This thesis exploits two water vapour products from the Microwave Limb Sounder (MLS), an instrument on the Upper Atmosphere Research Satellite, to investigate the spatial distribution and temporal variation of water vapour in the tropical region of the upper troposphere and lower stratosphere. Measurements are centred on 215 hPa in the upper troposphere and at 68 hPa and 100 hPa in the lower stratosphere. MLS provides almost global measurements with good vertical resolution, and offers many advantages over other measurement techniques. Focusing on the tropical region, time-series analysis techniques have been employed to investigate the variability of upper tropospheric water vapour on different timescales. The construction of longitude-time sections has allowed the annual cycle to be investigated in more detail than previously. The amplitude is found to vary greatly with longitude and to be most pronounced over the land-masses of South America and Africa. The oceans are found to have a moderating effect on the annual cycle but secondary maxima occur in regions influenced by the Asian monsoon, and over Indonesia and the central Pacific. An intraseasonal cycle with a period of 30-85 days is evident over the western Pacific at latitudes from 10°N to 20°S. This is associated with eastward moving moist features apparent in longitude-time sections. Through comparison with meteorological fields from the European Centre for Medium Range Weather Fore-casting's re-analysis project, the intraseasonal cycle in water vapour is identified with the convective anomalies associated with the Madden-Julian Oscillation.

The AVAS and ACE techniques allow us to model the possible nonlinearity suggested by linear models. Comparison of AVAS and ACE led us to concentrate on application of AVAS, for theoretical and practical reasons. Results indicate that the effect of precession is linear, with perihelion near vernal equinox corresponding with warmer climate. Obliquity's effect is also generally linear, though with a stronger warming for very high values. The role of eccentricity is more difficult to interpret. Northern hemisphere summer insolation variations were prescribed as a regression variable but proved to be redundant. Interaction modelling, using Generalised Additive Models, allows us to visualise the combined effect of two orbital variables. This confirms the combined role of high obliquity and vernal perihelion in the deglaciations, and also permitted the last 200 ka to be reproduced independently of eccentricity. The cause of the deglaciations, and therefore the Earth's 100 ka ice age cycle, is the combined warming effect due to precession and obliquity variations (during a glacial period), and not eccentricity (other than through its modulation of the precession effect). We therefore conclude that the Earth's ice age cycle operates as follows: • The Earth naturally accumulates ice; • When the ice sheets have reached a certain threshold size then their feedback mechanisms become strong enough to precipitate a glacial collapse; • This collapse is triggered by the warming pulse which results from high obliquity and autumnal perihelion, possibly due to the extended summer experienced in the Northern hemisphere during such an orbital configuration.

Interhemispheric differences in the zonal mean of water vapour mixing ratios are studied on equivalent latitude. Results indicate that the southern vortex of late-winter 1992 was fairly well contained at 530K, and not acting like a "flowing processor". The northern vortex of late-winter 1993 also appears to have been well contained, but a significant amount of air appears to have been mixing across the edge of the northern vortex of late-winter 1992. The southern vortex of late-winter 1992 is shown to contain significant internal structure in the lower and middle stratosphere, unlike the northern vortices which appear well mixed. The evolution of the southern vortex is studied over the 35 days of observations available during southern late-winter. The structure inside the southern vortex in the lower and middle stratosphere degrades almost completely over this period. There is a significant change in the difference between the average mixing ratios in mid-latitudes and those of the vortex mixing in the middle-upper stratosphere. This is interpreted as being due to mixing of mid-latitude air into the vortex, on a timescale with an upper bound of ˜50-70 days. The variation of the meridional gradient in PV at the equator is studied, and compared with the variation of zonal wind and water vapour mixing ratio. The strength of meridional gradients in PV at the equator is shown to be directly related to the meridional shear in zonal wind, as modulated by the QBO. In the upper stratosphere, meridional PV gradients at the equator are related to SAO winds. Significant cross-equatorial differences in the signal of the "tropical tape-recorder" are revealed in water vapour mixing ratios. These cross-equatorial differences indicate there may be significant inhibition of mixing across the equator, but they are not obviously dependent on the strength of the meridional PV gradient.

