Chapter 5 | EXPERIMENT DESIGN |
Although these last two constraints limit the flight path along the west to east region of maximum SST gradient, the observational program should allow the primary objectives of the CEPEX experiment to be accomplished.
Experiment Location and Period
This section provides an overview of the experiment location and duration. Specific details vary
depending upon the particular platform, each of which is described in Chapter 6.
The supporting platforms will span the entire CEPEX domain, which is roughly the region
between 20ºS and 20ºN and between 150ºE and 150ºW, as depicted in Figure 3. This domain
covers the expected maximum SST gradient and provides temporal and geographical
interconnection with TOGA-COARE.
Observations from the primary platforms (i.e., the three aircraft and the ship) will be taken
primarily along a path at 2ºS latitude. Although climatology suggests that any parallel of latitude
between 4º and 4ºN with sufficient gradient of SST would meet the design objectives, the
constraints imposed by the ER-2 operations dictate the choice of 2ºS for the primary track. As
shown in Figures 3 and 20, the Vickers will gather data between 160ºE and 160ºW; the ER-2 and
the Learjet flight tracks will span the 2ºS parallel between 165ºE and 170ºW; and the P-3 track will
span the 2ºS parallel between 160ºW and 155ºE.
The primary base of operations for CEPEX will be Nadi, Fiji. The P-3 will operate from
Honolulu, Hawaii, and the Vickers will sail from Honiara in the Solomon Islands. The ER-2 and
Learjet measurements will be made along two triangular paths, originating in Fiji (see Figure 3).
The endpoints of the western and eastern legs along the equator are largely constrained by the
range of the ER-2.
CEPEX operations will begin on or about 7 March 1993 and will extend through approximately 7
April 1993. Research flights will begin 7 March 1993 and will continue for one month. The
cruise of the Vickers will begin on 7 March 1993, from Honiara, in the Solomon Islands, and will
arrive at Christmas Island on 20 March 1993. These dates may change by a few days, depending
upon the availability of the aircraft or the ship. Data from the supporting platforms will be gathered
from 1 December 1992 to 30 April 1993.
Impact of the El Niño Phase on the Flight Plans
The west-east location of the transition region from the warm (SST > Tc) to the cold (SST < Tc)
ocean varies with the phase of the El Niño (see Figure 20). The transition region is slightly east of
the dateline during a normal year (e.g., March 1985, as shown in the top panel of Figure 20), and
moves west of the dateline during an intense cold period (e.g., the La Niña year of 1989, as shown
in the bottom panel of Figure 20). This transition region is eliminated altogether during a fully
developed warm event (e.g., the 1987 El Niño event, as shown in the middle panel of Figure 20).
For March 1993, the transition region is predicted to be very similar to that of March 1985
(Barnett, private communication). However, observed SST for early January 1993 reveals that the
central equatorial Pacific is warmer than normal by about 1 to 2 K. This is quite surprising, since
an El Niño event just occurred in 1992. As a result of the anamolous nature of this warming
immediately following an El Niño event, it is premature to conclude that this warming will persist.
Nevertheless, flight plans will have to remain very flexible, within the constraints imposed by
aircraft range and safety requirements.
Figures A.1 through A.3 show regional patterns of atmospheric greenhouse effect and cloud
forcing fields for March 1985 and March 1987, as obtained from ERBE observations. Figures
A.4 and A.5 show the observed monthly mean SST, Ga, and long-wave and short-wave cloud
forcing fields along the proposed flight tracks of the ER-2 and the Learjet. Figures A.6 and A.7
show the corresponding daily fields. The proposed flight and ship tracks are based upon a careful
review of the results shown in Figure 20, as well as in the figures in Appendix A.
Along the 2ºS parallel, between the western and eastern boundaries of the flight and ship tracks,
the SST ranges from slightly over 302 K to about 299 K. The atmospheric greenhouse effect
decreases systematically from 190 Wm-2 to 170 Wm-2 (Figure A.1). Similar ranges of SST and
Ga are seen from Fiji to the equator (Figures A.4 and A.5). The corresponding cloud forcing
fields (Figures A.2, A.4, and A.5) reveal that along the equator the short-wave cloud forcing
varies from -20 Wm-2 to -80 Wm-2 and the long-wave forcing varies from about 20 Wm-2 to 60
Wm-2, with slightly larger ranges between Fiji and the equator. The variation in evaporative fluxes
between the warm pool and the eastern boundary of the P-3 flight track is about 50 Wm-2 (Figure
13).
