Phenomenal Hailstorm with Thunderstorm, Sydney 1st January 1947
By B.W. Newman B.Sc.
As far as any historical records are available the southern portion of the Sydney metropolitan area was subject to a hailstorm on 1st January 1947, the severity of which was apparently the greatest experienced in that area. From newspaper reports it seems that hail in some cases was in solid lumps up to 4 lb. in weight, but such may have been masses of ice compacted after reaching the ground. There can, however be no exaggeration in the extent of damage caused long it's route, and the wounds inflicted on people in the Bondi Beach area where many surfers were caught unprotected. Numerous houses lost between 20 and 30 tiles from the roof, windows were shattered and motor car hoods holed. In most cases ice lumps were reported to be as large as a cricket ball and the writer is in possession of a mould of one stone almost the size of a tennis ball, the weight of ice in which would have been about 4 ozs. One of our own meteorological personnel also found one columnar-shaped piece approximately 7" x 4" x 4", and the estimated weight of this would be about three pounds.
The storm was first reported from Liverpool (Fig. I) at about 1400K and apparently travelled east-north-east, passing over Mascot at 1425K and Rose Bay at 1435K. Its outer edge passed over the Weather Bureau, Sydney at 1435K. The first definite indication of cloud development at Sydney occurred towards 1300 when cumulus commenced to build in the south and west. By about 1400 this covered the south-west quadrant of the sky and appeared to be moving east, but keeping south of the city. The underpart of the cloud was mottled and serrated or curtained, rather than mammilated and looked angrily black while false cirrus tufts were discernible at the top. Shortly before the rain commenced at the Weather Bureau, shallow cumulus was observed moving from the north-east below the main cloud structure, which was coming from the westward, and between this and the overlaying cloud, considerable turbulence was apparent. At this time there was a terrific noise which appeared to come from the Harbour Bridge as though several trains were passing over. It was definitely not the sound of hail or rain to the south, and it is reasonable to assume its origin was in the cloud.
At 1435K the wind changed from north-east to a short south-west squall of 34 m.p.h. Rain amounting to 15 points fell in five minutes, after which sun came out and the wind returned to the north-east. During the shower at the Bureau there was only a little hail the size of which was about normal, none being larger than one half to three-eighths of an inch. Mr. Fitzgerald, our rainfall observer at Liverpool, reported 29 points of rain in a short period, but no hail fell in the district. Four miles to the west of Liverpool no rain whatever was recorded, while the locality four miles to the east was on the fringe of the storm. At Mascot, up till noon, there was only scattered large cumulus and a trace of Ac., but towards 1300K, conditions had become cloudy and the storm struck the aerodrome at 1425K with heavy thunder and hail lasting until 1450K. The highest wind gust was 33 m.p.h from the south-west at 1440K, and a total of 39 points of rain fell during the storm. The change in the meteorological elements at Mascot are shown in Fig II. and II. Capt. S. Owen of Butler Transport who was carrying out gliding exercises in the Mulgos district states the origin was apparently in the Camden district, as he could see the build up of cloud going on in that area during the forenoon. Mr R.R.C. Porter, Observer at Mascot, saw the cloud which gave rise to the storm forming at approximately 1000K between approximately Fairfield and Liverpool. It had then reached a fairly advanced stage of development with a peculiar formation.
The storm broke over Rose Bay at 1435K and lasted until 1448K. Rainfall amounted to 25 points. Hail was up to 2 inches in diameter (billiard ball size). The maximum wind gust was 46 m.p.h from the south, the wind being fresh north-east before and after the storm. The hail made holes in the hanger roof at the Flying Boat Base. Samples and or descriptions of actual hailstones available to the Weather bureau are as follows:–
Mr R.C.L. White, Weather 'Officer'.
Samples from Edgecliff 2.1/16 inch diamater with four layers
Mr. Lurie (southern suburb)
Oval shaped 7" x 4" at least 12 layers.
Mr .G. Clarke, dentist, of Earlwood, made moulds of some which, however, were not the largest that fell. The largest one of which a mould was made was roughly 2½" x 2" x 2" and weighd 4 ozs.
|Max. Temp. 91°F. (32.8°C.)|
|Max. Temp. 90.8°F. (32.7°C.)|
|9 a.m. Dry 77.2°F (25.1°F.)|
|9.a.m. Wet 71.0°F|
|9 a.m Rel.Hum. 73%|
|Mix. Ratio 15.3 g./kg.|
|Min. Temp. 671°F. 0504K|
|Max. Temp. 80.5°F (26.9°C) 1335K.|
|9 a.m. pressure 1002 mbs (SL.)|
|9 a.m. Dry 75.7°F. (32.4°C.)|
|Rel. Hum. 67%|
|Mix. Ratio 13.0 g./kg.|
|1000 ft||2000 ft||3000ft||4000ft||5000ft|
Copies of autographic records have been given (Fig. II and III) for Mascot only. Records at the Weather Bureau and Rose Bay were of similar form.
