Lagrangian Experiments Linking Small and Large Scales

The Lagrangian Concept

The temporal and spatial scales over which cloud fi elds evolve and interact with the general circulation in which they are embedded are typically days and thousands of kilometers. Scales involved in the general circulation exceed those accessible to single aircraft flights and limit seriously our ability to understand the two-way interactions between microscale processes and the general circulation.

This need to observe the evolution of meteorological fi elds and the processes that influence their development over long timescales led to the development of the Lagrangian observational concept, which seeks to make quasi-continuous observations of an air mass (ideally over several days, although as we shall see there are considerable benefits from fights lasting only a few hours) often by employing multiple observation platforms which overlap in time. The Lagrangian observational approach aims to study processes within a frame of reference that moves with the air, an approach necessary to study time-evolving processes in the atmosphere without resorting to making assumptions about poorly constrained advective processes.

Despite the high costs associated with fully realizing the potential of the Lagrangian approach using in-situ aircraft observations, a number of such observational studies have been conducted over the last few decades. These studies have yielded a tremendously rich understanding of a wide variety of different processes, as we discuss below.

Lagrangian Boundary Layer and Cloud Evolution

Lagrangian sampling of cloud-related processes on timescales of a day or more has primarily focused upon the marine boundary layer (MBL), because many of the key processes relevant to climate occur there. In addition, the logistics associated with conducting a Lagrangian study in the lowest level are simpler (not least of which is the presence of a solid and uniform boundary at the base of the air mass). Although earlier Lagrangian studies involved balloons and tracers, the first multi-flight Lagrangian experiments devoted to understanding cloud processes were conducted during the Atlantic Stratocumulus Transition Experiment/Marine Aerosol and Gas Experiment (ASTEX/MAGE) in June 1992 (Albrecht et al. 1995; Businger et al. 2006). Two such studies were carried out, each for a period of 36-48 hours (~800 km of spatial advection). During these studies, constant-density balloons floated downwind with the air mass and radioed their GPS-derived locations to a relay of three sampling aircraft. The balloons served as markers for the boundary-layer air mass, which was continually being modified by chemical and energy fluxes at the surface, entrainment of free tropospheric air, wind shear within the boundary layer, horizontal dispersion, chemical reactions, and aerosol transformations. Here we will focus on those aspects most pertinent to clouds.

The ASTEX/MAGE Lagrangian experiments were revolutionary for a number of reasons. By continuously sampling the MBL over a period approaching two days, it was possible to make observations of the transition from a shallow, well-mixed stratocumulus-capped MBL to a much deeper, decoupled MBL containing trade cumuli of less extensive coverage (Bretherton and Pincus 1995). Such air mass transitions are a critical component of the transition from the subtropical to the tropical MBL, and the physical mechanism by which the MBL undergoes this transition was deduced directly as a result of the Lagrangian observations. Previous hypotheses for the transition followed ideas based upon the work on diurnally modulated decoupling of the boundary layer that had been carried out at fixed locations. The ASTEX Lagrangian data were able to demonstrate that although decoupling of the MBL is a key process in the transition, it is actually the increasing latent heat flux as the air mass moves over warmer water, rather than solar radiation, that is the key driver of the subtropical transition (Bretherton and Wyant 1997).

Lagrangian Entrainment Rate Estimates

A second novel aspect of Lagrangian studies is the ability to derive direct estimates of the rate of MBL deepening, and thus the entrainment rate. It is now widely appreciated just how important the entrainment rate is, not only for its physical impacts upon the MBL thermodynamic and cloud climatology (Stevens 2002), but also because entrainment is a critical mediator of the response of cloud fi eld to perturbations in atmospheric aerosols (Ackerman et al. 2004). Because microphysically induced changes (e.g., suppression of drizzle) impact the entrainment rate by changing the MBL turbulent structure, this introduces much longer timescales than the Twomey effect (Wood 2007), which cannot be quantified observationally using measurements from fixed locations.

ASTEX allowed Lagrangian estimates of the entrainment rate over the two days of observations (Bretherton et al. 1995). Recent studies have employed the Lagrangian technique using flight durations of only a few hours. For example, in DYCOMS-II (Stevens et al. 2003), single 6-hr Lagrangian flights using a C-130 aircraft were used to ascertain the entrainment rate by determining the observed rate of boundary layer deepening and subtracting the subsidence rate from reanalysis data (Faloona et al. 2005). This was then compared with the entrainment rate estimated using the standard flux-jump method to provide closure on the time evolution of the entrainment rate that can be used to constrain model ability to represent the entrainment process correctly.

Aerosol-Cloud-Chemistry Interactions using Lagrangians

Although aerosol indirect effects were not a specific focus of the ASTEX Lagrangians, the aerosol and gas phase chemistry measurements taken as part of the MAGE component of the campaign provided the first concrete demonstration of the benefits that the Lagrangian observational concept could bring to our understanding of chemical process rates in the atmosphere (Zhuang and Huebert 1996; Clarke et al. 1996), a need that had been identified since the early 1970s. One might argue that the development of the sampling concept for the early Lagrangian studies led to the realization that aerosol particles and their precursors are themselves a fundamental component of the atmospheric general circulation, which thereby provides further possibilities for interactions between these scales that were not previously appreciated.

Lagrangian experiments that focused specifically upon aerosols and aerosol precursors were conducted as part of the First and Second Aerosol Characterization Experiments (ACE-1, ACE-2) in 1995 and 1997, respectively (Bates et al. 1998; Johnson et al. 2000). They provided a wealth of data on the coupling between aerosols, aerosol precursors, clouds, and the MBL, in both pristine and polluted MBLs. From the ACE-2 Lagrangian studies it has been possible to assemble a conceptual model for continental pollution outbreaks that serves as an invaluable basis for the planning of future sampling of the modified continental plume. Timescale analysis applied to the ACE-2 Lagrangian datasets (Hoell et al. 2000) demonstrated the importance of meteorological factors (e.g., dilution) in controlling the properties of aerosols in the MBL. The example from the third Lagrangian experiment during ACE-2, shown in Figure 21.3, demonstrates the ability of Lagrangian experiments to connect micro- and mesoscale characteristics with those of the large-scale flow.

