Sunglint observations over land from ground and airborne L-band radiometer data
ABSTRACT 1] This study quantifies the effects of Sun reflection over land surfaces on radiometric measurements at L-band. The impact of the reflected Sun on radiometric measurements over a grass field and over an agricultural area reached 25 K and 17 K respectively. A model that predicts the impact of Sun reflection over land is developed and tested for two different radiometer configurations and spatial resolutions.
- SourceAvailable from: Jeffrey Walker[Show abstract] [Hide abstract]
ABSTRACT: The spatial and temporal invariance of Soil Moisture and Ocean Salinity (SMOS) forward model parameters for soil moisture retrieval was assessed at 1-km resolution on a diurnal basis with data from the National Airborne Field Experiment 2006. The approach used was to apply the SMOS default parameters uniformly over 27 1-km validation pixels, retrieve soil moisture from the airborne observations, and then to interpret the differences between airborne and ground estimates in terms of land use, parameter variability, and sensing depth. For pastures (17 pixels) and nonirrigated crops (5 pixels), the root mean square error (rmse) was 0.03 volumetric (vol./vol.) soil moisture with a bias of 0.004 vol./vol. For pixels dominated by irrigated crops (5 pixels), the rmse was 0.10 vol./vol., and the bias was -0.09 vol./vol. The correlation coefficient between bias in irrigated areas and the 1-km field soil moisture variability was found to be 0.73, which suggests either 1) an increase of the soil dielectric roughness (up to about one) associated with small-scale heterogeneity of soil moisture or/and 2) a difference in sensing depth between an L-band radiometer and the in situ measurements, combined with a strong vertical gradient of soil moisture in the top 6 cm of the soil.IEEE Geoscience and Remote Sensing Letters 11/2009; · 1.82 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The Land Parameter Retrieval Model (LPRM) has been successfully applied to retrieve soil moisture from space-borne passive microwave observations at C-, X-, or Ku-band and high incidence angles (50deg-55<sup>deg</sup>). However, LPRM had never been applied to lower angles or to L-band observations. This letter describes the parameterization and performance of LPRM using aircraft and ground data from the National Airborne Field Experiment 2005. This experiment was undertaken in November 2005 in the Goulburn River catchment, which is located in southeastern Australia. It was found that model convergence could only be achieved with a temporally dynamic roughness. The roughness was parameterized according to incidence angle and soil moisture. These findings were integrated in LPRM, resulting in one uniform parameterization for all sites. The parameterized LPRM correlated well with field observations at 5-cm depth ( r = 0.93 based on all sites) with a negligible bias and an accuracy of 0.06 m<sup>3</sup>middotm<sup>-3</sup>. These results demonstrate comparable retrieval accuracies as the official SMOS soil-moisture retrieval algorithm (L-MEB), but without the need for the ancillary data that are required by L-MEB. However, care should be taken when using the proposed dynamic roughness model as it is based on a limited data set, and a more thorough evaluation is necessary to test the validity of this new approach to a wider range of conditions.IEEE Geoscience and Remote Sensing Letters 11/2009; · 1.82 Impact Factor
Sunglint observations over land from ground and airborne L-band
M. J. Escorihuela,1K. Saleh,2P. Richaume,3O. Merlin,4J. P. Walker,4and Y. H. Kerr3
Received 18 June 2008; revised 4 September 2008; accepted 15 September 2008; published 23 October 2008.
land surfaces on radiometric measurements at L-band. The
impact of the reflected Sun on radiometric measurements
over a grass field and over an agricultural area reached 25 K
and 17 K respectively. A model that predicts the impact of
Sun reflection over land is developed and tested for two
different radiometer configurations and spatial resolutions.
Citation: Escorihuela, M. J., K. Saleh, P. Richaume, O. Merlin,
J. P. Walker, and Y. H. Kerr (2008), Sunglint observations over
land from ground and airborne L-band radiometer data, Geophys.
Res. Lett., 35, L20406, doi:10.1029/2008GL035062.
