Skip to main content

Full text of "Clear-Sky Surface Solar Radiation During South China Sea Monsoon Experiment"

See other formats


Clear-Sky Surface Solar Radiation During South China Sea Monsoon Experiment 



Po-Hsiung Lin', Ming-Dah Chou^, Qiang Ji^, Si-Chee Tsay^ 



Department of Atmospheric Sciences, National Taiwan University, Taipei, Taiwan. 
TJASA/Goddard Space Flight Center, Greenbelt, Maryland. 
Science Systems and Applications, Inc., Lanham, Maryland. 



Xo ^^-^ April, 2000 

Submitted to the Geophysical Research Letter 



Abstract: 

Downward solar fluxes measured at Dungsha coral island (20°42'N, 116°43'E) 
during the South China Sea Monsoon Experiment (May- June 1998) have been calibrated 
and compared with radiative transfer calculations for three clear-sky days. Model 
calculations use water vapor and temperature profiles from radiosound measurements and 
the aerosol optical thickness derived from sunphotometric radiance measurements at the 
surface. Results show that the difference between observed and model-calculated 
downward fluxes is <3% of the daily mean. Averaged over the three clear days, the 
difference reduces to 1%. The downward surface solar flux averaged over the three days 
is 314 Wm'^ from observations and 317 Wm'^from model calculations. This result is 
consistent with a previous study using TOGA CAORE measurements, which found good 
agreements between observations and model calculations. This study provides an extra 
piece of useful information on the modeling of radiative transfer, which fills in the puzzle 
of the absorption of solar radiation in the atmosphere. 



1. Introduction 

The validity of radiation model calculations of atmospheric solar (shortwave, or 
SW) heating has long been an unsettled issue. Traditionally, this issue concerns primarily 
the excess atmospheric heating due to the presence of clouds that is not accounted for in 
radiation model calculations [Cess etal., 1995; Pilewskie and Valero, 1995; Ramanathan 
et ah, 1995]. Other studies have shown that there is no clear evidence of the enhanced 
solar heating of the atmosphere due to clouds [Imre et al, 1996; Li et ai, 1997]. 
Uncertainties in SW heating of both clear and cloudy skies could contribute to the 
uncertainty in the estimation of the cloud effect on atmospheric SW heating (or cloud 
radiative forcing, CRF). In a study of the global radiation data sets derived from surface 
measurements, satellite retrievals, and climate model simulations, Arking (1996) 
suggested that clouds had little effect on the solar heating of the atmosphere. Rather, the 
large atmospheric CRF of model calculations was caused by the underestimation of water 
vapor heating in clear atmospheres. Subsequently, there were a number of studies on the 
clear-sky solar heating of the atmosphere and the surface, which used various types of 
measurements (total, spectral, direct, diffuse radiation) at various geographical locations. 
Some studies have suggested that, for different reasons, radiation models highly 
underestimate the clear-sky atmospheric heating and, hence, overestimate the surface 
heating [Kato et al, 1997; Halthore, et al, 1998; Arking, 1999; Pilewskie et al, 2000]. 
Other studies have found agreement between observations and model calculations [Chou 
and Zhao, 1997; Conant et al, 1998; Fu et al, 1998; Mlawer et al, 2000]. Resolving 
this issue is very important because our ability to model the absorption of solar radiation 
affects the reliability of climate model simulations and remote sensing of a wide range of 



geophysical parameters. 

In May and June 1 998, there was an intensive field experiment, South China Sea 
Monsoon Experiment, conducted in the South China Sea. The SCSMEX is an 
international field experiment to study physical processes and evolutions of the water and 
energy cycles of the East Asian monsoon system [Lau et al, 2000]. There was a suite of 
instruments set up at Dungsha (20°42'N, 116°43'E) measuring surface radiation and 
atmospheric temperature, humidity, and aerosols. Dungsha is a small coral island with a 
length of ~1 km and a width of -0.7 km. We use the data measured at Dungsha to study 
the surface SW radiation and compare the observations vdth radiation model calculations. 

