Showing papers in "Quarterly Journal of the Royal Meteorological Society in 1969"

A major low-level air current near the Indian Ocean during the northern summer

Abstract: It is demonstrated that high-energy flow, in the form of low-level southerly jet streams which have been reported earlier over Kenya, is only one part of a much more extensive current of air which flows rapidly around the western half of the Indian Ocean during the northern summer. The high speeds are associated with the concentration of the cross-equatorial airflow into the zone from longitude 38 deg E to about 55 deg E instead of being rather evenly distributed from 40 deg to 60 deg or 80 deg E as illustrated in many climatological atlases and charts of mean flow during the season. The high-speed current is shown to flow intermittently from the vicinity of Mauritius through Madagascar, Kenya, eastern Ethiopia, Somalia and thence across the Indian Ocean to the west coast of India and beyond. The stream is occasionally reinforced by northward flow through the Mozambique Channel. The high-speed air current, or system of low-level jet streams, is closely associated with the Intertropical Convergence Zone over the Arabian Sea and western India, and variations in the strength of the stream over Kenya during a two month period were related to the rainfall which western India received from the south-west monsoon.

Momentum transport by gravity waves

Abstract: SUMMARY A review of theories concerning the propagation of internal gravity waves in a horizontally uniform shear flow concludes that an upward transport of horizontal momentum inevitably accompanies the generation of such waves in the atmosphere, the mean flow being affected only at precisely those levels where the waves are dissipated. If the mean wind depends on horizontal position there may be a continuous transfer of momentum from the waves to the mean flow during propagation. In the absence of intense clear air turbulence or a critical level where the intrinsic frequency vanishes, many waves propagate upwards to great heights, even if the mean wind is time dependent. Computations of the wave drag in a 19 m s-I gradient wind over hilly terrain in north Wales show that on that occasion it amounted to 4 dyne cm-*, of which 3 dyne cm-* probably acted on the atmosphere above 20 km.

A comparison between laboratory and atmospheric density currents

Abstract: Some known results on density currents in liquids are discussed, and the geometry of the frontal regions and of the upcurrents in the region ahead of them is shown to be similar to that of some thunderstorm outflows and sea-breeze fronts. The detailed structure of the frontal region of laboratory density flows is described. An internal Froude number has about the same value in the model and in the atmosphere.

Eigenvectors for representing the 500 mb geopotential surface over the Northern Hemisphere

Abstract: A set of empirical orthogonal functions is produced which enable the large-scale features of the 500 mb topography over the Northern Hemisphere to be represented with near optimal efficiency. The more important functions are described. Some of the properties of the coefficients of these functions which best represent the 500 mb charts for each day of the years 1965 to 1967 are summarized. The question of the number of functions required to obtain the most satisfactory representation of the 500 mb fields is discussed.

Modification of mean flow and turbulent energy by a change in surface roughness under conditions of neutral stability

Abstract: A theory is developed which describes the adjustment of the flow of a hydrostatically neutral fluid in the lower portion of a fully-turbulent boundary layer, after an abrupt change in surface roughness. The model is based on the hypothesis that the horizontal shear stress is proportional to the turbulent energy. The theory postulates that the flow is primarily governed by the dominant terms of the horizontal-momentum, continuity, and turbulent-energy equations. The model was solved by numerical techniques on a digital computer. Unlike previous models there are no a priori assumptions about the distribution of velocity or stress, the behaviour of the nondimensional wind shear, mixing length, or momentum-exchange coefficient in the transition region. The theory, in contrast to earlier theories, suggests the distribution of turbulent energy, as well as velocity. An inflection point is predicted in the transition velocity-profile. The nondimensional wind shear is found to differ significantly from unity in the transition region. These predictions agree with observation.

