Some forecasting explanations

Weather patterns
Forecasting temperature
Forecasting rain
Forecasting wind
Interpretation

This page is intended to help users obtain the maximum out of the Mosfiloti Weather Forecasts and to extend the data to other localities. Every place on land has its own weather patterns and these foothills are no exception. As and when I see patterns developing, I'll add them to this page.

Weather patterns

There are three types of pattern that concern us here, which we shall call synoptic, mesoscale and macroscale (there is microscale, as well, but this is too small to be of interest; there are probably at least half a dozen microclimates in Mosfiloti alone). Roughly speaking, we can consider the three patterns to represent what is happening around us to a radius of about 1500 km, 100 km and 20 km respectively, although these are not fixed rules. They each influence the weather differently.

This is a partial synoptic barometric and precipitation weather chart of the E. Mediterranean basin, as an example. At the time of this map, the weather in Cyprus is mainly determined by the low pressure over Iran. The isobars (contour lines) are widely spaced over Cyprus, meaning relatively stable weather. The wind follows these lines and, indeed, we had a moderate NNW wind at that time, averaging 2 m/s and gusting to about 7 m/s. Experts can predict that the high pressure centred over Istanbul will keep these conditions stable for some time, and the zone of precipitation over the eastern side of the Black Sea will be pushed eastwards. It seems probable that this high pressure may dominate over the following days and form a ridge with the ones in the Red Sea and the northern Balkans.

Full synoptic maps will include much more data, including temperature, humidity, wind speed and direction, cloud cover etc. Such a map of the area in the one above would be about 1 metre wide and could not be reproduced here. However, even more important is to know what is happening at altitude (this is why weather balloons are used). Traditionally, the data is collected at altitudes where the barometric pressure is 200, 500, 700 and 850 hPa. Without knowing this, forecasting would be impossible.

The following map shows mesoscale precipitation on a certain day in 2003. I use it to illustrate the fact that you can see that the precipitation at Troodos was >35 mm, but it is insufficiently detailed to show what the exact rainfall was at any given place. For example, Nicosia is somewhere in the greeny-blue 15-20 mm zone, but it is quite conceivable that E. Nicosia had considerably less rain than W. Nicosia. In terms of forecasting, the mesoscale maps are typically what you see on newspaper or TV forecasts, where they give general data for the major towns, based on projected surface conditions, with a few icons. 

As we zoom in to a macroscale, let's have a look at what the area round Mosfiloti looks like from a satellite:

We see that we are in the bottom of a saucer, with hills surrounding us at 5-10 km. These hills protect us to some extent from the synoptic winds. Note particularly the winding valley south-east of Ayia Anna; this has considerable influence on our weather, as we shall see. Now let's have a closer look with enhanced colouring:

Here, we can see Mosfiloti has Mount Pipis to the south, separating it from Pyrga and some lower hills to the north. These can also influence the weather. Because the main part of the village is fairly exposed, although slightly less so from the East, the weather can come from any direction, according to the prevailing wind. The eastern part of the village, from the cemetery to the Psevdas border, is less exposed to the north and south because of the hills and has a slightly different microclimate. My weather station is in this part of the village.

Forecasting temperature

For the software I use to do the forecasting, the first data it will need are what the conditions are in Mosfiloti. My weather station measures a host of variables, including the barometric pressure, temperature, relative humidity, wind speed and direction, rain etc. From these data, it can calculate many other variables such as dew point, wet bulb temperature, cloud height and the trends of each variable over time. However, these are far from sufficient; we also need the synoptic data, and we collect this from weather stations from Italy to Iran and from Russia to Saudi Arabia, to see what is going on, at altitudes of 10 m and at pressures of 200, 500, 700 and 850 hPa. All these data are processed to produce a general forecast, which is then further processed to take into account the macroscale conditions.

This is a simplified sketch graph of a typical diurnal cycle between sunrise and sunset of what the temperature (green line) and relative humidity (blue line) would be without macroscale influences. The units don't matter, it is only the general shape that need be considered. As the sun rises, the temperature increases to reach a peak in the early-to-mid-afternoon, after which it drops and continues to drop after sunset. If the dew point remains substantially constant, the RH curve will assume an opposite shape. The general forecast, based only on synoptic data, may assume a curve of this general form.

