Cloud Phase Classification in the Arctic using Far-Infrared Radiances

Abstract:

Cloud thermodynamic phase, i.e., whether a cloud is composed of liquid, ice, or both, has a significant influence on the Arctic surface radiation environment. Liquid-water clouds tend to be opaquer in the infrared compared to ice-water clouds, and thus tend to emit more downwelling radiation. The spatial and temporal distribution of Arctic liquid-, ice- and mixed-phase clouds is not well understood. Given a lack of ground observations in the Arctic, satellite measurements are crucial. The far-infrared (far-IR) comprises >50% the Arctic outgoing longwave radiation (OLR) and the spectral variation of ice- and liquid-water absorption across the far-IR differs substantially. Two upcoming satellite missions, NASA PREFIRE and ESA FORUM, will measure far-IR spectral radiances, and therefore it is pertinent to investigate the potential of using the far-IR radiances for cloud phase determination.

A cloud phase classification approach is developed based on brightness temperature differences (BTDs) between sets of far-IR channels. Using an assortment of Arctic summer and winter cloudy profiles as inputs to the Principal Component Radiative Transfer Model, a set of simulated BTs spanning 100 to 600 cm^-1 at 0.5 cm^-1 spectral intervals are leveraged to produce BTD thresholds for phase classification. A series of ice- and liquid- phase classification tests are applied to a large set of far-IR BT spectra generated from ECMWF ERA5 cloudy profiles for the Arctic and Antarctic. The successful classification rate of ice clouds is 95%, while for liquid clouds it is 55%. A relatively lower classification success for liquid clouds is related to generally low altitudes of liquid clouds, also because liquid-water far-IR absorption has much less spectral variation than the ice-water does. Synthetic PREFIRE spectra are generated to investigate the impacts of spectral resolution on the far-IR cloud phase classification. A preliminary cloud phase classification approach for the PREFIRE mission is also discussed.