The polar low as a baroclinic disturbance

Abstract: This paper describes the three dimensional airflow and precipitation within small ‘polar’ depressions which sometimes cross Britain during northerly outbreaks and which in winter can be responsible for heavy snowfalls. The data used in the study consist of Doppler and conventional radar information, together with routine synoptic data and sequential radiosonde ascents from the radar station. Three dimensional airflow was derived from the radiosonde data assuming that wet bulb potential temperature was conserved. Horizontal and vertical air velocities were also derived from the Doppler radar measurements. Previous knowledge of polar lows is meagre; they are generally thought to be shallow features resulting from enhanced convection within cold air flowing over a warm sea. However, the well-developed polar low which is the main subject of this paper is shown to have been an essentially baroclinic disturbance. Although enhanced small-scale convection occurred in one sector, the main area of widespread precipitation associated with the polar low was produced not by small-scale convective overturning but rather by slantwise convection within a narrow tongue of air ascending steadily at about 10 cm s−1. The speed of travel and short wavelength (900 km) of the polar low in this study are consistent with its having formed in a region of enhanced baroclinicity within the polar air below 850 mb rather than in the major baroclinic zone bounding the polar air mass. Considerable low-level baroclinicity within the polar air is also shown to have been present during the formation of other intense polar lows.

Interhemispheric transport of air in the lower troposphere over the western Indian Ocean

Abstract: Equatorial cross-sections of monthly mean meridional flow in the lower troposphere over the western Indian Ocean have been prepared for January and July. These sections have allowed new estimates to be made of the mean transport of air across the Equator in the area.

Air motion and precipitation growth in a wave depression

Abstract: A case study is presented to illustrate the small and medium scale air motion and precipitation patterns within a travelling wave depression. Air motion has been studied by means of isentropic and Doppler radar analyses; precipitation patterns have been studied by means of meso-analysis of autographic rate-of-rainfall records and conventional 3-dimensional radar data. The results confirm and elaborate upon earlier studies in which precipitation generation is classified according to whether it is due to large-scale ascent alone or is modified by vertical instability. The influence of orography is also discussed.

Fall velocities of plate‐like and columnar ice crystals

Abstract: The Reynolds number of cylinders and discs falling in a viscous liquid was determined experimentally for values of the Best number (the product of the drag coefficient and the square of the Reynolds number) ranging from 1 to 104. The results were used to calculate the terminal velocities of plate-like and columnar ice crystals.

Tropical cloudiness and rainfall related to pressure and tidal variations

Abstract: A definitive statistical relationship is established between tropical cloudiness and rainfall and the semidiurnal solar (S2) atmospheric tide, as manifested in the semi-diurnal surface pressure variation. Pressure and weather data are used from Batavia (seventy years) and Wake Island (twelve years). The S2 tidal effect is shown to enhance cloudiness and rain near sunrise and sunset, and to suppress them shortly after midday and midnight. The analysis is based on (a) the fact that the S2 amplitude varies by 15–20 per cent between months and by more than 100 per cent from day to day and (b) the amplitude of the S2 wave as computed from the pressure data at a station is closely related to the 5–6 hourly pressure changes during the periods around 4–5 a.m. to 10 a.m., from 10 a.m. to 4 p.m., etc. The crux of the analysis is the demonstration that days with large 5–6 hourly pressure changes during these periods have large cloudiness changes (in the sense described above) during these same periods relative to days with small 5–6 hourly pressure changes. Possible mechanistic connections between S2 pressure tendencies and cloud properties are examined. The varying convergence-divergence field is suggested as the main link. It is shown how the concentration of active cloud updraughts in the Tropics can permit cloudiness to be extremely sensitive to small divergence fields. Finally, large-scale simultaneous pressure changes over the Pacific Ocean area are shown and related to cloudiness changes. A need for re-examination of the nature and origins of tropical disturbances is shown to exist, using the concept that possible small (terrestrial or extra-terrestrial) triggers may set off significant changes in weather both locally and on the synoptic and global scales.

On wind and shear stress profiles above a change in surface roughness

Abstract: The problem of determining the distribution of velocity and shear stress in the flow above a surface with an abrupt change in roughness is considered by using the ‘mixing-length’ theory to relate the shear stress to the velocity profile and solving the resulting system of partial differential equations numerically. The results are compared with those obtained by Panofsky and Townsend (1964) and Taylor (1967), by assuming special forms for the velocity or shear stress profiles.

Absorption and emission by carbon‐dioxide in the mesosphere

Abstract: Theoretical and experimental evidence is presented which leads to a vibrational relaxation time appropriate for the v2 vibration of CO2 at 15 μ under atmospheric conditions at 210°K and standard pressure of 6·0 × 10−6sec. The effect of this on the atmospheric cooling rate near 90 km due to emission by CO2 is discussed. It is shown that absorption of solar radiation by the v3 band (at 4·3 μ) and the combination bands of CO2 (at 2·7 μ) leads to a heating rate of about 2°C (12 hr)−1 near 80 km, this being one of the largest contributions to the radiative heating rate at this altitude. The processes by which relaxation from the v3 vibration of CO2 occurs involve vibrationally excited oxygen and the v2 vibration of H2O. The magnitude of heating-rate depends considerably, therefore, on the H2O concentration. For thermal radiative exchange by the v3 band, thermodynamic equilibrium begins to break down at 30 km; its contribution to the radiative budget of the mesosphere is consequently very small.

