180
W.B. Rossow
stratospheric aerosols (see references in lraci et al., 1995) and about half of tropospheric aerosols
(Charlson. 1993), (2) stratospheric aerosols, in turn, playa key role in stratospheric ozone loss
processes (Solomon. 1988). and (3) the OH radical from the decomposition of water vapor plays
key catalytic roles in the chemistry of several trace gases including ozone and methane.
8.2 Measurement Methods
Routine water vapor abundance measurements are made as part of weather observations by
hygrometers at surface weather stations and in rawinsonde balloons (called RAOBS for short)
that are launched a few times every day (Elliott and Gaffen, 1991). These are direct in situ
measurements that are quite accurate at relatively high temperatures and vapor abundances.
There are several ways to make this measurement; the most common are to measure the
temperature difference between a dry and a moist, evaporating thermometer, to measure the
change in size of a sensitive element caused by moisture absorption. to measure the temperature
at which condensation begins, and to measure changes in electrical resistance caused by moisture
absorption. The most common instruments today use the last method. These data, collected
for more than 30 years at hundreds of locations, have provided most of what we know about
the distribution and variations of water vapor in the atmosphere (Peixoto and Oort, 1992).
However, there are three shortcomings of these observations. (1) Current instrument sensitivity
significantly reduces the reliability of measurements of water vapor abundance in the upper
troposphere (below temperatures of -40°C, cf. Elliott and Gaffen, 1991), particularly in the
tropics (Rind et al., 1992), and in the whole polar troposphere, particularly in wintertime. (2)
Geographic coverage of rawinsonde balloon launch sites is poor, particularly over oceans. (3)
Most sites launch only two balloons per day at synoptic reporting times, which does not provide
a longitudinally consistent nor complete representation of diurnal variations in water vapor.
The advent of weather satellites in the 1960's-1970's offered the possibility of making remote
sensing water vapor measurements that are more complete than obtained from surface weather
stations; however, this opportunity has not generally been exploited. Routine water vapor abundance profiles have been produced since 1979 from multi-wavelength infrared measurements
by the High Resolution Infrared Radiation Sounder (HIRS) flown on the polar orbiting weather
satellites operated by the National Oceanic and Atmospheric Administration (NOAA) in the
United States (Smith and Woolf, 1976; Smith et al., 1979). Although providing global coverage
(with somewhat reduced vertical resolution), these measurements are no more reliable in the
upper tropical and polar troposphere than are the RAOBS: no results are reported at pressures
below about 300 mb. Even though NOAA has generally operated two polar orbiters, the analyzed dataset does not provide measurements more frequently than about once per day. Hence,
satellite water vapor measurements have only improved geographic coverage. A re-analysis of
this dataset could provide more information about diurnal variations of temperature and water
vapor.
Water vapor abundances can be determined from radiation measurements by determining the
amount of absorption by water vapor at wavelengths selected to be relatively free of other
effects. Sunphotometers on the surface, which point directly at the sun, can determine the total
amount of water vapor in the column of atmosphere by comparing the strength of radiation
in and out of absorption bands in the water vapor spectrum (Figure 8.4). This technique is
not generally used from satellites because the surface absorbs most of the sunlight inside and
outside the water vapor bands, greatly reducing the difference in radiation with wavelength.
At infrared wavelengths, water vapor profiles can be inferred from the amount of emitted
radiation observed at wavelengths where the absorption strength varies from weak to strong
W.B. Rossow
stratospheric aerosols (see references in lraci et al., 1995) and about half of tropospheric aerosols
(Charlson. 1993), (2) stratospheric aerosols, in turn, playa key role in stratospheric ozone loss
processes (Solomon. 1988). and (3) the OH radical from the decomposition of water vapor plays
key catalytic roles in the chemistry of several trace gases including ozone and methane.
8.2 Measurement Methods
Routine water vapor abundance measurements are made as part of weather observations by
hygrometers at surface weather stations and in rawinsonde balloons (called RAOBS for short)
that are launched a few times every day (Elliott and Gaffen, 1991). These are direct in situ
measurements that are quite accurate at relatively high temperatures and vapor abundances.
There are several ways to make this measurement; the most common are to measure the
temperature difference between a dry and a moist, evaporating thermometer, to measure the
change in size of a sensitive element caused by moisture absorption. to measure the temperature
at which condensation begins, and to measure changes in electrical resistance caused by moisture
absorption. The most common instruments today use the last method. These data, collected
for more than 30 years at hundreds of locations, have provided most of what we know about
the distribution and variations of water vapor in the atmosphere (Peixoto and Oort, 1992).
However, there are three shortcomings of these observations. (1) Current instrument sensitivity
significantly reduces the reliability of measurements of water vapor abundance in the upper
troposphere (below temperatures of -40°C, cf. Elliott and Gaffen, 1991), particularly in the
tropics (Rind et al., 1992), and in the whole polar troposphere, particularly in wintertime. (2)
Geographic coverage of rawinsonde balloon launch sites is poor, particularly over oceans. (3)
Most sites launch only two balloons per day at synoptic reporting times, which does not provide
a longitudinally consistent nor complete representation of diurnal variations in water vapor.
The advent of weather satellites in the 1960's-1970's offered the possibility of making remote
sensing water vapor measurements that are more complete than obtained from surface weather
stations; however, this opportunity has not generally been exploited. Routine water vapor abundance profiles have been produced since 1979 from multi-wavelength infrared measurements
by the High Resolution Infrared Radiation Sounder (HIRS) flown on the polar orbiting weather
satellites operated by the National Oceanic and Atmospheric Administration (NOAA) in the
United States (Smith and Woolf, 1976; Smith et al., 1979). Although providing global coverage
(with somewhat reduced vertical resolution), these measurements are no more reliable in the
upper tropical and polar troposphere than are the RAOBS: no results are reported at pressures
below about 300 mb. Even though NOAA has generally operated two polar orbiters, the analyzed dataset does not provide measurements more frequently than about once per day. Hence,
satellite water vapor measurements have only improved geographic coverage. A re-analysis of
this dataset could provide more information about diurnal variations of temperature and water
vapor.
Water vapor abundances can be determined from radiation measurements by determining the
amount of absorption by water vapor at wavelengths selected to be relatively free of other
effects. Sunphotometers on the surface, which point directly at the sun, can determine the total
amount of water vapor in the column of atmosphere by comparing the strength of radiation
in and out of absorption bands in the water vapor spectrum (Figure 8.4). This technique is
not generally used from satellites because the surface absorbs most of the sunlight inside and
outside the water vapor bands, greatly reducing the difference in radiation with wavelength.
At infrared wavelengths, water vapor profiles can be inferred from the amount of emitted
radiation observed at wavelengths where the absorption strength varies from weak to strong