Investigating the Effects of Viewing/Solar Geometry and Precipitable Water Vapor on Passive Radiometeric Cloud Detection and Property Retrievals

Abstract

Clouds are well established as a major source of uncertainty in climate and are the dominant modulators of radiation both at the surface and at the top of the atmosphere. Their impact on the Earth’s radiation budget mainly depends on the bulk cloud properties, such as cloud amount, height, and microphysical/optical characteristics. Large-scale satellite data are needed to both verify and improve general circulation model (GCM) parameterizations of clouds and radiation for climate prediction. For reliable application of satellite datasets in cloud processes and climate models, it is important to have a reasonable estimate of the bias and uncertainty in the derived cloud properties. Ideally, the calibration and evaluation of passive radiometric-based cloud products should be done via in-situ aircraft. This is often impractical; thus, satellite retrievals are often validated with long-term ground-based or space-borne active remote sensing instruments. This dissertation makes use of both types of platforms for validating passive radiometric-based cloud products. In the first study, the daytime single-layered low-level cloud properties retrieved by instruments onboard GOES are compared with ground-based observations and retrievals over the DOE ARM SCF from June 1998 through December 2006. Comparisons are made for monthly means, diurnal means, and one-to-one GOES and ARM collocated pairs. GOES cloud effective temperature (Teff) is highly correlated with ARM cloud-top temperature (Ttop), having an R2 value of 0.75, though GOES exhibits a cold bias. GOES retrieved cloud optical depth (τ) and liquid water path (LWP) have very good agreement with ARM retrievals with R2s of 0.45 and 0.47, while GOES retrieved cloud-droplet effective radius (re), on average, is about 2 µm greater than ARM re. An examination of solar and viewing geometry has shown that GOES retrieved mean re and τ values are impacted by solar zenith angle (SZA) and especially scattering angle (SCA). In the second study, a quantitative evaluation of maritime transparent cirrus cloud detection, which is based on GOES-16 data and developed with collocated CALIOP profiling, is performed. First, the relationships between the clear-sky 1.378 µm radiance, viewing/solar geometry, and precipitable water vapor (PWV) are characterized. Next, detection thresholds are computed using the Ch. 4 radiance (λ= 1.378 µm) as a function of viewing/solar geometry. In addition to bulk statistics, an example application and case study are shown for validation. The algorithm detects nearly 50% of sub-visual cirrus (COD or τ < 0.03), 80% of transparent cirrus (0.03 < COD < 0.3), and 90% of opaque cirrus (COD > 0.3). This study lays the groundwork for a more complex, operational algorithm to detect GOES transparent cirrus clouds. In the third and final study, this algorithm is modified and applied for detecting transparent cirrus clouds over land for removing some potential false-alarm pixels. Clear-sky false alarm rates over land are examined as a function of PWV, and a threshold for pixel-rejection is determined. Then, thresholds for removing pixels with low- and mid-level clouds underneath are devised by integrating the water vapor mixing ratio between the top-of-the-atmosphere (TOA) and a specific altitude. The total-column and layer PWV thresholds are applied to the full over-land sample to determine their effectiveness in reducing the number of false alarms. This study suggests that for cirrus applications, lower-tropospheric clouds are a much more significant source of contamination than the land surface, due to the difficulty of removing them without the addition of downstream or level-2 products.