2. Surface Measurements and Radiative Transfer Model 

Propagation of the East Asian summer monsoon (EASM) system during the spring- 
summer transition period influences the annual rainfall variation in the South China Sea. 
To understand the role of the EASM in the global energy and water cycle and to improve 
the simulation and prediction of East Asian monsoon and regional water resources, 
observations were conducted during two SCSMEX intensive observing periods (lOP). 
The first lOP was conducted in 5-25 May 1998 to observe atmospheric and oceanic 
circulation before the monsoon passing through the South China Sea. The second lOP 
was conducted in 5-25 June 1998 to observe tropical weather under the influence of 
EASM. 

In addition to the weather stations operated by the Taiwanese navy, other advanced 
facilities were also operated on Dungsha during the SCSMEX lOP. These facilities 
included Australian C-band polarization radar system, remote pilot vehicle "Aerosonde", 
National Center for Atmospheric Research (NCAR) Integrated Sounding System (ISS), 



and a NASA radiation measurement system. Radiative fluxes were measured in the 
period from 17 April through 6 July 1998. Three Epply Precision Spectral Pyranometers 
(PSP) and one Yankee Total Spectral Pyranometer were used to measure surface 
downward SW fluxes. The Epply pyranometers measured fluxes in the ultraviolet (0.3- 
0.4 |J.m), photosynthetically active radiation (0.4-0.7 |xm), and infrared (0.7-2.8 )im) 
spectral bands. Two Epply Precision Infrared Radiometers (PIR) were used to measure 
the downward longwave fluxes. A CIMEL 318-1 sunphotometer and a Yankee six-band 
Multi-Filter Radiometer were used to measure direct- and sky-radiation. A single data- 
acquiring system processed and stored all radiative flux measurements with a one-minute 
sampling rate. 

Except the CIMEL component, all of the radiation measurement facilities and the 
data acquiring system were new products with functions checked by the manufacturer in 
February 1998. After the SCSMEX campaign, all instruments were brought back to 
NASA/Goddard Space Flight Center for re-calibration. The pyranometer current 
equivalent to zero solar radiation was obtained by applying the dark-current checking 
procedure [Ji and Tsay, 2000]. The methodology involved the use of aluminum-made 
caps to cover the outer glass dome of the pyranometers during daytime operation. The 
overall uncertainty of the radiation measurements including data-logger performance is 
estimated to be 3%. 

The aerosol optical thickness (AOT) was retrieved from the radiances measured by 
the Cimel Electronique CE318-1 automatic sun-tracking photometer. This instrument had 
seven filters centered at 340, 380, 440, 500, 675, 870, and 1020 nm. Two collimators 
with 1.2 degrees were used to measure direct- and sky-radiances every 15 min. The 



measured-radiances were sent to the NASA Aerosol Network office [Holben, et al, 
1998] for AOT retrieval. The uncertainty of AOT under a clear-sky situation was 
estimated to be <0.01 for wavelengths >440 nm and <0.02 for shorter wavelengths. 

Integrated Sounding System (ISS) GPS-based balloon sounding was launched twice 
a day at 0600 UTC and 1800 UTC. Vaisala RS80-15G radiosonde was used in this 
balloon sounding system to measure the atmospheric temperature and humidity profiles. 
The temperature sensor, THERMOCAP, has a 0.2% accuracy up to 50 hPa, and the 
humidity sensor, HUMICAP, has a 3% accuracy. It is found that the measured 
precipitable water agrees well with that retrieved from the Special Sensor Microwave 
Imager [Wentz, 1994] and the CIMEL radiance measurements at 940 rmi. 