Recent observations of trace constituents were used as a basis for comparing the circulation of the UGAMP General Circulation Models with that of the real atmosphere. Measurements of two different constituents were used; upper-tropospheric water vapour and the volcanic aerosol injected into the stratosphere by Mt. Pinatubo. The study of upper-tropospheric water vapour in the UGCM was motivated by reports of an interhemispheric asymmetry in aircraft measurements; the northern winter hemisphere contained 2 to 4 times more <I>H<SUB>2</SUB>O</I> than the southern winter hemisphere [<I>Kelly et al., 1991</I>]. The UGCM did exhibit an interhemispheric asymmetry in water vapour fields between 200 and 300mb, but to a lesser degree than claimed by <I>Kelly et al.</I> A further analysis of water vapour in the UGCM revealed that, due to the longitudinal variability of the distribution, a different choice of measurement sites could significantly alter claims made about interhemispheric asymmetry. Confidence in the severity of the asymmetry was further decreased after examining MLS retrievals of <I>H<SUB>2</SUB>O</I> AT 215MB. The MLS measurements, with a near global coverage, showed very little difference in water vapour mixing ratios between the two winter hemispheres. The UGCM and EUGCM were used to simulate the dispersal of the Mt. Pinatubo aerosol cloud. With a lifetime of years in the stratosphere a well documented evolution, the Pinatubo cloud serves as a good test for model dynamics. Both models showed evidence of the two different mixing regimes seen in the real atmosphere; material mixing into the northern mid-latitudes below 20km and material mixing into the southern mid-latitudes above 20km. The two Pinatubo simulations developed a tropical stratospheric reservoir, similar to that in the real atmosphere. The tropical confinement of material was maintained in the EUGCM by steep PV gradients, acting as a barrier to isentropic transport, situated in the low latitudes. Material only escaped from the tropics into the mid-latitudes via a channel of low PV gradient.

Chemical depletion of polar stratospheric ozone occurring during periods of enhanced reactive chlorine concentration in the winter and spring is investigated using both models and observations. A computationally-cheap and easily initialised photochemical model utilising ClO measurements from the Microwave Limb Sounder (MLS) on the Upper Atmosphere Research Satellite is developed. With this model, ozone destruction rates within the polar vortices due to the ClO + ClO, ClO + BrO and ClO + O catalytic cycles are evaluated. The method involves calculating local reactive chlorine concentrations from individual ClO retrievals, and then inferring the diurnal cycle of ClO from a quadratic expression using the relevant kinetic parameters. In test integrations this simple method is shown to give good agreement with more detailed calculations, but its speed of operation and the ease with which the ClO measurements are assimilated make it highly suited to dealing with the large amounts of data generated by MLS. Application of the method to the 1992-1993 Arctic and 1993 Antarctic winters yields maximum vortex-averaged ozone loss rates at 465 K potential temperature of ˜1% per day in both hemispheres. Time-integrated ozone destruction in the Arctic is less mainly because the duration of temperatures sufficiently low to sustain polar stratospheric clouds (PSCs) is shorter, and hence enhanced reactive chlorine concentrations are less persistent. The estimated chemical destruction on isentropic surfaces in the lower stratosphere is broadly similar to the observed change in ozone distribution, implying that the ozone change is dominated by chemical destruction, with dynamics playing a lesser role. An Antarctic winter-vortex is simulated in a chemical general circulation model (GCM) for the months of August and September. Chemical and dynamical impacts on the ozone change in the model are resolved by contrasting the temporal evolution of the 'chemical' ozone field with that of an inert tracer having the same initial distribution. It is found that the model results are consistent with the MLS-based chemistry-only calculations in indicating that there is very little dynamical replenishment of ozone on isentropic surfaces lying below 500 K. At higher altitudes the model implies a somewhat greater role for the transport than does the chemistry-only analysis.

