A large number of cumulus cloud studies [e.g. Baker et al., 1980; Telford and Chai, 1980], and stratiform cloud studies [e.g. Telford and Wagner, 1981; Considine and Curry, 1998] have shown that cloud microphysical properties are influenced by entrainment and mixing. Most of these studies have focused on the turbulent mixing of dry, warm and presumably cleaner air from above with the cloudy air and its impact on the droplet spectra, without considering the CCN properties of the entrained air. Using a one-dimensional explicit mixing parcel model, Su et al. [1998] included CCN in the entrained air and found drop spectral broadening resulting from entrained CCN growing into small droplets.
Entrainment of cleaner free tropospheric air has been observed to result in a decrease in CCN concentration in tropospheric clouds [e.g Clarke et al., 1996]. Cases with polluted air above the inversion are not uncommon [e.g. Clarke, 1993; Clarke et al., 1997; Raes, 1995]. and the entrainment of such air, together with associated condensation nuclei (CN), could influence the number of CCN that activate at a given supersaturation.
Variations in CCN have the potential to alter cloud microphysical and optical properties through modifying the cloud particle phase, size and concentration [e.g., Twomey, 1991; Curry et al., 1993]. The response of the cloud microphysical, radiative, and dynamical structure to a variation in CCN concentration has been investigated numerically in warm-season arctic stratus [e.g. Olsson et al., 1998] and marine stratocumulus clouds [e.g. Feingold et al., 1994]. These studies support observational studies that show that an increase in CCN concentration produces a more reflective cloud composed of a larger number of smaller droplets for a given amount of liquid water.
Pollution of North American and Eurasian continental origin is believed to be one of the sources of Arctic aerosols during the winter months. Barrie [1986] showed that particulate pollution in spring was found throughout the depth of the arctic air mass, but was highest in the lowest 2 km. During the SHEBA/FIRE spring IOP (May 18, 1998) a high concentration of cloud nuclei in the air above the cloud was observed to over-lie a fairly clean boundary layer. This situation is also fairly common in Californian boundary layers where polluted continental air sometimes over-lies cleaner marine boundary layer air. In a numerical study, Duda et al. [1996] simulated such a case and showed that when air containing a six-fold higher concentration of CCN above cloud is entrained into cloud, a 6 to 9 % increase in cloud albedo is simulated.
Because of the strong stability of the Arctic boundary layer, it has been hypothesized that the importance of cloud entrainment effects are only secondary, whereas the background air chemistry and CCN characteristics are the principal determinant of the drop size distribution [Curry, 1995].
To understand how the microphysical and dynamical structure and radiation properties of Arctic stratus are modified by the entrainment of ``polluted'' air at cloud top in the arctic environment, we have performed a number of numerical simulations in a three-dimensional/LES framework with liquid-phase, bin-resolving microphysics [Feingold et al., 1996a]. The key issue to be addressed is how long it takes for differences in cloud dynamics and microphysics to develop in response to variations in CCN concentrations due to entrainment, and how this process affects the Arctic cloud evolution. Numerical simulations of the FIRE/SHEBA May 18 case are used to address this question. Where possible, comparison between the simulations are made with observational data, however the case study is used more as guidance for studying the contamination of the cloud by polluted air overriding the boundary layer than as a rigorous intercomparison with observations.
The paper is organized as follows: Section 2 contains a brief summary of the LES version of RAMS and the explicit microphysics model. Section 3 describes the chosen case and experiment design. Section 4 presents results from two three-dimensional runs: one using the CCN sounding observed on 18 May 1998 during the FIRE/SHEBA spring IOP, and the other using a vertically uniform CCN as a sensitivity test. Section 5 discusses and summarizes the results.