Atmospheric and Climate Sciences, 2011, 1, 9-17
doi:10.4236/acs.2011.11002 Published Online January 2011 (http://www.SciRP.org/journal/acs)
Assessment of Long Term UV Radiation Measured by the
Brewer Spectrophotometer in Hong Kong during
1995-2005
T. J. Wang1, K. S. Lam2, Q. Liu2, X. M. Wang 2
1School of Atmospheric Science, Nanjing University, Nanjing, China
2Department of Civil and Structural Engineering, the Hong Kong Polytechnic University, Hong Kong, China
3School of Environmental Science and Engineering, Sun Yat-sen Un iversity, Guangzhou 510275, China
Received December 20, 2010; revised January 14, 2011; accepted January 21, 2011.
Abstract
Time series of daily UV radiation measured by the ground-based Brewer spectrophotometer #115 in Hong
Kong during 1995-2005 were studied through statistics analysis, with focus on the variability and long term
changes in relation to total ozone, clouds and AOD (Aerosol Optical Depth). The 11-year mean UV daily
dose is 2644±262 J/m2, with maxima(3311 J/m2) in 2000 and minima (2415 J/m2) in 2002. The data were
compared with that from TOMS (Total Ozone Mapping Spectrometer) Version 8 and show general agree-
ment between the two. However, the Brewer UV measurement is about 10% lower compared to TOMS data.
Apart from the common-known strong seasonal cycle, 26 month periodical was resolved by use of wavelet
analysis, which was believed to be associated with quasi-biennial oscillation (QBO) of general circulation. In
cloudy days, the annual mean UV daily dose decrease 3.5% to 44.5% compared to clear days. It was also
found that surface UV irradiance has close relation to air pollution. Under clear sky condition, 1% AOD in-
crease will lead to 0.2% UV decrease. While global UV radiation increase due to the worldwide observed
ozone depletion, investigations indicate that this trend is not significant in Hong Kong during the last 11
years. The possible causes can be attributed to the compensative effect from two aspects. One is the increase
of UV resulting from the reduction of clouds with rate of 0.56/10 yr. The other is the decrease of UV due to
the enhancement of total ozone and AOD with a rate of 4.23 DU/10 yr and 0.33/10 yr, respectively.
Keywords: UV Radiation, Brewer Spectrophotometer, TOMS, Hong Kong
1. Introduction
The ultraviolet (UV) radiation refers to solar radiation in
the spectral band of 280nm to 400 nm, consisting of
UV-A (315 to 40 0 nm), UV-B (280 to 315 nm) and UV-C
(less than 280 nm), of which UV-B is the most dangerous
portion of UV radiation that can reach the ground level.
Atmospheric ozone shields life at the surface from most
of the UV-B and almost all of the UV-C. Over the past
decades, the observed decline in total ozone column has
led to an increase in surface UV-B does,
which has adverse effect on health and environment
[1]. The ground-based UV observations have been
available in the world to measure the long term variabil-
ity of UV radiation. However, such monitoring sites are
very scarce in Asia.
As one of the important mega-cities in East Asia,
Hong Kong has involved in UV study since 1990s. The
Hong Kong Polytechnic University first set up a YES
broadband UVB-1 pyranometer and 1 spectral Brewer
MKIV Ozone Spectrophotometer (spectral range: 290-325
nm) in 1994 and 1995, respectively to measure UV irra-
diance at the regional air monitoring station at Cape
D’Aguilar. Lam [2] analyzed the UV data measured
during Jan.1995 to Apr.2001 and found that surface UV
radiation exhibits seasonal cycles with maxima in July
and minima in January. In 1999, the Hong Kong Obser-
vatory installed two Yankee Environmental System
broadband UVB-1 pyranometers with a spectral range of
280-320 nm to measure the erythemal UV radiation at
King’s Park meteorological station. [3] reported that the
UV Index often rises above 10 in a summer day from

T. J. WANG ET AL.
10
June to August with variation primarily affected by
clouds rather than total ozone.
However, due to the limited data in several years, the
trend and periodical characteristics of UV radiation in
Hong Kong is not clearly understood and the causes of
the variability need to be further studied. Therefore, in
this paper, UV data of more than 10 years as well as
cloud, total ozone, AOD data are collected and analyzed
to investigate the characteristics of long term trend and
discuss the factors which might have influence on such
variations.
