the new sun sky lunar cimel ce318 t. a step forward in ... · the new sun‐sky‐lunar cimel...
TRANSCRIPT
The new sun‐sky‐lunar Cimel CE318‐T. A step forward in monitoring aerosols and water vapour.
A. BARRETO 1,2, E. CUEVAS1, M. J. GRANADOS‐MUÑOZ3, L. ALADOS‐ARBOLEDAS3, P. M. ROMERO1, J. GRÖBNER4, N. KOUREMETI4, A. F. ALMANSA1,2, T. STONE5, C. TOLEDANO6,
R. ROMÁN6, M. SOROKIN7, B. HOLBEN7 and M. CANINI2 1Izaña Atmospheric Research Center (IARC), Agencia Estatal de Meteorología (AEMET), 38001, Santa
Cruz de Tenerife, Spain. 2Cimel Electronique, 75011, Paris, France.
3Department of Applied Physics, University of Granada, 18001, Granada, Spain. 4Physikalisch‐Meteorologisches Observatorium Davos and World Radiation Center (PMOD/WRC), 7260,
Davos, Switzerland. 5US Geological Survey, 86001, Flagstaff, USA.
6Grupo de Óptica Atmosférica, Universidad de Valladolid, 47002, Valladolid, Spain. 7NASA/Goddard Space Flight Center, 20771, Greenbelt, USA.
ABSTRACT We introduce in this study the new multiband photometer CE318‐T, which permits a complete cycle of diurnal atmospheric aerosol and water vapor monitoring. This new instrument allows us to extend the current photometric information, restricted to light period, to night‐time using both the Sun and the Moon as light sources. That means a substantial improvement in aerosol and water vapour monitoring compared to that provided by the previous AERONET (Aerosol RObotic NETwork) standard CE318‐N photometer. Taking the advantage of the new CE318‐T features, we propose four different calibration techniques nocturnal calibration: the Lunar‐Langley method for absolute calibration, the Ratio Moon and Ratio Sun methods for calibration transference (using night‐time and daytime master vs. secondary coincident measurements, respectively), and the Sun‐Moon gain factor method. The latter method requires a reference instrument daytime calibrated and an integrating sphere (not necessarily calibrated), and it is not dependent on any lunar irradiance model. The two last techniques are new and significantly reduce the previous calibration complexities inherent to nocturnal calibration. A comprehensive performance evaluation is carried out in this study, including a quantitative estimation of Aerosol Optical Depth (AOD) uncertainty by means of error propagation theory, a daytime AOD evaluation at Izaña by comparing CE318‐T measurements with collocated information from independent reference instruments, including the World Radiation Centre (WRC) Precision Filter Radiometer (PFR) and the new WRC‐PSR (Precision Spectroradiometer), a nocturnal AOD evaluation at Granada using a collocated star photometer, and a nocturnal AOD evaluation performed at Izaña by means of a day/night coherence transition test using the master CE318‐T and daytime data from the CE318‐N AERONET master. Our results show the ability of the new CE318‐T to capture the diurnal cycle of aerosols as well as short‐term atmospheric variations, critical for climate studies.
1. INTRODUCTION
In the literature remarked efforts have been made to study the diurnal dynamics and
evolution of aerosols in the atmosphere, with the aim of filling the current gaps in the
aerosol optical depth (AOD) time series at night‐time. In this regard, star photometry
and moon photometry have appeared as feasible solutions for aerosol nocturnal
monitoring. The most outstanding works have been developed by Leitener et al.
(1995), Pérez‐Ramírez et al. (2015) and Baibakov et al. (2015) in star photometry, and
by Berkoff et al. (2011) and Barreto et al. (2013a,b) in moon photometry.
This work is based in the recent results reported by Barreto et al. (2016), who have
presented the new sun‐sky‐lunar photometer CE318‐T. The ability of the CE318‐T
device to perform day and night measurements, in addition to the new improvements
introduced in the CE318‐T, make this new instrument suitable to improve our current
knowledge of aerosol dynamic and distribution.
