Titan Solar Occultation Observed by Cassini Vims Gas Absorption and Constraints On Aerosol Composition Bellucci Ebook Full Chapter
Titan Solar Occultation Observed by Cassini Vims Gas Absorption and Constraints On Aerosol Composition Bellucci Ebook Full Chapter
Titan Solar Occultation Observed by Cassini Vims Gas Absorption and Constraints On Aerosol Composition Bellucci Ebook Full Chapter
PII: S0019-1035(08)00452-1
DOI: 10.1016/j.icarus.2008.12.024
Reference: YICAR 8862
Please cite this article as: A. Bellucci, B. Sicardy, P. Drossart, P. Rannou, P.D. Nicholson, M. Hedman,
K.H. Baines, B. Burrati, Titan solar occultation observed by Cassini/VIMS: gas absorption and
constraints on aerosol composition, Icarus (2009), doi: 10.1016/j.icarus.2008.12.024
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Titan solar occultation observed by Cassini/VIMS: gas absorption and constraints
on aerosol composition.
A. Bellucci1
B. Sicardy1,2
T
P. Drossart1
IP
P. Rannou3,4
P.D. Nicholson5
CR
M. Hedman5
K. H. Baines6
US
B. Burrati6
1
AN
Laboratoire d'Etudes Spatiales (LESIA), Observatoire de Paris, Univserité Pierre et
Marie Curie, Université Paris-Diderot, 5 Place Jules Janssen, 92195 Meudon, France
M
2
Institut Universitaire de France
ED
3
GSMA (UMR 6089) Université de Reims Champagne-Ardenne, 51687 Reims
cedex, France
T
4
Service d'Aéronomie (UMR 7620) Université de Versailles St Quentin,
EP
USA
AC
6
Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109, USA
58 Manuscript pages
17 figures
3 tables
1
ACCEPTED MANUSCRIPT
Running head: Titan solar occultation observed by Cassini/VIMS
T
IP
CR
US
AN
M
T ED
C EP
AC
2
ACCEPTED MANUSCRIPT
Abstract
A solar occultation by Titan's atmosphere has been observed through the solar port of the
Cassini/VIMS instrument on January 15th, 2006. Transmission spectra acquired during solar
egress probe the atmosphere in the altitude range 70 to 900 km at the latitude of 71°S. Several
T
molecular absorption bands of CH4 and CO are visible in these data. A line-by-line radiative
IP
transfer calculation in spherical geometry is used to model three methane bands (1.7 µm, 2.3
CR
µm, 3.3 µm) and the CO 4.7 µm band. Above 200 km, the methane 2.3 µm band is well fit
with constant mixing ratio between 1.4% and 1.7 %, in agreement with in-situ and other
US
Cassini measurements. Under 200 km, there are discrepancies between models and
observations that are yet fully understood. Under 480 km, the 3.3 µm CH4 band is mixed with
a large and deep additional absorption. It corresponds to the C-H stretching mode of aliphatic
AN
hydrocarbon chains attached to large organic molecules. The CO 4.7 µm band is observed in
the lower stratosphere (altitudes below 150 km) and is well fit with a model with constant
M
mixing ratio of 33±10 ppm. The continuum level of the observed transmission spectra
ED
provides new constraints on the aerosol content of the atmosphere. A model using fractal
Fractal aggregates with more than 1 000 spheres of radius 0.05 µm are needed to fit the data.
EP
Clear differences in the chemical composition are revealed between tholins and actual haze
particles. Extinction and density profiles are also retrieved using an inversion of the
C
continuum values. An exponential increase of the haze number density is observed under 420
AC
3
ACCEPTED MANUSCRIPT
Introduction
Titan’s atmosphere has been studied for a long time through ground-based observations and
more recently from space. Since 2004, the Cassini/Huygens mission has provided new
insights on that atmosphere through combined in situ and remote sensing observations. The
complex chemistry in this environment leads to the formation of haze particles that are
T
responsible for the orange color of Titan. The methane rich atmosphere exhibits several
IP
absorption bands in the near infrared, making the surface visible only in narrow spectral
CR
windows. The study of Titan's atmosphere from the Cassini orbiter and Huygens probe has
provided a large amount of data on the composition of gases and aerosols. The atmospheric
US
composition in the upper atmosphere has been studied by the Composite Infrared
Spectrometer (CIRS) in the far infrared (Coustenis et al., 2007; de Kok et al., 2007), while the
VIMS instrument (Visible and Infrared Mapping Spectrometer) has provided information in
AN
the near infrared (Griffith et al., 2006, Baines et al., 2005 & 2006). Huygens observations
have provided temperature profiles as well as aerosol distribution (Fulchignoni et al., 2005,
M
Tomasko et al., 2008). Nevertheless, constraints on the composition of the haze are still
ED
elusive. New dedicated observations of aerosols are therefore welcome to enhance our
This paper presents observations of Titan taken by the Cassini/VIMS instrument in the solar
EP
occultation mode, which allow us to sound the atmosphere of Titan over a large altitude
range. The second section of this paper describes the instrument and observations, while data
C
reduction is presented in the third section, before modeling and interpretation of molecular
AC
absorption (section IV). The fifth part deals with the aerosol content of the atmosphere.
The VIMS instrument is an imaging spectrometer that spans visual and infrared wavelength
from 0.3 to 5.1 µm (cf Brown et al., 2004). For Titan, this instrument has been mainly used
4
ACCEPTED MANUSCRIPT
for limb observations to study the atmosphere, particularly atmospheric emissions (Baines et
al., 2006) and nadir detection of the surface in the so-called CH4 windows where the
atmosphere is less opaque. By contrast, this paper presents data obtained in a VIMS
occultation mode. Stellar occultations have proved in the past to be a powerful tool to probe
T
the planet's atmosphere is usually the dominating factor that shapes the lightcurves. Inversion
IP
methods then provide a density profile at typical pressure levels of a fraction of mbar to a
CR
fraction of μbar. In the case of ground-based stellar occultations by Titan, some information
can be gathered on the distribution and properties of the haze, as well as the zonal wind
US
regime if the central flash is detected (see Sicardy et al., 2006, and references therein for
details). With the Cassini spacecraft, it is possible to observe occultations of the Sun or bright
AN
stars by Titan’s atmosphere, while being very close to the satellite. The deepest levels probed
by refracted rays have densities that vary roughly as 1/D, where D is the distance of the
M
observer to Titan. Consequently, close observations by a spacecraft probe denser, thus deeper,
altitude levels than Earth-based observations. This is illustrated in Fig 1, where the levels
ED
Titan. During the solar occultation presented here, the Cassini spacecraft was at a distance
T
between 8 000 and 10 000 km from Titan's center. For this range of D, the refractive 50 %
EP
(resp. 1%) attenuation occurs below 100 km (resp 150 km). However, we will see that in the
actual VIMS occultation lightcurves, the drop occurs at much higher altitudes because of
C
[Figure 1]
Here we report about the first solar occultation by Titan observed with the VIMS instrument.
On January 15th, 2006, during Cassini's 10th flyby of Titan (T10), the VIMS instrument
observed the ingress and egress of the Sun through Titan's atmosphere. Ingress occurred at a
5
ACCEPTED MANUSCRIPT
latitude of about 41°S and egress near 71°S. Because of the brightness of the Sun, the
observations are not done through the main aperture of the instrument. A special "solar port"
was designed for observations of solar occultations. This port is offset from the boresight
direction by 20° and is aligned with the UVIS solar occultation port. Its goal is to attenuate
the solar flux, which is achieved through a series of reflections inside the solar port. The light
T
beam that exits the solar port is then focused by the telescope optics into the VIMS-IR
IP
entrance slit, and subsequently follows the same path in the IR spectrometer as the beam that
CR
directly enters the main port. Ground calibration has shown that the attenuation factor is about
2.5x10-7. Some sunlight is scattered within the optics of the solar port. This induces some
US
stray light around the image of the Sun. As a consequence, the background level is different
from zero even when we observe far from Titan. We checked that this signal has the same
spectrum as the Sun, and concluded that these solar photons should not be removed from the
AN
signal. We also note that the image of the Sun is elongated and not circular as it was reported
During this observation, only the 256 infrared spectral channels where used, covering the
interval 0.85 to 5.1 µm with a spectral resolution of 16 nm. The standard IMAGE operating
T
mode was used with 12x12 spatial pixels at an angular resolution of 0.5x0.5 mrad per pixel,
EP
and an exposure time of 40 msec. The acquisition time of each cube is 6.9 s. With a total field
of view (FOV) of 6x6 mrad and a distance of observation between 8 000 and 10 000 km, each
C
size of the solar disk in the atmosphere perpendicular to the line of sight is about 7 to 10 km
(depending of the distance of VIMS to Titan). This size is smaller than the typical scale height
(about 40 km), therefore doesn't affect our analysis. In total, 617 cubes were acquired during
the observations. Among them, 63 cubes cover the Sun egress and are studied more
6
ACCEPTED MANUSCRIPT
eliminated (in particular, the ingress cubes, see below) or used to compute a reference solar
spectrum. Thus VIMS data cubes of this observation can be analyzed in two ways: Each of
the 256 spectral channels provides an occultation lightcurve. Conversely, the 63 cubes of
interest provide 63 spectra of the Sun observed through different levels of Titan's atmosphere.