2. Methodology
2.1. Measurement
Since 1995, Brewer spectrophotometer #115 has been
installed to measure the ozone column at Cape D’ Agui-
lar in Hong Kong (22o13’N, 114o15’E, 60m mean sea
level) from the direct sun observations at five isolated
wavelengths in the UV-B: 306.3nm, 310.1nm, 313.5nm,
316.7nm and 320.1nm with a resolution of 0.6 nm. The
narrow band UV-B irradiance is obtained by scanning
forward and backward from the 290 nm to 325 nm at 0.5
increments. More details about the measurement and
calibration can be referred to Lam’s [2]work.
2.2. Data Set
Two sets of UV data were used for this study. One is
daily UV radiation (280-360nm) measured by the
ground-based Brewer spectrophotometer. The other set
of UV data are released by NASA/USA using Version 8
of the TOMS algorithm. This data set includes daily
Erythemal UV dose with wavelength ranging from 280
to 400 nm. The time series of daily AOD and total ozone
observed by the Brewer spectrophotometer are also
adopted to study their relation to UV radiation. The daily
cloud amount during 1995-2005 is provided by Hong
Kong Observatory.
2.3. Statistical Algorithm
Apart from liner regression, the wavelet transform (WT)
was chosen for analysis of the variability of monthly UV
radiation anomaly series in Hong Kong. WT is a power-
ful tool well suited to the study of multi-scale and non-
stationary processes occurring over finite spatial and
temporal domains [4-6]. The uniqueness of WT is the
capability of simultaneously localizing the variability of
a signal in both time and frequency domains. By localiz-
ing in time and frequency of fluctuations in a signal, a
quantitative representation of its spectral makeup as a
function of time can be acquired. Th er ef or e, WT ca n b e a
valuable supplement to other conventional tools, such as
Fourier analysis and empirical orthogonal function anal-
ysis of geophysical signal.
3. Results and Discussions
3.1. Trend and Periodical Characteristics
The daily integral UV dur ing the period of 1995 to 2005
was illustrated in Figure 1(a). The annual mean daily
UV dose is 2644±262 J/m2. The maxima of 3311 J/m2
was found in 2000 and minima of 2415 J/m2 in 2002.
The daily integrated UV irradiance from TOMS version
8 was also shown on Figure 1(a). We compared the
Brewer spectrophotometer UV with that from TOMS.
There is general agreement between the two data sets.
Overall, the Brewer UV measurement is about 10% low-
er compared to TOMS data during 1997-2003.
For the periodical characteristics of UV radiation,
strong seasonal cycle with maximum in Jul. and mini-
mum in Jan. was observed due to variation of solar ze-
nith angle and total ozone. To study th e trend of UV rad-
iation, the annual periodical signal was removed from the
time series. We first calculated monthly UV radiation
from the daily value. Then, UV daily dose in each month
in 11 years was averaged to obtain the monthly mean UV,
which was abstracted from monthly UV series to obtain
the monthly UV anomaly. By use of wavelet analysis,
the period ic of 26 month was reso lved (s ee Figure 1(b)).
This periodic may be related to ozone variation con-
trolled by the quasi-biennial oscillation (QBO) of general
circulation. As expected, the 11 year solar cycle is not
resolved due to the limited data length.
It is interesting that trend is not significant for UV
radiation during the last 11 years, although UV increase
due to ozone depletion was observed at many sites in the
world [7]. In fact, ozone depletion is more significant in
middle-high latitude region, for a subtropical region as
Hong Kong, ozone depletion is not clearly observed.
Actually, a slight enhancement of total ozone column
was found in recent years (see Section 3.2.1), which will
theoretically induce the decline trend of UV dose in
Hong Kong. Therefore, more factors other than total
ozone should be investigated to understand the long term
UV variability observed in Hong Kong.
3.2. Possible Causes for Long term UV Variability
The surface UV irradiance is mainly governed by several
factors: (1)solar zenith angle, (2)sun-earth distance,
(3)altitude, (4) total ozone, (5) cloud, (6) atmospheric tur-
bidity, as exemplified by aerosol optical depth, (7) surface

T. J. WANG ET AL.
11
albedo [8]. The former three factors determine the length
of the path of solar beams through the Earth’s atmosphere.
For a special site, its altitude is fixed. Solar zenith angle
and sun-earth distance have influence on the seasonal and
diurnal cycles of UV radiation. Surface reflection can af-
fect UV radiation both through direction reflections to-
ward a target, and by enhancing the diffuse downwelling
radiation. Since reflectivity of many surfaces is lower in
the UV than at longer wavelengths, they are reported to
have little influence on the UV reaching the ground [9].