If March 1993 is a normal year, then the CEPEX flight tracks will capture a large range of
systematic variations in evaporation and the radiative fluxes. A central Pacific that is warmer than a
normal year would require more eastern than western flight trajectories for the ER-2 and the
Learjet. In addition, the P-3 flight track may have to extend to the east of 160ºW by another 10º
longitude, and the Vickers may have to sail along the 2ºS latitude parallel as far east as 130ºW to
capture the required range of SSTs.
Design of the Composite Observing System
The CEPEX composite observing system (Figures 3 and 19) is designed to serve a variety of
purposes. In developing the thermostat model, Ramanathan and Collins (1991) make several key
assumptions that must be verified through observations. Further, a complete examination of the
processes involved in a thermostat mechanism can only be done with complex, sophisticated,
radiation models and GCMs. These models and their parameterizations must be validated or
constrained by appropriate observations.
Most required parameters will be estimated from a variety of instruments and platforms (see Tables
6—10 for a listing of the instruments in each of the various platforms). Key tasks for CEPEX are
the intercomparison and integration of the data taken with instruments from the different platforms.
Sea surface temperature. SST will be obtained from (1) direct measurement from the Vickers, (2) the IR radiometer onboard the P-3, (3) the window channel radiometer onboard the Vickers as well as from the Learjet and the ER-2, and (4) COARE equatorial moorings. None of these will provide the SST data on the spatial and temporal scales required by CEPEX. However, the National Meteorological Center (NMC) SST analyses, derived from AVHRR (advanced very high resolution radiometer) satellite and buoy data, will be validated with the in-situ CEPEX data.
Water-vapor and temperature from the surface to about 25 km. Flight-level measurements will be obtained by the ER-2 (Lyman-alpha hygrometer), the Learjet (cryogenic hygrometer and chilled hygrometer), and the P-3 (Lyman-alpha and dew-point hygrometers). These measurements will be complemented by dropsondes deployed from the Learjet and balloon sondes launched from the Vickers to obtain vertical profiles along 2ºS. The balloon sondes include 15 frostpoint hygrometer sondes, which will determine water vapor to an accuracy of 5—10% up to about 25 km. In addition, water-vapor concentrations will also be obtained from upper-air stations and the special integrated sounding system (ISS) on the island stations.
Radiation. The most accurate broadband and spectral measurements will be made from the
ER-2 and the Learjet. The rest of the measurements will be made from solar radiometers and
pyrgeometers, whose errors are on the order of 15 Wm-2. Broadband solar and long-wave
irradiances in the lower troposphere will be measured by the P-3; solar and long-wave irradiances
at the sea surface will be obtained from radiometers and pyrgeometers on the Vickers; and solar
irradiance at the sea surface will be obtained as well from radiometers on two equatorial moorings
at hourly intervals.
Table 6
Evaporative cooling of the sea surface. CEPEX must observe directly (1) evaporative and sensible heat-fluxes in convectively disturbed and undisturbed conditions in the equatorial Pacific and (2) east-west gradients in evaporative flux. These measurements will be made from the P-3, using the eddy correlation technique. Because only a limited number of flight hours (less than 100) are available, fluxes will also be estimated using the bulk aerodynamic formula and boundary- layer measurements of humidity and winds from buoy sensors. The exchange coefficient will be determined from COARE data. COARE, which directly precedes CEPEX, has a boundary-layer flux program that will measure the variability of the surface moisture, heat, and momentum fluxes over the warm pool and adjacent waters. Observations will be made over the full range of wind structures, from the ambient trade wind regime through the episodic westerly bursts. Multiple aircraft will provide area-averaged flux estimates. COARE will provide estimates of both convectively disturbed and undisturbed conditions. CEPEX will employ the same aircraft and instruments to estimate the variation of evaporation from the warm pool to the colder ocean in the central and eastern Pacific under a variety of cloud conditions. Table 7
Microphysics. The Learjet and the P-3 both have microphysical instruments. The measured and derived quantities include ice-water content, size distribution of particles larger than a few microns, shape of the particles, and ice-particle collections. The focus will be cloud microphysical parameters (shape and size distribution of crystals) that contribute to the high albedo of cirrus. Because its maximum altitude is limited to about 44,000 ft, the Learjet will be able to observe these only for cirrus whose tops are below about 13 km. For cirrus with tops higher than 13 km, the Learjet will measure the properties near the base.