We use the solar radiative transfer model (CLIRAD-SW) developed at the NASA 
Goddard Climate and Radiation Branch [Chou and Suarez, 1999] to compute the 
downward surface SW flux at Dungsha. The model has been applied to various 
atmospheric models used in the Goddarftd Laboratory for Atmospheres, including a 
general circulation model, a mesoscale model, and a cloud ensemble model. It includes 
the absorption due to water vapor, O3, O2, CO2, clouds, and aerosols. Interactions among 
the absorption and scattering by clouds, aerosols, molecules (Rayleigh scattering), and 
the surface are fully taken into account. Fluxes are integrated virtually over the entire 
spectrum, from 0.175 |j.m to 10 |im. Integrated over all spectral bands and all absorbers, 
the surface heating is computed accurately to within a few watts per meter squared of 
high spectral-resolution calculations. 

The pyranometer measurements of SW flux at Dungsha did not include radiation in 
the spectral region 2.8-10 |im. For comparisons between measured and computed surface 



fluxes, radiation in this spectral region has to be taken into consideration. Line-by-line 
calculations show that the range of the surface flux in this spectral region is small. It 
ranges only from 10.8 W m'^ to 12.8 W m'^ for the column water vapor amount ranging 
from 2.8 cm to 5.6 cm when the sun is overhead. Therefore, we fit the surface flux 
computed for a coltmin water vapor amount of 3 cm as a function of the solar zenith 
angle. The total flux computed using the SW radiation model is then reduced by an 
amount derived from this function to remove the radiation contained in the spectral 
region 2.8-10 |J,m. 

3. Comparisons of measured and calculated surface SW fluxes 

In the tropical western Pacific and the South China Sea, clouds are v^ddespread, and 
it is difficult to identify those radiation measurements which are free of cloud influence. 
In studying the surface radiation in the Pacific warm pool during the Tropical Ocean and 
Global Atmosphere, Coupled Ocean-Atmosphere Response Experiment (TOGA 
COARE), Chou and Zhao [1997] used both the direct and diffuse components of the SW 
radiation to identify clear-sky surface fluxes. It is based on the facts that in a cloud-free 
atmosphere the direct radiation is large, the diffuse radiation is small, and the diurnal 
variation of the total surface radiation is in accordance with the incoming radiation at the 
top of the atmosphere. During the SCSMEX lOP, we measured only the total flux but 
not separately the direct and diffuse components of the radiation. We examined the 
diurnal variation of the total surface radiation and subjectively identified three days (2 
May, 22 May, and 29 June 1998) as being mostly clear. For these three clear days, the 
total surface radiation (dashed curves in Figure 1) is high and varies smoothly with time, 
following the radiation at the top of the atmosphere. 



When clouds block the sun, the surface radiation is greatly reduced. When clouds 
do not block the sun but scattered over the observation site, the surface radiation is 
greater than that of clear skies. Thus, clouds could either increase or decrease the surface 
radiation depending upon the relative locations of clouds, the sun, and the surface site. 
These situations can be clearly seen in Figure 1 (dashed curves). To estimate the clear- 
sky surface downward SW radiation, F , of those mostly clear days, we make the 
following adjustments to the measured surface radiation. First, we delete those data 
which are obviously affected by clouds. For example, the data in the early morning and 
late afternoon on 2 May (Figure la). Second, the remaining data are fit by a third-order 
polynomial function of the solar zenith angle, \io, separately for morning and afternoon 
data. It is found that F^ varies rather linearly with ^o, and the third-order polynomial 
ftmction fits well the surface radiation. Third, we further delete those data that deviate 
from the regression curves by >15 W m'^. The remaining data are further fit by a third- 
order polynomial function of f^o, again separately for morning and afternoon data. 
Finally, we replace the data deleted in the first and third steps by that computed from the 
regression curves. The diurnal variations of the reconstructed F^ are shown by the solid 
curves in Figures la-c. 