However, this is not what happens in Mosfiloti. There are two factors that influence it, one from the sea and one from the mountains. In summer, a sea breeze may come in and, depending on the direction and speed of the prevailing wind, this sea breeze may reach us at any time from about 1000 to 1600 hours, most often from 1200 to 1400 hours. Because it is coming from the sea, it is cooler and more humid than the land conditions. Most often, it comes up the valley through that gap SE of Ayia Anna and takes the form of an E wind. There is a narrower gap at Klavdia and this is sometimes more important, making the sea breeze SE or even southerly.

The effect of this is shown schematically above by the dotted lines: the humidity rises suddenly and the rate of rise of the temperature drops or it may even fall. This means that we have a reduced peak temperature and, in fact, the peak may even occur in the late afternoon, after the sea breeze dies down and it reverts more or less to the theoretical curve. This makes forecasting very difficult, as we never can tell, in advance, exactly when, or from what direction or with what force, the sea breeze will arrive. We can therefore make only an inspired guess as to by how much the peak of the normal curve will be cut off, depending largely on our forecast temperature of the sea and the land.

In practice, things aren't always so simple. This chart shows the actual conditions on the 17, 18 and 19 July 2007. In the bottom graph, the temperature is green and the RH blue as in my schematics above. On the 17th, the temperature rose normally and nearly peaked when the sea breeze arrived in mid afternoon, quite a violent one, the RH suddenly rising and the temperature dropping. The strong, gusty, wind (top and middle graphs) changed from N to S in minutes and then dropped as it veered to west and the sea breeze diminished. On the 18th, the effect was less pronounced and it occurred earlier, just after midday. On the 19th, there was no pronounced sea breeze and the wind slowly veering from N to S did not show any significant change of temperature or humidity.

This graph is an example of a sea breeze coming in at about 1130 and stabilising the temperature rise. This breeze has died down by 1500 and the temperature starts to rise again to a peak at about 1540. Notice the wind direction during the time that all this is happening. The sea breeze veers slowly from east to south but, when it stops, it reverts to a prevailing northerly wind.

The other effect that I mentioned is a tendency towards a katabatic wind in the afternoons. A katabatic wind is produced by convection. The cooler air at the top of a mountain is heavier than the warmer air in the valley. It rolls down the hillside until it meets the warmer air which rises, cools and the cycle is repeated. With the hills round Mosfiloti, the effect is relatively small, compared with the southern slopes of the Troodos massif, where the winds can be quite significant in the afternoon. Around Mosfiloti, the average wind speed from this effect is unlikely to exceed 2 or 3 m per second. The direction of a katabatic wind depends on the location of the nearest hill. In the case of my weather station site, it comes mainly from Mount Pipis, southwest of us.

Because katabatic winds are rather unpredictable, it is very difficult to forecast them. We can make an attempt with diurnal corrections, but the accuracy may leave to be desired. However, their influence is very rarely more than a drop of 1°C.

Our weather forecasts are generally uploaded to the Internet at just before 0930, 1330 and 1830 hrs. the data for these forecasts are collected 30 minutes beforehand. Temperature forecasting at 0930 is the most tricky, because we do not know whether a sea breeze will cool the natural cycle or at what time. By 1300, we can already have a good idea about whether a sea breeze has come in or is likely to come in. As a general rule, the 0930 forecast can usually predict to the high temperature to within about ±1°C with a probability of about 90%. The 1330 forecast should be better than half this.

So far, we have touched only on the summer daytime temperature. Unless some specific weather front passes, the night-time temperature is usually fairly predictable with little macroscale influence. The most important factor is whether there are clouds or not. With a clear sky, the heat accumulated by the soil during the daytime can radiate outwards more easily than if there are clouds. This means that cooling is more rapid and the minimum temperature will be lower.

