Microphysical Processes in Cirrus and Their Impact on Radiation: A Mesoscale Modeling Perspective

The impact of cloudiness on the global radiative budget and its climatic consequences have been widely discussed during the last three decades. It was gradually recognized that the climatic effect of cloudiness depends on its height: low- and middle-level cloudiness have a total cooling effect on the Earth climatic system, while the upper-level clouds, cirrus, may have mostly a warming effect (IPCC 1995). The net effect of cirrus (i.e., warming or cooling), is much less clear because neither their microphysical and optical properties, nor the processes that govern their formation, are well understood and parameterized in climate models. These uncertainties have stimulated several major field projects performed within the International Satellite Cloud Climatology Project (ISCCP; Rossow and Schiffer 1991) with subsequent data analysis reports [e.g., FIRE IFO-I (1990), FIRE IFO-II (1995), and EUCREX (Raschke et al. 1996)]. The relevant theoretical works, and even the simplest climate models, indicate that the climatic impact of cirrus depends on their microstructure: clouds composed of small crystals with effective radii less than about 16 μm have a total cooling effect, but clouds of larger crystals have a warming effect (Stephens et al. 1990). It was shown that the total cloud forcing at the top of the atmosphere (TOA) is positive from a few to a few tens of watts per square meter for the large crystals and decreases with decreasing crystal radius (Fu and Liou 1993). Most of the previous theoretical studies of cirrus radiative properties, after choosing some model of microphysics and some values for the mass extinction and absorption coefficients, then prescribed them to the whole cloud, neglecting any vertical variations. Simulations with general circulation models (GCMs) showed that cirrus clouds with their optical properties parameterized in such a way (i.e., constant with height) have a total warming effect and positive feedbacks with respect to greenhouse gas-induced global warming (e.g., Ramanathan et al. 1983; Wetherald and Manabe 1988). Today, the estimation of the warming/cooling effect of cirrus has become even more complicated due to two factors. First, for many years the usual in situ probes allowed the measurement of ice crystals with radii only larger than 25-50 μm, so the smallest and most optically and radiatively active crystals were unresolved.

Dynamic Processes in Cirrus Clouds: Concepts and Models

Advancement in the understanding of cirrus clouds and their life cycle comes through symbiotic use of models, observations, and related concepts (fig. 18.1). Models of cirrus clouds represent an integration of our knowledge of cirrus cloud properties and processes. They provide a capacity to extend knowledge and enhance understanding in ways that complement existing observational capabilities. Models can be used to develop new theories, such as parameterizations, and focus science issues and observational requirements and developments. For example, early model results of Starr and Cox (1985a) and Starr (1987b) predicted that fine cellular structure (~lkm or less) would be found in the upper part of extended stratiform cirrus clouds. This prediction was confirmed when high-frequency sensors were deployed both for active remote sensing (Sassen et al. 1990a, 1995) and later for in-situ measurements (Quante and Brown 1992; Gultepe et al. 1995; Quante et al. 1996). Sampling rates of 10Hz, or better, are now accepted as a minimum requirement for resolving cirrus cloud internal structure and circulation where 1-Hz or coarser measurements were previously used. Similarly, discrepancies between observed cloud radiative properties and calculations (theory) based on corresponding in-situ observations of cloud microphysical properties (Sassen et al. 1990b) led to the development of improved observing capabilities for small ice crystals (Arnott et al. 1994; Miloshevich and Heymsfield 1997; Lawson et al. 1998). Such sensors are now regarded as part of the standard complement when doing in-situ microphysical measurements in cirrus. At the same time, observations are absolutely essential in developing and evaluating cloud models. No cloud modeler wants to apply a model or theory too far beyond the limits of what can be observationally confirmed, at least in gross terms. The third aspect of this triad is concepts. Although models and observations can lead to predictions or diagnosis of unexpected relationships, they are each limited by the concepts that were used in their design and/or implementation. In the end, new concepts arising from analogy to other phenomena and/or from synergistic integration of existing knowledge can lead to new understanding, new models, new instruments, and new sampling strategies (fig. 18.1). Chapter 17 focuses on observations of internal cloud circulation and structure.

