Everyone has some level of interest in the weather, but I’m pretty sure I can lose most of you well before the end of this post. I might even have trouble making it through it myself.
(Alternative title: Crystal balls, and you are feeling sleepy)
After working for BOM for about two minutes, I have found that people love to complain about inaccurate forecasts. But understanding more about the weather will give you an appreciation for the challenge of distilling a two line report on sun, rain, temperature and wind at a specific location from a three dimensional system of swirling gases under a myriad of influences. This post will hopefully introduce some of those concepts.
If we want to make weather forecasting models, we need to gather the initial data to feed into the simulations. The simplest observations are those made at the surface – temperature, humidity, wind speed and direction, and sunshine. In Antarctica we don’t measure evaporation or take ground temperatures – snow makes things rather difficult.
These measurements aren’t as simple as they sound. It’s that three dimensional thing again (well, four if you include time!) – more on that later. But basically we have a standard reference that we use to make comparisons. Temperature and humidity is measured inside a Stevenson screen 1.5 metres above the ground. A barometer is placed at a surveyed height. Care is given into selecting sites for anemometers and wind vanes so they are not obscured by buildings or terrain. We also take cloud (amount, height, type, and direction), weather (100 weather codes, for example #38 is slight to moderate blowing snow, with every code being defined), and visibility observations on a three hourly basis.
Now it’s time to talk about the famous crystal ball. Outsiders suspect this is how we obtain our weather forecasts. In fact it is a simple device to measure hours of sunshine. Sunlight passes through the glass sphere to focus on a strip of cardboard, which burns to indicate sunshine. It will make a scorch mark with a certain amount of cloud obscuring the sun but you have to draw the line somewhere. In Antarctica we have a north facing ball for daytime sunshine and a south facing ball for night time sun.
Finding north with the sun
Like anywhere in the southern hemisphere, the sun is at due north at solar noon (actually about 1:45pm at Davis, a little late) and it rotates through the full 360 degrees around the sky every day. So at 7:45pm it is due west, 1:45am due south, and 7:45am due east. Whether it is above or below the horizon at these times depends on the season. But if you can see the sun, there is a trick you can use to figure out north. With an analogue watch, point 12 o’clock to the sun, and north is half way between that and the hour hand.
Half way because, the sun rotates once while your watch hour hand does two laps in 24 hours. Since I just said solar noon is at 1:45pm, you’ll actually have to line up where the hour hand would be at 1:45pm to the sun. Sorry about that – it means we can sleep in a bit longer without wasting winter light. Oh, during the summer night, at Davis, 7:45pm – 7:45am, this technique will find due SOUTH!
Physics of weather
Understanding weather for me is a lot about understanding physics. When I was in school and to this day I generally don’t believe anything anyone tells me. Instead I have to be convinced of its truth by understanding it from first principles. This can be annoying for others, sure, but the two advantages I see are that you tend to make less mistakes with assumption and misunderstanding the scope of rules, and you know you don’t have to remember anything because you can always figure it out.
The main driver behind weather is the sun. Sunlight hits the earth at the equator straight on, and the poleward areas only get a glancing blow. Some of this heat is reflected by clouds, snow, water, and ground to varying extents. This depends on the reflectivity of the surface, and (for smooth surfaces like water) the angle. Water especially is a lot more reflective at low angles – which is why you see the bright sparkles on a rippled water surface. What is not reflected is absorbed as heat. In summary the tropics are receive a lot more heat from incoming solar radiation (shortened to insolation) than the higher latitudes.
The tropics are obviously a lot hotter than the poles. However they aren’t actually as hot as they would be, if heat from the sun wasn’t redistributed from the tropics to the poles. Physics again – when a gas is heated, it expands. Expanding upwards it spills out at the top of the atmosphere. With that extra weight off the top, there is now lower pressure at ground level. This sets up a circulation from high pressure to low pressure, outwards at the top of the atmosphere and inwards at the surface. This is the driver behind large scale synoptic winds around the planet. These drive ocean currents which help move the heat as well but the majority of heat (about three quarters) is transferred through the atmosphere. The equatorial doldrums, tropical trade winds, sub tropical light variable winds, and mid latitude westerlies are related to this – and Coriolis.
The Coriolis effect is the reason why, in the southern hemisphere, wind blows anticlockwise around a high pressure system and clockwise around a Low. Coriolis is a result of the earth spinning. At the equator the circumference is about 40,000km, so the rotation is nearly half a kilometer a second. If air flows from (a high pressure on) the equator southward (to a lower pressure), the earth towards the poles doesn’t rotate as far, so it is not moving as fast, so the air gets ahead – and curves to its left (anticlockwise). This is more prevalent at altitude, away from ground friction. Yes, more physics…
Speaking of physics… and still on the topic of heat. We have heat conductors, and heat insulators. Air is a good insulator. Snow is mostly made of air and is a good insulator. Rocks are reasonable conductors. Water is a good conductor, ice is even better. Metal is an excellent conductor – metal body cameras are great for sucking the heat off your fingers.
