3.6 Conclusions

Water vapour is unique among atmospheric trace constituents in that conditions for saturation are common in the atmosphere. This property is the most important factor governing the distribution of water vapour in the atmosphere, both in the troposphere, where it varies by as much as four orders of magnitude in a vertical profile, and in the stratosphere, where variations are much smaller but still significant.

The global time-mean distribution and large-scale variations of water vapour are fairly well characterised by satellite data sets, especially HALOE and MLS (in the stratosphere) and MLS and HIRS (in the upper troposphere). In situ and ground-based data sets augment satellite information and present a picture consistent with satellite observations. In situ and ground-based data sets are also essential for revealing behaviour at smaller spatial scales, for long-term monitoring and for validation of satellite data sets.

In the stratosphere above approximately 100 hPa, the distribution of water vapour can be explained as a balance between dry air entering via the tropical tropopause and a source of water vapour from methane oxidation in the upper stratosphere. The "Brewer-Dobson" circulation then helps determine the distribution, along with wave-induced mixing and (in the lowest few kilometres of the stratosphere) upward extension of tropospheric circulation. Variations in the zonal direction are rapidly mixed so that water vapour is nearly constant following a fluid element. Nearly all air passing from the troposphere to the part of the stratosphere above 100 hPa enters through the tropical tropopause, where the removal of water vapour by low temperatures and a combination of other processes dries the air to around 3.5-4 ppmv in the annual mean. Some of this dry air rises slowly in the tropics, but most spreads poleward, or is mixed with mid-latitude air, especially in the lowest few kilometres of the stratosphere. Consequently, water vapour concentrations increase upward and away from the equator as methane is oxidised. Below approximately 100 hPa, the extratropical lower stratosphere is moistened by leakage from the troposphere, mostly by roughly horizontal transport across the subtropical tropopause. This horizontal transport has a strong seasonal component, being much stronger in the summer hemisphere. There is also a hemispheric asymmetry in the transport, with the South Asian monsoon in boreal summer significantly moistening the Northern Hemisphere, more than similar monsoon circulation in the Southern Hemisphere. Other important seasonal variations in the stratosphere occur in the winter and spring polar vortex, especially the Antarctic, where cold temperatures cause dehydration via the formation of ice clouds (which play a pivotal role in catalytic ozone destruction every spring).

In the tropics and subtropics, upper-tropospheric water vapour is strongly influenced by the three-dimensional time-mean circulation whose meridional projection is the Hadley circulation and whose zonal projection is the Walker circulation. The predominant source for moisture in the tropical and subtropical upper troposphere is convection, so that, on average, moist areas appear in the convective areas over the western Pacific, South America and Africa. Moist areas also appear seasonally in the region of the Asian summer monsoon and along the intertropical and South Pacific convergence zones. The seasonality of surface temperature and of convection, which roughly follow the sun, as well as seasonal variations in monsoon circulation, produce concomitant seasonal changes in water vapour in the troposphere. This relationship between convection and upper tropospheric moisture changes sign near the tropical tropopause, somewhere between 150 hPa and 100 hPa, so that convection dries the tropopause region. Water vapour is also influenced by fluctuations at both shorter and longer time scales, including the quasi-biennial oscillation in the stratosphere, and El Niño and the Southern Oscillation and the Tropical Intraseasonal Oscillation in the troposphere.

In middle to high latitudes in the upper troposphere, water vapour is highly variable and can be supplied by transport from the tropics, by mesoscale convection, or by extratropical cyclones. Dry air can be transported from the subtropics or from the extratropical lower stratosphere. These transport phenomena tend to be episodic rather than steady.

At the tropical tropopause, a complex mix of processes act to remove water vapour from air as it enters the stratosphere. Within the framework of large-scale mean ascent, the dehydration processes probably include smaller-scale ascent (convective and possibly non-convective), radiative and microphysical processes within clouds, and wave-driven fluctuations in temperature. The location, strength, and relative importance of these processes vary seasonally, but the seasonal variation in tropopause-level water vapour is influenced most of all by the seasonal variation in tropical tropopause temperatures. Air rising through the tropopause is marked with seasonally varying mixing ratio, and retains these markings as it spreads rapidly poleward and more slowly upward into the stratosphere.

In the upper troposphere and occasionally near sharp gradients in the stratosphere, filaments of water vapour with scales from tens to thousands of kilometres are prevalent. Sharp transitions between moist and dry air masses are produced naturally in the atmosphere by turbulent processes, and complicate quantitative comparisons between imperfectly collocated observations.