Lamb waves in a nearly isothermal atmosphere

Abstract: If the variations in sound speed c (z) and wind u (z) with height z are not too great, there is a mode of propagation of acoustic waves in the atmosphere in which the wave energy E (z) decreases exponentially with z with a scale c02/(2 – y) g (about 16 km), and the phase speed cm is given approximately by (about 300 m s−1).

Observational studies of the general circulation of the Tropics: long term mean values

Abstract: Data from 298 stations between 45°N and 30°S have been processed for the period July 1957 to December 1964 The mean wind and temperature fields are presented together with the momentum and heat fluxes for two three month seasons, December-February and June-August The tropospheric results were obtained from objective analysis of the long term station means and the stratospheric results from latitude band means giving equal weight to data from odd and even years so that the biennial component has been virtually eliminated The mean temperature cross-sections show little seasonal variation in the tropical troposphere and gradients are small The maximum temperature is in the vicinity of 5°S during the southern summer but moves to 20°N during June-August as a consequence of stronger heating over the continents In the tropical stratosphere the seasonal variation is from lower temperatures in December-February to higher temperatures in June-August The mean zonal wind pattern shows the tropical easterlies to be stronger and to occupy a greater range of latitudes in the upper troposphere during June-August The Hadley circulation obtained appears quite reasonable with the winter hemisphere cell predominating and extending over the Equator Although all three components of the momentum flux were evaluated for the troposphere, only the transient eddy contribution could be obtained for the stratosphere A strong flux is observed in the upper equatorial troposphere directed into the summer hemisphere The sensible heat-flux in the middle troposphere is equatorward in the Tropics of both hemispheres and a small transport into the summer hemisphere is noted This equatorward flux is countergradient in both hemispheres except during the northern summer The mean motion transport of total energy is about an order of magnitude larger than the eddy transport at the Equator and supplies energy to the winter hemisphere The zonal and eddy forms of kinetic energy and available potential energy have been evaluated for the tropical troposphere together with all conversions between them except for that by transient eddies between available potential energy and kinetic energy On the average, kinetic energy contents are about twice as large as available potential energy contents and the momentum flux conversions are larger than those involving heat fluxes The conversion by the mean motions, however, is largest of all over the range 24°N-24°S and leads to a rejuvenation time of some 3–5 days for the zonal kinetic energy This conversion appears to be sufficiently large to offset the destruction in middle latitudes Taken together, the results presented here give a consistent pattern for the tropical circulation The direct Hadley cell is seen as a source of both heat and momentum which are distributed by the eddy fluxes The energy supply for the mean cell is thought to be the latent heat of condensation in its ascending branch

Estimations from geostrophic trajectories of horizontal diffusivity in the mid‐latitude troposphere and lower stratosphere

Abstract: Zonal and meridional diffusivities Kxx and Kyy have been found at mid-latitudes from the spreads of sets of geostrophic trajectories at levels between 700 mb and 30 mb during each season of 1965. Values have been calculated using the relationship R = e−PT cos qr for the form of the corresponding Lagrangian autocorrelograms. In addition a similar set of Eulerian data has been analysed to obtain autocorrelograms and by comparison with the trajectory data Eulerian-Lagrangian scale relationships have been obtained. The results show values of Kxx ranging from 0·2.106 to about 20.106 and Kyy from 0·2. 106 to about 6.106 m2 sec−1. In both cases the lowest values are found in the summer stratosphere and the highest in the spring stratosphere and in the upper troposphere in winter. Their distributions principally follow that of the wind variances. Integral time scales are mainly between 5 and 30 hours meridionally and 10 and 50 hours zonally and are greatest in the stratosphere and lowest in the mid-troposphere. The contributions of damping and harmonic effects to these time scales are discussed. Comparison of the Lagrangian and Eulerian time scales indicates a mean value β of about 0·6 for their ratio with the meridional values smaller than the zonal values.