Subvisual Cirrus

Starting during World War II, pilots flying high over the tropics reported “a thin layer of cirrus 500ft above us”. Yet as they ascended, they still observed more thin cirrus above them, leading to the colloquialism “cirrus evadus.” With the coming of lidar in the early 1960s, rumors and unqualified reports of subvisual cirrus were replaced with validated detections, in situ sampling, and the first systematic studies (Uthe 1977; Barnes 1980, 1982). Heymsfield (1986) described observations over Kwajalein Atoll in the western tropical Pacific Ocean, where pilots and lidars could clearly see the cloud but DMSP (U.S. Defense Meteorological Satellite Program) radiance measurements and ground observers could not. The term “subvisual” is a relatively recent appellation. Prior terminology included cirrus haze, semitransparent cirrus, subvisible cirrus veils, low density clouds, fields of ice aerosols, cirrus, anvil cirrus, and high altitude tropical (HAT) cirrus. Subvisual cirrus clouds (SVC) are widespread (Winker and Trepte 1998; see chapter 12, this volume) and virtually undetectable with existing passive sensors. Orbiting solar limb occupation systems such as the Stratospheric Aerosol and Gas Experiment (SAGE) can detect these clouds, but only by looking at them horizontally where the optical depths are significant. SVC appear to affect climate primarily by heating the planet, though to what extent this may happen is unknown. Much of what we know is based on work by Heymsfield (1986), Platt et al. (1987), Sassen et al. (1989, 1992), Flatau et al. (1990), Liou et al. (1990), Hutchinson et al. (1991, 1993), Dalcher (1992), Sassen and Cho (1992), Takano et al. (1992), Lynch (1993), Schmidt et al. (1993), Schmidt and Lynch (1995), and Winker and Trepte (1998). SVC are defined as any high clouds composed primarily of ice (WMO 1975) and whose vertical visible optical depth is 0.03 or less (Sassen and Cho 1992). Such clouds are usually found near the tropopause and are less than about 1 km thick vertically. SVC do not appear to be fundamentally different from ordinary, optically thicker cirrus. They do, however, differ from average cirrus by being colder (-50-90°C), thinner (<0.03 optical depths at 0.694 μm), and having smaller particles (typically about <50μm diameter).

Ground-based Remote Sensing of Cirrus Clouds

Cirrus clouds have only recently been recognized as having a significant influence on weather and climate through their impact on the radiative energy budget of the atmosphere. In addition, the unique difficulties presented by the study of cirrus put them on the “back burner” of atmospheric research for much of the twentieth century. Foremost, because they inhabit the frigid upper troposphere, their inaccessibility has hampered intensive research. Other factors have included a lack of in situ instrumentation to effectively sample the clouds and environment, and basic uncertainties in the underlying physics of ice cloud formation, growth, and maintenance. Cloud systems that produced precipitation, severe weather, or hazards to aviation were deemed more worthy of research support until the mid- 1980s. Beginning at this time, however, major field research programs such as the First ISCCP (International Satellite Cloud Climatology Program) Regional Experiment (FIRE; Cox et al. 1987), International Cirrus Experiment (ICE; Raschke et al. 1990), Experimental Cloud Lidar Pilot Study (ECLIPS; Platt et al. 1994), and the Atmospheric Radiation Measurement (ARM) Program (Stokes and Schwartz 1994) have concentrated on cirrus cloud research, relying heavily on ground-based remote sensing observations combined with research aircraft. What has caused this change in research emphasis is an appreciation for the potentially significant role that cirrus play in maintaining the radiation balance of the earth-atmosphere system (Liou 1986). As climate change issues were treated more seriously, it was recognized that the effects, or feedbacks, of extensive high-level ice clouds in response to global warming could be pivotal. This fortunately came at a time when new generations of meteorological instrumentation were becoming available. Beginning in the early 1970s, major advancements were made in the fields of numerical cloud modeling and cloud measurements using aircraft probes, satellite multispectral imaging, and remote sensing with lidar, short-wavelength radar, and radiometers, all greatly facilitating cirrus research. Each of these experimental approaches have their advantages and drawbacks, and it should also be noted that a successful cloud modeling effort relies on field data for establishing boundary conditions and providing case studies for validation. Although the technologies created for in situ aircraft measurements can clearly provide unique knowledge of cirrus cloud thermodynamic and microphysical properties (Dowling and Radke 1990), available probes may suffer from limitations in their response to the wide range of cirrus particles and actually sample a rather small volume of cloud during any mission.

