DESCRIPTION OF THE INDIVIDUAL PLATFORMS
The ER-2 and the Learjet
- Measure, by direct atmospheric observations, the vertical structure of the water-vapor
- Measure the effect of cirrus on radiation fluxes over the equatorial Pacific.
- Measure the east-west gradients of vertical distribution of water vapor along the equatorial
- Explore the microphysical factors contributing to the high albedo of widespread tropical
- Determine the variation of upwelling long-wave broadband (450µm)and window
region(812µm) flux at the flight altitude of the ER-2 and Learjet for clear skies below the
aircraft. From these upwelling fluxes, compute the greenhouse effect at flight altitudes as the
difference between sea surface blackbody emission and the upwelling fluxes.
- Identify clear skies from the lidar and NFOVR on the ER-2. Identify clear skies below
the Learjet from NFOVR and the solar radiometer on the Learjet. Use microphysics instruments to
identify whether the Learjet is within or below the cirrus.
- Estimate the net long-wave and solar heating within the cirrus by differencing the net flux
measured by the ER-2 to the net flux measured by the Learjet.
- Difference the OLR between clear and cloudy skies to determine the greenhouse effect of
cirrus (long-wave cloud forcing). From the ER-2 radiation flux observations, difference the
cloudy-sky albedo from the clear-sky albedo to estimate the reduction in absorbed solar flux due to
clouds. From similar estimates from the Learjet, separate the effects of cirrus on outgoing long-
wave and solar fluxes.
- Estimate the profile of long-wave and solar heating rates during the takeoff and landing of
the ER-2 and Learjet for model validation.
- Measure the variation of water-vapor concentration along the flight altitude of the ER-2
and the Learjet.
- Measure the profile of water vapor below about 10 km with the Learjet's dropsondes.
- Estimate the optical depth of the cirrus at selected wavelengths using the TDDR (see Table
6) on the Learjet. Determine the albedo of the cirrus from the ER-2 and the lower-level clouds
from the Learjet. Albedo is defined as the ratio of reflected upward solar flux to downward solar
- Estimate cloud-top altitudes with the lidar and NFOVR on the ER-2. Estimate cloud-base
altitudes with the Learjet.
- Determine cloud-top temperature and estimate optical depth of thick cirrus (t > 1) and
cirrus emissivity by retrieval from the ER-2 NFOVR data at 6.7 Ám and 10.5 Ám. These
measurements will be taken only after flux-divergence-related objectives have been met and during
missions dedicated to microphysical observations by the Learjet.
- Measure the cirrus ice-particle-size spectra and characterize the ice crystal habits along the
equatorial tracks. Collect data from the cirrus cloud base (or -25ºC and lower temperatures) to as
high as the Learjet can penetrate, emphasizing the higher altitudes (colder temperatures).
- Simultaneously and at the same horizontal position, collect cloud brightness data (lidar)
and cloud structure with the ER-2.
- Collect ice crystals with the Learjet for later analyses of particle nucleus composition and
- Measure the winds and radar reflectivity structure with the P-3, coordinating
measurements with the Learjet and the ER-2.
Because several of the tasks will involve differencing the radiation flux measurements taken from
the two aircraft, the ER-2 and Learjet flights should be coordinated as much as possible. The
sampling times should coincide within a few minutes and within a few kilometers. The differences
in fluxes between the two aircraft should be obtained from running averages (longer than a few
minutes) of the instantaneous fluxes.
It is anticipated that roughly 80% of the time will be spent in measuring radiation fluxes from
anvils. For this purpose, both aircraft should make the measurements from their respective
constant altitudes. The ER-2 should maintain a constant altitude of about 60,000 ft (~ 18 km)
above the anvil tops. These measurements from the tropical tropopause will be critical for
validating the interpretation of radiation fluxes from ERBE by Ramanathan and Collins. CEPEX is
fundamentally interested in how the cirrus regulate energy below the clouds. Hence, the Learjet
should observe radiation fluxes below the anvil base and in cirrus, at a near-constant flight altitude
of between 35,000 and 40,000 ft (~ 1012 km). This constant altitude flight profile is very critical
for the following reason. The basic objective of CEPEX is to obtain the net flux divergence (or
convergence) across the anvil. However, radiation fluxes vary significantly with altitude. A
change in altitude, if it occurs, will have to be offset by correction factors that are poorly known.
Water-vapor distribution. The Learjet and the ER-2 should overfly the P-3 (within a
few hours) and the Vickers. Coordination is critical for several reasons. First, the dropsondes
from the Learjet should overlap the upsondes deployed from the Vickers within 50 km and within
30 minutes. Because the Vickers will launch accurate frostpoint hygrometers, this coordination is
helpful to validate the dropsondes. Second, the greenhouse effect calculations require SST, and
these will be taken from the P-3 and the Vickers. Third, by coordinating and collocating cloud-top
and cloud-base radiation flux measurements from the Learjet and the ER-2 with those from the P-3
and the Vickers, we can examine the fundamental issue of whether the water-vapor and cirrus
greenhouse effect warm the atmosphere or the sea surface.
Microphysics. Microphysics instruments are carried only on the Learjet. To measure
the microphysical properties, the Learjet will sample the cirrus in situ and obtain vertical profiles.
The Learjet, however, can only reach an altitude up to about 41,000 ft (about 13 km). This may
limit the vertical profiling of many cirrus clouds, whose tops may exceed 12 km. In spite of this
limitation, it is extremely important to obtain the size distribution and shape of the ice crystals.
Roughly 20% of the time planned for Learjet flights will be spent in penetrating the cirrus.