As can be seen in Figures la-c, the atmosphere on 29 June is the clearest among the 
three days. Unfortunately, there were no balloon sounding and sunphotometer 
measurements, and humidity and aerosol information are not available on that day. After 
the first surge of the monsoon passing through Dungsha on 10 June 1998, there was 
nearly no rain on the island, and the standard deviation of the 1 2-hourly column water 
vapor amount was only 0.4 cm. Therefore, we use the water-vapor soundings on 22 June 



as a surrogate for 29 June. Figure 2 shows the measured downward surface SW fluxes on 
29 June and 30 June. Although the cloud effect on the surface radiation is large on 30 
June, the two curves overlap very well when there is no cloud interference. It indicates 
that the AOT is similar on both days. Therefore, we use the average AOT inferred for the 
morning on 30 June to compute the surface radiation on 29 June. 

The AOT measurements on 2 May, 22 May, and 30 June are shown in Figures 3a-c, 
respectively. The AOT pattern on June 30 (Figure 3c) depends weakly on wavelength, 
which is quite different from the other two clear days in May (Figures 3a and 3b). The 
weather before the development of South China Sea summer monsoon was dominated by 
a quasi-stationary frontal system along the coastline of China. One can expect that fine- 
sized pollutants transported southerly from China to the South China Sea. When the first 
transition period of the South China Sea monsoon developed in late May 1998, the 
weather system was dominated by the prevailing southwesterly wind. It is expected that 
most of aerosol particles were sea salt but not the anthropogenic sulfuric aerosols from 
China. The weak dependence of AOT on wavelength shown in Figure 3 c is related to the 
large sea salt particles. Furthermore, the AOT of the maritime aerosols is much smaller 
than that of the continental aerosols (Figures 3a and 3b). 

In computing fluxes, we divide the atmosphere into 75 vertical layers. The 
thickness of a layer in the troposphere is ~25 hPa. The radiosoundings of atmospheric 
temperature and humidity taken at 0600 UTC are used to represent the daytime 
conditions. Fluxes are computed at 1-min resolution. Information on the vertical 
distributions of aerosols and column-integrated ozone amount are not available for flux 
calculations. Therefore, we assume that aerosol optical thickness derived from the 



CIMEL sunphotometric measurement has a uniform vertical distribution below the 800- 
hPa level. Sensitivity tests show that the results are not sensitive to the assumed thickness 
of the aerosol layer. For an aerosol optical thickness < 0.3 at the visible spectral region, 
the daily-mean downward surface flux changes by only <0.5 W m'^ when the top of the 
aerosol layer is extended from the 800 hPa to 600 hPa. The surface SW fluxes are also 
not sensitive to the ozone amount. For a change of the ozone amount from 0.30 (cm- 
atm)stp to 0.35 (cm-atm)stp, the daily-mean surface SW flux reduces by <0.5 W m'l 
Therefore, we use an ozone profile typical of a midlatitude summer atmosphere, which 
has an column amount of 0.32 (cm-atm)stp, in all calculations. 

Figure 4 shows diurnal variations of the incoming solar flux at the top of the 
atmosphere (upper curves), the reconstructed (dashed curves) and the model-calculated 
(solid curves) F'''. Diurnal distributions of the difference between the reconstructed and 
the calculated F^ are shown in Figure 5. The large difference in the early morning and 
the late afternoon on 2 May is due to the extrapolation of the clear-sky flux to these hours 
when the sky was cloudy, as indicated in the measured surface flux shown in Figure la 
(dashed curve). The relatively large bias of the model-calculated F^ on 29 June is due 
primarily to the lack of direct information on water vapor and aerosols. Table 1 
summarizes the water vapor amount and the aerosol optical thickness used in the 
radiation model calculations, as well as the incoming SW flux at the top of the 
atmosphere, surface measurements, and model calculations. The daily-mean difference 
between the reconstructed and the model-calculated F''' is 1.2, 0.9, and 7.1 Wm"^ for 2 
May, 22 May, and 29 June, respectively. 