T
During this observation, thrusters were used for the stabilization of the spacecraft. This
IP
stabilization method is less stable that the reaction wheels usually used but it is necessary
CR
when the spacecraft is close to Titan where the drag due to the extended atmosphere is
relatively high. The limited torque provided by the wheels must be larger than the
US
atmospheric drag not to loose the attitude control of the spacecraft. This was not possible
during this flyby. Consequently, the Sun didn't stay fully inside the VIMS field of view
(FOV) during ingress and part of the solar flux is missing, rendering ingress data unusable.
AN
During egress, the Sun stayed inside the VIMS FOV but still moved inside of it. This motion
of the Sun in the VIMS may introduce small errors in the normalized spectra. Additional
M
errors will be due to variations of the part of the scattered light included in the 12*12 pixels
ED
FOV, as the scattered light spectrum is not quite the same as the solar image spectrum.
T
a. Altitude retrieval
Each VIMS cube is spatially summed to provide one spectrum of the Sun observed through
C
Titan's atmosphere. While the sunlight had to pass through the atmosphere at multiple
AC
altitudes, we define the level sounded by each cube as the deepest level probed by the light, as
this level has the largest contribution to absorption and refraction effects. This minimum
altitude is computed using navigation data and accounting for the refraction.
The relative positions of Cassini spacecraft, Titan and the Sun are computed using the
NAIF/SPICE system. Relevant quantities are the solar phase angle, θ, and the distance d
7
ACCEPTED MANUSCRIPT
between the Cassini spacecraft and Titan's centre (Fig 2). We call Z the distance from Cassini
to the Sun-Titan line (Z=d.cosθ), and D the distance between Titan's center and the projection
of Cassini on the Sun-Titan line (D=d.sinθ). Note that Z corresponds to the position of the
observer in the shadow in the case of an Earth-based observation. Thus, for each data cube,
T
[Figure 2]
IP
CR
These values are entered in a ray-tracing program that calculates the differential refraction of
a set of initially parallel light rays from the Sun (Sicardy et al., 2006). To model the
US
atmosphere, a density profile from Yelle (1991) is used instead of the Huygens Atmospheric
Structure Instrument (HASI) data (Fulchignoni et al., 2005), since the local inhomogeneities
AN
of this latter profile results in many spikes in the lightcurve. Note that we made this choice for
refraction models only, while the HASI profile is used later in this paper for radiative transfer
M
calculations. This choice doesn't affect much our result. In the ray tracing analysis, the
atmosphere is assumed to be composed of 98% N2 and 2% CH4, but refraction depends only
ED
weakly on these values. The atmosphere is divided in 100 m-thick layers between 1300 km
altitude and the surface. A ray coming from the Sun with an impact parameter r (the
T
minimum distance of the ray to Titan center) is refracted by an angle dω in each layer. For
EP
each position (D,Z), that is for each cube, the code provides the minimum altitude reached by
the light rays and the flux φ(D,Z) that would be received with only refraction taken into
C
account, i.e. assuming a transparent atmosphere. This calculation is done for each of the 256
AC
wavelengths, as the refraction depends slightly on the wavelength. The 256 values of the
minimum altitude are spectrally averaged to give z0, the minimum altitude for the
corresponding data cube. For the egress cubes, the vertical sampling is smaller than 20 km,
8
ACCEPTED MANUSCRIPT
b. Lightcurves
For the data reduction, we proceed in two steps. First, we compute the lightcurves and make
the appropriate corrections. This is presented in this subsection followed by a short analysis of
the ligthcurves. Then, we produce transmission spectrum by normalizing the lightcurves. This
T
The 617 data cubes were calibrated using the VIMS online calibration routine (McCord et al.,
IP
2004). The IR background was automatically subtracted, the current IR flatfield applied, and
CR
DN (Data Number) were multiplied by the instrument performance model, and finally
US
The 144 pixels of each cube in each band are summed, resulting in 256 lightcurves or 63
spectra obtained during egress. Some of the egress lightcurves are presented in Fig. 3. They
show the solar flux as a function of the minimum altitude z0 defined above. These initial
AN
lightcurves, Φ0,λ(z0), don't have a constant flux at high altitude: the solar flux appears to
M
increase monotonically with altitude. We check on the VIMS navigation software (DETOUR)
that no bright object was in the FOV of the main VIMS aperture. We don't expect any excess
ED
light when the instrument is pointing to the Sun at more than 1 000 km above Titan's surface:
only the solar flux is received at these altitudes. We don't expect any flux from Titan because
T
of the extremely large attenuation factor in the solar port. Thus we attribute these low
EP
frequency variations to some instrumental effects that are not yet fully understood. Perhaps
they are due to the motion of the Sun in the field of view and the resulting changes in the
C
For each wavelength, a linear fit in aλ z0 + bλ is performed on the lightcurve Φ0,λ(z0) between
1 000 and 2 000 km (this includes 49 points). These limits are chosen to be close enough to
the occultation, so that the linear fit remains a reasonable assumption. We note that the slope
decreases with increasing wavelength. In order to correct the flux variation outside the
occultation, but not the mean level, the slope is corrected by dividing each lightcurve Φ0,λ(z0)
9
ACCEPTED MANUSCRIPT
aλ
by z0 + 1. The corrected lightcurves are presented in Fig 3. (red curves). They were
bλ
normalized (see below) in order to be compared to each other. For lightcurves taken around
flux above unity for altitudes between 400 and 500 km (see for instance the lower left panel
T
of Fig 3.). This altitude range corresponds to the clear area between the main layer and the
IP
detached haze (Porco et al., 2005). This effect could be due to forward scattering of the
detached haze layer in the FOV while the Sun is shining in the clear zone. However we don't
CR
see this 'bump' at other wavelengths. One possibility is that it could be concealed by haze
absorption at shorter wavelengths, and hidden in the noise at longer ones. We leave this
US
potentially interesting feature for future work.
[Figure 3]
AN
In the lightcurves presented in Fig 3., we can see that the solar flux decreases at much higher
M
altitude than expected from the purely refractive model with a transparent atmosphere. A 5%
drop is reached at about 440 km at 1 µm, at about 340 km at 2 µm and at about 300 km at
ED
4 µm. These observed drops are caused by absorption by gas and haze. Thus, the Cassini
spacecraft observes occultations for which absorption prevails over refraction. When
T
EP
refraction becomes significant, the flux at all wavelengths shorter than 4 µm is already
The full set of lightcurves can be represented as a 2D-image (Fig. 4). Each horizontal cut of
this image represents a transmission spectrum at a given altitude. Each vertical cut is a
aerosols have larger extinction at shorter wavelengths, so that the light decreases at higher
10
ACCEPTED MANUSCRIPT
[Figure 4]
c. Spectra retrieval
Each spectrum must be divided by the solar spectrum observed free of atmospheric absorption
T
components appear in the transmission spectra.
IP
The most accurate reference solar spectrum is the one acquired just after egress, when the Sun
CR
was observed through the same optical system as during the occultation. As this method deals
with relative measurements, data doesn't need to be calibrated before the division. Thus we
US
eliminate uncertainties on instrumental calibration. The solar reference spectra is computed as
the mean of the 49 spectra between 1 000 and 2 000 km, that is in the same altitude range
[Figure 5]
T
The 49 spectra used to calculate the solar reference spectrum are also used to estimate the
EP
noise of the data. This estimation is done in each of the four order-sorting filters of the VIMS-
IR instrument. However, two noise levels were defined in the fourth filter because we note
C
that the last 39 channels ([4.80-5.12] µm) are noisier than the other channels of this filter. The
AC
standard deviation of each spectrum is calculated and the mean over the 49 values is kept. The
rms-noise level in each interval is 6.2x10-3 in [0.88-1.60] µm, 3.6x10-3 in [1.67-2.95] µm,
7.5x10-3 in [3.03-3.83] µm, 27.1x10-3 in [3.90-4.79] µm, 81.1x10-3 in [4.80-5.12] µm. These
values are used in all the χ2 tests for the least square fits presented later.
11
ACCEPTED MANUSCRIPT
III) Modeling of molecular absorption bands
Various molecular absorption bands can be seen in the spectra displayed in Figs. 4 and 5.
Methane bands clearly appear under 800 km altitude, while CO shows up under 180 km,
T
Gas absorption has been modeled by radiative transfer methods, using a line-by-line method
IP
in spherical geometry (see details below). Line-by-line model is preferred to band models
CR
because of a higher accuracy in the well-studied 2.3 and 3.3 µm domain. Only absorption is
taken into account, i.e. no refraction or scattering is included. The abundance of the studied
US
component is the only free parameter. Theoretical transmission spectra were calculated for a
large range of abundances in order to find the one that best fits the data.
a. The model
AN
In-situ measurements by the Huygens/HASI instrument are used to model the atmosphere
M
below 1 000 km (Fulchignoni et al., 2005). Temperature, altitude and pressure profiles were
ED
retrieved from the Planetary Data System. The HASI temperature profile measured at 10°S is
slightly different from the temperature profile at 71°S. However, comparison to results
T
obtained with CIRS measured temperature profiles has shown that no significant differences
EP
appear. The density of gas particles is derived from the ideal gas law. The atmosphere is
divided in layers of decreasing thickness with altitude. This thickness is calculated so that the
C
horizontal path in each layer is a constant number, dS. To model the absorption along the line
AC
of sight reaching z0 at the limb of Titan, the atmosphere above z0 is divided into L layers of
(R + z0 + ( )) − ( + ) ( + ( −1)) − ( + ) , where
2 2 2 2
dS(i) = 0 − + 0 0 is Titan's radius.