From such cirrus penetration runs, it will be possible to derive a number of cirrus properties
required by models. It will then be possible to compute cirrus cloud emissivity from the particle
cross sections obtained directly from two-dimensional imaging probe data collected through cirrus
layers and relate it to the path-integrated ice-water content. The correlation can be used to derive
optical depth from ice-water content (Heymsfield and Donner, 1989). Through these
measurements, it will also be possible to examine whether changes in ice-particle size distribution
can (or do) contribute to enhanced cloud reflectivity. The albedo of a cloud depends ultimately
upon the scattering behavior of individual ice particles (the phase function, which depends upon
particle size and shape), integrated over both the particle size distribution and the liquid-water path
length. There is observational evidence from stratiform clouds that ice-particle sizes decrease with
decreasing temperature, from millimeter sizes near 0ºC to only tens of Ám at tropopause
Flight Mission Profiles
The ER-2 and Learjet profiles consist of a western and an eastern trajectory. The triangular paths
indicated in Figure 21 are approximately 2,900 nautical miles in length for the ER-2, which is also
its endurance limit. As a result of logistical constraints, the ER-2 is required to return to the same
airstrip from which it departed (i.e., round-trips only) during scientific operations. The range of
the Learjet is approximately 2,100 nautical miles, and refueling stops are required at Wallis and
Tarawa on the eastern path and at Nauru on the western path. Simultaneous flights of both aircraft
have been planned along the 2ºS latitude line.
During the eastward trajectory. The Learjet will fly below the ER-2 during the
diagonal flight from Wallis Island to the eastern boundary at 2ºS and from there to the dateline.
Only during this leg of the eastern path can the cirrus radiative heating be estimated.
During the westward trajectory. Simultaneity of the two aircraft flights will be
accomplished during the northward flight from Fiji to 2ºS and the westward flight along 2ºS.
About six of each of the eastern and western flight legs are required to obtain monthly mean
estimates of the fluxes. For example, a simulation study with ERBE data (as shown in Figures
A.6 and A.7) reveal that an average of six samples, taken once every four days, captures the
monthly mean within about 510 Wm-2. Based upon this study, we have arrived at a tentative
sampling schedule of six west and six east triangle missions.
Back to the front
- Determine the large-scale, east-west gradient of sea surface temperature (SST) and
evaporative flux of the sea surface along the equatorial Pacific zone.
- Measure the variation of the vertical distribution of water vapor in the lower troposphere
along the equatorial Pacific zone.
- Measure the downwelling solar radiation near the sea surface.
- Determine the variation of
- turbulent fluxes of water vapor, sensible heat, and momentum
- thermodynamic and dynamic structure of the subcloud layer
- depth of the convective boundary layer as a function of SST and convective activity
along the equatorial Pacific zone
- Produce vertical profiles of water vapor, potential temperature, horizontal wind velocity,
and equivalent potential temperature along the aircraft track.
- Determine relative magnitudes of horizontal advection and flux divergence of water-vapor
along the equatorial Pacific zone.
- Intercompare airborne flux measurements and transfer coefficients derived by TOGA-
COARE with those derived from ships and buoys using bulk aerodynamic methods.
- Estimate the height of the cloud base of precipitating clouds by airborne radar.
The observational approach for the P-3 is to use an airborne platform carrying immersion-type
instruments (see Table 7) from 30 m above the sea surface to mid-tropospheric levels, along an
equatorial track (Figure 22) between Majuro in the west and Christmas Island in the east. The
aircraft flight missions will be coordinated with the other CEPEX aircraft as well as with the
Vickers' s equatorial atmospheric sounding mission. When possible, the P-3 will overfly
meteorological buoys and the Vickers along the equator.
Figure 22 also shows an additional option: a triangular flight track starting from and returning to
Majuro, which extends the equatorial transit into the center of the warm pool. Also shown are two
options to be flown out of Christmas Island, which extends the near-equatorial flight further east,
in case the SST gradient will be located in this area. These two P-3 tracks have the potential of
monitoring a 5,000-km-long equatorial zonecentered over the SST gradientwith respect to
evaporative fluxes, atmospheric state parameters, and SSTs.
If a gradient of surface evaporation is determined, it is important from a modeling perspective to
inquire into the physical processes that may contribute to it, such as the variation of SST,
entrainment flux, advection, etc. Occasionally throughout the expedition, the hypothesis that the
flux profiles for water vapor and sensible heat are linear will be checked. Using the airborne radar,
onboard video imagery, and satellite imagery, the amount and location of deep convective activity
for each traverse along the equator will be characterized. This will be important in determining the
effect of deep convection on the observed boundary-layer fluxes.
Flight Mission Profiles
The observing strategy consists of two types of flight mission profiles (Figure 23). One profile
addresses the estimation of a large-scale gradient in surface evaporation, and the vertical
distribution of water vapor below 700 mb, as a function of SST (Figure 23, top panel). The other
profile addresses the issues of the zonal variation of the depth of the convective boundary layer,
horizontal advection, and flux divergence (Figure 23, bottom panel). The first flight mission
profile will consist of 200-km legs flown 30 m above the sea surface, regularly interspersed with
ramp soundings to an altitude of approximately 3 km. The other flight mission profile consists of
100-km legs alternately flown at 30 m and just below the cloud base. As the aircraft moves from
30 m to higher altitudes, a short sounding will be made to monitor the total height of the convective
The flight missions will attempt to fly over or near research vessels and meteorological buoys
along the equator. This will enable comparison of airborne eddy correlation fluxes with surface
fluxes derived from applying bulk aerodynamics based on TOGA-COARE-derived transfer
coefficients. If that comparison is successful, then the buoy measurements can be used to
approximate evaporation values over longer time periods of the annual cycle(s).
The airborne soundings will yield profiles of winds, temperature, and water vapor. In some cases,
these soundings will be coordinated with Learjet dropsonde missions and Vickers upsonde
launches along the equator to provide a first-ever cross section of the central equatorial Pacific.
The sampling strategy addresses the question of how representative values of the horizontal flux
gradient along the equator can be achieved within a given accuracy. The problem has two
components. The first component concerns the averaging technique, whereas the second
component concerns the climatological generalization of data taken during a limited time period.
Regarding the first component, the expected small east-west gradient of evaporative cooling (order
of 50 Wm-2) makes it necessary to measure the water-vapor fluxes within about 10% accuracy.
This, in turn, requires a certain "averaging length." Following Lenschow and Stankov (1986), the
averaging length (L) is a function of the flight altitude (Z), the total depth of the convective
boundary layer (H), the required accuracy (a), and the characteristic "integration length" (l) of the
turbulence variable. Using this method, the required magnitude of L should be on the order of 4060 km for Z = 3050 m, H =0.60.8 km, a = 0.1, and l ~ 4050m. We will work with an
averaging length of 60 km. The center point of this length could then be moved, as a running
average, over the total length (near 3,000 km) of the equatorial flight leg. Enhanced fluxes
associated with convective activity would be treated as an independent data set.