The CE318‐T most important features are presented in Sect. 2. In this section we also
presented the instruments used to evaluate the CE318‐T performance at daytime and
night‐time: a Precision Filter Radiometer (PFR), a Precision Filter Spectroradiometer
(PSR), the Aerosol RObotic NETwork (AERONET) master Cimel (CE318‐AERONET) as well
as a star photometer. The measurements sites are described in Sect. 3, and in Sect. 4
we summarized the CE318‐T calibration methods, as well as the uncertainties involved
in each method. The results of the comparison analysis are presented in Sect. 5, and
the main conclusions of this work in Sect. 6.
2. INSTRUMENTS
We have presented in this study the new sun‐sky‐lunar Cimel 318‐T photometer,
previously presented in Barreto et al (2016). CE318‐T is a combination of the standard
CE318‐N reference instrument in AERONET network (CE318‐AERONET), extensively
described in Holben et al. (1998), and the prototype CE318‐U (Barreto et al. 2013a, b),
able to perform nocturnal measurements using the Moon as light source. Therefore,
this new device is able to perform daytime and night‐time photometric measurements,
allowing us to extract a complete cycle of diurnal aerosol and water vapor
measurements. This instrument performs three different measurements at an
approximate field of view of 1.29º at the same 10 wavelengths than the previous
CE318‐AERONET instruments: spectral direct sun and moon irradiance measurements
as well as spectral sky radiance measurements. The most outstanding feature of this
new device is the improved tracking system (a new four quadrant detector to track the
Sun and the Moon), new control boxes using micro‐stepping technology to control the
robots (smoothed movements and pointing resolution improved up to 0.003º), data is
stored and transferred with 32 bits (digital gains are no longer necessary) and
atmospheric pressure is measured using a barometer installed in the control box.
We have also used independent AOD measurements in order to evaluate the CE318‐T
performance. These instruments are presented in Fig. 1. AERONET AOD Level 2.0 data
at Izaña and Granada and AOD data from a four‐wavelength (367.6, 412.1, 501.0 and
863.1 nm) Global Atmosphere Watch (GAW) PFR at Izaña have been retrieved for a
daytime intercomparison. We have also included AOD extracted from a PSR prototype
installed at Izaña during April, May and June, 2014, providing coincident
measurements with PFR. This spectroradiometer measures direct solar irradiance
between 300 and 1020 nm, with a resolution ranging from 1.4 and 6 nm. More details
of this instrument can be found in Gröbner et al. (2012). The intercomparison at night‐
time has been carried out at Granada station by means of the EXCALIBUR star
photometer (Astronómica S. L.). See Pérez‐Ramírez et al. (2008, 2012) for further
details.
Figu
3. M
This
statio
locat
Rese
is a h
as a
level
affec
abso
Gran
south
Atmo
Rese
ure 1: Instrum
MEASUREM
work has b
ons. The GA
ted in Tene
earch Cente
high mount
result of th
. Due to its
cted by mi
olute sun an
nada station
h‐east of
ospheric Ph
earch (IISTA‐
mentation us
MENTS SITES
been carried
AW Izaña At
erife (Canar
r (IARC) fro
tain station
he strong i
s location i
neral dust
d moon cal
n (37.16N;
Spain in a
hysics Grou
‐CEAMA).
sed for the d
Izaña and
S
d out using
tmospheric
ry Islands,
om the State
n representa
nversion te
n the proxi
outbreaks,
ibration sta
3.60W; 68
a non‐indu
up (GFAT)
aylight and n
Granada, re
measurem
c Observato
Spain). It is
e Meteorol
ative of fre
emperature
mity of the
, especially
ation for AE
80 m a.s.l.)
strialized m
at the An
night‐time in
spectively.
ments perfor
ry (28.31N
s managed
ogical Agen
ee troposph
normally l
e Saharan d
y in summe
ERONET‐Eur
is an AERO
metropolita
ndalusian I
ntercomparis
rmed at Iza
N; 16.49W;
by the Iza
ncy of Spain
here backgr
ocated belo
desert, Izañ
ertime. Thi
rope.