In winter, the effect of position is important. There is a tendency for cold air to collect in the valleys on still, cloudless nights. This means that the occasional frost can be more severe in the lower ground round the village than slightly higher. I have observed this when sensitive plants in gardens are more frost-burnt than similar plants on higher land. I've also seen hoar frost on the ground in the vicinity of the Kokkinou restaurant, when there is only a dew where we are, just a few metres higher and a couple of hundred metres away. It is under these conditions that microclimates have the most effect. We have a row of bougainvillea plants along the side of our garden. The most southerly three plants invariable lose their leaves, each winter, while the others don't. This is because the cold air coming down from Mt. Pipis creates a microclimate at that end. I have not made scientific measurements but, judging from observations, I estimate that the minimum temperature in this patch of land drops to about 1½°C less than the rest of the garden under these conditions. Of course, if there is a strong wind during frosty weather, the mixing of the air ensures that altitude makes no difference.

Forecasting rain

In the Cyprus winter, we normally get a series of synoptic weather patterns where we may get rain, generalised over the whole island, lasting one to three days. This is generally foreseeable, because it comes in from the west or south west with a front that can be seen by our collected data usually two or three days in advance. The exact time of the arrival of rain is less predictable, to within half a day. If the forecast is for four or five days hence, the accuracy is much lower than forecast is for one or two days hence.

Much of the rain in Cyprus comes in the form of thunderstorms. These are utterly unpredictable as to where and when the rain will fall. Generally speaking, thunderstorms happen when humid air reaches land which has been heated by the sun. Initially, convection causes the humid air to rise and form cumulus clouds. Much of the time when this happens, the cloud dissipates without rainfall. If the effect continues, however, the cloud becomes more dense and changes into nimbocumulus, often reaching the bottom edge of the stratosphere at, say, 12 km altitude, with the winds carrying it to form the traditional anvil shape. Inside this cloud, violent up-draughts will take the humidity up to altitudes where the temperature may reach -50°C, allowing hail to form. This activity will also generate enormous electric charges, causing lightning to happen. A typical thunder cloud cell will be only 2 to 5 km across, and the rain or hail will fall only in the centre of the cell. This is why thunderstorms are often very local. On several occasions, I have seen thunderstorms in Mosfiloti, where the village has been soaked but no rain has fallen on the east side, where my weather station is situated.

Because of the random nature of the development of thunder cloud cells, it is impossible to forecast whether a given place will receive rain or not. We can forecast with some accuracy whether the conditions will be ripe for thunder clouds to form and whether they are likely to be widespread or isolated. In our weather forecasts, therefore, we give the probability of rain falling in any particular area, as a percentage and the possible quantity. For example, if you see that the probability is marked at 20%, it means that about 1/5 of the island will receive rain, on the macroscale. This does not mean the rain will fall in any particular area although, of course, there is a greater chance that it will fall on higher ground, such as in the Troodos massif. The same applies to other forms of showers.

To calculate the quantity of rain that will fall, measured in mm, is very difficult. We would need to know, in advance, all the data as to cloud formation, including the quantity of water held in it, the type of cloud, the wind and numerous other factors.

Forecasting wind

Forecasting winds is the most difficult part of the whole procedure, if it is required at a local scale. We can guess with a reasonable accuracy the effect of the synoptic isobars, both in terms of direction and force. This is because we can see where the high and low pressures are over distances of up to 1500 km. As we have seen above, the mesoscale influences are less predictable and may be more important than the synoptic winds.

It is evident that the sea breezes, which we have already evoked, are one of the most important influences on local winds. The katabatic winds also play a role and cannot be ignored under some conditions. When thunderstorms are active, violent winds of unpredictable direction are an inevitable accompaniment. Very occasionally, such winds may develop into a tornado, causing damage. Another local phenomenon is the dust devil, a small whirlwind, 1 to 10 m in diameter and rarely more than 15 m high. These last only a few minutes at the most.

Interpretation

 

Approximate correction factors for weather conditions compared to Mosfiloti
	Time  h.	Temperature °C		Rain  %	Wind m/s
		Summer	Winter
Larnaca	0	-1 to -2	+½ to +1	90	+1 to +4
Limassol	-1	-1 to -2	+½ to +1	110	+1 to +3
Nicosia	0	0 to +2	-1	100	0 to +2 (to +5 if W)
Paphos	-2	-2 to -3	+1 to +3	125	+1 to +3 (to +4 if W)
Protaras	+1	-1 to -2	+½ to +1	75	+1 to +2