Further, from analysis of TOMS version 8 data in
1995-2005, there is not significant change in the surface
reflectivity in Hong Kong. Therefore, for long term trend
analysis, total ozone, cloud and aerosol optical depth are
the most important factors which should be concerned.
The variations of three factors and their relations to UV
irradiance are discussed below.
3.2.1 Total Ozone
Figure 2(a) gives the total ozone measured by Brewer
(a)
(b)

T. J. WANG ET AL.
12
(c)
Figure 1. (a) Time series of UV measured by Brewer during 1995-2005; (b) Monthly UV radiation and anomaly series; (c)
Transform coefficient from wavelet analysis of monthly UV anomaly series.
spectrophotometer in the past 11 years. Strong seasonal
variations were observed as reported elsewhere. The total
ozone measured by satellite (TOMS) were also illu-
strated in Figure 2(a). They are in general agreement
during the period of 1995-2005. The difference between
the two data sets generally ranges from -10 DU to 10 DU.
Variation of UV irradiance with total ozone was illu-
strated in Figure 2(b). It shows that the negative correla-
tion between daily UV dose and total ozone is not sig-
nificant, which can be attributed to the cloud effect. In
Comparison between Brewer#115 (Calibrated) and TOMS V8 Total O3, 1995-2005
150
170
190
210
230
250
270
290
310
330
350
31-Jan-9315-Jun-9428-Oct-9511-Mar-9724-Jul-986-Dec-9919-Apr-011-Sep-0214-Jan-0428-May-05 10-Oct-06
Date
Total Ozone (in Dobson Unit)
Cali. DATA
TOMS
(a)

T. J. WANG ET AL.
13
6000
5000
4000
3000
2000
1000
0
UV daily dose (J/m2)
320300280260240220200
Total ozone
(
DU
)
(b)
y = 0.0358x - 43.544
-35
-25
-15
-5
5
15
25
1995/1/1 1995/7/1 1996/1/1 1996/7/1 1997/1/1 1997/7/1 1998/1/1 1998/7/1 1999/1/1 1999/7/1 2000/1/1 2000/7/1 2001/1/1 2001/7/1 2002/1/1 2002/7/1 2003/1/1 2003/7/1 2004/1/1 2004/7/1 2005/1/1 2005/7/1
Date
Monthly total ozone anormaly(D.U)
(c)
Figure 2. (a) Total ozone measured by Brewer during 1995-2005; (b) Variation of UV irradiance with total ozone; (c) Varia-
tion and trend of monthly total ozone anomaly series.
cloudless sky, the absorption of UV due to ozone is
dominated, leading to UV reduction with total ozone
enhancement. The clear sky UV and total ozone rela-
tionship has been discussed by Lam [2], where, one per-
cent increase in total ozone is associated with one per-
cent decrease in surface UV irradiance in Hong Kong.
Here, we pay much attention to the trend of total ozone
on UV long term variability. To eliminate the influence
of seasonal signal on trend analysis, the time series of
monthly anomaly of total ozone was constructed using

T. J. WANG ET AL.
14
the method similar to UV data(see Figure 2(c)). It shows
that total ozone has a trend of increase during the study
period. Simple linear regression shows that the increas-
ing rate is 4.23 DU/10 yr. The trend of total ozone en-
hancement may result in trend of UV reduction.
3.2.2. Cl o u ds
Clouds can absorb, reflect and scatter portions of incom-
ing radiation, the more prominent effect is the reduction
of UV amount reaching the ground. Since clouds change
greatly in spatial and temporal, they will introduce tre-
mendous variability into surface UV radiation. The vari-
ation of UV daily dose with cloud amount at day time
was shown on Figure 3(a). During the cloudy days
(cloud amount > = 4), UV shows decline with cloud
amount. However, for clear days (cloud amount < 4), the
anti-correlation between UV and cloud amount is not
significant. This suggests that cloud amount is not the
only dominating factor effecting UV. In fact, UV irra-
diance also depends strongly on cloud types. It has been
reported that attenuation of UV radiation by cirrus is
weak and broken clouds can sometimes increase UV
irradiation by reflection from their sides [9]. Due to the
lack of cloud type data, the effect of clouds on UV radia-
tion can not be further investigated. Here, the cloudy and
clear days are cataloged and UV irradiance under these
two conditions was estimated. Here, the clear days were
selected with UV diurnal profiles having a smooth bell
shape and the cloud amount being less than 3. The results
are shown in Table 1, which indicate that the annual
mean UV daily dose in cloudy days decrease 3.5% to
44.5% compared to clear days during 1995-2005. Con-
sidering the total data sets, UV irradiance is negatively
correlated with cloud amount. The monthly cloud
amount anomaly was shown on Figure 3(b), which ex-
hibits a trend of decrease with the rate of 0.56/10 yr. The
trend of cloud amount reduction will lead to the UV in-
crease trend.