Precipitation-induced mesoscale wind perturbations in the melting layer

Abstract: Doppler radar observations during widespread frontal precipitation in England on 20 February 1967 show significant perturbations of the wind in and near the melting layer. The analysis suggests the following hypotheses: (i) Mesoscale oscillations of the wind near the melting layer are produced by pressure perturbations due to horizontal variations in cooling by melting snow associated with the precipitation pattern. A simple theoretical analysis of the effect is presented. In simple patterns, the perturbed wind has a maximum on the low pressure side of the rainfall. The perturbations propagate with the velocity of the precipitation generators and their magnitude increases with the precipitation rate. This hypothesis is in accord with the finding that the oscillation in the tangential wind across hurricane rain bands is generally a maximum on the low pressure side of the bands. (ii) Smaller scale wind perturbations were also observed in the melting layer with scale of a few kilometers comparable to those reported by Harrold and Browning (1967). However, unlike their observations, these perturbations did not appear to move with the melting layer winds. A tentative suggestion is made that these may represent wave motions.

Relations of radiation fluxes over natural surfaces

Abstract: Radiation fluxes over surfaces of bare soil and short grass sod were measured on clear days at St. Paul, Minnesota. Widely diverse types of daily cycles describing the hourly dependence of net upon net solar radiation were found. Heating coefficients depended as much upon atmospheric radiative properties as upon surface properties. Causes for these relations are proposed and tentatively verified experimentally. The accuracy of regressions of solar on net radiation is shown to depend upon the uniform distribution of data points extending to the region of the origin. Controversies in the literature are illuminated by this analysis, both theoretically and experimentally. Two equations relating the slope of the regression of solar on net radiation to albedo are also considered. Finally, a relation between the long-wave radiation balance and the heating coefficient proposed by Ekern (1965) is qualitatively substantiated.

Evaporation from a papyrus swamp

Abstract: Evaporation from an old stand of papyrus was estimated by the Bowen ratio method. It was found that about 30 per cent of the net radiation reaching the top of the canopy was converted into sensible heat. The layer formed by the heads of the papyrus also seemed to act as a major resistance to turbulent exchange of water vapour from the undergrowth. Evaporation from an old stand of papyrus was estimated to be 60 ± 15 per cent of Penman estimates of evaporation from open water.

Microwave observations of sea state from aircraft

Abstract: Microwave thermal emission by water measured from aircraft, studying radiation dependence on surface roughness

Charge distribution in thunderstorm clouds

Abstract: The negative charge involved in lightning flashes to ground is found to be distributed in a manner strongly dependent upon the direction of movement of the storm, and does not, in general, constitute a nearly vertical column as proposed by Malan and Schonland. Based on a study of electric field-changes measured at two stations 10 km apart involving 539 return strokes from 84 flashes in 10 storms, we conclude that the horizontal component of the in-cloud channel on the average exceeds the vertical component, and points in the direction of storm motion. An analysis is given of the five methods of Malan and Schonland, on the basis of which we suggest that significant horizontal components are also compatible with their observations of a vertical column, and that the ‘nearly vertical’ aspect of the charge distribution has been over-emphasized.

The disintegration and electrification of charged water drops falling in an electric field

Abstract: The spheroidal assumption employed by Taylor (1964) has been applied to a theoretical treatment of the deformation and disintegration of charged water drops falling through an electric field of strength F. Aerodynamic and hydrostatic pressure-difference terms were incorporated into equilibrium equations from which the fields required to produce instability were determined. Values of F (Ro/T)1/2 required for disintegration were computed as a function of the applied charge Q, undistorted drop radius Ro and surface tension T. The predicted values of F (Ro/T)1/2 have been tested experimentally for positively or negatively charged water drops of Ro ranging from 0·10 to 0·17 cm falling through horizontal or vertical fields at temperatures varying from + 20 to −18°C. The agreement between theory and experiment was found to be good; the maximum discrepancy of about 3 per cent was obtained with the largest drops utilized and is probably a consequence of the oscillation factor discussed by Ausman and Brook (1967). As predicted, the field required to disintegrate a falling drop was found to be markedly dependent upon the magnitude of the charge that it carried. It is shown that the field, F, required to effect the disintegration of the great majority of charged raindrops falling through the intense electric fields of a thunderstorm can be expressed in electrostatic units to within an accuracy of 2 per cent by the equation F (Ro)1/2 = 13·6 – 8Q. The mass loss accompanying the disintegration of falling drops was found to bear a similar relationship to the field strength as was previously observed by Latham (1965) and Matthews (1967). Measurements on suspended drops of water showed that the values of charge and mass loss, q and δM respectively, accompanying the disintegration are close to the theoretical maxima. The close parallelism between the experimental curves of q against Q and δM against Q suggests that instability occurs over a large fraction of the surface area of the lower hemispheroid.