This study quantifies the effects of Sun reflection over
 The Sun is an important emitter at L-band, with the
Sun brightness temperature ranging from 100000 K up to
several million Kelvin, depending on solar activity. Sun
radiation can impact microwave remote sensing in different
ways. Sunlight can get directly into the antenna or indirectly
by reflecting over the different surfaces, thus increasing the
overall brightness received by the radiometer.
 Soil moisture plays a key role in hydrological cycle. It
is consequently a key variable in weather forecast models,
climate studies, water resources, crop management, and
forecasting extreme events. The best way to access to soil
moisture is through the use of L band radiometry [Kerr,
2007]. In the near future two new satellite missions, the Soil
Moisture and Ocean Salinity (SMOS) and the Soil Moisture
Active Passive (SMAP) will be providing for the first time
global mapping ofsurface soilmoisturebased onradiometric
measurements at L-band. In the case of the SMOS (Soil
Moisture and Ocean Salinity) mission [Kerr et al., 2001],
Sun direct radiation will get directly into the instrument in
approximately 96% of the images and cancelation techni-
ques have already been developed [Camps et al., 2004]. The
reflected Sun is located, by assuming a perfectly smooth
surface, outside the SMOS useful field of view and conse-
quently it should not be a problem. It is important, however,
in order to evaluate remote sensing data at L-band the
effects of sunglint be assessed.
 The effects of Sun reflection (or sunglint) over ocean
have been the focus of several studies and Sun scattering
models have been developed [Reul et al., 2007; Le Vine et
al., 2005]. However, because of the low reflectivity of land
surfaces, sunglint effects over land are expected to be much
lower than over the ocean and no studies have addressed
this issue at L-band.
 The aim of this paper is to (i) investigate the effects of
Sun reflection over land surfaces and (ii) develop a simple
model that predicts impact of Sun reflection on radiometric
and NAFE’06) and radiometer configurations (ground
based and airborne respectively) are used. The SMOSREX
experimental site allows investigation of sunglint effects at
the field scale where the surface characteristics are well
known. The NAFE experiment has monitored surfaces
equivalent to a SMOS pixel, providing the unique opportu-
nity to validate previous results at the SMOS scale.
2. Sun Brightness Temperature
 The Sun radio flux (Fl) is the brightness (at a certain
wavelength l) integrated over the entire solar disc. For
Earth remote sensing, the solar disc is small (at L-band qSun=
0.293?) and Flcan be approximated as:
where the mean brightness Bl is expressed in
Wm?2Hz?1sr?1and WSunis the Sun solid angle as seen
from the Earth:
WSun¼ 2p 1 ? cosqSun
½? ¼ 8:2210?5sr
 Ground monitoring stations provide solar flux meas-
urements at different wavelengths. The daily data (one-
second resolution) from the Radio Solar Telescope Network
(RSTN) with observatories in Learmonth Australia 22S
114E, and San Vito, Italy 41N 18E were used in this study
for the NAFE and SMOSREX analysis respectively.
 Assuming that the Sun is a blackbody, the Sun
brightness temperature (TSun) can be calculated using the
Rayleigh-Jeans approximation [Le Vine et al., 2005]:
where Flis expressed in solar flux units (sfu) (1 sfu = 10?22
 The contribution to the antenna temperature due to
the Sun radiation reflected on the land surface can be
GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L20406, doi:10.1029/2008GL035062, 2008
1Department of Hydrology and Water Resources, IsardSAT, Barcelona,
2Department of Geography, University of Cambridge, Cambridge, UK.
3Centre d’Etudes Spatiales de la Biosphe `re, Toulouse, France.
4Department of Civil and Environmental Engineering, University of
Melbourne, Parkville, Victoria, Australia.
Copyright 2008 by the American Geophysical Union.
1 of 5
where Wais the beam solid angle, Fnis the normalized
antenna pattern and G is the surface reflectivity.