Using vertical temperature profile T(z0), the gravitational factor g(z0) varying with altitude,
the Boltzmann constant kb, and the mean molecular weight M of the atmosphere, the scale
12
ACCEPTED MANUSCRIPT
height, H(z0)=kb*T(z0)/M*g(z0), is calculated at each altitude. H(z) is averaged above z0 to
give H, an intermediate parameter used to define the size of the layer. The horizontal paths
dS(i) are then fixed to a constant fraction of H, dS=H/α, where α is a prescribed parameter.
The absorptions we want to calculate depend on the column density integrated over the line of
sight wich is a quantity that mustn't depend on the number or the size of the layers. The
T
integrated column density is calculated for different values of α. It appears not to vary
IP
significantly for α equal or greater than 7. Thus dS is fixed to the value dS=H/7, which is the
CR
largest size of the layers that ensures a correct calculation (that is a result that doesn't depend
on the layering of the atmosphere) without spending too much computing time. One can note
US
that the altitudes z(i) are changing with the value z0 of the minimum altitude, but that doesn't
AN
With this layering of the atmosphere, the number of layers is N=389 for a light ray passing at
z0=49 km above the surface, N=218 for z0=421 km, N=25 for z0=987 km.
M
The composition of the atmosphere in our model is 98% N2 (volume mixing ratio), a fixed
value of 4.35x10-5 of Ar (Niemann et al., 2005) and a varying value for CH4. When modelling
T
the 4.7 µm CO band, a value of 1.6% of CH4 (CIRS value) is used in the computation of the
EP
mean molecular mass of the atmosphere. However these particular values have little effect on
the final result, since the mean molecular mass of the atmosphere is only used in the
C
computation of the scale height, which is used in turn in the definition of the atmospheric
AC
layering.
Three band systems of CH4 are modeled: the ν3 system at 3.3 µm, the ν3+ν4 system at 2.3
µm and the 2ν3 system at 1.7 µm. The lines lists we used are issued from the work done in
Dijon University (Boudon et al., 2004). They include 12CH4 for these three bands and 13CH4
for the 3.3 µm band (with an isotopic ratio 12C/13C=82.3±1 (Niemann et al, 2005)). A Voigt
13
ACCEPTED MANUSCRIPT
profile was adopted for each line. The Lorentz broadening coefficients due to CH4-N2
collisions in the three methane bands are respectively 5.3x10-2 cm-1 (1.7 µm), 5.7x10-2 cm-1
(2.3 µm) and 6.0x10-2 cm-1 (3.3 µm). A Lorentz halfwidth of 4.8x10-2 cm-1 was used for CH4-
Ar collisions broadening for the three methane bands (Jacquinet-Hudson et al.,, 2005).
Finally, the exponent for the T dependences are 0.6 for N2 and 0.05 for Ar, respectively.
T
IP
To model the CO absorption band at 4.7 µm, we used the line list from GEISA database
CR
(Jacquinet-Husson et al, 2008) that includes all the isotopes of CO. We include CH3D as this
molecules has absorption lines in the wings of CO. Their absorptions are computed
simultaneously. A 6.5x10-2 cm-1 Lorentz halfwidth was used for CO-CH4 collision broadening
US
with a T-0.75 dependence.
AN
The high-resolution spectra resulting from line-by-line calculations are then convolved with a
set of Gaussians functions, where the FWHM equal those of the 256 individual spectral
M
c. Results on CH4
T
However, the CH4 abundance was not precisely determined before the Cassini and Huygens
C
measurements. The Gas Chromatograph Mass Spectrometer (GCMS) on board the Huygens
AC
probe made in situ measurements of the atmospheric composition (Niemman et al., 2005),
to 140 km. A rapid increase of the mole fraction is observed below 32 km, where it reaches
4.92x10-2 at about 8 km, and then stays constant below that level. Stratospheric measurements
by the CIRS instrument indicates a CH4 mole fraction of (1.6±0.5)x10-2 (Flasar et al., 2005).
14
ACCEPTED MANUSCRIPT
A consensus on the CH4 stratospheric abundance has been established around 1.4-1.6%. We
cannot re-improve this mixing ratio here. Instead, we validate our method on the methane
abundance and then apply it to the CO molecule for which the abundance is much less
constrained.
As the VIMS-IR spectrometer covers a wavelength range from 0.8 to 5.1 µm, we can observe
T
several CH4 bands, at 1.15 µm, 1.4 µm, 1.7 µm, 2.3 µm and 3.3 µm. The central features of
IP
the 3.3 µm band appears just below 800 km. The 2.3 µm band appears at about 700 km and
CR
the 1.7 µm band at about 500 km.
Our radiative transfer code was used to model these 3 bands with an abundance of CH4
US
varying from 1% to 2.5 %. Although other bands are identified, they are not modeled as
precise line lists for 1.15 and 1.4 µm CH4 bands are not yet available for line-by-line
calculations. A selection of transmission spectra with models of these three bands (1.7, 2.3
AN
and 3.3 µm) with 1.6% of CH4 is presented in Fig 6.
[Figure 6a]
M
[Figure 6b]
ED
There is a general agreement between our model and the VIMS spectra, especially for the
T
2.3 µm CH4 band. Below 480 km, there is a large discrepancy between the model and the
EP
observations for the 3.3 µm band because of an additional absorption (see below). The
agreement between the model and the observations of the 1.7 µm band is not as good as for
C
the 2.3 µm. But this discrepancy might be due to the lack of accurate laboratory data for this
AC
band. We will therefore concentrate on the model of the 2.3 µm CH4 band. An enlargement
on this band at several altitudes with the models at 1.4 %, 1.6% and 2.0% is presented in Fig
7. The differences between these predicted spectra are small compared to the uncertainty of
the data. Although consistent with previous results, our measurement is not sensitive enough
to improve the error bars for the CH4 abundance already given by other experiments..
15
ACCEPTED MANUSCRIPT
[Figure 7 a]
[Figure 7 b]
A least square fit is performed to determine at each altitude the mixing ratio of CH4 that best
fit our observations of the 2.3 µm band. These values are represented as a vertical profile in
T
Fig 8.
IP
[Figure 8]
CR
This figure indicates that under 200 km, a mixing ratio of more than 2 % would be necessary
to adjust the depth of the 2.3 µm band. This increase is unrealistic as the vertical profile of
US
CH4 is known to be uniform in the stratosphere, as shown by HASI and theoretical models.
However, the small differences between each model, as shown in Fig 7, reveals that our data
are not very sensitive to variation of methane abundance of a few tenth of percent. This
AN
increase of the CH4 abundance below 200 km is a systematic effect that we also observed in
other occultation data sets. Our model has been carefully tested and compared to other
M
existing models. So we don't think this increase is due to a modeling or instrumental effect.
ED
However, it could be due to a haze effect, especially because the haze density is important at
these altitudes. It could be an optical effect of the haze, such as diffraction, that makes the
T
path in the atmosphere longer than what we considered. It could be an absorption band of the
EP
haze mixed with CH4 so that the observed band is deeper that a methane only band. It could
also be due to a refractive effect. The refractivity of a gas changes in the absorption bands of
C
this gas but this variation hasn't been taken into account in our calculation of the refraction.
AC
This effect might be negligible but it's difficult to quantify it. However, above 200 km, our
observations are well fit with mixing ratios between 1.4 % and 1.7 %. These values are in
good agreement with previous measurements by Huygens and CIRS (e.g. Niemman et al.,
16
ACCEPTED MANUSCRIPT
d. Unknown absorption mixed with CH4.
The CH4 3.3 µm band can't be modeled, since an additional absorption is superimposed on the
methane absorption. This feature is centered at 3,4 µm and appears under 480 km and the
observed absorption is deeper than the CH4 absorption itself. From the actual knowledge of
Titan's atmosphere composition, no reasonable gaseous candidate can be found to explain this
T
deep absorption at this wavelength. Instead, we attribute that feature to the signature of solid
IP
particles present in the atmosphere.
CR
Observations by VIMS of a Procyon occultation through Saturn’s atmosphere show a similar
feature (Nicholson et al., 2006). Because of the low abundance of nitrogen in Saturn’s
US
stratosphere, this observation suggests that the observed component might not be a
nitrogenous compound. The Titan and Saturn features are over-plotted in Fig 9. It can be seen
that there are very similar, although the fine structure at the bottom of the band is slightly
AN
different.
[Figure 9]
M
ED
The strongest absorption of the Titan feature is found in the 16.5 nm wide channel centered at
3.3656 µm (2 971 cm-1). For the Saturn feature, the strongest absorption is found in the same
T
spectral channel at high altitude (P<10-2 mbar) but for the spectra at the deepest altitudes
EP
(P>10-2 mbar), the peak is found in the 16.5 nm wide channel centered at 3.4155 µm
(2 928 cm-1). A shoulder spread on the two spectral channels centered at 3.4487 and
C
3.4648 µm is seen in both data sets,, but is more evident in the Titan data.