Cloud-top and base altitudes. Although prior field measurements (e.g., see GATE data summarized in Houze and Betts, 1981) indicate that anvils occupy more than 80% of the tropical cluster area, it must be verified that the reflected solar radiation observed by ERBE is due to cirrus, not to low and mid-level clouds. This requires knowledge of the cloud top, cloud base, and clouds beneath the cirrus and will be determined from the following platforms: (1) the ER-2 (lidar and NFOVR [narrow field of view radiometer]); (2) the Learjet (NFOVR, FSSP-300 [a cloud droplet spectrum sensor], and a particle imager); (3) the P-3 (radar); (4) the Vickers (spectral radiometer and radar); and (5) satellite measurements (cloud top only).
Link between SST and the greenhouse effect. The vertical structure of Ga above waters with a range of SSTs greater and less than 300 K has to be estimated from the ER-2, Learjet, P-3, and Vickers measurements. The change in Ga at the tropopause, relative to the change in downward long-wave radiation flux (from water vapor) at the sea surface, will be calculated; that is, ¶Ga(trop) / ¶Fa-(Z=0) can be determined (Ramanathan and Collins' model implicitly assumes this factor to be 1). Table 8
From SST observations obtained from the moorings, the Vickers, the P-3, and AVHRR, upward
emission, F+(Z=0), at the sea surface for each day is calculated. Ideally, this will coincide with the
ER-2 and Learjet flight paths. The ER-2 broadband long-wave flux measurement for clear-sky
conditions will yield Fa+(trop), from which the greenhouse effect of the entire troposphere is
estimated as, Ga = F+(Z=0) - Fa+(trop). Trapping by the atmosphere below the anvil decks will
be determined from long-wave upflux at the Learjet's altitude. As the Learjet observes only
radiances, these measurements have to be converted to flux using algorithms developed from
models and ER-2 radiometric data. If the anvil base altitude is Zb and F+(Zb) is the upflux at the
Learjet altitude, then the greenhouse effect of the column between sea surface and Zb is G(Zb) =
F+(Z=0) - F+ (anvil base). If the sky below Zb is clear, then G = Ga, the greenhouse effect of the
atmosphere. Also note that the difference in Ga estimated from ER-2 and Learjet measurements
yields the greenhouse effect of the upper troposphere. Likewise, the difference in Ga between
Learjet and P-3 measurements yields the greenhouse effect of the low and middle troposphere.
To estimate the trapping by the various layers, fluxes will be computed with a validated radiation
code using the observed temperature and humidity profiles as input. Direct measurements of
downward long-wave flux at the surface will be obtained from the Vickers. Depending upon
instrument availability, this flux may also be measured from the P-3. Fluxes will be computed
from a model, using observed soundings from the Learjet dropsondes and the balloon sondes from
the Vickers as input. Collocated temperature and humidity soundings from the Vickers with
radiation measurements onboard the ER-2 and Learjet will be used to stratify the greenhouse effect
estimates between convectively disturbed and suppressed conditions. The representativeness of
these estimates, with respect to spatial and temporal sampling, will be addressed with the aid of
satellite and analyzed meteorological data.
Cloud radiative forcing of the column and radiative heating within anvils. The ER-2 and
the Learjet will fly respectively above and below the anvils, except when taking microphysical
observations, to obtain the radiative heating of the anvils and the cloud radiative forcing.
Consequently, identical radiative instrumentation will be used on both aircraft to assure compatible
measurements. These radiometric systems are the most advanced available and have been used by
the NASA Ames radiation group for several years.
