Remote Sensing of Atmospheric Water Vapor
181
(Figure 8.5), if the atmospheric temperature profile is known (e.g., Stephens, 1994). A similar
technique can be applied at microwave wavelengths (e.g. Janssen, 1993). The infrared and
microwave methods can be used both from the surface and from satellites. Infrared-based
determinations of water vapor are limited by the occurrence of clouds that block the emissions
from lower levels; hence, the available observations represent "clear" conditions (e.g. Wittmeyer
and Vonder Haar, 1994), although the differences of total water vapor abundances between clear
and cloudy conditions are not too large (Gaffen and Elliott, 1993). Note that, although the
operational analysis of the infrared measurements from HIRS produces an estimate of water
vapor abundance profiles, the analysis procedure does not actually make use of direct infrared
radiation measurements as described here. Rather the method relates water vapor amounts
to temperatures by an empirical relationship developed from comparisons to RAOBS (Smith
and Woolf, 1976). Microwave observations available from current satellites limit retrievals to
a single wavelength which may be affected by the presence of precipitating systems; hence
only the total column amount of water vapor can be estimated (e.g. Tjemkes et aI., 1991).
Strong effects of land surface emissions have limited microwave studies to ocean areas outside
of sea ice-covered regions. Within the past year a new multi-wavelength microwave instrument
(called SSM/T2) has been flown, the first of an operational series that retrieve water vapor
abundance profiles in the lower troposphere (Wilheit, 1990). Comparisons of monthly mean
maps of total column water vapor amounts from RAOBS, HIRS and retrieved from the Special
Sensor Microwave/Imager (SSM/I) flown on military weather satellites shows agreement to
within about 20% (Stephens et aI., 1994; Wittmeyer and Vonder Haar, 1994).
Several experimental water vapor measurement techniques have been studied though none
has entered regular use. From the surface, use of a Raman lidar technique shows promise
of providing very high vertical and time resolution observations of the vertical distribution
of water vapor over individual surface sites (e.g. Melfi et aI., 1989). To measure from space
the very low water vapor amounts in the stratosphere (about 10- 3 to 10- 4 the amount of
water near the surface), both the infrared and microwave techniques have been employed but
with the instrument viewing Earth's atmosphere at the limb instead of pointed towards the
surface: SAGE II used near-infrared measurements (Rind et aI., 1992) and the Microwave Limb
Scanner on the Upper Atmosphere Research Satellite uses microwave measurements (Barath et
aI., 1993). Because of a lack of accurate water vapor measurements in the upper troposphere,
there have been several studies of the possible utility of radiation measurements at 6.7 Jim
wavelength from satellites to infer the total water vapor amount in the upper atmosphere (van
de Berg et a!', 1993; Wu et a!., 1993,; Soden and Bretherton, 1994; Schmetz et aI., 1995;
Stephens et aI., 1995; Bates et aI., 1995).
8.3 Current Research Challenges
Accurate satellite determinations of sea surface temperatures and other quantities using radiation measured at infrared wavelengths are still hindered by uncertainties in the absorption
by water vapor, particularly the so-called continuum absorption (cf. Figure 8.5). Retrieval of
water vapor abundances at higher altitudes and latitudes are also affected by uncertainties in
the absorption line spectrum of water vapor (cf. Stephens et aI., 1995). Moreover, these uncertainties limit the accuracy of determining the precise effect of water vapor changes on the
radiation balance in the atmosphere. There are some empirical models (e.g. Clough, 1993) that
fit available laboratory observations reasonably well (though these laboratory measurements do
not cover enough wavelength and temperature range and are particularly lacking at the temperatures encountered in the atmosphere) and, at least, two theories for the physical process that
creates the continuum absorption. Recent work (Ma and Tipping, 1991; 1992; 1994) provides a
181
(Figure 8.5), if the atmospheric temperature profile is known (e.g., Stephens, 1994). A similar
technique can be applied at microwave wavelengths (e.g. Janssen, 1993). The infrared and
microwave methods can be used both from the surface and from satellites. Infrared-based
determinations of water vapor are limited by the occurrence of clouds that block the emissions
from lower levels; hence, the available observations represent "clear" conditions (e.g. Wittmeyer
and Vonder Haar, 1994), although the differences of total water vapor abundances between clear
and cloudy conditions are not too large (Gaffen and Elliott, 1993). Note that, although the
operational analysis of the infrared measurements from HIRS produces an estimate of water
vapor abundance profiles, the analysis procedure does not actually make use of direct infrared
radiation measurements as described here. Rather the method relates water vapor amounts
to temperatures by an empirical relationship developed from comparisons to RAOBS (Smith
and Woolf, 1976). Microwave observations available from current satellites limit retrievals to
a single wavelength which may be affected by the presence of precipitating systems; hence
only the total column amount of water vapor can be estimated (e.g. Tjemkes et aI., 1991).
Strong effects of land surface emissions have limited microwave studies to ocean areas outside
of sea ice-covered regions. Within the past year a new multi-wavelength microwave instrument
(called SSM/T2) has been flown, the first of an operational series that retrieve water vapor
abundance profiles in the lower troposphere (Wilheit, 1990). Comparisons of monthly mean
maps of total column water vapor amounts from RAOBS, HIRS and retrieved from the Special
Sensor Microwave/Imager (SSM/I) flown on military weather satellites shows agreement to
within about 20% (Stephens et aI., 1994; Wittmeyer and Vonder Haar, 1994).
Several experimental water vapor measurement techniques have been studied though none
has entered regular use. From the surface, use of a Raman lidar technique shows promise
of providing very high vertical and time resolution observations of the vertical distribution
of water vapor over individual surface sites (e.g. Melfi et aI., 1989). To measure from space
the very low water vapor amounts in the stratosphere (about 10- 3 to 10- 4 the amount of
water near the surface), both the infrared and microwave techniques have been employed but
with the instrument viewing Earth's atmosphere at the limb instead of pointed towards the
surface: SAGE II used near-infrared measurements (Rind et aI., 1992) and the Microwave Limb
Scanner on the Upper Atmosphere Research Satellite uses microwave measurements (Barath et
aI., 1993). Because of a lack of accurate water vapor measurements in the upper troposphere,
there have been several studies of the possible utility of radiation measurements at 6.7 Jim
wavelength from satellites to infer the total water vapor amount in the upper atmosphere (van
de Berg et a!', 1993; Wu et a!., 1993,; Soden and Bretherton, 1994; Schmetz et aI., 1995;
Stephens et aI., 1995; Bates et aI., 1995).
8.3 Current Research Challenges
Accurate satellite determinations of sea surface temperatures and other quantities using radiation measured at infrared wavelengths are still hindered by uncertainties in the absorption
by water vapor, particularly the so-called continuum absorption (cf. Figure 8.5). Retrieval of
water vapor abundances at higher altitudes and latitudes are also affected by uncertainties in
the absorption line spectrum of water vapor (cf. Stephens et aI., 1995). Moreover, these uncertainties limit the accuracy of determining the precise effect of water vapor changes on the
radiation balance in the atmosphere. There are some empirical models (e.g. Clough, 1993) that
fit available laboratory observations reasonably well (though these laboratory measurements do
not cover enough wavelength and temperature range and are particularly lacking at the temperatures encountered in the atmosphere) and, at least, two theories for the physical process that
creates the continuum absorption. Recent work (Ma and Tipping, 1991; 1992; 1994) provides a