MSLP patterns

Isobars join points of equal atmospheric pressure. On one side of the isobar pressure is smaller than the isobar value, on the other side greater. A common metaphor is a topographic map with it's lines denoting height: mountains are equal to areas of high pressure, dales equal to lows. Similarly, we can talk about ridges and troughs (or valleys). Wind follows roughly the direction of the isobars so that if you walk along an isobar and the wind is blowing onto the back of your neck, the lower pressure is at your left hand side (in the Northern Hemisphere .. reverse the rule for the Southern Hemisphere) *. Due to friction, the wind is not exactly parallel to isobars, but turned about 10-30 degrees towards lower pressure.

Furthermore, there are several local effects (e.g. the sea breeze) which affect wind and make this general rule not valid in such cases.

The closer the isobars are to each other, the higher the wind speed. (Meteorologists talk about steep or strong pressure gradient.) There are rules and even scales of how to determine the wind speed from distance between the isobars. When using them, be sure to check that the isobars are between the intended steps (usually 5 hPa but sometimes 4, 8 or 10 hPa) and the scale and projection of the underlying map.

Between tropical and arctic areas, at so called mid-latitudes, the isobars tend to form closed circles round the lowest pressure. These are called frontal depressions, and most of meteorological interest at our areas is concentrated in them. Tracking the frontal depressions is the main reason to look at MSLP chart.

(* this was first defined & explained by a Dutch meteorologist, Buys Ballot in 1857, and the 'Law' is named in his honour.)

500 hPa patterns

Level of 500 hPa is roughly dividing the mass of the atmosphere in two. It lies near 5 km, and it's height is typically analysed at intervals of 40 or 80 m, corresponding to MSLP isobars at intervals of 5 or 10 hPa.

The shapes of 500 hPa isohypses are similar to those at the MSLP isobars: even here you can find lows, highs, ridges and troughs, though the latter two are more usual than closed circles. Formations are much smoother because the underlying surface (except in the high mountain areas of the world) doesn't have an effect at these high altitudes (and because observations are sparse).

If you compare 500 hPa and MSLP charts, you often find a family of small surface lows below an upper level low. Also, the weather under a 500 hPa low or trough tends to have more precipitation, at least showers, even is there are no fronts present.

For numerical models, 500 hPa height (and especially the stream pattern it reflects) is much easier to get right than the exact location of individual surface lows and fronts. So we look at it to get a general outlook of weather, especially to talk about 5-10 days ahead.

Spaghetti plots

So named because of their resemblance to a certain Italian food. Spaghetti plots are another way of viewing Ensemble forecasts. Each of the individual forecasts are merged into one image. To reduce the confusion a bit, only two or three contours are drawn. The purpose of a spaghetti plot is to give the user some idea of the uncertainty in the forecast. When the contours are drawn in the same place, a high level of confidence can be used in conjunction with the Ensemble forecasts; when the plot resembles it's namesake (a plate of spaghetti) then the forecast has a very low confidence level, and it is probably best not to make any plans based on this forecast. Generally the ensemble members diverge with time, making forecasts further into the future much less certain.

Cloud forecasts

The Cloud cover forecasts give an output relating to the percentage of sky covered with cloud. Hence 100% would be overcast, 50% (equivalent to 4 oktas) predicts a half cloudy, half blue sky. This is useful for picking out the frontal systems and determining where they are most likely to be the most intense from the higher percentages. Low cloud (roughly below 6000 ft / 1800 m) is notoriously difficult to forecast and global NWP models do not usually attempt this output. Therefore forecasts on these sites should only be used to infer cloudiness at medium and high levels. There are other representations of cloud forecast, which may show overlays of cloud at different levels, and some centres output 'boundary layer' cloud forecasts - not sure how accurate these are though, or how defined.

Precipitation (PPN) forecasts

[ NB: Precipitation is a collective term which includes rain, snow, sleet, hail etc. ] These charts are usually of rainfall accumulation (or the amount of rain that would have fallen if dealing with snowfall). That is, the amount of rain predicted to have fallen in the preceding 12 hours. Hence a value of 6 on a forecast chart for 1200 means that 6 mm of rain should fall between midnight and midday of the day in question. [NB: 1 mm of rain depth equates to 1 litre of rain per square metre - some centres will in fact output rainfall forecasts in this manner, i.e. 40 litres/m2.]