A Strategy for Future Lagrangian Experiments

Although they have provided invaluable information about air mass transformations and the physical and chemical processes within them, the Lagrangian experiments performed to date have not yet reached their full potential as important strategies for determining how clouds will change in a perturbed climate:

1. Sampling limitations: To date, Lagrangian studies have sampled with one aircraft at any one given time, whereas current state-of-the-art observational techniques require the use of multiple aircraft (e.g., column closure experiments discussed above) to link the aerosol, ther-modynamic, and dynamic properties of the atmosphere measured in situ with the column cloud radiative properties remotely sensed from above. Adopting this strategy within a Lagrangian experiment would be very powerful, especially if scanning radiometers could be used to

5° 10° 15°E Lagrangian 3 boundary layer evolution

Subsidence inversion

"stratu^cumulus . , u „ , Weak decouPling humulis (1-2/8) Marine internal boundary layer

12:00 July 23 18:00 Flight 1

00:00 July 24 06:00 Flight 2



Figure 21.3 Flight tracks (top panel) for the third Lagrangian experiment of ACE-2 (Wood et al. 2000) which sampled a polluted continental outbreak heading southward over the northeast subtropical Atlantic Ocean. Back trajectories, estimated using EC-MWF analyses, are shown ending at different altitudes at the location of the start of the sampling. The three flights allow us to construct a picture of the changing structure of the lowest 2 km of the atmosphere, the essence of which is detailed schematically (bottom panel). Modified from Wood et al. (2000).

"stratu^cumulus . , u „ , Weak decouPling humulis (1-2/8) Marine internal boundary layer

12:00 July 23 18:00 Flight 1

00:00 July 24 06:00 Flight 2



Flight 3

Figure 21.3 Flight tracks (top panel) for the third Lagrangian experiment of ACE-2 (Wood et al. 2000) which sampled a polluted continental outbreak heading southward over the northeast subtropical Atlantic Ocean. Back trajectories, estimated using EC-MWF analyses, are shown ending at different altitudes at the location of the start of the sampling. The three flights allow us to construct a picture of the changing structure of the lowest 2 km of the atmosphere, the essence of which is detailed schematically (bottom panel). Modified from Wood et al. (2000).

broaden the measurements to be representative of an area. Moreover nadir-pointing remote-sensing systems provide crucial information on how microphysical fields that are sampled horizontally by in-situ aircraft overlap vertically (Damiani et al. 2006).

2. New sampling techniques: Beyond piloted aircraft, new vectors are becoming available for in-situ sampling of boundary-layer clouds at low speed. Unmanned aerial vehicles have a typical sampling speed of 30 m s-1 and they can now fly autonomously a predefined track (Ramanathan et al. 2007). Blimps have been used in the past and new platforms (e.g., the Zeppelin) might be used for Lagrangian studies, although the logistics involved remain a serious obstacle.

3. Interdisciplinary spirit: To characterize the evolution of the cloud layer fully, it is important to have a detailed characterization of all the potential factors controlling it, including the large-scale meteorology, the surface and MBL top boundary conditions (both physical and chemical) as well as the cloud microphysical, thermodynamic, and turbulent information. This is a tall order for a single platform but can be achieved, provided there is effective interdisciplinary communication and adequate financial support. The forthcoming EUCAARI and VOCALS experiments provide good examples of the possible connections that can be made between research communities (oceanography, atmospheric chemistry, cloud physics, and climate dynamics).

4. Synergy between models and observation: The most serious obstacle in the characterization of a boundary layer is our inability to measure the thermodynamic conservative variables (total heat and moisture) and their budgets accurately enough to constrain the resulting LWP. However, LES models, which can simulate the evolution of a boundary layer, suggest observable signatures of the mechanisms that drive its evolution. For instance, Sandu et al. (2008) showed that the impacts of the aerosol on the diurnal cycle of stratocumulus clouds are reflected by observable differences in the vertical velocity variance at cloud top, in the vertical profile of liquid water content, and in the level of turbulent kinetic energy below cloud base.

5. Complementary information: With the current generation of satellite remote sensing, there is a tremendous amount of additional information that can be added from space to multi-day Lagrangian studies. Geostationary satellites, especially those with high-resolution (1 km), multi-channel and high-sampling frequency (15 min for SEVIRI on MSG) spectrometers can provide cloud microphysical, aerosol, and cloud-top temperature information. Polar-orbiting satellites with active remote sensing, such as in the A-Train, provide additional information, at least once a day, to calibrate geostationary satellite observations better. Promising techniques appear to derive accurate temperature, humidity, and boundary-layer depth information from spaceborne GPS limb sounding. A reanalysis focused on clouds, aerosols, and the boundary layer will be an excellent way to incorporate the satellite information into a useable framework. High-resolution numerical modeling could be calibrated using in-situ data to "fill in the gaps" in rather the same way that assimilation is used in the numerical weather prediction community.

In addition, aircraft observations do not provide complete information about the large-scale dynamic fields; key variables, such as the large-scale vertical motion field, are completely unknown. These fields are critically important for understanding the interactions between the small and large scales and for allowing a top-down assessment of the entrainment rate that can be compared with bottom-up measurements carried out by integrating over many turbulent eddies. This is currently accomplished by using standard reanalysis products, but these are not generally produced with the needs of the cloud-climate community in mind.

0 0

Post a comment