 The theoretical calculation requires the estimation of
the surface reflectivity. In the case of land surfaces, simpli-
fied zero order radiative transfer equations are usually
applied. The basic assumption is that, since the vegetation
refraction index is close to that of the air, the reflectivity of
the vegetation can be neglected. The scattering in the
vegetation layer is thus not taken into account and only
the contribution of scattering to energy loss is considered
[Mo et al., 1982]. In this case, surface reflectivity can be
calculated from brightness temperature measurements (TB)
TB¼ 1 ? G?
where G* is the effective surface reflectivity (which
implicitly takes into account attenuation due to vegetation
or surface roughness) and Tsthe surface temperature.
 In this study, the assumption that the standard
deviation of surface height is small compared to the
radiometer wavelength (l = 21 cm) is made (based on an
assessment of surface roughness measurements). Thus scat-
tering due surface roughness is also neglected and the Sun is
reflected specularly on the field.
3.1. SMOSREX Dataset
 The SMOSREX experiment was designed to con-
tribute to a better understanding of the different processes
affecting the microwave signal at L-band in the context of
the SMOS mission preparation. The experimental site is
located in the South of France (43? 23N, 1?18E), at an
altitude of 188m. Since January 2003 SMOSREX has pro-
vided accurate field measurements of dual polarized L-band
brightness temperature from a both bare soil and a fallow
plot together with ground measurements of meteorological
variables, soil moisture and temperature profiles. Further
details concerning the SMOSREX experiment are provided
by de Rosnay et al. .
 The L-band radiometer for Estimating Water In Soils
(LEWIS) is an L-band dual-polarized radiometer observing
at a frequency of 1414 with 20 MHz bandwidth. It is
equipped with a 1.3m diameter Potter horn antenna. The
antenna pattern was measured in the Centre National
d’Etudes Spatiales, Toulouse (France). The beam-width at
?3 dB is 13.6?, and the first side-lobes are at ?38 dB. The
calculated beam efficiency is greater than 98%. The instru-
ment resolution is 0.2 K for a 4 s integration time and the
estimated accuracy of the calibration is 0.5 K. The radiom-
eter is thermally regulated to 0.02 K [Lemaı ˆtre et al., 2004].
 LEWIS is mounted on a 15 meter high structure
centered on the experimental area. The grass field plot is
situated south of the instrument. In this paper, continuous
measurements of the grass field plot at 40? incidence angle
3.2. NAFE’06 Dataset
 The second National Airborne Field Experiment was
conducted in South-Eastern Australia in November 2006.
L-band microwave and infrared airborne measurements
together with ground soil moisture and temperature were
performed across three sites (Yanco, Kyeamba and Yenda)
in the western part of the Murrumbidgee catchment. The
NAFE’06 campaign primary ground verification focus was
on six farms located in the 60 x 60 km Yanco area. The
Yanco soil moisture network includes 13 monitoring sites
spatially distributed over the area. Landuse is mainly
agricultural including dryland and irrigated cropping, pas-
ture and grazing. The NAFE’06 experiment focused on
practical applications for the SMOS mission, further details
are provided by Walker et al.  and Merlin et al.
 Radiometric measurements at L-band were per-
formed with the Polarimetric L-band Multibeam Radiometer
(PLMR). PLMR measures both V and H polarizations at a
frequency of 1.413 GHz with a bandwidth of 24 MHz. The
PLMR antenna consists of an 8x8 array of dual polarized
microstrip patches with two Butler beam-forming matrices
providing six beams with pointing angles of ±7, ±21.5,
±38.5 degrees from nadir and a bandwidth of 14? across
track and 17? along track [ProSensing, 2005]. PLMR
antenna pattern is provided in the along and across axes,
therefore an interpolated antenna pattern was used to
estimate the gain at different incidence angles. The instru-
ment can be mounted on the aircraft in either across track or
along track configuration.
 PLMR flights were performed between 30th October
and 20th November 2006. In this study, regional flights
covering the entire Yanco area with the radiometer mounted
in the across track configuration are used. The aircraft was
flown at 3000 m height above the ground level providing
data at 1km spatial resolution. There is a one beam overlap
between northward and southward PLMR ground track to
ensure all the surface is covered and to allow for comparison
between up/down flights. For this study, measurements
performed by the outer overlapping radiometer beam in
northward and southward flight lines are used.