Field measurements of the surface – atmosphere exchange of SO₂ and NH₃ have been made using two very different micrometeorological methods, a long term averaging method based on flux-gradient principles and an eddy covariance method, using a tunable diode laser fast response gas analysis system. The field data are then applied in surface-atmosphere models to calculate deposition of pollutant gases at the regional scale. The Time Averaged Gradient system (TAG) has been developed following a prototype implemented during a pilot study in 1998. The main goal of the TAG is to provide direct long-term average (1 to 4 weeks) flux-gradient measurements for a range of trace gas species, between atmosphere and terrestrial surfaces. Over daily periods, atmospheric conditions can range from high stability, where the vertical gradients of ambient concentration are enhanced due to very small diffusivity, to highly unstable conditions, in which concentration gradients are small due to the intense turbulent activity of the surface layer. By sampling continuously over a long-term period, the large vertical gradients generated by high stability would lead to an over-estimate of the actual flux; therefore it is necessary to avoid the bias due to these processes. To overcome this problem, the TAG system operates conditionally, sampling the micrometeorological variables within a carefully defined range of stability. A data series of five years, from 1999 to 2003, for SO₂ and a data series of two years, from 2001 to 2003 for NH₃ has been obtained at Auchencorth Moss, a field site in Southern Scotland. The measurements provided a characterisation of the chemical climate, meteorology, turbulent characteristics, as well as deposition-emission rates at the field site. A comparison with a continuous flux-gradient system running in parallel on the same field site allowed the reliability of the TAG system to be quantified: the correlation coefficients for u_* and H show a very good agreement (above 90% in both cases) between TAG protocol and continuous system suggesting the removal of stable conditions from the sampling period doesn’t modify the evaluation of the turbulent fluxes, although it introduces a bias. The SO₂ fluxes calculated omitting stable conditions under-estimate the fluxes of SO₂ measured by the continuous system by providing fluxes that are the 77% of the values estimated with the continuous system. A second TAG system has been implemented to improve estimates of gradients and reduce uncertainty on the fluxes and to increase the data coverage. It has been tested on the Easter Bush field site (Southern Scotland) measuring NH₃ fluxes form autumn 2001 to spring 2003. Field measurements of NH₃ fluxes using an eddy covariance technique were made for a total of 60 days between July and October 2002 at intensively managed grassland in Southern Scotland. The collected data demonstrate the suitability of a Tuneable Diode Laser Absorption Spectroscopy (TDLAS) system coupled with a sonic anemometer for eddy covariance measurements.

Under atmospheric blocking conditions, the normal passage of storms is interrupted by a region of high-pressure which remains lodged at the end of the storm tracks for periods of a week or more, causing the jet to split and the storms forced to pass around to the north and south, causing anomalous weather conditions over this high-pressure region. Being able to predict when these events occur, how long they will persist, and their eventual decay would be of value to improve weather prediction. This work looks at a simplified idealisation of this situation, with a view to improve understanding of any precursors to such events occurring and their subsequent behaviour. A 2-layer, <I>β</I>-plane, quasi-geostrophic channel model is used to examine the interactions between an upper-layer jet and high-frequency eddies supplied from a wavemaker in the lower layer. For certain initial jets, a dipole similar to an atmospheric block is formed, which remains stable to large-amplitude. By adding a shear to the upper-layer jet, a low-frequency vacillation cycle is induced, whereby the high-frequencies excite a split in the jet, which breaks down due to instability. This instability is demonstrated using a local instability analysis technique, and is also reflected in energy diagnostics. The role of the high-frequency eddies through the various phases of the cycle is also examined. A spherical-geometry model is also used with an aim to help bridge the gap between this highly-simplified model and the real atmosphere. These results suggest that the meridional shear in the upper-level atmospheric jetstream may determine whether blocking would develop, persist or breakdown. The structure of the upper-level jet could be controlled by seasonal variations or large-scale teleconnection patterns.

Stationary waves are vitally important in the general circulation of the atmosphere. These planetary scale waves are generally attributed to large scale orography and diabatic heating. The purpose of the present study is to investigate stationary waves as a linear response to the large scale orography and steady-state diabatic heating by developing a linear steady-state spectral model with the primitive equations in the global domain, in which Rayleigh friction and Newtonian cooling as well as biharmonic horizontal diffusion are included. We use perturbation theory to linearize the nonlinear system of primitive equations with respect to the zonally symmetric component of the dependent variables. The structure of the dependent variables is described by truncated series of spherical harmonics in the horizontal and orthogonal functions in the vertical. The model solution is obtained by way of linear superimposition of zonal waves calculated in the spectral domain for each zonal wave. The three dimensional wave activity flux, which was derived by Plumb (1985) for linear quasi-geostrophic stationary waves on a zonal mean flow, as well as the EP flux are used as a diagnostic method for studying the three dimensional structure of wave propagation. The Plumb flux is a conservable measure of the wave activity flux which reduces to EP flux in the zonal mean, and is non-divergent for steady, conservative, linear waves. The results of this thesis may present new insights into forcing mechanisms of the stationary waves. The numerical experiments for the response to idealised orography in northern middle latitudes and actual global orography show that the northern middle latitude orographic forcing, in particular the Tibetan Plateau, plays the most important role in the maintenance of the orographically forced stationary waves in the middle and upper troposphere as well as lower stratosphere, while forcing by the orographic effect of the Antarctic plateau is also very important, but restricted only to the lower troposphere of the southern high latitudes.

No description available.

No description available.