AC
A similar absorption is observed since more than 20 years in the Interstellar Medium (ISM)
(Sandford et al.,1991, Pendleton et al., 1999 & 2002). This feature is considered as a tracer of
the solid state organic component of the diffuse interstellar medium (DISM). This 3.4 µm
band probes the C-H stretching mode of aliphatic hydrocarbons. However, this band doesn't
17
ACCEPTED MANUSCRIPT
provide much information about what these aliphatic chains may be attached to. In the ISM,
these chains are supposed to be attached to large organic molecules. The different stretching
modes (asymmetric or symmetric) of –CH2 and –CH3 groups determines the position and
shape of this absorption. The assymmetric C-H stretching of –CH3 and –CH2 groups are
T
D'hendecourt et Allamandola, 1986) . With a mean resolution per spectral channel of 16.6 nm
IP
for VIMS instruments, there is a good agreement between the observed position of the
CR
observed sub-peak and the laboratory values. Laboratory values are measured for saturated
US
gases or liquid for n up to 10. But we suppose that the observed feature comes from aliphatic
chains attached to larger molecules. If electronegative groups (OH, NH) were associated close
to the –CH3 and –CH2 groups, the observed peaks would be shifted to shorter wavelengths
AN
from their nominal positions. So that OH or NH groups might not be chemically bond
adjacent to the C-H bond responsible for the observed feature. We may alos note in passing
M
that the signature for symmetric C-H stretching is at 3.48 µm which is slightly beyond the
ED
As the –CH3 stretch feature is more visible than the –CH2 one in Titan data, that might
T
suggest quite short aliphatic chains. In Saturn's atmosphere, the main peak is at the position of
EP
the –CH3 stretch at high altitudes and at the –CH2 stretch position at the lowest altitude. This
However, this analysis is preliminary. The VIMS spectral resolution is not high enough to
AC
draw assertive conclusions. The observation of this absorption with higher spectral resolution
is necessary to study the structure of this feature, enable the measurement of the CH2/CH3
ratio and thus give indications on the length of the observed aliphatic chains.
Many lab experiments have been performed to identify the observed feature in the DISM
( Pendleton et al., 2002 ). It should be useful to compare precisely these data to our own
18
ACCEPTED MANUSCRIPT
observation. The visual comparison indicates that some of these laboratory-produced
materials could match our feature. The needed precise comparison of our data with laboratory
Finally, it might be possible that the material producing this 3.4 µm feature has weaker
T
overtones in the 2.3 µm band. We have considered so far that this band is only due to
IP
methane. But such an overtone would explain the apparent rise of the methane mixing ratio.
CR
In figure 6 a and b, we can see that at the altitudes where our model underpredicts the depth of
the 2.3 µm band, our model overpredicts the depth of the 1.7 µm band. An additional
US
absorption in the 2.3 µm band under 300 km would thus decrease the CH4 ratio needed to fit
e. Results on CO
AN
Carbon monoxide was first detected in Titan's atmosphere in 1983 by Lutz et al., with a
M
mixing ratio of 60 ppm. The CO photochemical lifetime in Titan's atmosphere is about 500
ED
Myrs to 1 Gyrs (Lellouch et al., 2003; Wong et al., 2002), much longer than the typical
transport timescale (160 yrs for a typical eddy diffusion coefficient of 1000 cm2 s-1). Note that
T
CO has the same molecular weight as N2, the most abundant gas, and that the temperature is
EP
never low enough for CO to condense. Thus CO should be uniformly mixed throughout
Titan's atmosphere. The CO vertical profile is important because it addresses the question of
C
an internal or an external origin for CO. Actually, if CO were only formed by oxygen
AC
photochemistry, it's abundance would be much lower than data currently published. The
steady state was found to occur at CO = 10 ppm in Lara et al., 1996, and and at only 1.8 ppm
19
ACCEPTED MANUSCRIPT
Many observations with different techniques have been made to determine the CO abundance.
A tropospheric measurement comes from VLT observation by Lellouch et al. (2003) in the
5 µm window and indicates a 32±10 ppm for CO/N2 mixing ratio. Observations by
Hidayat et al. (1998) indicates a CO mixing ratio decreasing with altitude: 29+9/-5 ppm at 60
km, 24±5 ppm at 175 km and 4.8 +3.8/-1.5 ppm at 350 km. Observations of the CO
T
fluorescence by Lopez-Valverde (2005) support the tropospheric value of 32 ppm and
IP
indicates an increase to 60 ppm in the stratosphere. This latter value is consistent with
CR
successive values published by Gurwell et al. (1995, 2000, 2004). Their observations are
presented as evidence of a well mixed vertical profile with a mixing ratio of 50±10 ppm
US
(Gurwell et al., 1995 a and b), 52±6 ppm (Gurwell et al., 2000), and 51±4ppm (Gurwell,
2004). Space observations by the CIRS instrument on Cassini indicate a mole fraction of
45±15 ppm (Flasar et al., 2005) and 47±8 ppm (De Kok et al., 2007). Nighttime emissions of
AN
CO were discovered by the VIMS instrument and indicate an abundance of 32±15 ppm
In our occultation data, the 4.7 µm absorption band of CO is seen at altitudes below 180 km
(it is not detectable above that altitude due to noise). Our model assumes a uniform mixing
T
ratio with values between 10 and 80 ppm with steps of 2 ppm. Because CH3D has absorption
EP
lines in the shorter wavelength wing of the CO band, it was included in the calculation with
The continuum level was estimated using data points in the intervals 4.132-4.166 µm (3
AC
interval were corrected for CH3D absorption. A linear fit to these 13 points was made and the
results taken as the continuum level. A second order polynomial fit was also used to evaluate
the impact of the continuum level on the determination of the CO abundance. Observations
20
ACCEPTED MANUSCRIPT
[Figure 10 a]
[Figure 10 b]
The mixing ratio of CO was evaluated on 6 spectra between 127 and 72 km using a least
square method. The χ2 tests were performed on the 24 points spanning the CO band. The
T
noise of the data in this wavelength range was estimated using spectra at higher altitudes as
IP
described above. The rms deviation is σ=27.1x10-3 in [3.90-4.78] µm and σ=81.1x10-3 in
CR
[4.80-5.12] µm intervals. Individual measurements are presented in Table 1 with their 3-σ
error bars. The measured value in the spectrum at 82 km altitude is quite far from the other
US
ones because of noisier data points. Using those six values, the mean value for CO mixing
ratio is 33.6 ppm with a formal error bar of ± 3.6 ppm. However, one can note that individual
AN
measurements are more scattered that this theoretical error. Furthermore, when using a second
order polynomial fit to estimate the continuum, the mean value of CO mixing ratio is
M
29.7 ppm with individual error bars between 6 and 12 ppm. A reasonable estimate of the CO
mixing ratio from this data set is thus 33 ± 10 ppm. Our measurement represents the
ED
abundance of CO in the lower cold stratosphere, in the altitude range 70 to 130 km. The
absorption band is seen in our data above 130 km to 180 km, but they are too noisy to be
T
included in the fit. So that no conclusion on the abundance can be drawn from the detection of
EP
CO at these altitudes. Our model with uniform mixing ratio predicts that absorption by CO
should be visible up to 300 km, but this faint absorption is then dominated by noise. In other
C
words, if a CO absorption is present between 180 an 300 km, we can't detect it with the
AC
[Table 1]
In this work, the observation of the CO absorption in the lower cold stratosphere yields a CO
mixing ratio in good agreement with the value measured with the observation of CO
21
ACCEPTED MANUSCRIPT
nighttime emission in the upper warm stratosphere (200-300 km) (Baines et al., 2006). The
CO abundance presented here is also in very good agreement with CIRS measurements in the
same altitude range (Flasar et al, 2005, De Kok et al., 2007). Tropospheric measurements
from Earth based observations are also in the same range of value (Lellouch et al., 1998,
Lopez-Valverde, 2005). These different observations, probing different altitude ranges, all
T
yield similar values of the CO mixing ratio, so that CO appears to be relatively constant with
IP
altitude.
CR
f. Another absorption
US
In figure 10 a and b, an absorption feature is visible between 4.2 and 4.3 µm. It was first
attributed as a CO2 signature. However our radiative transfer model, using a mixing ratio of
1.5x10-8 according to CIRS measurements (Coustenis et al., 2007), clearly dismisses this
AN
possibility. The observed depth of this feature is about 5 times deeper than the predicted CO2
shorter than the actual position of the CO2 band. Therefore, the identification of this feature
ED
remains to be done.
T
We now turn to the study of haze properties. We recall that egress observations are made at
about 70° S in the summer southern hemisphere, i.e. far away from the winter polar hood,
C
Comparison of the different transmission spectra taken at different altitudes shows that the
continuum level of those spectra decreases with decreasing altitude. This general drop is due
to the absorption by haze particles present in Titan's atmosphere, and responsible for its
orange color. The "2-D" image of the occultation (Fig 4.) also indicates that this absorption is
22
ACCEPTED MANUSCRIPT
more important at shorter wavelengths. That is the expected behavior for light absorption by
haze particles.
Thus, in this part, we present a model of haze absorption. First, we derive a vertical profile of
the extinction coefficient in continuum wavelengths. Assuming fractal aggregates for the
structure of haze particles, density profiles are retrieved for different size of aggregate. Then
T
the corresponding transmission spectra are computed and compared to observation to
IP
constrain haze structure and optical constants. In a last part, we also look at the spectral
CR
behaviour of the haze and compare it to previous results.