4.Sunglint Observations Over Land
 In the SMOSREX experiment, radiometric measure-
ments show a sudden increase of brightness temperature
around midday, followed by a rapid decrease to values close
to those in the morning. This fluctuation cannot be attrib-
uted to variations in parameters that affect the emission,
such as vegetation water content, soil moisture or temper-
ature. Consequently, the increase of brightness temperature
measured by the radiometer is attributed to the Sun reflec-
tion over the grass field.
 The amplitude of the Sun reflection varies through-
out the year and depends on the solar elevation. At the
diurnal scale, maximum Sun reflection occurs at noon when
the Sun is exactly in the azimuth measuring direction of
LEWIS. The impact of solar reflection on radiometric
measurements is maximum when the Sun and the radiom-
eter are aligned both in zenith and azimuth, which occurs at
noon close to the solar equinox. When close to the solstices
(winter or summer) the impact of Sun in radiometric
measurements is minimum and reaches about 2K at noon.
When close to the solar equinox, the increase in brightness
temperature due to solar reflection on the grass can reach
25 K (depending on the reflectivity).
ESCORIHUELA ET AL.: SUN-GLINT OBSERVATIONS OVER LAND
2 of 5
 The impact of Sun reflection at the SMOSREX field
was simulated by using equations (3) and (4). The surface
reflectivity (Gp) was estimated by using the averaged
polarized brightness temperature and surface temperature
before and after Sun reflection in equation (5).
 The effects of Sun reflection for two arbitrarily
chosen days during the year are shown in Figure 1. For
each plot the theoretical calculation of the impact of the Sun
is also plotted. The following values were used in the
calculation of the surface reflectivity on DOY 77, TBV=
247.5, TBH= 222.5 and Ts= 290 and on DOY 252 TBV=
278.5, TBH= 270 and Ts= 300.
 The agreement between measurements and simula-
tions is very good in terms of the duration and amplitude of
the Sun effects. Should vegetation or soil roughness scat-
tering play a role, measured Sun effects will spread over a
wider time span than those simulated by the specular model.
This is not the case here, actually measured effect seem to
be somewhat narrower than the simulated one, which
validates the hypothesis of specular reflection. However,
measurements exhibit a more fluctuating behaviour than the
simulations. The fluctuation pattern is different between non
consecutive days although it is very similar in consecutive
days (not shown). Thus it cannot be attributed to the antenna
pattern (whose effect remains constant) or variations of TSun
(very similar fluctuation pattern on DOY 77, DOY 78 and
DOY 79). A constant value of reflectivity, derived from
radiometer data, is used to calculate the impact of Sun
reflection. However, because of the differences in solid
angle between the Sun (qSun= 0.293?) and the radiometer
(qr= 13?), the surface that contributes to Sun reflection is
smaller than that seen by the radiometer. Consequently, the
Sun scans the surface reflectivity at a higher spatial resolu-
tion than the radiometer. Soil moisture and vegetation
heterogeneity within the radiometer FOV, and a non-
constant reflectivity as a result, could explain the fluctuating
behaviour of measurements. Apart from the oscillating
behaviour, this simple model provides a good estimation
of the effects of Sun reflection on radiometric measure-
ments. This model is next tested with the NAFE dataset.
 During the NAFE’06 experiment the effects of
sunglint clearly show up as lines with higher than average
brightness temperature in Regional flights performed on
November 13th, 14th and 16th by Merlin et al. [2008,
Figure 2], despite the relative position between the Sun and
antenna not being optimal for sunglint to occur.
 To quantify sunglint effects, the brightness temper-
ature of the outer overlapping radiometer beam in north-
ward and southward flight lines are compared. The
overlapping ground tracks are observed at the same inci-
dence angle (q) but the relative azimuth to the Sun changes
by 180? between northward and southward lines. The
brightness temperature of the outer right beam (3R) for
northward and southward lines as a function of the latitude
is shown for November 14th in Figure 2. The Sun position
changes during measurements; at the beginning of measure-
ments (11h30 LT), the Sun position is fSun= 0 qSun= 16,
while at the end of measurements (13h30 LT), the Sun
position is fSun= 304 and qSun= 26.