4. Conclusion 



During the SCSMEX Intensive Observing Period, which covers 50 days in May and 
June 1998, only three days are found to be clear with minimal cloudiness at Dungsha. 
Diurnal cycles of the clear-sky surface downward SW flux, F^, are reconstructed by 
removing the effect of clouds based on the near-linear relationship between F^ and the 
cosine of the solar zenith angle. The reconstructed F^ on the three clear days are 
compared with radiative transfer model calculations. The input data to the model 
calculations include the temperature and humidity profiles from radiosoundings and the 
aerosol optical thickness inferred fi-om sunphotometric radiance measurements. The 
difference between the measured and the model-calculated F"" is <3% of the daily means, 
which is comparable to the estimated uncertainty of the surface measurements. The result 
is consistent with a previous study using TOGA CAORE measurements, which found 
good agreement between observations and model calculations. Averaged over the three 
clear days, F*" is 3 14 Wm'^ from observations and 317 Wm"^ from model calculations. 

Previous studies on F"*" by other investigators gave mixed results. Some showed 
good agreement between model-calculated and measured F^. Others showed significant 
disagreement. Those studies covered different geographic locations in the tropical 
western Pacific and the ARM sites. Whether radiative transfer models overestimate the 
surface radiation, or equivalently underestimate the atmospheric absorption, is likely to 
remain an issue for some time to come. The results of this study provide an extra piece of 
useful information on the modeling of radiative transfer in a clear-sky atmosphere, which 
fills in the puzzle of the absorption of solar radiation in the atmosphere. 



Acknowledgment. The work conducted at the National Taiwan University was 
supported by National Science Council in Taiwan. The work conducted at NASA 
Goddard Space Flight Center was supported by the Radiation Processes Program, NASA 
Office of Earth Science. The NASA Aeronet group processed the Cimel data. Prof Pay- 
Liam Lin of National Central University, Taiwan, provided the ISS data set, and the 
Dungsha Weather Station, Naval Meteorological Center, Taiwan, provided logistic 
support. 



10 



References 

Arking, A., Absorption of solar energy in the atmosphere: discrepancy between model 

and observations. Science, 273, 779-792, 1996. 
Arking, A., The influence of clouds and water vapor on atmospheric absorption, Geophys. 

Res. Lett. , 26, 2729-2732, 1 999. 
Cess, R. D., M. H. Zhang, P. Minnis, L. Corsetti, E. G. Dutton, B. W. Forgan, D. P. 

Garber, W. L. Gates, J. J. Hack, E. F. Harrison, X. Jing, J. T. Kiehl, C. N. Long, J.- 

J. Morcrette, G. L. Potter, V. Ramanathan, B. Subasilar, C. H. Whitlock, D. F. 

Young and Y. Zhou, Absorption of solar radiation by clouds: Observations versus 

models. Science, 267, 496-499, 1995. 
Chou, M. D., and W. Zhao, Estimation and model validation of surface shortwave 

radiation and cloud radiative forcing using TOGA COARE measurements, J. 

Climate, 10,611-620,1997. 
Chou, M. D., and M. J. Suraez, A shortwave radiation Parameterization for atmospheric 

studies. Volume 15, Technical Report Series on Global Modeling and Data 

Assimilation,!^ AS ArYU-\999-\0A6Q6. pp40, 1999. 
Conant, W. C, A. M. Vogelmann and V. Ramanathan, The unexplained solar absorption 

and atmospheric H2O: a direct test using clear-sky data, Tellus, 50A, 525-533, 1998. 
Fu, Q., G. Lesins, J. Higgins, T. Charlock, P. Chylek and J. Michalsky, Broadband water 

vapor absorption of solar radiation tested using ARM data. Geophy. Res. Letter, 25, 

1169-1172,1998. 



