GFDL) GCM when compared to benchmark line-by-line computations. The temperature dependence of gas absorptivity, a factor neglected in previous studies, is found to modify the radiative transfer within the atmosphere-surface system but has little net impact on the OLR. It is also shown that the OLR anomaly can be linearly decomposed into contributions from temperature and water vapor anomalies by using the Jacobians. Hence, the Jacobians offer a computationally efficient method to perform climate feedback analysis.
We configure a moderate spectral-resolution model to simulate OLR spectrum from GCM-generated profiles of the meteorological variables. The simulated spectra are compared to newly available satellite observations; this constitutes a new method of model validation. It is demonstrated that compensatory biases of opposite signs in different spectral regions can lead to a fortuitously good agreement in the broadband longwave flux between model and observation. By spectrally decomposing the OLR response to surface temperature change in the context of seasonal cycles, it is demonstrated that different feedback processes can be distinguished from the signatures appearing in different potions of the spectrum; thereby, different model biases can be diagnosed. Hence, spectrally resolved radiation is proven to be important for understanding the impacts of temperature, water vapor and clouds upon OLR variation. It affords a stricter and more insightful metric for model validation than the conventional broadband flux.
Finally, we investigate the variability and change in the OLR spectrum. The change of OLR spectrum in the context of forced climate change is simulated using the GFDL coupled atmosphere-ocean-land GCM, CM2.1; both the recent temporal evolution and long-term differences are compared to the variations obtained for the unforced climate. Spectrally resolved radiances are shown to have more pronounced (and thus potentially more detectable) changes than broadband fluxes. Band-by-band analyses of the simulation results offer guidance for designing satellite instruments that could be used for long-term climate monitoring, especially regarding the characteristics of spectral resolution and accuracy. Moreover, it is noted that the change in the difference between clear-sky and all-sky radiances, which is often used to measure the change in the radiative effect of clouds, is influenced by changes in the vertical temperature profile.