3.1 Introduction

Knowledge of the distribution and temporal variability of water vapour in the UT/LS is important for a variety of current scientific problems. However, our knowledge shows some surprising gaps, given the wide range of observational and modelling power now available to us. Previous attempts have been made to summarise and understand this time-dependent distribution (e.g. see the book by Ludlam, [1980]; see also, Harries, [1976]; Peixoto and Oort, [1992]), with some success, though the fundamental questions of precisely what mechanisms control the distribution of water vapour in the UT/LS have in many cases remained unsolved. This chapter brings together a summary of our understanding of the mechanisms that control the distribution and variability of water vapour in the UT/LS, based on the findings of Chapters 1 and 2. The remainder of this Introduction illustrates some of the scientific issues of radiation, dynamics and chemistry that depend on this understanding.

3.1.1 Radiation

The principal impact of water vapour on the radiative balance of the Earth is through the long-wave, infrared part of the spectrum. Water vapour has an intense pure rotation band stretching from about 50 to 400 cm-1 (200-25 m m wavelength), and a strong n 2 vibration-rotation band centred at about 1590 cm-1 (» 6.3 m m), extending for over 100 cm-1 in both directions [Goody, 1964]. In addition, water vapour exhibits the curious ‘continuum’ absorption across wide reaches of the infrared spectrum, being most noticeable in the ‘window’ region between about 800 and 1250 cm-1 (12.5-8.0 m m; see Bignell, [1970]; Burch, [1981]). Because of the strong absorption and emission by these bands, water vapour is the dominant greenhouse gas in the atmosphere. Figure 3.1 shows a calculation of the spectrally resolved greenhouse parameter for a tropical and a sub-Arctic winter model atmosphere [Brindley and Harries, 1998]. The greenhouse parameter shows the difference between the flux of radiation leaving the Earth’s surface (at an assumed temperature), and the flux leaving the top of the atmosphere. This is a measure of the energy trapped in the earth system by the atmosphere, as a function of wave number. The importance of the pure rotation band of water vapour (50 to 400 cm-1), highlighted in work by Doherty and Newell [1984] and Clough et al., [1992], is clear from this diagram. Figure 3.2, also taken from work by Brindley [1998] shows the cooling rate for the same two model atmospheres, showing the importance of the far infrared emission to space, between about 100 and 500 cm-1 (100-20 m m). This cooling is largest in the mid to upper troposphere. Clearly, this mechanism of planetary cooling is important in maintaining the energy balance of the Earth, and so is of great importance in climate models. However, due to the considerable uncertainty over the concentration and variability of water vapour in the upper troposphere, it is unclear at present whether or not these models correctly represent these cooling rates. More work is required to establish upper tropospheric concentrations and variability, and to measure radiative flux divergences in situ in the atmosphere.

Current radiative problems that require an accurate knowledge of water vapour distributions include:

 

Figure 3.1. Water vapour greenhouse effect as a function of wave number for an atmospheric profile typical of (top) the tropics and (bottom) sub-Arctic winter. The dashed curve in each panel shows the Planck blackbody radiation for the indicated temperature, and the solid curve shows downward long-wave emission from the atmosphere.

Figure 3.2. Variation of long-wave cooling rates with pressure and wave number for the Tropical (top) and sub-Arctic winter (bottom) model atmospheres depicted in Figure 3.1. Adapted from Brindley and Harries [1998]. Units of K per day per wave number.

In all these areas, key uncertainties include an accurate knowledge of the distribution, variability and trends of water vapour, and how these effects interact with the radiation field and the climate.

3.1.2 Dynamics

Water vapour distributions are strongly influenced by atmospheric dynamics, but also influence them in turn. Thus, observed water vapour distributions can serve as useful diagnostics of the atmospheric circulation. Water vapour condensation also provides a substantial reservoir of heat in the atmosphere, so that accurate knowledge of water vapour is also mandatory for accurate predictions of the future development of the circulation.