US
a. Profile of the extinction coefficient
Along its path through the atmosphere, sunlight is attenuated by aerosol scattering by a factor
layer i, ds(i) the elementary path of the light in that layer along the sunlight's path , and σN(λ)
ED
the scattering cross-section of the aggregate. The optical depth is derived from the observed
spectra; thus, the extinction is the only unknown in this set of Ns=63 equations
T
EP
altitudes increasing with i. Layer i is between z0(i) and z0(i+1) and the 63th layer expands up
AC
to 1000 km. The main assumption is that the extinction is constant within each layer. This is
reasonable as the 63 layers are each less than 20 km thick. We note here that this vertical
sampling is smaller than the resolution of each cube (50-60 km, see section II). The inversion
begins at the altitude z0(i_top=55)=854 km. We assume that above that level,
23
ACCEPTED MANUSCRIPT
⎛ ( ) − 0( _ )⎞
z 0 (i ≥ i _ top), k λ (i) = kλ (i _ top) × exp⎜ − 0 ⎟,
⎝ H ⎠
RT
where the scale height H = = 62 km , with T(800 km)=160 K, m=mN2=28 g.mole-1 and
mg
g(800 km)=0.76 m s-2. Using the observed transmission at 854 km, kλ(i_top) is calculated.
The values of kλ(i) for i<i_top are then determined using the transmission observed at z0(i)
T
and the value of kλ(i') for i' between i+1 and i_top. This "onion-peeling" process is repeated
IP
for each observed spectrum until the Ns=63 values of kλ are computed.
CR
This calculation is done for each VIMS spectral channel. But the relevant parts of the spectra
are those in the continuum intervals, where there is no molecular absorption. We have defined
US
6 of those intervals, whose limits are indicated in Table 2. No data points were chosen in the
high altitudes when the transmission can get artificially larger than unity because of noise. In
ED
Fig 11, the extinction appears to increase exponentially below 460 km. A linear fit to ln(kλ)
between 77 and 461 km for continuum wavelength below 4.7 µm give values for the scale
T
EP
[Figure 11]
C
AC
Several measurements have shown that the light scattered by Titan's atmosphere is highly
polarized. To explain this observation, it is supposed that aerosols are fractal aggregates of
spheres. (West and Smith, 1991) The radius of these spherical monomers is constrained by
DISR measurements of the phase function and degree of polarization. Tomasko et al., 2008
24
ACCEPTED MANUSCRIPT
shows that a radius of 0.05 µm or less is needed to explain DISR observations. In the same
paper, the number N of particles per aggregate was found to be about 3000 above 60 km.
In our model, haze particles are aggregates of N spheres of 0.05 µm radius, with a fractal
dimension Df=2 (Rannou et al., 1997). We consider ten values of the number N of spheres in
T
each aggregate between 1 and 30 000. Each kind of aggregate is characterized by its
IP
scattering cross-section σN(λ). They result from an optical model developed by Botet et al.
CR
(1995) and applied for Titan by Rannou et al. (1997). In 1984, Khare et al. produced
laboratory analogs of Titan's aerosol that they called tholins. Real and imaginary parts of the
US
complex refractive index of those produces were measured. These indices are the inputs of the
microphysical model.
AN
According to the previous section, the density in layer i is retrieved through n N ,λ (i) =
kλ i
σN λ
.
Using this equation, we have, for each spectral channel, 10 possible density profiles,
M
corresponding to the 10 values of the number N of spheres per aggregate. As for the
ED
extinction, we only use continuum wavelengths. However, all the intervals are not relevant at
all the altitudes. Only short wavelengths (< 2 µm) have significant absorption at high
T
altitudes. As the altitude decreases, longer wavelengths are also included. Then, for altitudes
EP
less than 140 km, the flux at short wavelengths is completely absorbed and not taken into
account. The use of each interval at each altitude is described in Table 2. In each layer, the
C
density is the median of the values nN,λ(i) in the relevant wavelengths interval. Negative
AC
The 10 resulting density profiles ñN(z), independent of the wavelength, are considered as an
initial guess. At each level, we explore 371 values around this initial result, from 1x10-4 ñN(z)
to 20 x ñN(z). The best value for the density is determined using a least square fit of the
corresponding transmission models to the VIMS observations. The χ2 test is made at each
25
ACCEPTED MANUSCRIPT
altitude and for each type of aggregate in the intervals defined above, avoiding molecular
absorptions. The values of the noise used are defined in section III.c. We determine the
density that minimizes χ2 and the corresponding formal 3-σ error bar. The models for the
transmission are computed iteratively, using for the altitudes above the level of interest the
T
IP
c. Results on transmission spectra
CR
In Fig 12, we overplot to the VIMS observed spectra the best fit models for several values of
N. Models with less than 100 particles per aggregate are excluded right away, as it is clear
US
that they cannot reproduce the continuum. Large numbers of particles per aggregate, typically
larger than 1 000, are needed to reproduce satisfactorily the observed slope of the continuum.
AN
[Figure 12]
M
To improve the determination of N, we look in each layer for the value of N that leads to the
smallest minimum of χ2. Because of the uncertainty of the models above 500 km (see
ED
discussion below) we only present this result under 500 km in Fig 13. Horizontal lines
represent the range of value of N with 3-σ level error-bars. This range covers all of the 10
T
possible values of N for altitudes above 500 km. This figure indicates once more that large
EP
values of N are more satisfactory. Our data don't allow us to determine a best number N in the
interval 1 000-30 000, but we can exclude values of N lower than 1 000 spheres per
C
aggregate. However, it's important to note that our near-infrared observations lead to the same
AC
conclusion as for the visual data from DISR (Tomasko et al., 2008): both data sets need large
numbers of particles (N>1 000) per aggregates to fit the data. It must be underlined that our
measurements involve higher altitudes than DISR measurements. DISR measurements are
made below 150 km while our result extend up to 450 km. Finally, our results about the size
26
ACCEPTED MANUSCRIPT
of the aggregates is also in good agreement with the model developed by Bar-Nun et al., 2008
that predicted at the altitude of 100 km a number of monomer between 2 400 and 2 700.
[Figure 13]
Whatever the value of N, we can see in Fig 12 that the predicted transmission spectra have
T
two large absorption bands at 3 µm and 4.6 µm. They result from two peaks at those
IP
wavelengths in the refractive index of the Tholins produced by Khare et al. (see fig 4 of
CR
Khare et al., 1984). The 4.6 µm is due to vibrational transitions of C҂N. The 3 µm is
attributed to C-H bound in Khare et al., 1984 and to N-H bounds in Tran et al., 2003 (a) and
US
Imanaka et al., 2004. and The identification in this last paper is the most likely. Our data
don't show any evidence of the 3 µm absorption. At 4.6 µm, the predicted absorption in Khare
AN
et al’s tholins is right at the position of the CO absorption band (4.7 µm). Thus we can't be as
assertive about its absence. However, as we said before, we believe that the 3.4 µm feature
M
observed in our data and in Saturn occultations data is a haze signature. The presence of this
compound on Saturn implies that it might not contain any nitrogen. This would reinforce the
ED
non-observation of the 4.6 µm (C҂N signature) band of Khare et al’s tholin in our data.
These differences indicate that there are real chemical differences between the actual Titan
T
haze and the materials called tholins produced in the lab. These differences concern the
EP
nitrogen content of the haze that must be inferior in actual haze than in Khare's tholins. This
result added to the identification of aliphatic chains in the aerosols with the 3.4 µm absorption
C
bands suggest that actual haze might have little nitrogen, such as photochemical analogs
AC
Our inversion process begins at 854 km altitude. However, densities retrieved between
854 km and 470 km are somewhat dependent on the normalization of the lightcurves. The
27
Another random document with
no related content on Scribd:
SHUHBA, BATHS AND ROMAN PAVEMENT
Our guide took us straight to the house of the sheikh, who advanced
to meet us with profuse expressions of welcome. Dismounting in the
street, we followed him at his invitation. With difficulty we made our
way dry-shod over a huge pool of rain-water which had collected in
front of the arched doorway, through which we entered a wide
courtyard. To the left stood a rickety erection, in the construction of
which some of the finest materials from the old buildings had been
employed. A broad stairway of large lava blocks led up to it. A roof of
branches and brushwood rested upon gracefully hewn marble pillars,
which were tied together at the top by a rude architrave. These in
turn were supported upon beautiful capitals, turned upside down,
and on carved blocks of stone. The back wall was of the usual mud-
built character, and the pavement was rough in the extreme. A diwân
round three sides doubtless provided sitting accommodation for the
sheikh and his friends in fine weather. Nothing could better mark the
low level of the present inhabitants than their pride in such a bit of ill-
fitting, incongruous patchwork as this, in proximity to the magnificent
remains of a past civilisation.
A strange, rambling old house it was into which we entered by a
narrow winding passage from the left corner of the courtyard. First
we found ourselves in a series of great gloomy apartments
communicating with each other in a line east and west; then, turning
to the right, we scrambled through a doorway, the broken threshold
of which was some feet above the level of the floor; and, pushing
forward, we entered a second courtyard, much smaller than the first,
with rooms all round, on one side two stories high. Some remains of
ancient ornamentation were still visible on the walls, and the
pavement of the yard was evidently from of old. Here were our
quarters for the night, the gentlemen having two little rooms, one of
which served as dining-room, on one side, and the ladies a larger
room on the other. The stair leading up to the gentlemen’s
apartments had been failing for centuries, and now was nigh unto
falling; but, observing great caution, we all escaped without accident.