11 



Halthore, R. N., S. Nemesure, S. E. Schwartz, D. G. Imre, A. Berk, E. G. Dutton and M. 
H. Bergin, Models overestimate diffuse clear-sky surface irradiance: A case for 
excess atmospheric absortpion. Geophy. Res. Let., 25, 3591-3594, 1998. 
Holben, N. L., T. F. Eck, I. Slutker, D. Tanre, J. P. Buis, A. Setzer, E. Vermote, J. A. 
Reagan, Y. J. Kaukman, T. Nakajima, F. Lavenu, I. Jankowiak and A. Smimov, 
AERONET - a federated instrument network and data archive for aerosol 
characterization. Remote Sens. Enviorn., 66, 1-16, 1998. 
Imre, D. G., E. H. Abramson, and P. H. Daum, Quantifying cloud-induced shortwave 
absorption: An examination of uncertainties and of recent arguments for large excess 
absorption, J. Appl Meteorol, 35, 1991-2010, 1996. 
Ji, Q. and S. C. Tsay, On the dome effect of Epply Pyrgeometers and Epply 

Pyranometers, J. Geophys. Res., 2000. 
Kato, S., T. P. Ackerman, E. E. Clothiaux, J. H. Mather, G. G. Mace, M. L. Wesely, F. 
Murcray and J. Michalsky, Uncertainties in modeled and measured clear-sky 
surface shortwave irradiances. J Geophy. Res., 27, 25881-25898, 1997. 
Lau, K. M., Y. Ding, J. T. Wang, R. Johnson, T. Keenan, R. Cifelli, J. Gerlach, O. Thiele, 
T. Rickenbach, S. C. Tsay and P. H. Lin, A report of the field operations and early 
results of the South China Sea Monsoon Experiment (SCSMEX). Bull Amer. 
Meteor. Soc. , xxxx-xxxx, 2000. 
Li, Z., L. Moreau and A. Arking, On solar energy disposition: a perspective from 
observation and modeling. Bull. Amer. Meteor. Soc, 78, 53-70, 1997. 



12 



Mlawer, E. J., P. D. Brown and S. A. Clough, Comparison of spectral direct and diffuse 

solar irradiance measurements and calculations for cloud-free conditions, Geophy. 

Res. Letter, 2000. submitted. 
Pilewskie P., F. P. J. Valero, Direct observations of excess solar absorption by clouds, 

Science, 267, 1626-1629, 1995. 
Pilewskie, P., M. Rabbette, R. Bergstrom, J. Marquez, B. Schmid, and P. B. Russell, The 

discrepancy between measured and modeled downwelling solar irradiance at the 

ground: Dependence on water vapor. Geophy. Res. Lett, 27, 137-140, 2000. 
Ramanathan, V., B. Subasilar, G. J. Zhang, W. Conant, R. D. Cess, J. T. Kiehl, H. Grassl, 

and L. Shi, Warm pool heat budget and shortwave cloud forcing: A missing 

physics? Science, 267, 499-503, 1995. 
Wentz, F. J., User's Manual SSM/I -2 Geophysical Tapes. Tech. Rep. 070194, 20 pp, 

1994. [Available from Remote Sensing Systems, Santa Rosa, CA.] 



13 



Table 1. Daily mean values of the column water vapor (w), aerosol optical thickness in 
the UV (Tuv), visible (Xv), and infrared (tir) spectral regions, insolation at the top of the 
atmosphere (Stoa), reconstructed surface downward SW flux (F^r), niean atmospheric 
transmittance (7), and model-calculated surface downward SW flux (F^m). 



May 2 May 22 June 29 



w(cm) 5.6 5.9 5.6 

Tuv 0.45 0.42 0.09 

Tv 0.27 0.27 0.06 

Xir 0.10 0.13 0.04 

S,oa(Wm-^) 450.6 457.5 458.7 

F\(Wm-^) 308.1 312.6 319.7 

r(%) 68.4 68.3 69.7 

F^,n(Wm-^) 309.3 313.5 326.8 



14 



Figure Captions: 

Figure 1: Diurnal variations of the measured surface downward SW flux (dashed curve), 

and the reconstructed clear-sky downward SW flux (solid curve). 
Figure 2: Diurnal variations of the surface downward SW flux measured on 29 June 

(solid curve) and 30 June (dashed curve) 1998 at Dungsha. 
Figure 3: Diurnal variations of the aerosol optical thickness at 340, 500, and 1020 nm 

measured (retrieved) at Dungsha on 2 May (a), 22 May (b), and 29 June (c) 1998. 
Figure 4; Diurnal variations of the insolation at the top of atmosphere (circle-dashed 

curve), the model-calculated (solid curve) and the reconstructed (dashed ciirves) 

surface downward SW fluxes. 
Figure 5: Model-calculated surface downward flux minus reconstructed surface 

downward flux. 