 During northward flights, the radiometer is oriented
in such a way (i.e. the relative azimuth to the Sun is large)
that the Sun reflection is out of the field of view. Measured
Figure 1. Impact of Sun reflection on DOY 77 and DOY
252 on SMOSREX site. For each date, measured and
simulated vertical and horizontal polarized brightness
temperature are provided.
Figure 2. Horizontal polarized brightness temperature of 3R-beam during Yanco Regional flights for northward and
southward flights on November 14th as a function of the latitude. Three different overlapping flight lines are shown. The
angular difference in zenith (Dq) and azimuth (Df) between the Sun and the radiometer decreases during the flight time
and is also provided.
ESCORIHUELA ET AL.: SUN-GLINT OBSERVATIONS OVER LAND
3 of 5
brightness temperature is thus not affected by Sun reflection
and can be used as a reference. During the first and second
southward lines, the difference in zenith (Dq) decreases
from 40 to 24? and the azimuth difference (Df) from 25 to
21?. The Sun reflection is therefore outside the main lobe of
the radiometer and both overlapping brightness temper-
atures (northward and southward) compare well (see first
and middle second in Figure 2. During the third southward
line, the difference between the Sun and the antenna zenith
Dq decreases from 18 to 14? and in azimuth from Df = 22
to 20?. Consequently, the impact of the reflected Sun in
radiometric measurements (and the difference between
northward and southward lines) increases with time as it
is clearly shown in the third panel in Figure 2 (note that for
the southward flights the latitude decreases with time).
 The impact of the reflected Sun on the measured
brightness temperature depends basically on the Sun tem-
perature, the surface reflectivity and the antenna gain. The
aircraft pitch, roll and heading changes slightly as the
aircraft moves and therefore the gain at which the antenna
’sees’ the Sun changes with time. Because of the high value
of TSun, those slight differences in gain and reflectivity as
the aircraft moves translate into a ’noisy’ brightness tem-
perature in the southward line. The averaged increase in
brightness temperature at H-pol is 17 K and at V-pol is 7K.
 The effects of Sun reflection were simulated for the
14th November flight. For a given latitude, the radiometric
measurement from the northward flightline was used to
estimate surface reflectivity and as reference brightness
temperature. A constant value of Ts= 300 was considered.
The comparison between model and observed brightness
temperature is shown in Figure 3. The model predicts well
the increase of brightness temperature although there is in
general a slight underestimation. Two approximations used
in the calculation of the reflected brightness temperature
could lead to underestimated brightness temperature. First,
the antenna pattern multiplies directly TSun(see equation (4))
and thus a slight error in the antenna gain has an impact in
the calculation of the reflected brightness temperature.
Second, the reference temperature might increase from
northward (at 12h55) to southward flights (at 13h30).
Similarly to the SMOSREX case, measurements exhibit a
more oscillating behaviour than predicted by the model.
Because of the coarser spatial resolution of the airborne
radiometer (1km2) when compared to the ground radiometer
(20 m2), heterogeneity within the radiometer field of view is
expected to have a higher impact than at the field scale and
can partly explain the fluctuating behaviour of measurements.