The second aspect of the observing strategynamely, the climatological validity of the
measurements taken in March 1993will use a classification of convective activity encountered
along the flight track as developed for and applied during TOGA-COARE (Mapes and Houze,
1992; Chen and Houze, 1993). It measures the size and intensity of convective systems in terms
of contour areas of outgoing long-wave radiation of cumulonimbus tops on hourly satellite images.
The -65º OLR contour has been found to be the most representative parameter. The five categories
range in area from 0 to over 60,000 km2. They will be applied along the flight leg and
supplemented by the P-3's airborne radar from below. The convective fluxes associated with
various categories encountered will also be evaluated statistically in terms of bins. The
approximation for other seasons of the annual cycle will be based upon climatological satellite
information on convective categories available from the central Pacific.
Data analysis has three components: (1) relationship of fluxes, CBL (convective boundary layer)
structure, and CBL depth to SST and convective activity, including the important determination of
the near surface gradients of water-vapor flux; (2) production of the lower tropospheric profiles of
water vapor; and (3) determination of the zonal variation of entrainment flux, flux divergence, and
horizontal advection. Grossman's method (1984, 1992b) will be applied.
Gradient of water-vapor flux. This analysis is one of the most important for
CEPEX. In the tropics, water vapor and sensible heat fluxes are known to increase dramatically in
the vicinity of cumulus congestus and cumulonimbus convection. Because it is expected that these
cloud types (in various forms of organization) will be encountered on the P-3 traverses along the
equator, the analysis must take this into account to reduce data scatter. Data will primarily come
from the mission profile shown in Figure 23, top panel.
By integrating the aircraft ground speed, all data will be put in a distance series as opposed to a
time series. Using this distance series, two approaches will be used to estimate a gradient. The
first approach is the running mean, or low-pass filter, approach. To avoid introducing phase lags
between fluxes and the other data, all data will be filtered by a running mean (or low-pass filter).
The following will be done to accomplish this:
- Display distance series of all data, along with the water-vapor flux and sensible heat
flux. The CBL structure data will be plotted as profiles made during short soundings to monitor
the height of the CBL, so this will have a horizontal resolution from 100 to 200 km along the track.
The variation in profile structure will then be interpreted. The height of the CBL analysis will
follow that outlined in the steps below, but at much lower horizontal resolution than the normal in-
situ data measured on the level legs.
- Determine those portions of the distance series that are affected by active convection and
place them in a separate distance plot (now with gaps).
- Visually inspect the two distance series (one for suppressed conditions and one for
active convection conditions) for evidence of a consistent large-scale horizontal gradient.
- If a gradient is evident, then estimate the magnitude using curve-fitting techniques. It is
important to estimate as well the confidence limits of the gradient (Grossman, 1992a).
The second approach is the block average approach. Data other than fluxes will be unfiltered and
plotted. The fluxes will be block averaged and plotted at the center point of the block. Using 60
km as a guide, about three independent flux estimates can be made for each low-level leg. These
will be stratified into suppressed and convective categories and displayed on the distance series.
Apply techniques described in steps 3 and 4, above. Determination of the horizontal advection and
flux divergence of water vapor will utilize data from the mission profile shown in Figure 23,
bottom panel. It is also important to relate the estimates of horizontal advection and flux
divergence to the nearby state of convection. In this way, it may be possible to see if any variation
of these quantities is associated with observed changes in SST and CBL structure.
Profiles of water-vapor. Quality-controlled data will be used to display vertical
profiles of potential temperature, water vapor, horizontal wind velocity, and equivalent potential
temperature. Separate profiles of up/downwelling solar and infrared radiation will also be
produced. These profiles will be available as computer data files for input to radiation and cloud
models for other CEPEX scientists.
Cloud-base analysis from radar. The base of precipitating clouds will be determined
utilizing radar data from the P-3. With the tail radar in the vertical position, the radar is very
sensitive (~ -25 dBZ minimum detectable reflectivity) and should give a good indication of anvil
cloud base and cloud top, as well as of regions of precipitation.
Back to the front
The R/V Vickers
- Measure, by direct atmospheric observations, the water-vapor greenhouse effect at the sea surface.
- Measure the effect of cirrus on radiation fluxes at the surface over the equatorial Pacific.
- Measure the east-west gradients of water-vapor concentrations.
- Measure the east-west gradient in SST by direct and by radiometric methods.
- Measure the east-west cross section of vertical water-vapor and temperature distribution
from the surface to about 20 km.
- Measure the downward long-wave radiation and its east-west gradient.
- Determine the effect of clouds on the downward long-wave and solar radiation.
- Measure the window radiance to identify the cause of the super-greenhouse effect.
- Measure heights of the precipitating cloud base and precipitation rates.
- Measure emission spectra in the 520µm band.
- Measure surface wind speeds and humidity.
- Measure precipitation rates.
- Measure the three-dimensional distribution of aerosols from the surface to the PBL
(planetary boundary layer) with either an elastic backscatter lidar or the elastic channel of the
- Measure the cloud-base height from the elastic backscatter with the elastic lidar or the
- Measure the spatial distribution of water vapor between the surface and the lower
- Measure the vertical distribution of water vapor at high spatial and temporal resolution.
The Vickers will leave Honiara on 57 March, depending upon the readiness of the ship and its
scientific equipment and the flight plans of the CEPEX aircraft. At present there are two scenarios
for the beginning of the Vickers cruise (Figure 24):
1. Steam north along 160ºE for approximately 1.3 days to reach 2ºS, 160ºE and then turn
eastward along the 2ºS latitude.
2. Steam northeastward for approximately 1.8 days to reach 2ºS, 165ºE and then turn
eastward along the 2ºS latitude.
The first option lengthens the equatorial traverse, so that more time is spent in the warm-pool
region. This particular cruise-track option is favored if the P-3 is available for an overflight in the
western Pacific during the very early days of CEPEX. If there is no possibility for an overflight
by the P-3 when the Vickers is west of 165ºE, then the second option is favored because it will
allow for more central Pacific cruise time, as well as more overflight possibilities by both the
Learjet and the ER-2.