Satellite Remote Sensing of Cirrus

The determination of cirrus properties over large spatial and temporal scales will, in most instances, require the use of satellite data. Global coverage at resolutions as fine as several meters are attainable with instruments on Landsat, and temporal coverage at 1-min intervals is now available with the latest Geostationary Operational Environmental Satellite (GOES) imagers. Extracting information about cirrus clouds from these satellite data sets is often difficult because of variations in background, similarities to other cloud types, and the frequently semitransparent nature of cirrus clouds. From the surface, cirrus clouds are readily discerned by the human observer via the patterns of scattered visible radiation from the sun, moon, and stars. The relatively uniform background presented by the sky facilitates cloud detection and the familiar textures, structures, and apparent altitude of cirrus distinguish it from other cloud types. From satellites, cirrus can also be detected from scattered visible radiation, but the demands of accurate identification for different surface backgrounds over the entire diurnal cycle and quantification of the cirrus properties require the analysis of radiances scattered or emitted over a wide range of the electromagnetic spectrum. Many of these spectra and high-resolution satellite data can be used to understand certain aspects of cirrus clouds in particular situations. Intensive study of well-measured cases can yield a wealth of information about cirrus properties on fine scales (e.g., Minnis et al. 1990; Westphal et al. 1996). Production of a global climatology of cirrus clouds, however, requires compromises in spatial, temporal, and spectral coverage (e.g., Schiffer and Rossow 1983). This chapter summarizes both the state of the art and the potential for future passive remote sensing systems to aid the understanding of cirrus processes and to acquire sufficient statistics for constraining and refining weather and climate models. Theoretically, many different aspects of cirrus can be determined from passive sensing systems. A limited number of quantities are the focus of most efforts to describe cirrus clouds. These include the areal coverage, top and base altitude or pressure, thickness, top and base temperatures, optical depth, effective particle size and shape, vertical ice water path, and size, shape and spacing of the cloud cells.

Laboratory Studies of Cirrus Cloud Processes

A number of processes that play a role in the formation, evolution of microphysical properties, and radiative characteristics of cirrus clouds are amenable to investigation in a laboratory setting. These laboratory studies provide fundamental data for quantifying and validating theoretical concepts and help guide investigations involving direct and remote measurements of cirrus. Laboratory data also may be used for formulating parameterizations for numerical cloud models, especially where information is incomplete or full descriptions are not possible. This chapter reviews results from laboratory studies of ice formation, ice crystal growth, radiative transfer, and aerosol scavenging and transformation in the cirrus environment. Emphasis is placed on ice formation in cirrus conditions. The related topic of contrail formation is covered separately in this book. The formation mechanisms of lower stratospheric clouds are reviewed elsewhere (e.g., Tolbert 1994; Peter 1996; Carslaw et al. 1997; Koop et al. 1997a). Laboratory studies of cirrus ice formation are at a rapidly developing stage, so it is useful to provide significant background bases for current and needed studies. Key issues are aerosol composition, ice nucleation mechanisms, and the synergy between theory and laboratory measurements. Vali (1996), Baker (1997) and Martin (2000) discuss some of these issues in review papers. Upper tropospheric aerosol particles play an important catalytic role in the formation of cirrus. The nucleation process is important in determining the microphysical properties of cirrus. Numerical modeling studies (e.g., Jensen and Toon 1994; DeMott et al. 1994, 1997; Heymsfield and Sabin 1989) indicate that variation in the factors that drive the nucleation of ice and variations in the physical and chemical characteristics of aerosol particle populations lead to the formation of cirrus with different microphysical characteristics. Knowledge of the physics and chemistry of aerosols in the upper troposphere and lower stratosphere has evolved at a rapid pace. A detailed accounting of this topic is beyond the scope of this chapter. For the purpose of the present discussion, it is sufficient to note that the aerosol from which cirrus nucleate may vary significantly from place to place. Differences in aerosol properties in time and space occur because particles can arrive to the upper troposphere in so many ways and from so many sources.

