Let <(ct)^q>_F be the q-th moment of pathlength in transmission (F=T) or reflection (F=R). The upper left panel shows the ratio of <ct>_T to the product of cloud thickness H and scaled optical thickness, (1-g)tau, as function of SZA cosine (u_0) and log_10(tau); g is the phase function’s asymmetry factor. Knowing any 2, one can compute the 3rd quantity in the ratio. The upper right panel is the non-dimensional ratio of the root-mean-square (RMS) path, <(ct)^2>_T^(1/2), to its mean, further divided by its asymptotic value √7/5. We see here that the mean is a very good predictor of the RMS, so now new cloud parameters can be gleaned from higher moments for ground-based O2-DOAS, at least in this simple cloud geometry. Lower panels show RMS(ct)_R/<ct>_R and <ct>_R/H. The former ratio of O2-DOAS observables can be used to derive tau, knowing g and u_0, and the later can then be used to derive H from mean pathlength. O2-DOAS can thus be used as a standalone cloud probing modality from space.
Differential optical absorption spectroscopy (DOAS) of oxygen in its “A” band (~760 nm) has been demonstrated theoretically and observationally (e.g., field campaigns at the SGP ARM facility) as an exquisite diagnostic of spatial complexity in cloudiness. “Complexity” captures here any mix of multiple and/or horizontally broken layers, the essence of large-scale cloud “3Dness.” This makes O2-DOAS a powerful diagnostic of cloud-radiation interactions in the solar spectrum for the most challenging scenarios, e.g., for GCM shortwave radiation schemes. This has lead ARM to invest in the development of fieldable high-resolution A-band instruments, through both Science Team and SBIR efforts.
Overlooked in this development is the opportunity for O2-DOAS to become a new modality in cloud property remote sensing, either stand-alone or in synergy with other cloud-probing sensors.
The basic physics for the cloud 3Dness detection and remote sensing are the same. O2-DOAS is used to infer low-order moments of the integrated paths that sunlight takes between its source and its detection, either above or below the cloud. The length of this path is random, with a distribution determined primarily by the number of scatterings suffered in the clouds.
For a single unbroken deck, reasonably well-approximated by a plane-parallel slab, low-order moments of solar photon path length are known quantities. In diffusion regimes (optical thickness > ~10), we have analytical functions of cloud thickness and optical depth. The present author has recently added to these expressions the effects of solar zenith angle, the presence of an overall internal gradient in cloud opacity, small-scale random fluctuations of droplet concentration, and gross deviations from slab geometry.
The top panels in the figure demonstrate that, for ground-based O2-DOAS, one can confidently infer cloud thickness knowing its optical depth, or vice versa. When ARM acquires continuously operating A-band instruments, it will add robustness to its cloud profiling under all conditions.
In contrast, from above, using diffusely reflected rather than transmitted light, stand-alone cloud property remote sensing is a possibility, since various path-length moments bring new pieces of information (cf. lower panels in figure). We are therefore excited about the re-flight of NASA''''s Orbiting Carbon Observatory (OCO), which has its own reasons for having hi-res O2-DOAS capability. It could be a cloud probe as well as CO2 monitor.