ONET statio
an area, m
nstitute fo
son performe
aña and Gra
; 2373 m a.s
aña Atmosp
n (AEMET).
round cond
ow the stat
ña is sporad
s site is an
on located i
managed by
or Earth Sy
ed at
anada
s.l.) is
pheric
Izaña
itions
tion’s
dically
n the
n the
y the
ystem
4. CE318‐T CALIBRATION
CE318‐T calibration for daytime measurements is similar to the technique proposed by
Holben et al. (1998) for CE318‐AERONET. Barreto et al. (2016) performed a thorough
study about the CE318‐T calibration during night‐time. They proposed four different
calibration techniques to calibrate this new instrument.
4.1. CE318‐T absolute calibration at night‐time: Lunar Langley Method
The CE318‐T absolute calibration at night‐time is attained by means of the
Lunar Langley method (LLM), a modification of the previous Langley method
for daytime absolute calibration, especifically developed to be applied under
variable illumination conditions, as is the case of the Moon. This methodology
requires the use of the Robotic Lunar Observatory (ROLO) model (Kieffer and
Stone, 2005) to estimate the Moon’s extraterrestrial irradiance (I0). Once I0 is
estimated at each channel wavelength (I0,λ), we can extract the instrument’s
calibration constant (ҡλ) from the Eq. 1.
, , ∙ ҡ (Eq. 1)
4.2. CE318‐T calibration transference at night‐time
Barreto et al. (2013a, 2016) proposed different techniques to transfer the
calibration from a master instrument to a secondary instrument. The Ratio
Moon (RM) calibration technique uses the ratio of coincident
master/secondary nocturnal measurements (DCNM and DCN
S, respectively) to
do that, while the Ratio Sun method uses coincident measurements at daytime
(DCDM and DCD
S, respectively). Both techniques are based on the following
equation:
, , ∙ ~ , ∙ (Eq. 2)
with denoted as Ratio Moon and denoted as Ratio Sun.
The fourth technique for calibration transference is the Sun‐Moon Gain factor
technique, based on previous works developed by Berkoff et al. (2011) and
Barreto et al. (2013a). This technique involves the estimation of the different
amplification used in sun and moon measurements to transfer directly the
daytime calibration to night‐time. This estimation can be performed in
laboratory using an integrating sphere to estimate the moon/sun amplification
(G), using the daytime calibration (V0,λ) and the extraterrestrial solar irradiance
(E0,λ).
ҡ ,
,∙ (Eq. 3)
This calibration method only requires a sun‐calibrated instrument and
daytime measurements, not depending on the ROLO lunar irradiance model.
4.3. Uncertainty in CE318‐T calibration
Barreto et al. (2016) also estimated the CE318‐T uncertainty involved in these
four different calibration methods. Their results are presented in Table 1,
showing similar uncertainties in the four calibration methods, with maximum
AOD uncertainties for low illumination conditions of 0.017 in case of visible
spectral range, and 0.021 in case of longer wavelength range.
Table 1: Estimated AOD standard combined uncertainty (uAOD) extracted from Barreto et al.
(2015) for different moon’s illumination (full moon or FM and quarter moon or QM). We have
included the four different calibration techniques: Lunar‐Langley Method (LLM), Ratio Moon
(RM), Ratio Sun (RS) and sun‐moon gain factor method (SMG).
VIS 440 NM nIR
LLM FM QM
±0.011‐0.013 ±0.012‐0.014
±0.011‐0.013 ±0.012‐0.016
±0.012‐0.017 ±0.013‐0.018
RM FM QM
±0.011‐0.013±0.012‐0.014
±0.011‐0.013±0.012‐0.017
±0.013‐0.017±0.014‐0.018
RS FM QM
±0.012‐0.014 ±0.013‐0.015
±0.012‐0.014 ±0.013‐0.017
±0.015‐0.020 ±0.016‐0.021
SMG FM QM
±0.016 ±0.017
±0.016 ±0.017‐0.019
±0.016 ±0.017
5. R
5
Figur
perf
RESULTS OF
5.1. AOD da
We hav
AERONE
2014. W
perform
includes
showed
AERONE
errors (
and CE3
and MB
re 2: AOD sca
ormed at Iza
F THE INTER
aytime inter
ve intercom
ET, PFR and
We present
med with th
s the main
d a good agr
ET compari
RMSE) ≤ 0.0
318‐T/PSR
B and RMSE
atterplot at 5
aña during M
RCOMPARIS
rcompariso
mpared day
d PSR at Iz
in Fig. 2 th
he four inde
n statistics
reement be
son showed
002, with h
comparison
< 0.006.