3.6.1. Long-Term Variations

Because water vapour is a radiatively active and chemically important gas, its future concentration will have an important influence on future climate. Measurements of water vapour in the lower stratosphere indicate increases over the last 20 years of about 20%. Older measurements imply a 50% increase in water vapour in the lower stratosphere since the 1950s. Unfortunately, projections of future concentrations of water vapour in the UT/LS are hampered by several factors, most notably the lack of definitive understanding of the causes of these long-term changes. Also, as discussed in chapter 2, the instrumental record is not of sufficient quality to combine records across instruments to derive overall long-term changes for water vapour in the lower stratosphere.

In the upper troposphere, long-term relative humidity observations were made as part of routine meteorological soundings, but are of such poor quality as to be unusable for detection of change. Satellite measurements since 1979 show a small but insignificant increase in relative humidity. Satellite measurements of specific humidity have been made only since late 1991, and show no significant change. Changes of specific humidity can be derived from measurements of relative humidity and of temperature, both of which show small increases since 1979, implying increases of specific humidity, but the combined uncertainty of these two measurements is such that the uncertainty in specific humidity is large and masks any changes of significance to climate. Total column water vapour, which is dominated by the lower troposphere, has increased over much of the Northern Hemisphere in the past 20-30 years.

In the lower stratosphere, satellite measurements of sufficient quality and continuity for detection of long-term variations began much more recently. The longest continuous, reliable data set is at a single location (Boulder, Colorado) and dates back only to 1981. Over this period, variations in stratospheric water vapour are consistently positive with a value of about 1% year-1 at all levels between about 10 and 28 km, and these increases are statistically significant above 16 km. While a linear trend can be fit to this data, increases in the record are neither continuous nor steady. Global variations calculated over a much shorter period using the HALOE data set show an increase of 2% year-1 over several years (1993-1996) but also show a change in the sign of the variations around 1996. This underscores the limitations of using data records of this length for studying long-term linear trends.

Our understanding of factors controlling long-term changes in UT/LS water vapour is inadequate to explain the observed variations or to make good projections into the future. In the lower stratosphere, the observed 20-year increase at Boulder (and probably globally) is about twice what would be expected from methane oxidation alone. A simple explanation would be an increase in average saturation mixing ratio at the tropical tropopause; changes in this quantity are difficult to calculate, but appear to be in the opposite direction [e.g., Simmons et al., 1999]. Thus, it is likely that the expected increases in the concentration of atmospheric methane will lead to further increases in stratospheric water vapour, but the inability to explain past changes diminishes our confidence in the projection of future increases in water vapour.

It is fairly clear which processes (transport, convection, and clouds) are involved in determining the distribution of upper tropospheric water vapour, but their influences are very difficult to quantify. Since it is also difficult to predict how these might change in response to natural and human-induced climate change, future changes in the distribution and variability of upper tropspheric water vapour are unknown.

The combination of these uncertainties on the future of the water vapour content of the upper troposphere and lower stratosphere means that using available data we cannot answer fundamental questions about the influence of UT/LS water vapour on future climate.

3.6.2. Recommendations

  1. Upper tropospheric specific humidity should be monitored with a view to determining long-term variations. It is important to have complementary observations, not relying solely on one instrument or approach.
  2. Process studies of upper tropospheric water vapour and convection should be undertaken. These would include joint measurements of water vapour, cloud microphysical properties, and chemical species that can provide a history of the air.
  3. More observations of the tropical tropopause region (15-20 km), by both in situ and remote sensing methods, are needed in order to improve our understanding of stratosphere-troposphere exchange there.
  4. Stratospheric water vapour should be monitored at various latitudes. Maintaining current long-term measurement programs is necessary for any interpretation of long-term change. These measurements would have greater value if methane were simultaneously measured. Monthly balloon measurements like those over Boulder would be extremely useful for validation of satellite instruments in the upper troposphere and lower stratosphere and for measurement of long-term variations. Future satellite instruments should be planned so as to overlap with existing instruments.
  5. Further theoretical work is needed to make best use of existing observations. For example, the variations observered in the HALOE data set around 1996 need to be better understood.
  6. All the data used in this Assessment should be kept in a digital archive at the SPARC data centre. The purpose of this archive would not be on-line instant access, but research access to data in the future (e.g., for subsequent assessment activities). Therefore, data sets collected in the future should be added to this archive. Valuable data from the 1940’s, 1950’s, and 1960’s may already be lost, but some could and should be rescued.