3.2.3. Aerosol Opti cal Depth
Aerosol can affect UV radiation through extinction and
scattering process. It was reported that increased air
Table 1. Annual mean UV daily dose under clear and cloudy
conditions.
Year Clear days UV
(J/m2) Cloudy days UV
(J/m2) Reduction rate (%)
1995 3112.838 2515.687 -19.2
1996 3335.21 2398.401 -28.1
1997 3013.153 2241.636 -25.6
1998 4128.601 2292.941 -44.5
1999 2782.365 2684.82 -3.5
2000 4155.097 3115.136 -25.0
2001 3271.7 2624.202 -19.8
2002 2916.931 2353.281 -19.3
2003 3067.99 2482.839 -19.1
2004 2920.975 2635.289 -9.8
2005 3552.64 2536.069 -28.6
6000
5000
4000
3000
2000
1000
0
UV daily dose (J/m
2
)
86420
Cloud amount
(a)

T. J. WANG ET AL.
15
y = -0.0047x + 5.7693
y = -0.0041x + 0.2702
-4
-2
0
2
4
6
8
1611 1621 26 3136 4146 5156 61 66 71 76 81 8691 96101106111116121126131
Date
Monthly average cloud amount in day and abomornly
Cloud amount in day
Anormaly
Linear (Cloud amount in day)
Linear (Anormaly)
(b)
Figure 3. (a) Variation of UV irradiance with cloud; (b) Monthly average cloud amount in day during the day during 1995-2005.
pollution, as exemplified by high aerosol loading, are
capable of reducing the erythemal solar irradiance at the
ground as much as 30-40% [10]. Here, the aerosol opti-
cal depth at 320.0nm is regularly retrieved and recorded
through Brewer Spectrophotometer measurement. The
average optical depth is 0.58 with a standard deviation of
0.29 during 1995-2005. Figure 4(a) is the scattering
plots of UV radiation and AOD. Calculations show that
UV irradiance is negatively correlated with AOD. Al-
though the correlation coefficient 0.124 is not very high,
it reaches 95% confidence level due to large number of
sample.
To investigate the influence of air pollution on UV
dose, clear days (cloud amount < 4) were selected. Then
UV data with solar zenith angle (45 ± 5 degree) and total
ozone (270±20DU) were filtered. The scatting plot of
UV and AOD under clear sky was shown on Figure 4(b).
Calculations suggest that 1% AOD increase will lead
6000
5000
4000
3000
2000
1000
0
UV daily dose(J/m2)
3.02.52.01.51.00.50.0
Aerosol o
p
tical de
p
th
(a)

T. J. WANG ET AL.
16
2500
2000
1500
1000
UV daily dose under clear sky (J/m
2
)
0.80.70.60. 50.40.30. 2
Aerosl O
p
tical De
p
th
(
AOD
,
320nm
)
UV=2413.2 - 887.3AOD
(b)
2.5
2.0
1.5
1.0
0.5
Aerosol Optical Depth(AOD)
01/01/1996 01/01/1998 01/01/2000 01/01/2002 01/01/2004
Date
(c)
Figure 4. (a) Variation of UV irradiance with AOD; (b) Variation of UV irradiance with AOD under clear conditions; (c)
AOD measured by Brewer during 1995-2005.
to 0.2% UV decrease. Figure 4(c) gives the increasing
rate of AOD during 1995-2005. The significant increas-
ing rate of 0.33/10 yr was exhibited from the AOD series,
which will lead to the decreasing trend of UV.
As discussed above, measurements have shown that
both total ozone and AOD exhibit trend of enhancement
during 1995- 2005. On the o ther hand, reduction of cloud
amount has also been observed. It is believed that in-
crease of total ozone as well as AOD and decrease of
cloud amount will have compensative effect on UV radi-
ation, which may explain the insignificant trend of UV
observed in Hong Kong.