Mid-latitude and Tropical Cirrus: Microphysical Properties

Cirrus, a principal cloud type that forms at low temperatures in the upper troposphere, is composed almost always of ice crystals (Heymsfield and Miloshevich 1989) and on average cover about 20% of the earth's surface (Hartmann et al. 1992). The purpose of this chapter is to characterize the microphysical properties of cirrus clouds. The Glossary of Meteorology (Huschke 1970) defines cirrus clouds as detached clouds in the form of white, delicate filaments, or white or mostly white patches, which are composed of ice crystals. This cloud type forms primarily in the upper troposphere, above about 8km (25,000 feet), where temperatures are generally below -30° C. There are a number of types of cirrus clouds, with the most frequent ones occurring in layers or sheets with horizontal dimensions of hundreds or even thousands of kilometers. Because horizontal dimensions are much greater than vertical extent, this particular type of cirrus cloud is called cirrostratus. Cirrus can also form in a patchy or tufted shape, when the ice crystals are large enough to acquire an appreciable fall velocity (the rate at which ice crystals fall in the vertical) so that trails of considerable vertical extent may form. These trails curve irregularly or slant, sometimes with a commalike shape, as a result of changes in the horizontal wind velocity with height and variations in the fall velocity of the ice crystals. A wispy, layered cloud that forms at the top of a cumulonimbus cloud, termed an “anvil” because of its shape, is a cirrus that consists essentially of ice debris which spreads outward from the convective parts of the storm. Anvils do not include the white, dense portions of thunderstorms or the active convective column. Anvils can spread to form large, widespread cloud layers. Tropical cirrus clouds are thought to arise primarily from cumulonimbus clouds. Unlike the thin, wispy cirrus typifying mid-latitudes, the high altitudes and extensive lateral and vertical development that often characterize tropical cirrus impose substantial large-scale radiative effects in the atmosphere and at the earth's surface (Hartmann et al. 1992; Collins et al. 1996). The cirrus-like low-level ice clouds and ice fogs of the Arctic are not considered cirrus.

Cirrus, Climate, and Global Change

Understanding the climate of Earth and the way climate varies in time requires a quantitative understanding of the way water cycles back and forth between the atmosphere and at the Earth's surface. The exchanges of water between the surface and atmosphere establish the hydrological cycle, and it is the influence of this cycle on the energy budget of Earth that is central not only to understanding present climate but also to the prediction of climate change. Processes relating to the smallest of the reservoirs of water—namely, the atmospheric branch of the hydrological cycle—play an especially critical role in climate change. Water in vapor phase is the critical greenhouse gas (e.g., Chahine 1992) providing much studied feedbacks on climate forcing (Lindzen 1990; Rind et al. 1991; Stephens and Greenwald 1991; Inamdar and Ramanathan 1998; Hall and Manabe 1999). Water in the form of condensed, precipitation-sized particles is an important source of energy fueling circulation systems and is the fundamental supply of fresh water to life on Earth. Liquid water cloud droplets significantly modulate the radiative budget of the planet (e.g., Wielicki et al. 1995). Water that exists as ice particles suspended in the atmosphere is perhaps the smallest of the water reservoirs of the atmosphere, yet these ice crystals when distributed as part of large-scale cirrus clouds exert a disproportionate influence on the energy and water budgets of the planet. This chapter briefly speculates on the important ways cirrus clouds affect the Earth's climate. The topics discussed are central to what is referred to as the cloud-climate problem, which might be schematically represented in terms of the coupled processes represented in figure 20.1. The two most critical scientific questions associated with the cloud-climate problem are also stated in figure 20.1. Answers to these questions require a clearer understanding of how the large-scale circulation of the atmosphere governs cloud formation and evolution, how these clouds heat and moisten the atmosphere, and how this heating and moistening effect in turn feeds back to influence the dynamical and thermodynamical properties of the atmosphere.

