Measurement and Analysis of Complex Layered-Cloud Systems Project (Task 10)

We have estimated from more than ten (10) years of low spatial resolution satellite observations from the International Cloud Climatology Experiment (ISCCP) that of the global, annual fractional cloud cover on Earth (60 percent) that approximately one-third or 20 percent of the areal cloud coverage consists of extended layers of non-precipitating middle and upper tropospheric cloud. These clouds may be totally liquid, totally ice or a combination of ice crystals and water droplets. Pre-satellite climatologies of these mid-layer clouds are vague or silent regarding the distribution of such clouds because of the difficulty of observing them well from the surface (due to obscuring lower clouds and hazes) even where a surface network may have existed prior to the advent of operational satellite coverage in the 1970's.

Of course the regional and seasonal variability of mid-layer clouds is also quite high as evidenced by recent encounters during Desert Shield and Desert Storm in a December through March period and over Bosnia throughout the year. Based upon forecaster reports to us and upon satellite observations it appears that less than 50 percent of the occurrence of these cloud layers can be attributed to an association with mid-latitude cyclones, jet streams and/or upper level vorticity maxima. These layers of cloud - above the land and marine boundary layers and thus excluding fogs and marine stratus - apparently form, exist and dissipate in response to processes and interactions within adjacent atmospheric layers.

To explore these cloud systems we posed three science questions to be addressed under Task 10. Preliminary results on each of these questions were included in Vonder Haar et al. (1996).

Question 1: Can cloud bases and layers be estimated using only satellite techniques?

In the referenced reports we describe a physically-based method which combines multi-spectral satellite remote sensing data with in situ rawinsonde (or climatology) data to estimate cloud tops and bases in eight atmospheric layers. Preliminary results were encouraging and USAF needs were great enough to apply this unvalidated method to a high resolution satellite data set [(CHANCES, see Reinke et al. (1995)] for a 30-day period in July, 1994 (Figure 2.1.1) as part of Center for Geosciences research.

The CLEX experiment provides the first new data set on cloud tops, bases, and water contents (and their variabilities) to allow a test of this and other layer-cloud estimation methods. We will apply the Geosciences Center 8-layer estimation method to satellite radiances and rawinsonde data collected during CLEX at the same time the aircraft and ground-based systems are measuring details of each cloud layer. Results of this research will allow uncertainty values to be assigned to satellite estimation methods for specific cloud situations. In addition, we expect the CLEX data to help improve certain aspects of the estimation methods based upon better physical understanding of the layer cloud "targets" and their temporal variation.

Question 2: Can existing layer cloud conditions be detected and input into atmospheric models?

During our research to date, we have explored four (4) methods to answer Question 2. These are discussed in Vonder Haar et al. (1996) and are closely related to 4-D VAR assimilation methods being tested and studied at ECMWF, NCEP and FNOC. They are related also to the

new water vapor profile determination research of Task 8 (see below) and to the "coupled-analysis" model/observation methodology of Lipton and colleagues at the USAF Phillips Laboratory.

CLEX data will strongly support diagnostic and prognostic research on Question 2 because an extensive, time-dependent layer cloud data set will be obtained within a network of extensive concurrent weather observations, model analyses and model forecasts. Our CLEX experiment area has been located over NOAA's wind profiler network which is augmented by numerous newly-commissioned ASOS observing sites. We have made special arrangements to obtain NCEP, NWS, and ECMWF model analyses and forecast results for the CLEX period and region.

Question 3: How does cloud layer-induced radiative heating/cooling impact the layer cloud lifetime?

We hypothesize that in the absence of major synoptic-dynamic effects the morphology of extensive single and multiple cloud layers is significantly influenced by both solar and infrared heating/cooling processes. Figure 2.1.2 shows the infrared situation. To explore this idea, Olsen (1996) recently completed a M.S. thesis as part of Geosciences Center research. She found, for cases of cloud layers over tropical and subtropical oceans, that cloud-induced radiative processes were significantly strong to influence layer conditions and cloud morphology when major precipitation was not occurring (Figure 2.2.3). Tiedke (1993) and Morecrete (1994) had assumed these effects were important and had developed their ECMWF methodology accordingly.

Figure 2.1.3 Heating and cooling rate profiles with and without clouds at each site

at 0000Z (after Olsen, 1996).

CLEX will provide new cloud top, base, liquid water, droplet size data set over land so that our Center for Geosciences diagnostic radiative heating/cooling approach can be improved and extended to answer Question 3 for a major land data set. All the satellite, rawinsonde and surface data required by the approach will be collected. CLEX also brings a major improvement in that cloud radiative properties will be measured both in situ and remotely by the non-satellite instrumentation. Figure 2.1.4 describes how the CLEX cloud radar and lidar measurements will "capture" the radiatively significant cloud layers. In Figure 2.1.4, blue envelopes show the estimated occurrence space for typical liquid clouds (left) and ice clouds (right). Two "boundary" lines are superimposed on each chart. The estimated -30 dBZ limit of the JPL Airborne Cloud Radar is shown to depict that the cloud radar will measure cloud layers composed of liquid and ice clouds falling in occurrence space to the RIGHT of the -30 dBZ lines. These clouds to be measured during CLEX by the radar are indicated by "R." Remaining cloud layers - such as thin cirrus and water clouds with very small droplets - will be measured by the cloud lidar and those occurrence spaces are labeled "L." The other "boundary" shown is that of clouds having a visible optical of less than 0.01 (equivalent to a cloud transmittance of approximately 99%). We are using this as a limit of radiative significance of clouds in CLEX insofar as they would or would not be involved in heating/cooling processes notes in Figures 2.1.2 and 2.1.3. Overall, we see that most of the clouds we will measure during CLEX will be radiatively significant - they fall in occurrence space above the optical depth limit.