500 nm for C
March, April,
re
SON ANALY
on at Izaña
ylight AOD
zaña in the
he quasi‐co
ependent in
of the in
etween the
d mean bia
igh regress
n analysis s
CE318‐T, CE3
May and Jun
egression lin
YSIS
measureme
60‐days pe
incident (±1
nstruments
ntercompar
four instrum
as (MB) ≤ 0.
ion coefficie
showed reg
18‐AERONET
ne 2014. Dot
ne.
ents from C
eriod from
1 min) AOD
at 500 nm.
rison analy
ments. CE3
.001 and ro
ents > 0.99
gression coe
T, PFR and P
tted line repr
CE318‐T, CE
March to
D measurem
. This figure
sis. Our re
18‐T and CE
oot mean sq
. The CE318
efficients >
SR measurem
resents the l
E318‐
June,
ments
e also
esults
E318‐
quare
8/PFR
> 0.99
ments
inear
5
Figur
(trian
5.2. AOD nig
We ha
measur
nights in
using st
ones si
evolutio
and sta
betwee
500 nm
higher d
0.018. T
star pho
CE318‐T
problem
re 3: AOD at
ngles), CE318
ght‐time int
ave comp
ements per
n August, 2
tar photom
milar to th
on in this tim
ar photom
n CE318‐T
m, with MBs
discrepancie
This latter
otometry an
T field instr
m in the sta
870 nm, 500
8‐T at daytim
m
tercompari
ared noct
rformed at
014 (7‐8, 8‐
meter measu
he CE318‐T
me period w
eter data
and star p
s up to 0.0
es in case o
result show
nd also high
ruments. W
r photomet
0 nm and 440
me (asteriscs)
moon is mar
ison at Gran
turnal CE3
Granada in
‐9, 11‐12 an
urements a
measurem
with CE318‐
(for night‐
hotometer
13 and RM
of 440 nm c
wed higher
her than the
We attribute
ter in this sp
0 nm measu
) and CE318‐
rked with the
nada
318‐T and
a tempora
nd 14‐15). T
t 880, 500
ments. We
‐AERONET d
‐time). The
in this per
MSEs up to
channel, wit
r difference
e maximum
ed these dis
pecific chan
red in Grana
‐T night‐time
e vertical line
star ph
l window of
This analysis
and 440 n
present in
data (for da
ere is a g
riod in case
0.009. How
th MB = ‐0.
es than tho
m uncertaint
screpancies
nnel.
ada using the
e (circles) in A
e.
otometer
f ±15 min in
s was perfo
m channels
Fig. 3 the
aytime), CE3
good agree
e of 870 nm
wever, we f
033 and RM
ose expecte
ties expecte
to a calibr
e star photom
August, 2014
AOD
n four
rmed
s, the
AOD
318‐T
ment
m and
found
MSE =
ed for
ed for
ration
meter
4. Full
Table 2: Main AOD statistics of the CE318‐T and star photometer intercomparison performed
in Granada during August, 2014. Mean bias (MB), root mean square error (RMSE), Pearson
coefficient (r) and number of coincidences (N) are included. Extracted from Barreto et al.
(2016)
Channel (nm) MB RMSE r N
870 ‐0.001 0.003 0.946 15
500 0.013 0.009 0.937 15
440 ‐0.033 0.018 0.911 14
5.3. AOD day/night transition coherence test at Izaña
We have completed the CE318‐T performance evaluation by means of a
day/night transition coherence test performed at Izaña in four different moon
cycles in March, April, May and June, 2014. In this 60‐days period we have
compared CE318‐AERONET daylight and CE318‐T nocturnal measurements in a
1‐h period during moonset‐sunrise (MS‐SR) and sunset‐moonrise (SS‐SR) for
different illumination conditions, assuming stable AOD conditions to assure
the day/night/day comparison. In this respect, we have excluded the AOD
information for June 12‐13, 14‐15 and 15‐16, in which a significant change in
AOD is observed (three dust outbreaks affecting the station). We present in
Fig. 4 an example of the AOD evolution extracted in June, 2014. It is a good
example of the coherence between CE318‐AERONET daylight AOD and CE318‐
T nocturnal AOD. Our results showed MB ≤ 0.015 for channels with λ ≤ 870 nm
in case of those nights before full moon (MS‐SR). Higher differences MB <
0.022 were retrieved in case of nights after full moon (SS‐MR). In case of
longer wavelengths we found significant discrepancies (MB > 0.6), especially in
1640 nm channel. Only MBs < 0.02 were found when moon illumination was ≥
90%. We observed a nocturnal cycle, zenith angle and phase angle dependent.