While sun does partly heat the atmosphere on the way to the ground, most of the sunlight passes through to the surface, which then heats the air by conduction. But since air is a poor conductor, further heat transport above the surface is via convection – physical movement of air. Thermals move hot air from the surface higher into the atmosphere. Solid rock (or water!) will conduct the suns heat deeper, meaning it takes longer to heat up and also to release its heat when it cools. On the other hand, light porous gravel will only be heated on the surface so the temperatures will be more extreme and therefore the air will be heated (or cooled) more.
Antarctica is a cold place, and most of the time it is losing heat into space through surface radiation. In winter we get strong surface inversions – the other day I released a weather balloon into -24.2 degrees Celcius and in less than a minute (about 250 metres altitude) it had warmed to -11.4 Celcius. Inversions like this means there will be no thermals and the air is very still. When we get a blow, the wind forces the air at different heights to mix and the surface temperature climbs by 10-15 degrees or more.
In addition to radiation, conduction, and convection, heat can be transferred with latent heat. Evaporation contributes to wind chill – it has a cooling effect. Latent heat disappears when water evaporates into vapour, and reappears again when it condenses in a cloud (tiny water droplets). Similarly with ice or snow sublimation into vapour or deposition into ice crystals. The moist atmosphere from the southern ocean brings Antarctica latent heat as clouds form and drop snow around the coast. Again Antarctica does not do its share, and loses heat – overall precipitation is greater than evaporation.
Every summer sea ice and icebergs are blown away from Antarctica to melt further north, and are replaced by seawater. This is another way Antarctica is warmed through latent heat. World wide, latent heat transfer is a few times higher than heat transfer by conduction and convection. This isn’t surprising with water on two thirds of the earth’s surface.
Latent heat is major driver of weather, allowing energy storage (via the latent heat of evaporation) and release in the forms of thunderstorms and cyclones. The positive feedback nature of heat release promoting more uplift gives rise to rather violent weather, and is why a warmed planet (warmer air can carry more water vapour) will have more extreme weather events.
Aside from global influences, the physical properties of the local environment affect the local weather. A seabreeze circulation is a local wind driven from an expanded column of warmed air overflowing at altitude to return air to the cooler surroundings. It even is affected by coriolis, with the wind direction backing (going from east to northeast to north etc) as the afternoon progresses. (In the northern hemisphere the wind veers instead).
Local terrain plays various roles – ground friction and venturis, orographic lift, the reflectivity and heat absorbing properties of the rock, ice, or water, evaporation levels (which depends on how dry the air is, how warm or wet the surface is, and wind speed). Of course, we can’t forget the sun – from paragliding I know how often a simple cloud can shut down the thermals in an area and put you on the deck.
Many of the above are affected by wind. Wind can be very three dimensional! Generally there will always be some wind at altitude, and wind speeds will be lower on the ground due to the friction against the surface. Have you noticed how winds drop off at night? This is due to the ground cooling the air, leaving a pool of heavier denser cold air underneath a lighter warmer layer of air which slides over the top. It won’t happen as much on the coast as the air doesn’t cool as much (due to the warm water nearby) so the wind can blow all night. But if you’re observant you can often notice a drop in the wind even when a shady cloud passes.
At Davis, we are spared from most of the katabatic winds that plague Mawson station every morning. Katabatic wind is cold dense surface air sliding downhill. Mawson is located on the edge of the ice, but at Davis we have a 20 kilometer buffer from the edge of the dome shaped ice plateau – the Vestfold hills. Aside from offering surface friction, the bumpy rocky surface will build up lots of cold heavy air throughout the night. So the katabatic air can slide right over the top. If you go to the fiords near Davis, however, you will feel the katabatic wind faithfully draining the ice shelf, a river of cold air.
Terrain can influence the flow of air well above the surface during windy frontal periods as well. In areas where wind converges towards a point there tends to be uplift. At Davis we often see showers in the same place in the distance all day, but they never reach us. This could be due to air forced up by a tightened constriction, or it could simply be air rising orographically as it meets the ice plateau.
Fog is a phenomenon caused by saturation of the air at the surface – the air has been cooled until it can’t hold water anymore, just like a dripping air conditioner unit. We get fog occasionally at Davis when moist (relatively) warm air from the open ocean drifts onto the land.