The meteorological situation at 0600 hours and also at 0900 hours (Fig. IV) was that of a wide trough extending from beyond the N.S.W. coast almost to South Australia. On the previous day a fresh north-east gradient existed along the coast and Rathmines radiosonde showed practically saturated conditions up to 820 mbs. On the 1st (or more correctly 1400Z of 31st Decr.) the Rathmines radiosonde still showed appreciable moisture and a comparatively steep temperature lapse rate up to 4000 ft. but a disconcerting small inversion at about 8000 ft. A light north-west wind prevailed at Sydney up to 9.a.m., after which the sea breeze developed moderate strength. Without the upper air soundings, but under the existence of a weak pressure field with warm and moderately humid conditions the normal forecast would have been "A local thunderstorm, but weather mainly fine and warm, tempered by a fresh N.E. breeze." With such a weak gradient the sea breeze would have been expected to check any marked rise in temperature. Frontal analysis suggested a quasi stationary front at 9 a.m. extending from about Walgett, Dubbo, Canberra to Gabo Is. The nearest cold front was across Western Victoria to the west of Tasmania. The Weather Bureau, Mascot, and A.M.F.A. were all agreed on these two facts. The quasi-stationary front was 150 miles from Sydney at 9 a.m. and there is no evidence from pilot balloon data to suggest that this front regenerated and moved to Sydney in the five hours available. The fact that the wind returned to the north-east immediately after the storm also precludes this front being involved. It is true that Goulburn reported a south-west wind at noon, but this is a frequent occurrence there.
It seems that the origin of the storm was definitely localised convection inland. It drifted seaward under a light westerly component above and additional energy was generated by the seabreeze. Temperatures at Rathmines and Sydney were sufficiently close to justify the application of Sydney's readings to Rathmines radiosonde. As a further point of interest a composite pseudo-adiabatic chart for 9 a.m (Fig. VI.) has been constructed using temperature and humidity for the surface and then joining up with Rathmines data from 860 millibars. To completely overcome the inversion at 8000 feet by auto-convection a surface temperature of 36 degrees centigrade (97°F.) was necessary. The maximum temperature reached at the Weather Bureau was 80.5°F or 29.6°C. Applying this temperature,condensation occurs at 880 mbs. or 4000 feet but ascent would be arrested at the inversion (actually about 7800 feet). We must remember, of course, that the thunderstorm did not originate over the city but rather some 20 miles inland and drifted eastward. On the other hand conditions at Sydney should have been such that the lifting of the air and the activity of the storm could be maintained. In this case some other means in addition to thermal convection were necessary to lift the condensation beyond the inversion. We are again assuming the inversion was real - if not the temperature curve might have been taken as straight from 770 mbs. to 700 mbs. We would then have had condensation to unlimited height and a thunderstorm certain.
Returning, however to the actual pseudo-adiabatic chart with its inversion, and examining Mascot's pilot balloon observations, a light NNW wind at 8 a.m. up to 2000 feet (probably 3000 feet) was replaced at 2 p.m. by north-east up to probably 3000 feet; possibly it may have been deeper during the forenoon. The question of how much an overlaying mass of compressible fluid is lifted by an invading wedge below is probably not definitely known. It is reasonable to assume, however, that 3000 feet (possibly 4000 feet) of north-east sea-breeze would lift the upper air at least from 7800 to 9500 feet. In this case the condensing mass would have been carried along the wet adiabat to a condition of instability beyond 9500 feet or 720 mbs.
In regard to the conditions over the core inland locality where the storm originated we have since obtained from observers some useful data. Both Picton and Parramatta had a maximum temperature of approximately 91°F. (32.8°C.) At Parramatta the 9.a.m. relative humidity was 73% (actually higher than that of Sydney by 6%.) By using Parramatta's surface data and taking a composite pseudo-adiabatic chart with Rathmines data from 860 mbs (Fig. VII.) we find condensation at about 3700 feet when surface temperature has reached only 84°F or 28.9°C. and this would just surmount the inversion; with the temperature continuing to 90.8°F. unlimited instability was assured. The closeness of the Picton temperatures to those of Parramatta give some justification for assuming somewhat similar conditions in the Camden-Liverpool area. Parramatta's maximum temperature applied directly to Rathmines pseudo-adiabatic diagram is barely sufficient to surmount the inversion. Instability, however, would have followed very easily with only a slight additional mechanical lift or even if the Rathmines inversion were not real. The thunderstorm, however, did not develop over Parramata, and did not approach within seven miles of Parramata. It may be that not only was the inversion at Rathmines real, but there was probably a much greater inversion over the Parramatta area but no inversion to southward. Of course there may be other possibilities such as that the actual conditions at Parramatta were less favourable than in the composite adiabatic diagram. There was a second thunderstorm some two hours later over the northern suburbs (see Fig.I) but this was not severe and missed Parramatta by 8 miles.