Our host for this night formed a contrast in every respect to the
dignified and magnanimous chief of Damet el-ʿAliâ. A short, thick-set
man, with stubbly white beard, very red nose, and puffy cheeks, he
bustled about with the air of a man who does a very great favour
indeed. With evident pride he displayed his rooms, and fished for
compliments, suggesting that they were beautiful and clean, mithl
lokanda—“like a hotel.” Ideas of cleanliness differ, but we avoided
controversy by gently turning the conversation to the subject of our
entertainment This we were allowed to provide for ourselves, even to
the coffee, of which he seemed glad to drink a share. He was one of
the less noble sort; and, his people taking their cue doubtless from
their chief, our servants found it difficult to secure all necessaries at
reasonable prices. But as the night closed darkly around us, and the
mountains were alternately lit up by sheets of blinding lightning and
filled with loud rolling thunder, while the rain fell in torrents, and the
wind whistled eerily among the ruins, we were thankful, even with all
its drawbacks, to be under such substantial shelter. If, for reasons
which need not be specified, we slept but little, we could all the more
realise our good fortune, in that, on these high, open uplands, we
were not exposed to the full fury of the tempest.
The morning broke clear and beautiful, and we were out betimes to
make a rapid survey of the old remains. A few paces north of the
chief’s house we struck the main street running east and west. It
seems just possible, from the remains of bases here and there, that
this may once have been a pillared street like that at Jerash, so
striking even in its desolation; or that at Gadara, where the columns
lie prone and broken along the whole length. Following this street
eastward, it sinks rapidly, and passes under a long archway, which
might almost be called a tunnel, strongly built of dressed basalt. This
doubtless formed the substruction of some important public building.
A blacksmith has his workshop in one of the deep cellars in the side
of the archway, and his blazing fire sends cheerful gleams through
the gloom. Beyond this archway eastward lie all the ruins possessing
special interest for the visitor. To the south of the road stands the
great amphitheatre. Carefully built of massive stones, the walls and
tiers of seats are still almost entire. It is the best preserved of all
such structures to be seen east of the Jordan, and it appears to have
been one of the largest. Several poor Druze families were in
possession of the lower parts of the building when we visited it, and
very comfortable houses they made—superior certainly to any of the
modern erections around.
We visited in succession a great sunk octagonal building, as to the
use of which we could make no satisfactory guess; the ruins of
several temples, one of which must have been of no ordinary
splendour; and the remains of the tetrapylon which once graced the
crossing of the two main streets. Now only three of the original four
massive bases are to be seen, and the arches have entirely
disappeared. We scrambled over rickety walls and scattered stones,
and crawled into noisome crypts in search of sculpture and
inscription. We saw enough to persuade us that a rich harvest may
be gathered here by the patient explorer. Of the ancient baths which
stood in the south-eastern quarter not far from this crossing, very
large portions are still in a good state of preservation, and form,
perhaps, the most interesting part of all the ruins.
The material employed in their construction, like that of all the
buildings in the city, is basalt, and in parts the appearance is very
fine; but no adequate idea of their original splendour can now be
formed. The rows of gaping holes in the walls tell of the lining of
marble with which they were once adorned. The destruction of this
was doubtless dictated by the desire to possess the iron fastenings
by which the marble slabs were held in position, and the lead by
which these were fixed into the walls—a temptation which the
cupidity of the Arabs would make it extremely difficult for them to
resist The water channels are skilfully built into the walls, and from
the points at which they project we may guess where the baths were
placed; but the floors are now entirely heaped over with ruins. The
walls are still over thirty feet in height, and of great strength. Most
interesting of all, in connection with the baths, is the old aqueduct, by
which the water was conducted across the low valley to the eastward
from the hills beyond. Several of the substantial arches are still
standing, and the line can be traced away towards the eastern
uplands. Eleven or twelve miles was the water brought to minister to
the comfort of the splendid, luxury-loving Roman.
These and other similarly great structures we owe to the ancients’
ignorance of the principles of hydrostatics. Only when we gaze upon
such vast undertakings, where the channel was raised by artificial
means, so that the water might flow along a regularly inclined plane,
do we fully realise what an immense saving of labour the discovery
of these simple principles has proved to the modern world.
The ancients appear to have spent their strength in the erection of
public buildings. The houses of the common people seem to have
had nothing special about them. Built of the ordinary black basaltic
stone which abounds in the neighbourhood, they have long since
gone to ruin, probably in the shocks of earthquakes. West of the
town stand two beautifully formed conical hills. Some of our party
who ascended them found them to be extinct volcanoes—one
having a circular, cup-like crater in the top. Seen from a distance,
these hills bear a striking resemblance to heaps of grain on a great
threshing-floor. This resemblance has not escaped the sharp eyes of
the imaginative Arabs, who call them “the grain-heaps of Pharaoh.”
Local tradition associates them with the name of a notable oppressor
of the people, the builder of the Qanâtîr Firʿaun (“the arches of
Pharaoh”), the great aqueduct which stretches from the
neighbourhood of Nowa past Derʿat to Gadara. Having exhausted
the people with taxes for the completion of this work, he finally
seized all the grain in the land and stored it here, ready for his own
purposes. He sent a gigantic camel to fetch it, and just as the
unwieldy animal drew near, the wrath of God was kindled against
Pharaoh, and a bolt from the clouds blasted grain and camel
together, leaving two blackened heaps as monuments of the
impotence of all earthly tyrants before the King of heaven.
This town is believed by many to represent the ancient Philippopolis.
True it is that “Philip the Arabian,” a native of this region, having
been elected emperor by the army in Syria about the middle of the
third century—244-249 a.d.—founded a city in his native country,
and adorned it in Roman fashion. But so little is known with certainty
on the subject, that almost any considerable site in Haurân may
claim the honour, if honour it be. The modern name of Shuhba is
said to be derived from the noble Moslem family of Shehâb, who in
the early years of the Mohammedan era came northward from
Arabia Felix, and in their wanderings, before settling in Mount
Lebanon, made this city a temporary home. Relatives of the prophet
of Arabia, they received distinguished honour, and assumed a
leading part in the affairs of the Lebanon. The name of Emîr Beshîr
Shehâb was well known in Europe in the earlier half of the
nineteenth century. This prince of all the Lebanon fell in the year
1840; and the family, already shorn of much of its glory, went finally
down amid the bloody revolutions of 1860.
There is a prevailing belief among the uninstructed in all parts of the
country that the Franj—the name given to all Westerns—are literally
loaded with gold. To this belief we owed a somewhat unpleasant
experience. The avaricious old sheikh took counsel with a faithless
one among our attendants, who evidently wished to smooth the road
for his own return by satisfying the cupidity of the natives at our
expense. He advised the sheikh to demand a most outrageous sum
for our entertainment, in which demand the said faithless one should
support him. The arrangement was at once agreed upon. Meantime
a second attendant, who bore no love to the former, having
overheard the plot, revealed the whole. We decided the amount and
manner of payment, taking care that there should be no reasonable
ground of complaint. Finding himself detected, the sheikh’s
accomplice ignobly forsook him. When the money was put into his
hand, with expressions of thanks for shelter afforded, the old man
could not conceal his surprise, and it was some time ere he
recovered sufficiently to hint that the sum was small. Just before we
started, a few piastres extra were added, to save what little of dignity
he possessed. He, as well as we, wished everything done in secret,
knowing well that a report of his mean conduct spreading among his
brother sheikhs in Jebel ed-Druze would prove fatal to his reputation,
especially as Englîze were in the question. This was the only display
of meanness or stinginess we met with east of Jordan; and for even
this our own servant was chiefly to blame.
CHAPTER V
Ride to Kanawât (Kenath)—Impressive situation and
remains—Place-names in Palestine—Israelites and Arabs
—Education—A charming ride through mountain glades—
Suweida.
We left the city by the southern, the only double gate the city
boasted, as it is still the best preserved. Here also the city wall is
seen in something like its original proportions. Our way led straight
southwards from the gate, along a track lined on either side with
fallen and broken columns, which showed that the splendour of the
old city had been by no means confined within the walls. A large pool
had formed in the hollow to the right during winter, and, replenished
by the previous night’s rain, afforded refreshment to the horses ere
they faced the steep hill before them. By a zigzag path we soon
ascended to a considerable height, finding far more various
vegetation than we had thought possible.