Inversions, lapse rates, and the temperature trace
Most people have heard that hot air rises. But it is more correct to say that hotter air rises. It rises because it is less dense, just like our hydrogen filled weather balloons. (I thought I’d slot that in, since given the title of this post I should probably show you a weather balloon before I get to the end of it.)
The weather balloon sonde
Attached to the weather balloon is a sonde, which transmits pressure, temperature, humidity, and position (for winds) back to a ground station.
A few things to note about this picture:
- The battery pack (made up from six AA’s) on the left has two water capsules to help keep it warm as it goes to high altitudes
- The thin filament on the sonde sensor boom is the temperature sensor, bumpy spikes are humidity, and that other bump is the pressure sensor
- The helix thing on the right is the GPS antenna. The thin metal wire on the other end is the transmitting antenna
- The string unwinder is partly visible, after release the sonde trails the balloon by about 20 metres.
- The ground check unit which calibrates the unit prior to launch is the background.
A parcel of air
Most people also know that the air is thinner at high altitudes. The density of air depends on (besides its temperature) the amount of air sitting on top of it, so it decreases with height.
So, imagine we have a parcel of air, that does not mix with or transfer heat with the air beside it (this lack of heat transfer means the air parcel is adiabatic). This air happens to be hotter, so it rises – which means it expands – which means the heat is now more spread out, so it has adiabatically cooled. The rate at which it cools as it rises is the adiabatic lapse rate. There is in fact two lapse rates, dry (DALR) and saturated (SALR), because saturated rising air (ie a cloud) has extra latent heat released. The DALR is about 10 degrees Celcius per thousand metres, the SALR is less than that.
So we know air cools as it rises (yes, it also warms as it sinks). Now how far will it rise? This depends on the lapse rate of the environment. Basically, if our parcel of air remains warmer than its surroundings it will continue to rise. When it reaches a warmer (ie, lighter) layer of air it will stop. This layer, with temperatures increasing with height (rather than decreasing) is called an inversion (or, if it’s the same temperature, it is called an isotherm). In practice there are often several inversions at various levels of the atmosphere – it varies with the day and this is a major factor behind determining the weather of the day, and why we have weather balloons in the first place. A standard (or average) atmosphere has a lapse rate of 6.5 degrees Celcius per thousand metres.
So the graph above is called a skew-T or a trace and shows us all the information gathered from this morning’s weather balloon. A few things to note:
- This winter we release one weather balloon every morning at 6:15am local time (the 23Z release). In summer we also did a 11Z release before dinner. To view them look for aerological diagrams here.
- ICAO height on the vertical axis is approximately correct – it is based on pressure readings taken from the weather balloon, and how high the ICAO standard atmosphere is at that pressure. There will be small differences, especially as Antarctic air isn’t quite standard!
- Besides pressure, the sonde gives us the red lines – temperature, and dewpoint temperature, and wind speed and direction (GPS measures drift of the balloon).
- All the other lines on the skew-T (DALR, SALR) let us examine how stable each section of the atmosphere is. See the more detailed explanation on the BOM web site.
- Rising air cools, and sinking air warms. The strong subsidence inversion at 3500 feet is formed by sinking high pressure air. Note there are smaller subsidence inversions higher up as well.
- The levels of the absolute humidity lines show how little moisture cold air can hold. Below the strong subsidence inversion (about 3500 feet) the dewpoint line is parallel to the absolute humidity lines, this shows the air is evenly mixed through these levels (since a parcel of air maintains the same absolute humidity, and the relative humidity changes as it warms or cools).
- Looking at the temperature graph at the same height we see the environmental lapse rate is parallel to the DALR lines (neutrally unstable for dry air). This also means there will be mixing throughout these levels, which is consistent with the humidity distribution.
- The dewpoint and temperature lines touch just underneath the strong surface inversion. This correlates with 7 octas of thin stratocumulus cloud observed at the time and is quite a common scenario. If the dewpoint and temperature lines are close, relative humidity is nearing 100% and we can expect to see some cloud. They don’t have to be as close higher up (for cirrus).
- The surface inversion will be stronger during light winds (strong winds will mix the air with the air above) and after dark nights.
- During a normal day, the lower part of the temperature graph will wag to the right as it is heated by the ground – in Antarctic winter though not a lot happens!
- The tropopause is the ultimate inversion, all our weather occurs beneath it. On this day it was at 253hPa, 9.25km (30000 feet), with a temperature of -70.7 degrees Celcius. It is higher in the tropics.
- On this day the stratocumulus blew away, but other mid and high level cloud ended up creeping in. It was a beautiful and fine day though.
Now…. you are feeling sleepy….