The reason for this nocturnal cycle is unclear, and further investigations must
be carried out to clarify it.
Fig
AER
6. C
We
phot
Rega
and
inher
calib
Ratio
meas
meth
trans
rang
We h
analy
CE31
repo
for s
carri
gure 4: AOD
RONET is pre
CONCLUSIO
have summ
tometer an
arding calibr
the Sun‐M
rent to no
rate refere
o Sun techn
surements
hod avoids
sference by
ing from 0.0
have evalua
ysis at Izaña
18‐AERONET
rted discrep
sun photom
ed out an
day/night ev
esented with
NS
marized in
d the diffe
ration at nig
oon Gain f
cturnal cal
ence instrum
nique allow
(uAOD rangi
the use
y means of
016‐0.019).
ated the CE
a involving s
T. We have
pancies bet
metry in the
intercomp
volution at Iz
h asteriscs an
displa
this paper
erent calib
ght‐time, th
factor tech
ibration. W
ments with
ws us to tran
ing from 0.
of a lunar
a simple a
E318‐T perf
several refe
also used i
tween the f
e AERONET
parison at
zaña during
nd shaded in
ayed with sq
r the main
ration app
he Lunar La
niques allo
With the Lu
uncertaint
nsfer the d
.012 to 0.0
r irradiance
and easily i
formance a
erence inde
nformation
four instrum
T and GAW
night‐time
June, 2014.
yellow. Nigh
uares.
characteri
roaches fo
ngley meth
ow the red
unar Langle
ties ranging
aylight cali
021), while
e model to
mplemente
at daytime
pendent in
n from the P
ments simila
‐PFR netwo
at Granad
Daytime dat
ht‐time data
istics of th
r daytime
hod as well
uction of t
ey method
g from 0.01
bration usi
the Sun‐M
o perform
ed laborato
by means o
struments,
PSR prototy
ar to the pre
orks (≤0.01
da station
ta from CE31
from CE318
he new CE3
and night‐
as the Ratio
the comple
we are ab
11 to 0.018
ng only day
oon Gain f
the calibr
ory routine
of a compa
such as PFR
ype. This an
ecision expe
). We have
by means
18‐
‐T is
318‐T
‐time.
o Sun
exities
ble to
8. The
ytime
factor
ration
(uAOD
arison
R and
alysis
ected
e also
of a
nocturnal intercomparison of AOD extracted from a CE318‐T and a star photometer.
This evaluation showed similar discrepancies for 870 nm and 500 nm channels (≈0.01)
but higher for 440 nm channel (≈0.03), pointing to a calibration problem in the star
photometer in this channel.
Finally, we performed a coherence test between AERONET daylight AOD and CE318‐T
night‐time AOD at Izaña in the 1‐h period after and before sunset and sunrise. Our
results showed a good consistency between daytime and night‐time AOD values
(provided stable aerosol loads are assured) in case of channels at λ≤870 nm. For near
infrared channels this consistency was only observed for moon illumination ≥ 90%. This
test also showed the existence of a nocturnal cycle on AOD impacting more
significantly in 1640 nm channel. Further investigations are required to understand the
source of this problem and to minimize this cycle.
Our results demonstrate the ability of the new CE318‐T to capture the diurnal cycle of
aerosols as well as short‐term atmospheric variations, critical for climate studies and
crucial for high‐latitude locations, given the extended periods of darkness during
winter.