In either case, a principal objective of CEPEX is to have the Vickers spend as much time as
possible along the top of the triangular flight plan, either waiting for overflights or maneuvering to
facilitate the launching of the balloon sondes. From preliminary calculations, it appears that the
Vickers can spend approximately 59 days along 2ºS before reaching the extreme northeast vertex
of the triangular pattern, 2ºS and 170ºW. This should allow for two to three overflights by each
From the northeast vertex, the Vickers will sail eastward to approximately 163ºW and will then cut
northeast to Christmas Island. This part of the voyage should take approximately 2.6 days, during
which the P-3 may be able to overfly the Vickers again. Upon arrival at Christmas Island (on
approximately 20 March), the frostpoint hygrometer sonde, ozone sonde, and Vaisala sonde
equipment and personnel will be put ashore to continue observations for approximately one
additional week. The Vickers will not remain there but will continue on to Los Angeles, arriving
on approximately 30 March.
The cruise track of the Vickers is partly constrained by (1) the cost of operating the ship away from
its originally planned (and funded) direct return track from Honiara to Los Angeles; (2) the
necessity of traversing the SST and cloud-cover gradients between the warm pool and cooler areas;
(3) the CEPEX requirement that some aircraft and ship measurements be made coincidentally; and
(4) the desire to intercalibrate ship instruments with TOGA/TAO (tropical atmosphere ocean) array
of buoy instruments.
Atmospheric and oceanic energy budgets.
A central objective of CEPEX is to
observe and explain those energy fluxes at the sea surface that are responsible for the regulation of
upper-ocean temperatures. The Vickers will continuously measure SST, as well as downward
long-wave and short-wave radiation. The SST measurements will be used by themselves and as a
means to calibrate the SST remotely sensed by the CEPEX aircraft and satellite platforms. In turn,
the aircraft and satellite data plus SST from the moored array of TOGA/TAO buoys will be used to
extend the total CEPEX SST data set to cover a longer period of time and a greater area.
Upward long-wave energy flux (Wm-2) is accurately given by the Stefan-Boltzman equation,
R = e(ÓT4), where e = the emissivity of the sea, Ó = 5.67 x 10-8 Wm-2 k-4, and T is the SST in
degrees Kelvin. In the long-wave portion of the energy spectrum, e ~ 1 (i.e., the sea acts like a
blackbody). The downward long-wave energy flux from the atmosphere to the sea will be
measured by the upward-looking radiometers.
Further, the solar and long-wave flux measurements at the surface will be correlated with clouds
(from satellites) and ER-2 measurements to examine the link between the water-vapor greenhouse
effect and downward long-wave radiation at the surface, as well as the effect of cirrus on surface
Using the ship's own meteorological package and (tentatively) a more sophisticated package
operated by Los Alamos National Laboratory, the Vickers will make continuous measurements of
surface winds and humidity/temperature from an exposed location on the ships mast (and
tentatively, the bow). These data will be used to calculate sensible and latent heat fluxes using the
bulk formulas. These sensible and latent heat-flux values plus short-wave and long-wave flux data
will, in part, test the thermostat hypothesis for the regulation of upper-ocean temperatures. TOGA-
COARE will collect a very large amount of heat-flux data at the surface of the warm pool, but
CEPEX will extend this data set across the edge of the warm pool and into the cooler region to the
These calculated and directly observed heat fluxes will provide only some of the data necessary for
constructing a complete heat budget for the upper ocean. Unfortunately, there is no easy way to
observe or calculate the convergence of heat energy due to advection/mixing, nor will there be a
way to monitor temperature changes following a water parcel. An acoustic Doppler current profiler
(ADCP) will be operated continuously, and some expendable bathy-thermograph (XBT) profiles
will be collected (tentative). If the data from these instruments indicate horizontal uniformity over
long distances, this will give some rationale for dismissing the advection of zonal gradients as an
important temperature regulation mechanism. Fine structure in the near-surface temperature
profiles, if it exists, will be evidence for the absence of active vertical mixing as an important
Atmospheric profiles of temperature, water-vapor, and ozone. The CEPEX
aircraft will measure upward and downward short-wave and long-wave fluxes at more or less
fixed altitudes. In order to explain or model these measurements, however, it is necessary to
follow the variation of temperature, water vapor, and clouds vertically above the sea surface and
horizontally along the flight tracks. As explained in Chapter 5, the Learjet will drop a limited
number of sondes, which will profile temperature and humidity from an altitude of 35,00040,000
ft (10.512 km) down to the surface. The P-3 will complete intermittent temperature and humidity
profiles from altitudes of 30 m to 3 km using aircraft-mounted instruments.
In order to intercalibrate and extend the aircraft profiles, the Vickers will deploy a series of
temperature/humidity sondes, which are of a type characteristically more accurate than the
dropsondes. These upsondes, about 90 packages total, will be deployed along the track from
Honiara to Christmas Island, with special emphasis on deployment in the warm pool at the time of
aircraft overflights and at Christmas Island. One combination frostpoint, Vaisala, and ozone sonde
package will be deployed in Honiara, and three more combination packages will be deployed from
Christmas Island. The suggested sonde launch routine for the Vickers while under way includes
- 60 separate Vaisala (pressure, temperature, and humidity) sondes, one every 60 to 100 km
- 10 frostpoint hygrometers, with attached Vaisala and ozone sondes, one every 600 km
- 15 separate ozone sondes, one every 500 km
Adjustments to this deployment routine will be made in the field in response to environmental
conditions. To the extent possible, the CEPEX Field Operations Center in Fiji will advise the
Vickers regarding sonde launches, based upon near real-time analyses of measured or modeled
atmospheric conditions, satellite data, and aircraft observations. These sonde profiles plus profiles
from the P-3 will constitute the first extensive water-vapor section along the central tropical Pacific
ocean from the warm pool to the cooler, upwelling zone.