15 



(a) 



^ 1200 



8 qL 



20:00 




5:00 



10:00 15:00 

Local Time 



20:00 



Rj. I 



CM 

E 


1 ' 1 ' 1 ■' ] . ^ , , , 




^ 


■ 29 June 




30 June 




S 1000 







u. 


^,P^ '.'w^hriVv^ 




^ 


y^ r ' • ^''f|V 


w 


^^^ ' '^^ 


1 500 


- /^ %s. 




5 


/• * \ 




1 


yC'^' 'A. 




Q 


y^ ^v 




O 


ar \^ 







— '^ ' ' ' ' 1 . 1 . , V 


1 



5:00 7:30 



10:00 12:30 15:00 17:30 20:00 
Local Time 



^ 



,2- 



« 
w 
<a 

c 

o 



03 
O 

Q. 
O 

o 

(0 

o 



0) 

< 



1.0 



0.8 



(a) 



Ts 0.6 



0.4 



2 0.2 



0.0 



• AOT_1020 
+ AOT_500 
c AOT 340 



Q ^GanoOGC°a^c^cio, 



°ca°':^° 



ao^ 



oaaQD 



cx:g 
a 



.:>.. * — — * 



+ + 
+ 



8:00 



12:00 



16:00 



1.0 

<0 
<S 

5 0.8 
i 0.6 

CO 

o 

O 0.4 

o 
w 

2 0.2 
< 

0.0 



(b) 



— i 1 1- 

• AOT_1020 
+ AOT_500 
a AOT_340 



r-C 



++ 



-^++^++++++ ++++++++++ + 



+ +-^+ 



8:00 



12:00 



16:00 



in 
<D 

c 

o 



1.0 



0.8 



(c) 



H 0.6 



O 



5- 0-4 
O 



o 
w 

2 
< 



0.2 



0.0 



8:00 



• AOT_1020 
+ AOT_500 

□ AOT 340 



°t + 
it a 



..^^^$°°t t?° ' g ,9??°' + rr'tr*^ 



12:00 
Local Time 



16:00 



Up 



_ 1500 

'E 



3 1000 



"2 500 
CO 

5 

c 

o 

° 



(a) 



1 1 T" 






r" ■ 


-^ 


e 


- 


e- 


^ 


T 


1 ' '. ' ' 
2 May 


• 






;a 












s. 




• 




/ 
















\ 




;Z! 






^ 


r="- 


- 


^^>: 


^ 




5^ 


/ 




/< 


/'• 












'V 


N. ^ 


c/ 


/^ 


















X \ 






















\°N 


/ // 




















\ \ 


: iX 






















^^ . 


, 






, 


, 




, 




1 





5:00 



10:00 



15:00 



20:00 



1500 



X 1000 
iZ 

■S 500 

i 

o 



(b) 







1' ' ' ' r — ^ 

X 


1 1 \ — 

22 May 


~ 


■ ^ 

/ 


.''^'""""^''^ 




■ 






Si 


— 


/ y 


\ 


\ 




/ 






■ ' 


, , 


!.,,,!. 


N^ . 



1500 



§ 1000 



en 

^ 500 

i 

I 

° 



5:00 
(c) 



10:00 



15:00 



20:00 




29 June 



5:00 



10:00 15:00 

Local Time 



20:00 



Ffl f 



E 

o 

c 

0) 

© 



X 

u. 




5:00 



29 June 
22 May 
2 May 



10:00 15:00 

Local Time 



20:00 



f^,r