Summarising, it appears that the following primary predictions or estimates would have been necessary:-
(1) With Rathmines pseudo-adiabatic chart as the basis, a maximum temperature of at least 97°F. over inland suburbs was to be predicted.
(2) This would mean a temperature of about 85°F. at Sydney
(3) Obviously a sea-breeze would have been forecast, but even though the temperature as about 76°F. at 9 a.m.; a maximum of 85°F, under a sea breeze would not normally be expected - an estimate would have been about 82°F.
(4) A possible increase in humidity. This did not actually occur at the surface but another 10% may have been predicted. Generally, however, if we expect an increase in humidity with the sea breeze, we are inclined to lower our estimate of maximum temperature.
(5) The possible mechanical lift due to the sea-breeze. There is no doubt that this has often been the cause of thunderstorms. There was a case some years ago of a heavy hailstorm over the Mount Lofty ranges when a light northerly wind and apparently humid drift was lifted by a sea breeze.
(6) A comparison of surface temperature and humidity data from inland suburbs with the radio-sonde station and the possibility of constructions a composite pusdeo-diagram. This construction may be open to question. However we might say that at least a thunderstorm was likely over somewhere in the area. Even if this where not at the spot where it was thermodynamically predicted, such a forecast might be considered sufficiently useful. Such a procedure would be an advance on forecasting a local thunderstorm empirically without such analysis.
where is the mean acceleration.
R and u gas constants equal to 8.3 x 10 and 28.9 respectively
= Mean temperature
= Surface temperature
= Pressure oscillation
= Surface Pressure
h = height
is the pressure "bump" or sudden pressure rise as the the thunderstorm passes the locality. The values of determined from the barograph traces on this occasion were as follows:-
Under the temperature and pressure conditions prevailing, the maximum ascending velocities for the three stations were approximately as follows :-
The Variation in speeds is consistent with the relative intensity of the storm.
- Bulletin of the American Meterological Society,February 1942; "The effect of vertical acceleration on pressure during thunderstorms (Levine)""; March 1943; "The Determination of Vertical Velocities in Thunderstorms (Buell)."
Comments by D.Met.S Reviewing Committee on Papers in this Issue
1. "Does the stratosphere exist?" by Dr F. Loewe.
This is an interesting and stimulating paper by reason of its subject and its treatment. Swiss meteorologists at Payerne, both experimentally in a test chamber and in actual radio-sonde flight results found that there was a considerable thermal lag in the recordings of the temperature element. The critical condition under which this lag becomes important is the dependent on the velocity of ventilation and the density, and would in normal radio-sonde flights commence to be important near the level of the stratosphere in Europe. The Swiss from quantitative calculation of the lag arrived at the conclusion that the approximate isothermal condition found for the stratosphere was really due to this lag and that in reality a lapse rate of 0.5°C. per 100m is generally prevalent.
Dr. Loewe has reviewed the subject general and while he can find no flaw in the work of the Swiss, he gives convincing reasons for accepting the present temperature and lapse rated recorded in the stratosphere. He shows that is the Swiss contention in is fullest claims were true definite improbabilities would arise in connection with a number of sets of observations on record.
Mirages are only slightly treated in the usual meteorological text books so a reference to a full treatment and discussion of "cold" class mirages should be useful. This is given in two papers by Fujwhara and Hideka in the Geophysical Magazine of the Central Meteorological Observatory, Tokyo Vol IV, No 4. The problem is treated quantitatively and the diagrams, ray tracing and discussion are most instructive. The mirages there described "Sinkora" are more complex than those dealt with by Vollprecht, and Japanese authors consider the isopycnic surfaces which cause the refraction are not planes but a system of curved surfaces. The mirages appear over the bay on certain occasions when the air is colder than the water but cooler off-shore than near land, a situation which would produce flat isopycnic surfaces at sea but elliptical shaped near shore.