One of the most intriguing questions affecting the study of the energy balance of the atmosphere is
the apparent dryness of the stratosphere. A clear understanding of the processes responsible for
this remarkable phenomenon is of great potential importance to the thermostat hypothesis, because
the stratosphere may become more dry or more wet as a result of ocean-temperature and
tropospheric changes. Stratospheric water vapor is an important greenhouse gas and plays a
fundamental role in stratospheric ozone chemistry.
Two previous experiments, one in Panama and another in Micronesia, have supplied some
interesting but still inconclusive information on two opposite mechanisms that determine the water-
vapor content of the tropical lower stratosphere:
- hydrationprobably due to ice-particle evaporation from overshooting turrets forming
large stratospheric anvils above convective cloud systems
- dehydrationdue to the fallout of large, extremely cold, radiatively cooled ice particles
emanating from overshooting turrets, penetrating into the lower stratosphere, and thereby
entraining and drying stratospheric air
Another related question is, what determines the overall, net upward movement of air in the
tropical stratosphere up to altitudes of 2530 km? This movement is clearly indicated by previous
observations of chemical tracers in the stratosphere. Strong radiative cooling from the tops of the
thick anvil clouds may be expected to lead to downward motion in the stratosphere above. Does
the strong radiative warming in the thinner regions of the anvil cloudswhich are exposed to
upwelling IR radiation fluxesprovide a driving force for the upward branch of the reverse
Walker cell? If so, what influence does such an induced circulation have on the transfer of water
vapor into the stratosphere?
The combined ozone and frostpoint hygrometer sonde profiles to be collected by the Vickers
between Honiara and Christmas Island will identify regions of deep convection and will provide
information on their horizontal and vertical structure. Meridional sections of the moist static
thermodynamic energy over the Atlantic ocean have shown deep convection in the free troposphere
in a narrow +2-degree-wide region centered in the ITCZ (intertropical convergence zone).
Unfortunately, these data do not give information about the dynamical structure of the convective
cell in the upper troposphere, because the diabatic heating processes and large-scale subsidence
north and south of the upwelling region increase the enthalpy over and beyond the limits of
However, ozone can be used as a tracer to augment and amplify the information provided by moist
static thermodynamic energy. Ozone concentration in the lower troposphere, away from localized
sources, is very low. During deep convection, the ozone in an air parcel does not increase by
chemical processes, because these processes are rather slow when compared to the speed of
vertical ascent. Previous ozone profiles taken along an Atlantic meridional section show that ozone
is a perfect tracer for deep convection in the tropics. The combined CEPEX data setespecially
the frostpoint hygrometer and ozone profiles obtained from the Vickers and the Lyman-alpha
records from the ER-2will help to determine (1) if stratospheric air is entrained into the
troposphere in the turbulent regions of the anvils and (2) if the thin cirrus region can contribute
significantly to the transfer of tropospheric air into the stratosphere (as proposed by Gage et al.,
1991). If water vapor enters the stratosphere exclusively through the anvils, then water vapor
above 18 km in regions to the east of the active convection regions should be below 3 ppm. If it is
above 5 ppm, this is evidence for the significance of the Gage et al. mechanism. Relative humidity
above 10 ppm in the lower stratosphere plus ER-2 confirmation of thin cirrus and large radiative
heating will provide more conclusive evidence.
The upsondes from the Vickers, the ER-2 Lyman-alpha, and the Learjet dropsondes will be
calibrated as follows.
The frostpoint hygrometer sonde, frostpoint hygrometer, the Lyman-alpha hygrometer on the ER-
2, and the Learjet cryogenic hygrometer are the most accurate of this set.
- Use the frostpoint hygrometer sondes to calibrate the Vaisala sondes, as both will be on
the same balloon (1015 launches).
- Conduct simultaneous measurements with the frostpoint sonde, the Vaisala sonde, and
the dropsonde, when it is possible, for use in intercalibrating the different instrument
- Use the Lyman-alpha onboard the ER-2 and the frostpoint hygrometer onboard the Learjet
to calibrate the upper-troposphere water-vapor measurements by the Vaisala and
Mid-IR spectra of downward radiation flux. Mid-IR(520 µm) emission spectra
directed downward from a clear sky clearly show the signature of water vapor, carbon dioxide,
and ozone, thus demonstrating their individual contributions to the greenhouse effect. The Vickers
will carry an upward-looking FTIR (Fourier transform infrared) spectroradiometer, which will be
used to measure this spectrum about twice per hour during daylight and on a reduced schedule at
night. This radiometer is similar to the FTIR systems being developed by the University of
Wisconsin group for the Department of Energy (DOE) Atmospheric Radiation Measurement
(ARM) Program (e.g., Smith et al., 1990), but it is easily transportable for remote field
measurements. The radiometer will be calibrated at two points, using a blackbody temperature
source traceable to the National Institute of Standards and Technology (NIST). The FTIR will be
mounted on a gimbaled table and will look directly up through an open hatch located on the main
deck of the Vickers.
For FTIR data taken under cloudy skies, an algorithm has been developed to retrieve spectral cloud
emissivity and to estimate the spectral optical depth and the equivalent radius for the cloud
droplets. This algorithm is based on matching FTIR brightness-temperature spectra to radiative
transfer solutions provided by a discrete ordinates model (Stamnes et al., 1988). Interpretation of
the FTIR data will also make use of recent work on the signatures of cirrus clouds in the mid-IR
window (e.g., Ackerman et al., 1990).
Precipitating cloud-base altitudes, spatial frequency of precipitating clouds,
precipitation rates, vertical cross sections of cloud liquid- or ice-water content,
in-situ precipitation rates. As part of its complement of TOGA-COARE instrument systems,
the Vickers carries a 5-cm Doppler radar owned by the Massachusetts Institute of Technology
(MIT) and operated by Colorado State University. This radar, plus some automatic rain gauges,
will remain in use during CEPEX.
The radar data will be valuable to CEPEX for purposes of characterizing the vertical and horizontal
structure of the atmosphere around the ship and for providing precipitating cloud-base heights and
other observables of use in modeling the radiative balance. Quick-look data from the radar will be
of particular value in judging when to launch the sondes. This would enable CEPEX to distribute
the profile data uniformly over the various atmospheric conditions encountered.