Atmospheric edge waves

Abstract: In an isothermal windless atmosphere Lamb's wave, the energy of which decays exponentially with height, propagates non-dispersively with the speed of sound In the real atmosphere the speed of sound and the wind speed vary with height, but it is known that an edge wave similar to Lamb's wave is still possible Assuming the wind speed and variations in the speed of sound to be much less than some typical sound speed, this atmospheric edge wave is shown to have group velocity given approximately by Where ω is the wave frequency, c the mean of the sound and wind speeds weighted with the energy density of the basic Lamb wave, and D is a positive dispersion coefficient defined in terms of a simple integral of the departures of the sound and wind speeds from their weighted means Expressions are derived for the dispersive effect of a change in the sound or wind speed at any height, and these are evaluated for a particular model atmosphere It is shown that the effect of horizontal inhomogeneities in the atmosphere is merely to average the long wave speed and dispersion coefficient along the great circle path from source to receiver The theory is compared with the results obtained from microbarograms of pressure pulses from large atmospheric explosions, but it is found that the paucity of atmospheric data makes it difficult to use these results to estimate winds above sounding heights on the path from source to receiver Atmospheric edge waves are shown to be rather insensitive to the upper boundary condition in general, though the effect on microbarograms of coupling between the edge wave and waves propagating high in the atmosphere is discussed, and various decoupling mechanisms, including dissipative decoupling, are described

Surface films and the production of sea‐salt aerosol

Abstract: The number of condensation nuclei produced by the bursting of a single bubble in sea-water, and detectable at a few per cent supersaturation, is shown to be markedly dependent on the film pressure of naturally occurring surface-active material adsorbed on the surface. At low film pressures around 1 dyne cm−1 the presence of the film increases the variability of the number produced, but the number is always less than that expected for a clean water surface. With increase of film pressure the number of nuclei produced decreases markedly. The dependence of nucleus number on the state of any surface-adsorbed film has important consequences for the sea-to-air transport of organic material. This is discussed.

Airborne studies of the electrical properties of large convective clouds

Abstract: Studies have been made using instrumented aircraft of the electrical and microphysical properties of cumulus and cumulonimbus clouds over Flagstaff, Arizona. Measurements were made throughout the life-cycle of these clouds either by making horizontal penetrations through the cloud at intervals of 2 Centigrade degrees or by entering the cloud at the highest level attainable and spiralling down to cloud base. The aircraft could attain an altitude of around 32,000 ft (9·7 km), corresponding to a temperature ceiling of approximately −30°C. Continuous measurements were made of vertical electric field strength, F, the charge carried on larger hydrometeors, Q, the concentrations, types and dimensions of cloud particles, liquid-water content, temperature and other meteorological parameters. F was measured using a rotating cylindrical field-mill of the type devised by Kasemir (1964), which enabled fields ranging from about 100 Volt m−1 to corona values to be measured with an accuracy of about 10 per cent. Q was measured using a double-sheath induction technique which permitted charges ranging in magnitude from about 10−2 to 1·0 e.s.u. to be measured to within 10 per cent. A series of careful subsidiary experiments established the range of conditions over which these electrical instruments provided reliable results. The concentrations, types and dimensions of cloud particles were measured using the continuous particle sampler described by MacCready (1962). The measured electrical characteristics of the clouds studied varied markedly from cloud to cloud but it was possible to isolate certain recurrent patterns and correlations. The measured values of Q were found generally to be carried on rimed aggregates of ice crystals. The degree of electrification, represented by the values of F and Q, was highest in clouds where these aggregates coexisted with appreciable quantities of supercooled water droplets and ice crystals. At all levels within the clouds positive and negative values of Q were obtained, but the proportion of positively charged larger hydrometeors increased with increasing temperature within the cloud. Small clouds or clouds in their early stages of development usually possessed a simple electrical structure, with positive fields existing throughout most of their volume. Larger, more mature clouds possessed extremely complex electrical structures and a much higher degree of electrification. Sharp reversals of the dominant sign of Q were often observed to occur once or twice during the horizontal penetration of a precipitation shaft below cloud base at temperatures above 0°C. Lack of vital information, particularly concerning the cloud dynamics, rendered it impossible to assess the measurements in terms of all current theories of thunderstorm electrification. However, it can be shown that both the Reynolds-Brook process and the Muller-Hillebrand mechanism probably contributed significantly to the electrification of the clouds studied. It is difficult to distinguish the respective roles of these processes. The evidence suggests, however, that clouds possessing a simple structure, or clouds in the early stages of their maturity, have an electrical structure that is largely compatible with the predictions of the inductive process, whereas the gross electrical characteristics of more complex clouds are often explicable in terms of the Reynolds-Brook theory.