The diagnostic applications of water vapour observations are clear. First, imagery of upper tropospheric water vapour now available from geostationary satellites demonstrate structures which have been used to improve our understanding of the general circulation of the middle and upper troposphere [e.g. Woodberry et al., 1991]. Nimbus 7 provided quantitative near-global measurements of lower stratospheric humidity for the first time [Gille and Russell, 1984; Russell et al., 1984; Remsberg et al., 1984]. The nearly one decade of measurements of water vapour in the lower stratosphere from the HALOE instrument on UARS have been particularly valuable [Russell et al., 1993; Harries et al., 1996], as have the long time series of balloon frost point measurements by Oltmans, and before him, Mastenbrook [Oltmans and Hofmann, 1995].

Conversely, the influence of water vapour on dynamics is also well recognised by forecasters and climatologists. The atmospheric temperature structure would be dramatically different if not for water vapour condensation. Storm initiation and development have also been shown to depend on water vapour abundance, not only in the boundary layer [Normand, 1938] but also in the middle and perhaps upper troposphere [e.g. Stommel, 1947, Peppler and Lamb, 1989]. The ability of vapour to influence cloud populations also gives it a strong indirect influence on large-scale dynamics, including those in the UT/LS.

Despite this progress, the dynamical processes that control the dryness of the stratosphere have remained enigmatic for more than three decades, despite a huge volume of work, much too voluminous to list in this report with any adequacy. The original Brewer-Dobson hypothesis [Brewer, 1949, Dobson et al., 1946], and other early pioneering work [e.g. Mastenbrook, 1968] remains the foundation of thinking, but many other subtle effects have been identified. These include:

Despite the huge progress achieved, substantial uncertainties exist. Most prominent, of course, is the remaining uncertainty over the specific processes that keep the stratosphere so dry! Behind this and many related questions lies the difficulty that the large scale phenomenon that we observe ? the dry stratosphere - is probably due to the collective effect of one or more small scale processes occurring over a large regime of the tropopause region (see section 3.3.3). The processes that control the dryness of the stratosphere are almost certainly smaller than the scales represented in our models, and also difficult to parameterise. They are also extremely difficult to observe, maybe because of this small scale, or simply because of the difficulty of their detection (e.g. detecting the drying effect of penetrating thunderclouds).

3.1.3 Chemistry

Water vapour is important for atmospheric chemistry in several ways. It is the source of OH, the hydroxyl radical, in both the troposphere and the stratosphere. OH is of direct importance in many chemical cycles in both regions [Brasseur and Solomon, 1984, Wayne, 1985]. OH and more generally hydrogen oxides (HOX = OH + HO2) take part in an important catalytic cycle for regulating the production and destruction of ozone in both the troposphere and stratosphere [Wennberg et al., 1994]. OH also controls the oxidising capacity (cleansing) of the atmosphere for short lived gases and regulates the lifetimes of the longer lived species such as CO and CH4. Thus, it is important to understand the dynamics driving the concentration and distribution of water vapour, as well as the chemical and photochemical reactions transforming water vapour into OH.

The budget of hydrogen in the stratosphere has been studied by a number of workers, including Le Texier et al. [1988], the satellite work already mentioned by Jones et al. [1986], and by Pyle et al. [1983]. These papers, and many others, have discussed the oxidation of methane to produce water vapour, as an in situ stratospheric source of humidity.

In this chapter we discuss the variability of water vapour in the upper troposphere and lower stratosphere using the observations discussed in Chapters 1 and 2. We start with the annual mean distribution in section 3.2. The seasonal cycle, which is the dominant time scale of variability in both the UT and LS is discussed in section 3.3. In section 3.4 we discuss longer (interannual) and shorter (intraseasonal) variations in water vapour, along with small scale or ‘transient’ variations. Finally, in section 3.5 we discuss long-term variations in the distribution of water vapour in the upper troposphere and lower stratosphere. These various time and space scales require using a full suite of in situ, ground-based and satellite observations of water vapour. This analysis focuses on those data sets with significant time and space resolution to cover the various modes of variability. As a result, extensive use is made of satellite data where possible. In situ data are also shown where appropriate. Having available the analysis in chapter 2, we show only those data that best illustrate the variability discussed.