The Doppler radar antenna is stabilized up to an elevation angle of only 55º above the horizon;
therefore, vertical scanning is not possible. A military-grade inertial navigation system is used to
sense the antenna's movements. Information on roll and pitch is fed into the radar's computer,
which provides updates to the antenna to maintain the desired pointing angles. Corrections are
done in real time, so Doppler velocities are corrected for ship and antenna movements. A Global
Positioning System (GPS) receiver is used to provide precision time and backup navigation
The radar displays both reflectivity (returned power) and radial Doppler velocity for precipitating
and non-precipitating droplets and ice-crystals 200 Ám or larger. For CEPEX, the radar will be
operated 24 hours per day. It will provide continuous volume scans consisting of a series of
constant elevation angle sweeps from 0.8º up to 55º. The upper elevation angle will be dependent
upon the depth of convection and the range to the target. Very similar scans are being collected
during TOGA-COARE. The range of the radar in the volume-scan mode is typically 100150 km.
Such volume scans can be completed in about 8 minutes.
Every 30 minutes, low-level surveillance scans will be collected for a period of 2 minutes. These
scans will locate all precipitation areas within 300 km of the ship. The scan sequence will consist
of a surveillance scan followed by three volume scans; the sequence will then repeat. When
extensive thick anvil clouds are present, the radar will also do RHI (range-height indicator)-mode
scans in which the azimuth is held constant and the radar scans in elevation only. These scans will
provide vertical cross sections of the anvils.
In addition to providing CEPEX with precipitating cloud-base heights of anvil clouds and rainfall
rates, the operation of the Doppler radar in conjunction with the aircraft and shipboard radiation
measuring systems will benefit programs such as the Tropical Rainfall Measuring Mission
(TRMM), the Global Energy and Water Cycle Experiment (GEWEX) and NASA's Clouds and
Earth Radiation Energy System (CERES), which is a follow-up to ERBE. The collocated radiation
and radar data, along with the AVHRR and GMS data, should provide a rich source of information
for these large programs. For example, the estimation of rainfall from OLR images is expected to
be very difficult in marginally convective regions such as the central Pacific. Most of the TRMM
resources will be focused on the warm pool, where convection is intense.
Remotely sensed water vapor and water-vapor fluxes, aerosol distributions,
and cloud-base heights. Los Alamos National Laboratory will be using an elastic backscatter
lidar and/or an ultraviolet (UV) Raman lidar system aboard the Vickers to sense remotely both
water-vapor and aerosol concentrations up to altitudes of 510 km. If this system can be adapted
to conduct a horizontal and vertical scanning of water-vapor concentrations from a rolling and
pitching ship, there are techniques whereby the scan data can be used to derive vertical moisture
fluxes (Cooper et al., 1992).
The solar-blind water-Raman lidar uses a pulsed UV laser (248 nm) to generate light, which is then
Raman-shifted to 263 nm and 273 nm by atmospheric nitrogen and water vapor, respectively. The
Raman return light is collected during either the day or night by a 61-cm-diameter Cassegrain
telescope that is mounted horizontally on a self-contained enclosure. The light is subsequently
detected by three photomultipliers (N2, H20, and elastic backscatter) and digitized at a rate of 100
MHz, resulting in an intrinsic range resolution of 1.5 m. The Raman effect is useful, because the
shifted light is molecule specific and is proportional to the concentration of that molecule at a given
location. Therefore, an absolute measurement of water vapor is made by taking the ratio of the
water-vapor signal to that of nitrogen. This ratio results in a water-vapor-concentration
measurement (mixing ratio in g of water per kg of air) at all points along a given line of sight,
which is typically within 35% of psychrometric measurements. Using a periscope mirror
assembly with a steerable upper mirror, the system can scan +90º in azimuth and +90º in elevation.
The scanning process allows a large number of measurements to be made almost simultaneously,
and enables the observation of atmospheric processes in both horizontal and vertical dimensions.
A number of shots taken along several elevations at constant azimuth results in a measurement of
the two-dimensional water-vapor concentration in the vertical plane. Similarly, scans can be made
in the horizontal direction by maintaining a constant elevation and stepping across the azimuthal
plane. The angular step size of the scans is user-selectable for any step greater than 0.01º. The
range of the system is, in part, dependent upon the water-vapor concentration, ozone
concentration, and the number of laser pulses summed per time bin. Previous experiments indicate
that the horizontal range during the day might be up to 3 km and 4 km at night. At night, the
Raman lidar should increase due to decreased background noise.
Visual images of the clouds overhead. The Vickers will carry a video camera and a
35-mm camera with Polaroid filters to photograph cloud formations on a regular basis, in
connection with other observations. The video footage plus still images will be used to help
interpret events recorded by the other instrument systems on the Vickers.
Back to the front
- Obtain synoptic-scale gradients and monthly variations for
- greenhouse effect and cloud forcing
- cirrus and thin cirrus cloud cover
- cloud-cluster statistics (Houze classification)
to complement high-resolution aircraft and ship measurements
- Obtain synoptic-scale diurnal variations in
- cloud albedo
- cloud cover and cluster statistics
- cloud IR radiation
- Relate spectral satellite radiances to top-of-the-atmosphere broadband fluxes, using
coincident satellite and ER-2 radiometer measurements and model calculations. Based upon these
correlations, obtain the broadband top-of-the-atmosphere fluxes on regional (2.5º x 2.5º) scales
with good diurnal sampling.
- Derive cirrus and convective super-cluster statistics to help relate aircraft and ship
measurements to synoptic-scale cloud distributions.
- Calculate approximate SSTs for the CEPEX region using McCains multichannel SST
procedure (McCain et al., 1985). These estimates can be refined with Reynolds's method (1988)
to blend ship, buoy, and P-3 measurements with satellite retrievals.
- Obtain regional clear-sky climatology, greenhouse effect, and cloud radiative forcing.
- Obtain column-integrated water-vapor distribution(s) for the CEPEX region.
