Three-wavelength approach for aerosol-cloud discrimination in the SAGE III/ISS aerosol extinction dataset

The tropical upper troposphere and lower stratosphere (UTLS) region is dominated by aerosols and clouds affecting Earth's radiation budget and climate. Thus, satellites' continuous monitoring and identification of these layers is crucial for quantifying their radiative impact. However, distinguishing between aerosols and clouds is challenging, especially under the perturbed UTLS conditions during post-volcanic eruptions and wildfire events. Aerosol-cloud discrimination is primarily based on their disparate wavelength-dependent scattering and absorption properties. In this study, we use aerosol extinction observations in the tropical (15°N-15°S) UTLS from June 2017 to February 2021, available from the latest generation of the Stratospheric Aerosol and Gas Experiment (SAGE) instrument-SAGE III onboard the International Space Station (ISS) to study aerosols and clouds. During this period, the SAGE III/ISS provided better coverage over the tropics at additional wavelength channels (relative to previous SAGE missions) and witnessed several volcanic and wildfire events that perturbed the tropical UTLS. We explore the advantage of having an extinction coefficient at an additional wavelength channel (1550 nm) from the SAGE III/ISS in aerosol-cloud discrimination using a method based on thresholds of two extinction coefficient ratios, R ₁ (520 nm/1020 nm) and R ₂ (1020 nm/1550 nm). This method was proposed earlier by Kent et al. [Appl. Opt.36, 8639 (1997)APOPAI0003-693510.1364/AO.36.008639] for the SAGE III-Meteor-3M but was never tested for the tropical region under volcanically perturbed conditions. We call this method the Extinction Color Ratio (ECR) method. The ECR method is applied to the SAGE III/ISS aerosol extinction data to obtain cloud-filtered aerosol extinction coefficients, cloud-top altitude, and seasonal cloud occurrence frequency during the entire study period. Cloud-filtered aerosol extinction coefficient obtained using the ECR method revealed the presence of enhanced aerosols in the UTLS following volcanic eruptions and wildfire events consistent with the Ozone Mapping and Profiler Suite (OMPS) and space-borne lidar-Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP). The cloud-top altitude obtained from the SAGE III/ISS is within 1 km of the nearly co-located observations from OMPS and CALIOP. In general, the seasonal mean cloud-top altitude from the SAGE III/ISS events peaks during the December, January, and February months, with sunset events showing higher cloud tops than the sunrise events, indicating the seasonal and diurnal variation of the tropical convection. The seasonal altitude distribution of cloud occurrence frequency obtained from the SAGE III/ISS also agrees well with CALIOP observations within 10%. We show that the ECR method is a simple approach that relies on thresholds independent of the sampling period, providing cloud-filtered aerosol extinction coefficients uniformly for climate studies irrespective of the UTLS conditions. However, since the predecessor of SAGE III did not include a 1550 nm channel, the usefulness of this approach is limited to short-term climate studies after 2017.

Stratospheric Aerosol and Gas Experiment III cloud data product

The latest in a series of solar occultation satellite instruments, Stratospheric Aerosol and Gas Experiment (SAGE) III, was placed into orbit in December 2001, and data were obtained until March 2006. Measurements were made of the extinction attributable to aerosols and cloud at a number of wavelengths between 290 and 1550 nm. The analysis of data obtained by its predecessor, SAGE II, has shown that an intercomparison of such data at two or more wavelengths may be used to separate the effects of cloud and aerosol. This analysis has been done on a routine basis for many years using SAGE II data at 525 and 1020 nm and applied extensively to global studies of tropospheric cloud and aerosol. Here we describe the aerosol-cloud separation algorithm developed for use with the SAGE III data, which uses the extinction at 525, 1020, and 1550 nm. This algorithm is now being used to produce vertical profiles of cloud presence as a standard SAGE III data product. These profiles have a vertical resolution of 0.5 km and cover the altitude range from 6.0 to 30.0 km, and data are presently available from March 2002 onward. An outline is given of the development of this algorithm, the nature of the SAGE III data, and the algorithm performance. To maintain continuity with SAGE II cloud data, the relative performances of the SAGE II and SAGE III algorithms are also examined. An example of the application of the algorithm to SAGE III tropospheric data is shown and discussed.

Cloud identification in atmospheric trace molecule spectroscopy infrared occultation measurements

High-resolution infrared nongas absorption spectra derived from the Atmospheric Trace Molecule Spectroscopy (ATMOS) experiment are analyzed for evidence of the presence of cirrus clouds. Several nonspherical ice extinction models based on realistic size distributions and crystal habits along with a stratospheric sulfate aerosol model are fit to the spectra, and comparisons are made with different model combinations. Nonspherical ice models often fit observed transmission spectra better than a spherical Mie ice model, and some discrimination among nonspherical models is noted. The ATMOS lines of sight for eight occultations are superimposed on coincident geostationary satellite infrared imagery, and brightness temperatures along the lines of sight are compared with retrieved vertical temperature profiles. With these comparisons, studies of two cases of clear sky, three cases of opaque cirrus, and three cases of patchy cirrus are discussed.

Discrimination of cloud and aerosol in the Stratospheric Aerosol and Gas Experiment III occultation data

The Stratospheric Aerosol and Gas Experiment (SAGE) III, scheduled for a first launch in mid-1998, will be making measurements of the extinction that is due to aerosols and gases at many wavelengths between 385 and 1550 nm. In the troposphere and wintertime polar stratosphere, extinction will also occur because of the presence of cloud along the optical path from the Sun to the satellite instrument. We describe a method for separating the effects of aerosol and cloud using the extinction at 525, 1020, and 1550 nm and present the results of simulation studies. These studies show that the new method will work well under background nonvolcanic aerosol conditions in the upper troposphere and lower stratosphere. Under conditions of severe volcanic contamination, the error rate for the separation of aerosol and cloud may rise as high as 30%.