Recently assessed increases in lower stratospheric water vapour mixing ratio over the last few decades are likely to have caused a decrease in stratospheric temperatures by an amount comparable to that produced by ozone decreases (Forster and Shine, 1999; Smith et al., 2001) (see lower-stratospheric temperature trends in Section 2.2.3). These changes also impact on ozone chemistry (Chapter 4) and on radiative forcing of the atmosphere (Chapters 6 and 7). Data from over twenty-five instruments that measure water vapour concentration and relative humidity in the upper troposphere and stratosphere were recently compared and assessed in the international SPARC study (Kley et al., 2000). The purpose of the study, which included measurements made by both in situ and remote sensing techniques utilising balloons, aircraft and satellites, was to determine the data quality and to estimate the magnitude of any trends. The study showed that some stratospheric instruments have sampled over a long enough period that several overlapping time-series of intermediate length (8 to 15 years) can be used to help evaluate stratospheric changes. A reasonable degree of consistency was found among strato-spheric measurements made from near the tropopause up to as high as 50 km (about 1 hPa). Most observations were within ±10% of the grand mean of all measurements to which they were compared.
Accurate balloon observations of lower-stratospheric water vapour are available from 1964 to 1976 over Washington, D.C. and from 1980 to present over Boulder, Colorado, USA (e.g., Mastenbrook, 1968; Harries, 1976; Mastenbrook and Oltmans, 1983; Oltmans and Hofmann, 1995). The SPARC study shows that these point measurements are nevertheless representative of global stratospheric conditions above about 18 to 20 km, but not of the lowest stratosphere where there can be significant regional and seasonal changes. A positive lower stratosphere trend of about 1 to 1.5%/year in specific humidity (about 0.04 ppm/year) since the mid-1960s is indicated by the balloon data (Oltmans et al., 2000). The increase was not monotonic but showed several rapid rises with plateaux in between. Even though the recent satellite record is relatively short, these measurements have revealed changes of the same character. The satellite results show a spatial pattern of trends in the lower stratosphere, and suggest a slowing in the positive trend after 1996 (Smith et al., 2000). Although not definitive, these observations are consistent in suggesting that lower-stratospheric water vapour has increased globally on average at about 1%/year over at least the past forty years, but at a variable rate.
Although radiosondes have made observations of water vapour in the upper troposphere (i.e., above 500 hPa) since the 1950s, these observations have suffered from instrumental errors (Elliott and Gaffen, 1991). Peixoto and Oort (1996) have re-examined these observations for the period 1974 to 1988 and found large trends in upper-tropospheric humidity at the 300 hPa level. They concluded that these trends were unrealistically large and were likely to be due to instrument changes. Satellite observations of upper-tropospheric humidity (UTH) measurements made by TOVS (Television infrared observation satellite Operational Vertical Sounder) since 1979, and representative of a deep layer between 500 to 200 hPa, show very large interannual variability (Bates et al., 1996). The SPARC assessment of these observations (Kley et al., 2000) indicated that they were of sufficient quality for trend analyses. The SPARC study and an analysis by Bates and Jackson (2001) show large regional trends that are attributed to circulation changes associated with ENSO, decadal variability over equatorial Africa, and decadal variability of the Arctic Oscillation (see Section 2.6). Statistically significant positive trends of 0.1%/year are found for 10°N to 10°S, and a non-significant trend of 0.04%/year for 60°N to 60°S, but this includes a component negative trend of -0.1%/year for 30oS to 60oS. The trends in large zonal bands tend to be residuals from cancellations in sign and magnitude of much larger regional trends. These UTH trends should be treated with caution especially in the deep tropics because of significant interannual variability and persistence, both of which hamper trend detection.
In summary, in situ and radiosonde measurements tend to show increasing water vapour in the lower troposphere and near the surface, though this is not seen everywhere, and data quality is still an issue. The longer, more reliable data sets suggest multi-decadal increases in atmospheric water vapour of several per cent per decade over regions of the Northern Hemisphere. New analyses of balloon and satellite records indicate that strato-spheric water vapour above 18 km shows an increase of about 1%/year for the period 1981 to 2000 but with a slowing of the positive trend after 1996. Satellite observations of upper-tropo-spheric humidity from 1980 to 1997 show statistically significant positive trends of 0.1%/year for the zone 10°N to 10°S.
Other reports in this collection