Two types of satellite data will be collected at Fiji: (1) high-resolution visible and IR images in
narrow-window spectral bands and (2) IR sounding measurements. The two sources of image
data are four NOAA polar orbiting satellites (NOAA-9, 10, 11, and 12) and the Japanese GMS
satellite at 140ºE. The high-resolution infrared sounder (HIRS) is flown with the AVHRR
instrument on the NOAA satellites, and the soundings will be captured simultaneously with the
AVHRR data. The soundings contain 19 narrow-band measurements spanning the near and mid-
IR spectrum. Microwave soundings of water vapor from Defense Meteorological Satellite
Program (DMSP) instruments will be incorporated after completion of the experiment.
The NOAA platforms fly over a given point at two fixed local times each day, one in the daytime
and one at night, twelve hours later. The nominal overpass times for Fiji are shown in Table 11.
The overpass times for NOAA-9 and NOAA-10 are very approximate, as the orbits of these
satellites are steadily drifting. Because the width of the images is comparable to the east-west
extent of the CEPEX region, usually two images are required to cover the eastern and western
Because the orbital period is 100 minutes, the two images are taken 100 minutes
apart. The full-resolution AVHRR data will be captured by the SeaSpace high-resolution picture
transmission (HRPT) system at Fiji, with archival backup at Kwajalein.
The GMS satellite transmits full-disk visible and IR images every hour, at spatial resolutions
comparable to AVHRR. The full-resolution GMS data will be recorded on a SeaSpace GMS
system at Fiji, and the sectors of the visible and IR data covering the CEPEX region will be
automatically selected for further analysis. In addition, the Australian Bureau of Meteorlogy
(BOM) will archive GMS sectors spanning the central Pacific during the experiment.
The wavelengths, equivalent blackbody temperature accuracy (Gautier et al., 1991), and resolution
of the GMS and AVHRR data is given in Table 12.
Total column water-vapor and precipitation rates will be obtained from the special sensor for
microwave/imager (SSM/I) on the DMSP platforms. The water-vapor column footprint is
60 x 40 km at nadir, and each cross-track scan contains 64 samples spanning 1,400 km. The
presumed accuracy of the vapor measurements is 1015%. This information will be available,
retrospectively, through NASAs WetNET. Additional DMSP data, the optical line scanner (OLS)
visible, and IR images with roughly 5-km resolution may be used in conjunction with the AVHRR
and GMS imagery. These images will be available through the CEPEX Central Archive (CCA) in
Boulder, Colorado. The DMSP satellites are in polar orbits with periods similar to AVHRR, so
the platforms carrying the SSM/I-2 instrument will fly over the CEPEX region twice daily.
Satellite radiances will be collocated with broadband flux measurements from the ER-2.
Theoretical and observational studies suggest that the relationship between narrow-band and
broadband measurements should be nearly linear for both visible (Cess and Potter, 1986) and IR
(Minnis et al., 1991). Preliminary empirical relationships will be fine-tuned and then applied to the
satellite data to derive synoptic-scale top-of-the-atmosphere visible and long-wave fluxes. Clear-
sky regions in the data will be identified using one of several standard techniquese.g., spatial-
coherence (Coakley and Bretherton, 1982). With the clear-sky and average fluxes and regional
estimates of SST, the atmospheric greenhouse effect and cloud radiative forcing (from Cess and
Potter's Method I; the same method used for ERBE data) may be calculated for the entire CEPEX
The distribution and evolution of convective super-clusters will be obtained using equivalent
blackbody temperatures for the IR-window channel with the TOGA-COARE classification scheme
(Williams and Houze, 1987; Mapes and Houze, 1992). The distribution of thin cirrus can be
obtained from HIRS data using the cloud-slicing algorithm (Ackerman et al., 1992). This set of
information should be sufficient to meet the primary scientific objectives for the satellite data.
Back to the front
- Determine SST on hourly time scales.
- Determine the effect of cirrus on surface solar insolation.
- Determine the boundary-layer winds, humidity, and temperature.
- Determine precipitation.
- Correlate the effect of cirrus on solar-radiation flux at the tropopause with the effect of
cirrus on surface insolation. From the precipitation sensors on moorings, examine the relative
effect of precipitating and non-precipitating clouds.
- Estimate evaporation from surface measurements by employing drag laws and COARE
estimates for exchange coefficients.
- Use the moorings' SST values as ground truth.
An array of moored buoys (see Figure 25) is envisioned to span the entire tropical Pacific at zonal
intervals of about 15º longitude; actual installations as of the time of CEPEX will be concentrated
mostly to the east of the CEPEX flight patterns and to the west (for purposes of supporting TOGA-
COARE). The closest TOGA/TAO buoy locations are shown in Figure 25. Solid triangles denote
Atlas-style mooring equipment, and solid squares in circles indicate the more elaborate Proteus-
style moorings (Atlas: autonomous temperature line acquisition system; Proteus: profile telemetry
of upper-ocean currents). Most of the TOGA/TAO moorings shown in Figure 25 are expected to
be operational during the TOGA-COARE period and to continue operating through the CEPEX
period and after.
In addition to the TOGA/TAO arrays, R. Weller and his group at Woods Hole Oceanographic
Institution are scheduled to install a meteorological and oceanographic buoyIMET (improved
meteorological instrument)-style mooringnear the center of the TOGA-COARE array. The
IMET buoy is indicated by a star-in-circle symbol in Figure 25. The TOGA-COARE schedule
mentions installation in October 1992, and removal sometime during March 1993. In brief, all of
the buoys mentioned measure air and sea surface temperatures, wind speed, and direction.
The Atlas and Proteus buoys sample wind speed and direction at 0.5-second intervals and record
6-minute vector component averages once every hour. They sample air temperature and SST every
10 minutes and record them as hourly and daily averages. However, an ARGOS (the French/U.S.
satellite and data location system) transmitter on each of the Atlas and Proteus buoys is turned on
for two four-hour periods each day. The brief data message repeatedly sent out during these
periods consists of the previous day's averages for winds and temperatures plus the most recent
"spot" value of each variable. Data are available to outside investigators through an Internet
connection to the Pacific Marine Environmental Laboratory (PMEL) computer. Daily averages for
the previous month are easily available. Near-real-time access to spot values and daily averages is
more complicated, but Scripps Institution of Oceanography (SIO) is working with PMEL to
overcome the difficulty.