Riding thus along the western slopes of the mountain, a wonderful
panorama spread out before us: Shuhba, which we had just left,
black and desolate-looking on its blasted hill; the whole extent of
Haurân, el-Lejâʾ, Jaulân, and Gilead; Jebel esh-Sheikh, throwing
high his gleaming shoulders in the north-west; while once again we
could see the Safed hills and the uplands of Lower Galilee, with
Tabor’s rounded cone distinctly visible above his fellows. We could
almost trace all our wanderings from the point where we entered the
Haurân, through the scorched fields of el-Lejâʾ, on to the mountain
over which we were passing. And here it was impossible to avoid
noting once more the dark spots over the far-stretching plains,
marking the positions of ancient towns now waste and ruined. To the
traveller in this country, almost fabulously rich in agricultural wealth,
the phrase “a land of ruins” ever and anon returns like the refrain of
some sad song. A lower road from Shuhba leads by way of Suleim
and ʿAtyl, each with ruins of interest—the former of a temple,
subsequently a Christian church; and the latter of two temples. But it
was much longer, and we feared the hollows would be heavy from
the rain; and wishing to have as much time as possible in Kanawât,
we took the way across the mountain. The immediate surroundings
were dull, but descending a little, and turning a spur of the hill, a
scene of surpassing beauty met our eyes. The valley below opened
into a fair plain, embosomed among the mountains, where teams of
oxen, guided by peasant Druzes, in their white turbans and
tricoloured coats, drew furrows in the soft soil with wooden ploughs,
contrasting picturesquely with the brown and green of the
surrounding slopes. The southern edge of the plain is washed by a
little stream; beyond it the rising ground was covered with glancing
foliage, over which rose the tops of tall columns. Eastward the valley
narrowed, and the stream dashing down a precipice many feet high,
formed a delightful waterfall, on either side of which were gathered
the ruins of Kanawât. The mountains, grey in the changing light,
formed a pleasing background. Just as we swept round in full view, a
light shower drifted down the valley. The sun, striking through the
rain on glistening foliage, white waterfall, and stately ruins on the
brow of the hill, transformed the whole into a vision of fairyland. It
seemed as if the stream of time were suddenly turned back, and the
broken, hoary city on the height smiled again in the beauty and
splendour of her youth. So complete was the illusion, that the
passage of warriors long dead, with the kingly form of Herod in their
midst, hotly pursued by the wild Arabians, would have seemed so
natural as hardly to excite surprise.
We crossed the plain, waded the stream, and climbed the slope
towards the city. Leaving the ruins of a fine temple crowning a leafy
knoll, to the right, we pushed on through thickets of ground oak and
thorn, a strong prickly network of brambles covering all the
undergrowth. The lower part of the town presents nothing distinctive.
It is only partially inhabited by a colony of Druzes. Many of the empty
houses are quite perfect, stone doors and windows in position, and
swinging as easily as they did to the hands of their old possessors.
Going as far as we dared along the edge of the cliff, over which this
part of the town seems to impend, we obtained a fine view of the
gorge into which the waterfall descends, and also of the picturesque
old mill by which the water-power is utilised for the benefit of the
inhabitants. Turning cautiously, we retraced our steps, and entered
the street leading to the sheikh’s house. As he was absent we could
not pay him our respects. An easy ascent leads to the upper town,
where, in open spaces, all the great buildings were gathered. We
crossed the broken remains of a fine old aqueduct, just above the
waterfall, beside the ruins of a gigantic wall; and climbing over
shapeless heaps of stones, many of them beautifully cut and carved,
we entered the largest of all the structures that tell of glories long
waxed dim. It is variously called by the natives es-Seraiah—“the
Palace,” and Makâm Ayyûb—“the place of Job.” Thus, on either side
of the great plain, on which in the far past, as tradition hath it, his
flocks browsed and his husbandmen gathered the golden harvests, a
spot is consecrated to the patriarch’s memory.
KANAWÂT, RUINS OF TEMPLE
The Seraiah is a group of massive buildings, adorned with
colonnades and artistic sculptures. Around a doorway still almost
entire, opening on a wide paved space, are beautifully carved
bunches of grapes, leaves, and flowers. On the lintel of a door
leading from one part to another, a cross is cut in the stone,
indicating the presence of Christians at some period, while one of
the halls has evidently been used as a church. These apartments
are of spacious dimensions, the smallest of the three measuring
eighty feet by seventy. Most likely they were originally dedicated to
heathen gods. What information as to the ancient city and its noble
buildings may be buried under the great piles of debris no one can
say; but few places, I should think, on either side of Jordan would
better repay the excavator’s toil.
Our cloth was spread on the stump of a fallen column, in the
innermost shrine. Sitting around on huge blocks, finding shelter from
the sun, we enjoyed our mid-day meal. Troops of kindly Druzes
gathered about, ready to bring leban, cheese, milk, bread, or
whatever viands were at their command. The horses, having been
refreshed from the brook, seemed to appreciate the cool shade of
the middle chamber, haltered to the stately columns.
The remains of Kanawât might well engage attention for as many
days as we had hours to spend. On the opposite bank of the deep
valley is a small theatre almost wholly cut out of the solid rock, about
sixty feet in diameter, with a cistern in the centre of the area. A
Greek inscription intimates that it was built by Marcus Lysias,
probably a wealthy Roman officer, for the delectation of the
inhabitants of Canatha. A little higher up stands a temple of modest
proportions, and still further eastward a large tower, resting on
massive substructions, evidently of high antiquity. This is
approached by steps hewn in the rock. Close by is a lofty round
tower, probably sepulchral. Just visible over the oak thickets above
us on our way to Suweida, we saw several similar towers. If we
cannot fix their date, it is clear at least that they belong to a time in
the far past. Of the great reservoirs, whose arched roofs have in
many places been broken through, we could make no minute
inspection. They lie between the Seraiah and the remains of a noble
temple, of which the thick side walls are standing, while in front a few
columns of splendid proportions rise from a huge confused mass of
great stones. It was perilous climbing, many of the blocks being
ready to fall; but the view from the top justified the risk and toil. The
commanding situation of the ancient city is seen to advantage. On a
gentle slope of the mountain, overlooking at no great distance the
wide plain, then as populous as it is desolate to-day, with plentiful
natural supplies of water, rich soil, and thick embowering forests, it
was just such a spot as the splendour-loving Herod might well select
for lavish adornment. Traces of a hippodrome are found close to this
temple, and several of the gardens cultivated by Druzes are
surrounded partly by old walls and partly by new walls of old
materials. The grouping together of so many noble buildings, within
so small space, the graceful shafts of beautiful columns rising in
clusters here and there, reminded one of Athens; but the dark stones
lacked the dazzling effect of the white marbles on the Acropolis.
The name Kanawât probably points to that the city bore ere it fell into
the hands of the conquering Israelites, when it was called Nobah—a
name of which there is now no trace. Before the days of Christ the
old name had reasserted itself, and Josephus calls it Canatha—a
very slight change from the ancient Kenath. The identity of Kanawât
with Canatha is certain. It is interesting to observe, all over Palestine,
this reappearance of ancient names, and the practical obliteration of
those imposed by temporary rulers. The present Beisân is clearly a
modification of the old Bethshean, Scythopolis being forgotten.
Banias is simply the Arabic form of the Greek Panias, the Arabs
having no b; Cæsarea-Philippi is known only to strangers. Beitîn is
evidently another case, representing the ancient Bethaven; while
Bethel is locally unknown. It would be interesting further to inquire
how the characters of the trans-Jordanic tribes affected the
nomenclature of the land. They were essentially a pastoral people.
This tended to cut them off from the other tribes. They never took
kindly to the agricultural life prevalent on the west of Jordan. Their
nomadic habits would leave the captured cities more or less open for
the return of their inhabitants from the fastnesses to which they had
been driven; and of course they would bring the old names with
them. Thus Nobah and Bashanhavoth-Jair are names to be found
only in the Bible records.
The remarkable facial likeness to the Jews found among the people
east of the Jordan leads one to wonder if there is not a closer
relationship than that of cousinship between the two races—if, in
short, the eastern tribes did not in the end mingle freely with their
nomadic neighbours, and thus become gradually alienated in
sympathy from the people and religion of Israel, as they were
already separated from them by the mighty gorge of Jordan. It was
this very calamity the prophetic foresight of their fathers sought to
obviate, when they erected the gigantic altar of witness “in the
forefront of the land of Canaan, in the region about Jordan, on the
side that pertaineth to the children of Israel.” It should be an altar of
witness to succeeding generations of the unity of the people, lest the
children of the tribes westward should be tempted at any time to say,
“What have ye to do with the Lord, the God of Israel? For the Lord
hath made a border between us and you.” The real danger lay in
another direction. Thus there was a certain fitness in the fact that
these eastward tribes were the first to bear the brunt of the great
invasions from the north by which Israel was scourged.
KANAWÂT, SCULPTURED DOORWAY IN TEMPLE
A Druze villager who attached himself to our company proved a
pleasant and chatty companion. Bright eyes looked out from under
his spotless turban; black whiskers and shining white teeth combined
with a frank, open countenance to prepossess us in his favour. He
said he had been teacher in a school which the Englîze had
supported for some time in the village. By way of corroboration he
aired a few words of English picked up from his superior. Very
strangely they sounded from his lips, without any connection, and
seemingly so out of place amid these surroundings. His
acquaintance with English was like that of a Syrian gentleman friend
of mine, who occasionally in company announces that he knows
English. “What,” he will ask, “is English for Narghîleh?” And without
waiting for reply, exclaims, “Hubble-Bubble!” laughing heartily at his
own joke.
The school had been summarily closed by the authority of the
Government, to the sorrow of the villagers, who were beginning to
appreciate the advantage of a rudimentary education. There is a
great field for missionary enterprise—medical by preference—in all
this region. The missionary’s efforts would find assistance in the
generous instincts of the people themselves. They are yet
uncorrupted by the unhappy influences associated with the passage
of the great travelling public. These are often, unfortunately, all of
civilisation known to the untutored inhabitants; and the barriers thus
raised against the missionary and his work can be fully appreciated
only by those who have had them to face.