T. J. WANG ET AL.
17
4. Conclusions
Ground-based spectral ultraviolet radiation measurement
has been carried out by Brewer spectrophotometer #115
at Hong Kong since January 1995 . The annual mean UV
daily dose is 2644±262 J/m2. The maxima of 3311 J/m2
and minima of 2415 J/m2 were measured in 2000 and
2002, respectively. In general, the observed UV was 10%
lower than that from TOMS Version 8. The 26 month
periodic, related to the quasi-biennial oscillation of gen-
eral circulation, was resolved from the monthly UV
anomaly series by use of wavelet analysis. Compared to
the clear days, the annual mean UV daily dose decrease
3.5% to 44.5% under the cloudy weather conditions. It
was found that 1% AOD increase will lead to 0.2% UV
decrease in clear sky. Investigations show that UV radia-
tion did not exhibit significant trend during the period of
1995-2005. However, the changing rates of -0.56/10 yr,
4.23 DU/10 yr and 0.33/10 yr were observed for cloud
amount, total ozone and AOD, respectively. It is believed
that the compensative effect between the increase of UV
resulting from the reduction of clouds and the decrease
of UV due to the enhancement of total ozone and AOD
plays important roles in the long term UV variability in
Hong Kong.
5. Acknowledgement
This project was supported by the National Key Basic
Research Development Program of China (2010CB
428503, 2011CB403406) and the Research Grant Coun-
cil of Hong Kong. Thanks to Hong Kong Observatory
for providing cloud data. We also thank United States,
NASA, Goddard Space Flight Center for making the
TOMS data publicly available.
6. References
[1] H. Slaper, M. Velders, J. S. Daniel, F. R. Degruijl and J.
C. Vanderleun, “Estimates of ozone depletion and skin
cancer incidence to examine the Vienna convention
achievement,” Nature, Vol. 384, No. 6606, 1996, pp.
256-258.
[2] K. S. Lam, A. J. Ding, L. Y. Chan, T. Wang, and T. J.
Wang, “Ground Based Measurements of Total Ozone and
UV Radiation by the Brewer Spectrophotomer #115 at
Hong Kong,” Atmospheric Environment, Vol. 36, 2002,
pp. 2003-2012.
[3] Y. W. Chan, “A study of the characteristics of ultraviolet
radiation in Hong Kong,” Radiation Protection Dosime-
try, Vol. 91, No. 1-3, 2000, pp. 173-176.
[4] C. K. Chui, “An introduction to wavelets,” Academic
Press, 1992, p. 266.
[5] J. J. Benedetto and M. W. Frazier, “Wavelets: mathemat-
ics and applications,” CRC Press, 1994, p. 575.
[6] G. Kaiser, “A friendly guide to wavelets,” Birkhaser,
1994, p. 300.
[7] WMO, “Scientific assessment of ozone depletion,” Glob-
al Ozo ne Res. Monit. Project, Geneva, 2002 p. 47498.
[8] W. R. Burrows, M. Vallee and J. B. Wardle, “The Cana-
dian Operational Procedure for Forecasting Total Ozone
and UV Radiation, Meteorological Applications,” Vol.1,
pp. 247-265.
[9] R. E. Hester and R. M. Harrison, “Cause and environ-
mental implications of increased UV-B radiation,” Pub-
lished by the Royal Society o Chemistry, Thomas Gra-
ham House. 20-22, 2000.
[10] G. G. Palancar and B. M. Toselli, “Effects of metrology
and tropospheric aerosols on UV-B radiation: a 4-year
study, ” Atmospheric Environment, Vol. 38, 2004, pp.
2749-2757.
[11] K. Buttner, “Physik. Bioklimat, quoted in M. L. Nack &
Green, A. E. S., 1974. Influence of clouds, haze, and
smog on the middle ultraviolet reaching the ground,”
Appl. Optics, Vol. 13, No. 10, 1938, pp. 2405-2415.
[12] F. Carvalho and D. Henriques, “Use of Brewer Ozone
Spectrophotometer for Aerosol Optical Depth Measure-
ments on Ultraviolet Region,” Adv. Space Res, Vol. 25,
No. 5, 2000, pp. 997-1006.
[13] den Outer, P. N. Slaper and R. B. Tax, “UV radiation in
the Netherlands: Assessing long-term variability and
trends in relation to ozone and clouds,” Journal of Geo-
physical Research, Vol. 110, No. D02203, 2005.
doi:10.1029/2004JD004824,2005
[14] J. Grobner and C. Meleti, “Aerosol optical depth in the
UVB and visible wavelength range from Brewer spec-
trophotometer direct irradiance measurements:
1991-2002,” Journal of Geophysical Research, Vol. 109,
No. D09202, 2004. doi:10.1029/2003JD004409,2004
[15] F. Marenco, A. Di Sarra and D. J. Luisi, “Methodology
for determining aerosol optical depth from Brewer
300-320 nm ozone measurements,” Applied Optics, Vol.
41, No. 9, pp. 1805-1814, 2002.