Cirrus: History and Definition

The most distinguishing physical property of cirrus (cirrostratus and cirrocumulus) is their composition. Cirrus are made predominantly or wholly of ice, whereas the majority of clouds (both in name and number) are composed of water droplets. That most clouds were composed of water droplets was probably well known to the ancients, who must surely have encountered fog in valleys and mountains. Yet the presence of ice in cirrus is not easily experienced in everyday life. To answer the question Who discovered that cirrus are made of ice? we have to trace developments in meteorology back almost 2500 years. Anaxagoras of Clazomenae (c. 500-428 B.C.) might have deduced that cirrus were made of ice. Using an inductive approach based on measurements and observations, Anaxagoras knew that clouds were made of water and that air was colder aloft. He believed that warm, moist air convected upward and that the water vapor cooled, condensed, and ultimately froze at great heights to become hail. We do not know if Anaxagoras considered cirrus explicitly because what little is left of his writings do not mention any cloud recognizable as cirrus (Gershenson and Greenberg 1964). Two thousand years passed before any substantial progress was made on cirrus. In 1637 Descartes (1596-1650) published Discours de la methode (Descartes 1637) in three parts: Dioptrics, Meteorology, and Geometry. In Dioptrics he set forth the law of refraction (Snell’s law) and in Meteorology he applied the law to the rainbows by performing numerical ray traces. Although he almost certainly knew the principle of minimum deviation, there is nothing in his writings that explicitly refers to it. In the ninth discourse on Meteorology, Descartes conjectures that the common 22° halo was due to refraction through ice crystals. . . . around the heavenly bodies there sometimes appear certain circles . . . they are round . . . and always surround the sun or some other heavenly body . . . they are colored, which shows that there is refraction. But the circles are never seen where it rains, which shows that they are not caused by the refraction which occurs in drops of water or in hail, but by that which is caused in those small little stars of transparent ice . . . those that we have observed most often have had their diameters at around 45° . . . (Olscamp 1965) . . .

Dynamic Processes in Cirrus Clouds: A Review of Observational Results

Local dynamical processes are a key factor determining the microphysical characteristics and typically heterogeneous macroscopic structure of cirrus cloud fields. The internal and background flow fields are correspondingly heterogeneous, albeit only weakly turbulent in most instances, as is discussed here. Nucleation processes and ice crystal growth and habit are intrinsically governed by the local temperature and humidity (saturation ratio) conditions that, in turn, are strongly regulated by the intensity and duration of local updrafts and downdrafts. The microphysical result of equivalent lift by a 50cm/s updraft over a cell width of 200m is quite different from that by a 0.5 cm/s updraft over a 2-km width, even though the overall mass fluxes are equivalent. The great degree of horizontal structure seen in fallstreaks emanating from cirrus likely reflects corresponding variability in microphysical properties, primarily ice crystal size, resulting from variability in the dynamical conditions in the ice-crystal-generating layer. The ice fallout process is a first-order effect in determining overall cloud ice water path. Entrainment of noncloudy environmental air and internal mixing processes are other dynamical aspects that likely play a significant role in cloud life cycle. Dynamical processes provide an important coupling between cirrus cloud microphysical and radiative processes, as described in chapter 18 and illustrated in figure 17.1. Cirrus cloud microphysical properties and macroscopic structure strongly affect the overall radiative properties of a cirrus cloud field and thus the important radiative effect of cirrus in the climate system. Knowledge of the dynamical processes influencing cloud macrophysical properties and microphysical structure is important to understanding the origin of these characteristics. Moreover, cloud-resolving models of cirrus cloud systems must be evaluated in these respects due to the importance of cloud dynamical processes in determining overall cloud properties. Dynamical processes in cirrus are linked to the state of the background flow field that, in general, is characterized by significant wind shear and a stable thermal stratification. Gravity waves are ubiquitous and occur over a range of scales. Upper tropospheric turbulence tends to occur intermittently in patches, likely a result of sporadic shear generation (Kelvin-Helmholtz instabilities) or breaking gravity waves. Turbulent mixing in stratified shear flows is a notoriously difficult subject, and advances in its description have been obtained only recently (e.g., Fernando 1991; Schumann and Gerz 1995; Vanneste and Haynes 2000).