In addition to air and sea temperatures and winds, the IMET and Proteus buoys, together with
some of the Atlas buoys, are equipped with sensors. The IMET has two Vaisala Humicap-type
sensors feeding into the two separate data-averaging and recording systems. The sampling,
recording, and broadcasting specifics are as noted for winds and temperatures. R. Weller has
estimated that the humidity sensor is accurate to 23%.
Atlas and Proteus moorings with humidity sensors (Rotronic type) are noted by the leter H
attached to the solid triangles and solid squares in circles in Figure 25. It is not presently known if
humidity sensors are on any of the moorings east of the dateline, other than the Proteus mooring at
140ºW and 0ºN.
Rotronic humidity sensor output is sampled at 10-minute intervals. Previous daily averages and
spot hourly averages are broadcast via ARGOS on the same schedule as noted for SST, etc.
Humidity sensors are a new feature (since 1991) of the Atlas and Proteus buoys, and their
accuracy will undergo testing during TOGA-COARE. The Rotronic sensor is, however, similar to
the Humicap sensor extensively tested by Weller.
Four of the moorings in the western tropical Pacific are carrying or will be carrying rain gauge
sensors. These buoys have been marked (see Figure 25) with the letter P for the precipitation. The
presence or absence of rain gauges on buoys in the eastern Pacific is not known. The rain will be
sampled at 4 Hz and averaged over 1-minute periods for 20 minutes every six hours. One Atlas
buoy and two Proteus buoys will carry optical rain gauges built by Science Technologies. The
optical rain gauge will have a five-second sample rate, with statistics computed for one-hour
intervals (average, maximum, and standard deviation of rainfall rate and percentage of time it is
raining). Some hourly values will be transmitted via ARGOS, and all data will be recorded for
later recovery by ship.
There are Eppley PSPs (precision spectral pyranometers) on the equatorial Proteus buoys at
140ºW, 165ºE, and 156ºE. According to Eppley, the WG7 clear glass hemispheres used in (most
of) the PSPs are transparent from 285 to 2,800 nm. The cosine response is +1% from 0º to 70º
and +3% from 70º to 80º zenith. It should be noted that on the Proteus buoys the PSPs are
probably located below the level of some of the other instruments, especially the anemometers.
This location will cause a certain amount of shadowing of the PSP.
Back to the front
Island Station Measurements
The island stations (see Figure 26) consist of four categories:
- ISS at selected stations
- wind profiler at Christmas Island
- routine upper-air island stations
- special balloon sondes during the CEPEX field-phase
Integrated sounding system. The ISS consists of a UHF/VHF Doppler wind profiler, a radio
acoustic sounding system (RASS), an Omega NAVAID-based sounding system, and a surface
meteorological station. The Doppler wind profiler operates at 915 MHz, in conjunction with
RASS, to obtain half-hourly virtual temperature and wind profiles (as well as signal to noise fields
for each beam). Vertical resolution of RASS profiles are 60 m from 112 m AGL (above ground
level), up to a maximum of 1,553 m AGL. Vertical resolution of wind profiles are 238 m from 91
m AGL up to a maximum of 11,755 m AGL, depending upon atmospheric conditions. The
NAVAID upper-air sounding system utilizes a lightweight radiosonde to telemeter measurements
of pressure, temperature, and humidity back to the ISS station. Winds are computed using Omega
navigation signals. Vertical resolution of these soundings are 10 seconds from the surface to 400
mb, and 250 m thereafter to balloon burst. The surface meteorological station performs one-
minute averages of surface-layer winds, temperature, pressure, humidity, radiation (total solar,
net, diffuse, and incoming infrared), and rainfall (one-minute total accumulation). (Direct radiation
is measured at Manus and Kavieng only; incoming infrared radiation is measured at Nauru and
The high-resolution ISS data sets are recorded on-site and later transmitted to the National Center
for Atmospheric Research (NCAR) for final processing and quality control. A subset of these data
(one-hour RASS/wind profiles, mandatory/significant levels, half-hour surface averages) will be
inserted onto the Global Telecommunication System (GTS) in real time for incorporation into
forecast and analysis models. CEPEX will collect ISS data from four sites (Manus, Kavieng,
Nauru, and Kapingamarangi). Twice daily soundings (00 and 12 UTC [universal time
coordinate]) are planned from these sites.
Wind profiler at Christmas and Penrhyn islands. The NOAA Aeronomy Research Laboratory
(ARL) operates similar VHF wind profiler systems at Christmas and Penrhyn islands. Each
system consists of a 915 MHz (0.156-km range) and a 50 MHz (2.515-km range) profiler,
which measures and records high vertical resolution wind data every two minutes. Half-hour
average wind profiles are derived from the minute data. No RASS capability is included with these
systems. From the 00 and 12 UTC wind profiles, standard PIBAL (pilot balloon) observations are
created and transmitted via GTS. The high-resolution data are forwarded to ARL on a monthly
basis for final processing and archival. These wind profiles are needed to examine the link
between radiative heating due to thin cirrus, vertical velocity in the lower stratosphere, and water
Routine island-based rawinsonde releases. Once to twice daily rawinsonde releases are performed
from numerous tropical island stations in the Pacific Ocean. These stations are a combination of
WMO and nation-operated sites, using both manual and automatic data recovery systems (see
Table 13 for a listing of those sites pertinent to CEPEX) Following each release, the
mandatory/significant-level data are coded and transmitted via GTS. Higher vertical resolution data
will be stored on-site and used in the final archive. The soundings will provide high-resolution
thermodynamic and wind profiles for defining environmental conditions.
Special island-based rawinsonde releases. Five island-based rawinsonde stations (Fiji, Kanton,
Funafuti, Honiara, and Nauru) in the CEPEX domain have been designated for special scheduled
releases to support CEPEX activities. Scheduled release times are shown in Table 13. All special-
release high vertical resolution data will be stored on-site and used in the final archive. A portable
Vaisala Marwin system provided by the Naval Postgraduate School will be available in Fiji for
CEPEX as a backup for the existing system operated by the Fiji Meteorological Service and to
provide special releases to support CEPEX aircraft operations (i.e., intercalibrations, etc.). All
soundings will provide high-resolution thermodynamic and wind profiles for defining
Back to the front
Back to Main Page