Our cheery companion waited until we were all mounted, then led
the way, by many tortuous windings, through the old town, to an
opening which had once been a gate, on the road to Suweida. Few
traces are left of the ancient Roman road, and soon we were on a
track of the usual kind, very soft in parts, from the recent rains. We
passed between fruitful vineyards and cultivated patches, where the
white turbans of the vine-dressers moved to and fro among the
green with pleasing effect. Our ride that afternoon along the hillsides,
through oak and thorn thickets, the green interspaces sprinkled with
flowers, openings in the foliage affording glimpses of the wonderful
plains of Bashan, was the most agreeable by far of all we enjoyed in
Haurân. The freshness of the leaf, the music of the birds, and above
all, the cool breeze that met us, almost persuaded us that the Orient
was but a dream, and that we were traversing an upland in Bonnie
Scotland.
Through a break in the forest we descried our tents, pitched on the
green sward, and ready for our reception, beside a curious-looking
block of masonry. Then sweeping round into the open, we obtained
our first view of Suweida, lying darkly on the farther bank of a little
ravine, by which it was separated from our camping-ground. The
roofs were alive with men straining their eyes in our direction. Our
advent clearly caused no small stir in that remote town. Arriving at
our tents, we found a large company assembled to survey us. They
watched all our movements with an amused curiosity, like that of
children in a menagerie. We were in time to witness the sunset, and
in the calm cool air were tempted to watch how long he took to
disappear, from the instant when his under rim touched the horizon.
We looked earnestly, and seemed relieved when at last he vanished.
Our observers, I am sure, entertained a shrewd suspicion that some
remnants of sun-worship still lingered among these curious
westerns. Little thought they how our hearts followed the departing
beams to the land where, in the slant rays of the longer evening,
dear ones sat musing, drawing vague pictures of regions famed in
sacred story, and praying the Father of all, the light of whose eye
fades not from earth like the passing day, to guard the wanderers
from peril.
CHAPTER VI
Healing the sick—A strange monument—Telegraph and post
in Haurân—Cruel kindness—The Ruins of Suweida—
Turkish methods of rule—ʿIry—Sheyûkh ed-Druze—
Jephthah’s burial—Enterprise of Ismaʾîl el-ʿAtrash.
Here, as at every point touched in our journey, we had ample
evidence of the prevalence of sickness and suffering, and of the
crying necessity for competent medical aid. The weak and diseased
are a prey to every travelling quack, and they bore in their bodies
only too convincing proof of their simple-hearted confidence in men
who professed to be able to relieve them. Ruined eyes and maimed
limbs told only too plainly what havoc unscrupulous men work
among these trustful people. The quack hopes to pass but once in
any given way, and cares but little for the results of his operations if
only he make present gain. The name of the good doctor wrought
like magic. Almost before we could realise it the camp was
surrounded by patients; a motley gathering they were—Moslem,
Druze, and Christian; men, women, and children, of all ages, clad in
richly varied costumes; they came forward, one by one, to tell of their
sufferings, and receive what help was possible. Not unpleasantly the
time passed, examining antique coins, making cautious purchases,
and engaging the more intelligent in conversation about their town
and district, until the cheerful voice of the dinner-bell summoned us
within.
With the morning we were able to see the strange tower under
whose shadow we had slept. It is reputed one of the oldest
monuments in the country. According to inscriptions, Greek on one
side and Palmyrene on another, it was built by one Odainatus to the
memory of his wife Chamrate. The building is over thirty feet square,
and rests on a base, to which a couple of steps lead up. Between the
Doric pilasters that adorn the sides, the monument is ornamented,
as became the tomb of a soldier’s wife, with emblems, in relief, of
military accoutrements. The top of the monument is now a heap of
confused blocks, while many great stones, rolled down, lie in utter
disorder to the south-west. The name Debusîyeh, by which it is
known among the natives—“the pin-shaped”—shows that it was,
probably at no remote period, finished off in a pyramid. The evil that
has befallen it may be due to some thought that buried treasure
might be found there. In these circumstances no structure would be
safe from the destroying hands of the Arabs. It has been thought that
the monument dates from not later than the first century of our era,
and that therefore this Odainatus was not the warrior husband of the
famous Zenobia, ruler of Palmyra. The Odainatus known to history
was in these parts; and there is nothing impossible in the supposition
that the glories of the campaign may have been dimmed for the
chivalrous soldier by the death of his sweet companion, ere the star
of Zenobia arose in the heaven of his love. This would bring the date
down past the middle of the third century. The conjecture is so far
supported by the presence of the inscription in Palmyrene. Withal it
is the most interesting of all the remains of the past now to be seen
in Suweida and its neighbourhood.
Descending the steep bank, we crossed by an ancient bridge the
little stream that flows in the bottom of the ravine. With the advance
of summer this stream soon vanishes, and the town becomes
entirely dependent for water supply on reservoir and cistern. At the
gate of the town we found a little guide who conducted us to the
post-office. The quarters occupied as imperial post and telegraph
office would horrify the humblest of our Western officials. We
scrambled over several dunghills and broken walls, and but for the
telegraph wires it would have been impossible to distinguish the
“office” from a number of rude cattle-shelters around. The maʾmûr,
or official in charge, was all politeness and courtesy. Learning that a
mail was about to be despatched to the north, we set about writing
pencil-notes to our friends, while the maʾmûr, business being slack,
engaged in a conversation by telegraph with his brother operator in
Damascus, securing for us information on several points of
importance. The amount of telegraphing thus done for the friends of
the maʾmûrîn in Syria would not be readily credited in the West. A
message is sent to bring one to the office, when, if nothing special is
on hand, he may hold a long conversation on any subject with his
friend or man of business at a distance. These maʾmûrîn in Syria are
almost all Christians, Moslems possessing the requisite
qualifications in linguistic attainments and intelligence being seldom
available. This speaks volumes for the system of education
inaugurated and carried on chiefly by the missionaries, of which as
yet few Moslems have taken advantage. The position of clerk in very
many of the various Government departments is also occupied by
Christians. Moslems in the country are, however, slowly awakening
to realise the advantages of education, and are seeking in greater
numbers than ever to avail themselves of opportunities hitherto
despised.
The Druze sheikh of the town, who was also kaim makâm, or
lieutenant-governor of the district, we found in his own house near
the top of the quarter at present inhabited. He was in sore distress
over the apparently hopeless illness of his son, a lad of some twenty
summers, who sat suffering among his friends. The room was
crowded in every part by relatives and friends, who had come from
far and near to show their sympathy in the hour of trial. Anything
more completely opposed to all humane and civilised ideas of the
conditions that ought to prevail in a sick-room it is impossible to
imagine. The air was foul with many breaths, and laden with the
fumes of tobacco, in which all seemed to indulge, conversation being
carried on in manner and tone suggestive of the public market; the
dying youth, meanwhile, utterly wearied of the noise and confusion,
with difficulty attracted attention to have his few wants supplied. It
must not be thought that this conduct was the result of exceptional
thoughtlessness on the part of the sheikh’s sympathisers. It was all
done in obedience to custom, whose requirements are far more
stringent than those of written law in this country. The man whose
sick-room is not crowded with hosts of sympathising friends is held in
but little respect. To refrain from mingling with the crowd and adding
a quota to the hubbub is to prove lack of all interest in the case. So
firmly is the custom rooted, that the energetic efforts of enlightened
medical men in many parts have as yet produced almost no
appreciable result. We long for quiet in our time of trial, and true
friends jealously guard against intrusion upon our grief. Here trial
and sorrow must alike be borne practically in presence of the public.
When death enters a household the place is literally taken
possession of by so-called sympathising friends; and their well-
meant endeavours to divert the thoughts of the mourners from their
loss must nearly always have the effect of deepening the woe they
are intended to alleviate.
The sheikh’s house, less squalid perhaps than most in the town, was
built around a paved courtyard, entered from the street by an
imposing doorway. One large room had also a door opening upon
the street, approached by a flight of steps. Here we were entertained
with coffee. As a Government official who had received instructions
from his superiors to receive the travellers with all courtesy, the
sheikh bore himself with no little dignity; and only the haste of our
departure prevented his making a larger display of hospitality. The
diwân of the sheikh stands on the opposite side of the street a little
lower down, on the site of an ancient temple. Many of the columns
which once surrounded the latter are standing still, but serve only to
cast a dreary air of departed glory over the place. A few paces
farther down, the street is spanned by a triumphal arch, of Roman
workmanship. This street is paved throughout. We visited, in rapid
succession, the remains of a church, of a mosque, and of a building
called by the natives el-Mehkemeh—“the court of justice.” All of
these are in a completely ruinous condition. Suweida offers a rich
field for inscription-seekers. Only he who would make thorough work
must be prepared for risks and unpleasantnesses,—in hanging, for
example, over the top of a rickety doorway to read an inscription
placed upside down, or in creeping into holes and cellars where
one’s attention is almost entirely absorbed in the important but well-
nigh impossible process of breathing. Here are also the remains of a
nymphaeum and aqueduct dating from the time of Trajan. Two large
reservoirs afford the chief supply of water, there being no fountains
in or near the town. These are built round with solid masonry, and
the water is reached by means of stone stairs. When the summer is
well advanced, it must require a stout heart and no little usage to
enable one to conquer a natural repugnance to the unwholesome
liquid collected in these reservoirs. I imagine that the memory of the
oldest man does not carry him back to the time when they were last