 L-band radiometric measurements performed during
SMOSREX and NAFE’06 experiment were used to evalu-
ate the impact of Sun reflection over land surfaces. The use
of two different data sets allowed evaluation of the impact
of Sun reflection over land at two different scales (field and
 In the SMOSREX field experiment, Sun reflected on
the grass field increased the brightness temperature by 2 to
25K depending on the Sun-radiometer angular configura-
tion. During the NAFE’06 airborne campaign, the Sun
impact on radiometric measurements was easily observable
and appeared as lines of higher than average brightness
temperature. For a zenith angular difference of 14?, and an
azimuth angular difference of 20? between the radiometer
and the Sun, the averaged increase in brightness temperature
at H-pol and V-pol was 17 K and 7K respectively over an
 A simple model was developed to estimate the
impact of Sun reflection. This model considered specular
reflection on the land surface, by assuming the scattering
effect of vegetation and soil roughness negligible. This
model predicted well the amplitude of the impact of Sun
reflection. However the observations exhibited a more
fluctuating behaviour than simulations. The fluctuating
behaviour of measurements is attributed to soil moisture
and vegetation heterogeneity within the radiometer field of
Meteorology in Melbourne and specially Peter Steinle for hosting her for
three months. The authors thank Joe Tenerelli and Nicolas Reul for
developing the TRAP software that allows the calculation of the relative
zenith and azimuth angles between the Sun and the radiometer. The
National Airborne Field Experiment was funded by Australian Research
Council Discovery Project DP0557543.
M. J. Escorihuela thanks the Bureau of
Camps, A., M. Vall-llossera, N. Duffo, M. Zapata, I. Corbella, F. Torres,
and V. Barrena (2004), Sun effects in 2-D aperture synthesis radiometry
imaging and their cancelation, IEEE Trans. Geosci. Remote Sens., 42,
de Rosnay, P., et al. (2006), SMOSREX: A long term rield campaign
experiment for soil moisture and land surface processes remote sensing,
Remote Sens. Environ., 102, 377–389.
Kerr, Y. (2007), Soil moisture from space: Where we are?, Hydrogeol. J.,
(2001), Soil moisture retrieval from space: The soil moisture and ocean
salinity (SMOS) mission, IEEE Trans. Geosci. Remote Sens., 39, 1729–
Lemaı ˆtre, F., J. Poussiere, Y. Kerr, M. Dejus, R. Durbe, P. de Rosnay,
and J. Calvet (2004), Design and test of the ground based L-band
radiometer for estimating water in soils (LEWIS), IEEE Trans. Geosci.
Remote Sens., 42, 1666–1676.
Le Vine, D., S. Abraham, F. Wentz, and G. Lagerloef (2005), Impact of the
Sun on remote sensing of sea surface salinity from space, Atmos. Res., 87,
Merlin, O., et al. (2008), The NAFE’06 data set: Towards soil moisture
retrieval at intermediate resolution, Adv. Water Res., in press.
Figure 3. Comparison between measured and simulated
Sun effects during the NAFE experiment. Measurements
were performed on November 14th during the third flight
ESCORIHUELA ET AL.: SUN-GLINT OBSERVATIONS OVER LAND
4 of 5
Mo, T., B. Choudhury, T. Schmugge, J. Wang, and T. Jackson (1982), A
model for microwave emission from vegetation-covered fields, J. Geo-
phys. Res., 87, 11,229–11,237.
ProSensing (2005), Polarimetric L-band Multi-beam Radiometer (PLMR)
technical report, Amherst, Mass.
Reul, N., J. Tenerelli, B. Chapron, and P. Waldteufel (2007), Modeling Sun
glitter at L-band for sea surface salinity remote sensing with SMOS,
IEEE Trans. Geosci. Remote Sens., 45, 2073–2087.
Walker, J., O. Merlin, and R. Panciera (2006), National Airborne Field
Experiment 2006 (NAFE’06): Experiment plan, Univ. of Melbourne,
Parkville, Victoria, Australia. (Available at http://www.nafe.unimelb.
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?
M. J. Escorihuela, Department of Hydrology and Water Resources,
IsardSAT, Barcelona, Spain. (email@example.com)
O. Merlin and J. P. Walker, Department of Civil and Environmental
Engineering, University of Melbourne, Parkville, Vic 3010, Australia.
P. Richaume and Y. H. Kerr, Centre d’Etudes Spatiales de la Biosphe `re,
F-31401 Toulouse CEDEX 9, France.
K. Saleh, DepartmentofGeography,UniversityofCambridge,Cambridge
CB2 3EN, UK.
ESCORIHUELA ET AL.: SUN-GLINT OBSERVATIONS OVER LAND
5 of 5