It would not seem that this shape of density surface could be present in most directions at Port Hedland when the phenomena described by Vollprecht were observed and probably this accounts for the simpler appearance. His account and discussion is most interesting and deals with as matter still requiring further treatment. Meteorologists who may desire to take a closer interest in this subject would find a study of Japanese articles helpful.
Forecasting of exceptionally heavy hailstorms is impracticable at present, and, as this article shows, even the forecasting of the occurrence of thunderstorms on the day in question would not have been undertaken with a strong likelihood of success. Adiabatic diagrams constructed from the upper air temperature and humidity recordings give very good indications of probable thunderstorms, but, unfortunately, these records are taken during the night and not near the time of thunderstorm formation. It is therefore necessary, while retaining as most probable the upper level conditions revealed by the earlier ascents, to estimate lower level conditions (such as surface maximum temperature etc.) later in the day about the time of thunderstorm formation. The author in this paper shows that this procedure would have lead to the likelihood of thunderstorm formation inland but not on the coast. This actually occurred, and the thunderstorms move Eastwards to the coast. Thus, with this kind of analysis of conditions, a forecast of probable thunderstorms would have been justifiable. The uncertainty factor in this analysis is, of course, evident but the procedure is still practically helpful.
Several interesting matters are raised by this article. They are ―
(1) The influence of sea-breezes on the uplift of the overlying air and, hence on convectional cloud and thunderstorm formation in certain favourable cases;
(2) The maximum size of hailstones;
(3) The vertical currents associated with hailstones and with the pressure rise in thunderstorms; and
(4) The modification of present technique in forecasting convectional cloud from adiabatic diagrams.
Regarding the first matter, there is little information available, but the general physical effects and nature of (illegible text) sea-breezes are of first class importance in sub-tropical lands, because of the climatic weather and even probable radii effects. It is not improbable that moist sea air arriving early over the land might with diurnal heating, be a decisive cause of thunderstorms, or that its arrival might conceivably cause sufficient uplift of upper air to touch off a thunderstorm.
The maximum size of hailstones is discussed in a paper by Belham and Relf in 1937 (Q.J.R. Met. Soc.) There the relations between terminal velocity and diameters are deduced from values of the drag coefficient obtained from observations on a sphere towed by aeroplanes. Values of terminal velocities are calculated for various mean specific gravities, and it is concluded that an upper limit of 1.5 lbs, is set by aerodynamical considerations to the mass attainable by spherical hailstones. This follows from the conclusion that under natural conditions there is a sudden and very large increase in the terminal velocity when the diameter of a hailstone reaches a certain value. Taking a hailstone of s.g. 0.6 at 10,000 ft., the terminal velocity increases steadily from 88 f/s for diameter 1.5" to 153 f/s for 4.5", then there is a rapid increase to 168 f/s at 4.8" when it jumps suddenly to a value of 380 f/s for somewhat larger stones, The sudden increase is due to the change in the nature of the turbulence flow at this stage. Thus it is concluded that the maximum spherical hailstone is likely to be about 5 inches in diameter and weigh about 1.5 lbs. Possibly the "record" hailstone is that is noted by Potter and found in Nebraska. It weighed 1.5 lbs and was 5.4" in diameter, being formed in concentric spheres around a single centre, proving it to be a single hailstone.
The vertical currents associated with hailstones have just been touched on, but the calculation of vertical currents from the pressure rise in thunderstorms has been used by the author of the article by Newman in his article under comment to estimate the vertical currents involved in the thunderstorm, He uses formulae given by Levine and Buell. In this Bulletin for July 1946 Loewe in an article on "A remarkable pressure variation" quotes reasons for doubting the decisive effect of vertical accelerations on the pressure change and further indicates that in the calculations of Levine and Buell have not taken into account the deceleration effects, so that their formulae are unreliable.
A more important matter is that the present technique (parcel method) for forecasting the tops of convectional cloud from tephigrams greatly overestimates these heights (Convection in Theory and Practice - M.O. S.D.T.H. No. 102 Petterssen) by amounts in the order of 4,000 feet. Strictly virtual temperatures should replace dry bulb temperatures, and virtual saturated adiabatics replace the ordinary saturated adiabatics in evaluating convective processes on the tephigram. As the virtual saturated adiabatic laps rate is steeper than the ordinary saturated-adiabatic rate, the use of the latter, which is present practice, results in an estimated cloud top which would be higher than that derived from the more correct procedure, which already gives tops too high. Hence especially in low latitudes or with moist air, estimated cloud tops may apparently be some thousands of feet too high, materially affecting the forecasting of thunderstorms. Forecasters should keep this in mind. Some checks against figures provided by planes would be valuable.