Two independent methods will be used to obtain the desired result: (1) the direct method and (2)
the modeling approach. By the direct method, the net flux from the downward- and upward-
looking radiometers will be obtained on the ER-2 above the cirrus tops (altitude of about 16—18
km) and on the Learjet below the cirrus base (altitude of about 6—12 km). Under ideal conditions,
the difference between the two net fluxes should yield energy converging into the anvils. This
method requires simultaneous observations by the ER-2 and the Learjet. Because of slight
differences in their cruise speeds, we will not have exact simultaneity all the time. To account for
this, a running average (for example, over 10 km or more) of the net flux measurements from the
two aircraft will be calculated. With the modeling approach, the cirrus microphysical properties
(size and number distribution by Learjet spectrometers and particle collectors) will be measured,
and the cloud-base and cloud-top altitudes will be determined by lidar. The optical depths can be
obtained from the total-direct-diffuse radiometer (TDDR) instrument, which along with temperature
and humidity data will be incorporated in a multiple-scattering code to obtain radiative heating
rates. The two methods will be compared for consistency.
Estimation of the f factor along the flight track. The f factor is simply the ratio of the long- wave cloud forcing at the surface, Cl (Z = 0), and the long-wave cloud forcing at the tropopause, Cl, such that Cl (Z = 0) = F- (Z = 0) - Fa- (Z = 0). Likewise, Cl = Fa+ (Z = 18 km) - F+ (Z = 18 km). The ER-2 will measure F+(Z = 18 km); the Vickers will measure F- (Z = 0 km); and the P-3 will measure F at flight-level in the boundary layer (about 100 m). Table 9
Reduction in sea surface solar radiation along the flight track. Direct measurements will be made from upward-looking solar radiometers onboard the Vickers, moorings, and the P-3. In addition, sea surface solar fluxes will be computed from scattering codes, constrained with the Learjet and ER-2 measurements at the upper boundary.
Estimates of Ga, Cl, and f for the tropical Pacific between 20º and 20ºS. Satellite AVHRR (cloud imager) radiances from the NOAA polar orbiters, and VISSR (visible and IR spin scanning radiometer; cloud imager) radiances from GMS will be employed for this purpose. Collocated AVHRR and GMS-VISSR radiances will be used to calibrate VISSR with respect to AVHRR. Then, collocated AVHRR and GMS radiances with in-situ broadband fluxes measured by the ER-2 and the Learjet will be used to relate top-of-the-atmosphere narrow-band radiances with fluxes at the tropopause. There are several existing algorithms that relate AVHRR satellite radiances to fluxes at the surface and at the top-of-the-atmosphere. One of these algorithms will be validated with CEPEX measurements; then, AVHRR and GMS radiances will be used to map out the broadband fluxes for the entire tropical belt. Table 10
Heating due to thin cirrus. Optically thin cirrus are prevalent in the tropics, but their effects are unknown. The albedo effect of these thin cirrus are negligible, but their long-wave trapping effect (and potential net heating effect) can be appreciable (Prabhakara et al., 1993). It is not clear if they result from downwind blow-offs from the convective cirrus anvils or if they are formed in situ. If they result from the anvils, then they have to be accounted for as part of a thermostat: the key issue is the effect of these thin clouds outside the warm pool. Although this issue is not central to the validity of a thermostat, it is important for relating the thermostat to tropical mean climate and thus to the global warming issue. CEPEX will make preliminary attempts to explore this issue.
Climatological Generalization of CEPEX Data
The question has arisen, how can data taken during the limited time period of the CEPEX field
phase be extrapolated to other seasons of the annual cycle? TOGA-COARE has adapted a
classification of convective activity encountered in the COARE experimental area (Chen and
Houze, 1993). It measures the size and intensity of convective systems in terms of OLR contours
of cumulonimbus tops and anvils. The 208 K contour has been found to be most representative in
matching the radar echoes from airborne platforms. The five categories range from zero (clear) to
over 60,000 km2 extension (super-cluster). All data taken during CEPEX from the aircraft and the
ship will be categorized according to this scheme and evaluated statistically in terms of "bins." The
extrapolation to other seasons of the annual cycle will be based upon climatological satellite
information on convective systems over the central Pacific and upon the corresponding SST fields
at other times.
In this chapter the needed measurements were described and summarized. Descriptions of the
different measurement instruments, by platform, and the strategy employed for obtaining these
measurements are provided in Chapter 6.
Back to Table of Contents