Antarctic stratospheric warming reviewed in the light of 1967 observations

Abstract: Unlike the Northern Hemisphere where a number of extremely well documented case studies to very high levels facilitate the study of polar stratospheric warming, most investigations so far in the Southern Hemisphere have been directed mainly to the 100 mb, or occasionally 50 mb level. In 1967 it was possible to draw a series of daily 30 mb contour charts through the spring period, and 30 mb isotherm charts for the month of October over the Antarctic. The contour charts showed that the polar vortex: changed little in depth until the end of September then filled steadily; was initially single centred, became double-centred, then finally single centred again; developed a double-wave pattern, the perturbations moving progressively and regularly around the vortex. The movement of these waves was accompanied by short period, but marked, temperature changes at a given point (warm ridges and cold troughs), but even after the major warming was completed at this level over the continent, the circulation was still predominantly cyclonic. From the limited information available it appears that in 1967 the short period temperature changes may have been more pronounced than in some earlier years, but the general warming trend was normal. Wave movement around the polar vortex could also be detected in an autumn and winter season, and apparent mid-winter stratospheric warmings previously reported or observed may be ascribed to the passage of migratory warm ridges, although this explanation is not advanced for the mid-winter warming event inferred over McMurdo in 1963.

A theoretical study of cup and vane anemometers

Abstract: The predictions of a simplified theory of anemometers are compared with observation. The theoretical response of cup and vane anemometers to steady and varying flows are derived analytically and by using an analogue computer. It is found that the mean indicated windspeed can be overestimated by as much as 20 per cent and that the form of fluctuations is distorted and their amplitude considerably reduced.

The spreading of accreted droplets

Abstract: Measurements of the spreading factor of accreted droplets (the ratio of the final radius to the initial radius) have been determined over a wide range of impact speeds and ice deposit temperatures for ambient temperatures of −11 and −22°C and droplet radii of 17 and 54μ. The spreading factors vary from 2 to 6 and depend primarily on the impact speed and deposit temperature.

Energy flux and convergence patterns in the lower and middle stratosphere during the IQSY

Abstract: The vertical and horizontal fluxes of energy by large-scale standing waves have been estimated on a monthly basis for the IQSY for the 100–10 mb region. A large upward flux into the region occurs in winter some of which passes into the atmosphere above 10 mb, where it may exceed the energy generated there by radiative processes. The meridional wave energy flux diverges from the 40°–60°N region and converges in both high and low latitudes in winter. The total energy flux convergence pattern shows that the polar night jet in the lower stratosphere in winter is a source of kinetic energy; when the jet collapses in the spring the same region turns into a strong sink and it is noted that while this was a gradual process in 1965 and occupied three months, there was a large absorption in March 1964 that dominated the energetics pattern in spring. The largest energy absorption occurs with the largest energy flux into the region, thus supporting previous suggestions that the spring warming is ultimately dependent on tropospheric events. The adiabatic approximation was used in the flux computations and the role of radiative processes in the observed events is briefly discussed.

The coreless winter at Scott Base, Antarctica

Abstract: The temporary early-winter reversal of the downward trend in monthly mean surface air temperatures, which has been noted for other stations in the Ross Sea – South Pole region, is a particularly marked and persistent feature at Scott Base. An examination of temperature records shows that this reversal in the trend is probably due to a more frequent destruction of the low-level surface inversion in these months, rather than a temporary increase in the level of warm air advection into the region.