7. ACKNOWLEDGEMENTS
This work has been developed within the framework of the activities of the World
Meteorological Organization (WMO) Commission for Instruments and Methods of
Observations (CIMO) Izaña Testbed for Aerosols and Water Vapor Remote Sensing
Instruments. The AERONET sun photometers at Izaña have been calibrated within the
AERONET‐Europe TNA, supported by the European Community‐Research Infrastructure Action
under the FP7 ACTRIS grant agreement no. 262254. The GAW‐PFR network for AOD at WMO‐
GAW global observatories has been implemented by the World Optical Depth Research and
Calibration Center (WORCC). The Granada GNSS station belongs to the Instituto Andaluz de
Geofísica. This work has also been supported by the Andalusia regional government through
projects P12‐RNM‐2409 and P10‐RNM‐6299, by the Spanish Ministry of Science and
Technology through project CGL2013‐45410‐R; and finally by the EU through ACTRIS project
(EU INFRA‐2010‐1.1.16‐262254).
8. REFERENCES
Barreto, A., Cuevas, E., Damiri, B., Guirado, C., Berkoff, T., Berjón, A. J., Hernández, Y.,
Almansa, F., and Gil, M.: A new method for nocturnal aerosol measurements with a
lunar photometer prototype, Atmos. Meas. Tech., 6, 585‐‐598, doi: 10.5194/amt‐6‐
585‐2013, 2013a.
Barreto, A., Cuevas, E., Damiri, B., Romero, P. M., and Almansa, F.: Column water vapor
determination in night period with a lunar photometer prototype, Atmos. Meas. Tech.,
6, 2159‐‐2167, doi:10.5194/amt‐6‐2159‐2013, 2013b.
Barreto, Á., Cuevas, E., Granados‐Muñoz, M.‐J., Alados‐Arboledas, L., Romero, P. M., Gröbner,
J., Kouremeti, N., Almansa, A. F., Stone, T., Toledano, C., Román, R., Sorokin, M., Holben, B.,
Canini, M., and Yela, M.: The new sun‐sky‐lunar Cimel CE318‐T multiband photometer – a
comprehensive performance evaluation, Atmos. Meas. Tech., 9, 631‐654, doi:10.5194/amt‐9‐
631‐2016, 2016.
Berkoff, T. A., Sorokin, M., Stone, T., Eck, T. F., Hoff, R., Welton, E., and Holben, B.:
Nocturnal aerosol optical depth measurements with a small‐aperture automated
photometer using the moon as a light source, J. Atmos. Ocean. Tech., doi:
10.1175/JTECH‐D‐10‐05036.1, 2011.
Gröbner, J., Kouremeti, N., de Coulon, E., Dürig, F., Gyo, M., Soder, R., and Waser, D.:
Spectroradiometer for Spectral Aerosol Optical Depth and Solar Irradiance
Measurements, Annual Report PMOD, p.13, 2012, available at:
http://www.pmodwrc.ch/annual_report/annualreport2012.pdf, 2012.
Holben, B. N., Eck, T. F., Slutsker, I., Tanré, D., Buis, J. P., Setzer, A., Vermote, E.,
Reagan, J. A., Kaufman, Y. J., Nakajima, T., Lavenu, F., Jankowiak, I., and Smirnov A.:
AERONET ‐‐ A federated instrument network and data archive for aerosol
characterization, Remote Sens. Environ., 66, 1‐‐16, 1998.
Leiterer, U., Naebert, A., Naebert, T., and Alekseeva, G.: A new star photometer
developed for spectral aerosol optical thickness measurements in Lindenberg, Contrib.
Atmos. Phys., 68, 133‐141, 1995.
Kieffer, H. H. and Stone, T. C.: The spectral irradiance of the moon, Astron. J., 129,
2887‐‐2901, 2005.
Pérez‐Ramírez, D., Aceituno, J., Ruiz, B., Olmo, F. J., and Alados‐Arboledas, L.:
Development and calibration of a star‐photometer to measure the aerosol optical
depth: Smoke observations at a high mountain site, Atmos. Environ., 42, 2733‐‐2738,
2008.
Pérez‐Ramírez, D., Lyamani, H., Olmo, F. J., Whiteman, D. N., Navas‐Guzmán, F., and
Alados‐Arboledas, L.: Cloud screening and quality control algorithm for star
photometer data: assessment with lidar measurements and with all‐sky images,
Atmos. Meas. Tech., 5, 1585‐‐1599, doi: 10.5194/amt‐5‐1585‐2012, 2012.