NOTE: Some of the forecasts are of 6 hour accumulations, 24 hour accumulations, or even of 1 hours accumulation. The charts are usually superimposed on another field, usually surface pressure, although the COLA-MRF images (in Wetterzentrale site) are shown with vertical velocity.

850hPa temperature fields

850 hPa level is roughly at 1.5 km, usually above the atmospheric boundary layer. That means there is no diurnal temperature variation, and the underlying surface such as cool sea doesn't affect it's temperature. That is why 850 hpa temperature is used to distinguish air masses and thus to locate cold and warm fronts.

Because the models have had several problem in surface parametrisation, 850 hPa temperature forecasts have been more accurate than those for lower levels. So we used to bring the 850 hPa to surface by adding 15 (or 10) degrees, and use it instead of surface maximum temperature. [***]
WARNING: all this works only at low altitudes, not in the mountains. It works only when the sun is heating the ground: not at the sea or windward coast, not at night-time, not at winter. 15 is the continental value (for dry adiabatic lapse rate); in Ireland they use 10 (for moist adiabatic lapse rate) and I guess at the Alps the 850 hPa temperature is more or less equal to the surface temperature.

You can see from all these warnings, that using the 850 hPa temperature to infer the screen temperature is problematic at best!

[ *** To a first approximation, you can also use T850 to assess the surface temperature in persistent precipitation, provided the layer 850 - surface is near-saturated: in this case, add ~7 degC (i.e. obeying the Saturated Adiabatic Lapse Rate) to the forecast temperature at 850 hPa. ]

Vertical motion fields

The magnitude and sign (i.e. whether "upward" or "downward") of motion in the vertical is obviously important in operational meteorology. All NWP models output such guidance - usually expressed in terms of the variable 'omega' (w), which is defined as the rate of change of pressure with time. Pressure is used, rather than something more intuitive like vertical velocity, because upward/downward motion in the atmosphere due to dynamic forcing (e.g. frontal ascent), comes about as a result of a disturbance of the "hydrostatic equilibrium" - or the balance of forces in the vertical between gravity trying to collapse the atmosphere & the upward, opposite - directed pressure gradient force.

The 'Omega equation' which is used to calculate vertical motion is made up of two terms: an element due to the rate of change of vorticity advection with height, and an element due to horizontal thermal advection. It turns out that solving these terms gives upward motion (ascent) a NEGATIVE value, and downward motion (subsidence) a POSITIVE value: this can be confusing when first looking at these charts, but it might be easier to remember that 'minus' values are tied to falling pressure values or '-ve' pressure tendency (i.e. potential development) & 'positive' values with rising pressure values or '+ve' pressure tendency (i.e. weakening systems, or building anticyclones). (A little more on this HERE).

Vorticity fields

We live on a spinning planet and air in motion tends to pick up a 'turning' moment due to this rotation at every point, except right on the equator (where the local horizontal component of the earth's angular rotation is zero). This natural 'twisting' motion, or local vorticity, is due to the Coriolis acceleration at any particular latitude (f).

Any developments in the atmosphere which act to enhance this effect will increase the likelihood (other factors being right - e.g. enough humidity, sufficient thermal contrast) of development leading to wind, rain etc. Any developments which act to reduce this action negate development. Vorticity due to atmospheric developments, or flow patterns, is termed relative vorticity(z) [ relative to earth 'normal' ].

The sum total of vorticity at any one point (Absolute Vorticity or zA) = f + z. For the vast majority of cases, Absolute Vorticity is positive, but Relative Vorticity can be positive or negative, and it is usually (though not always) this latter quantity that is seen on WWW sites.

... and why are these changes in vorticity important? Well, anything that encourages air to spin more quickly (Advection of Positive [relative] Vorticity, or PVA) will lead to a narrowing of an atmospheric column, and it's expansion upwards, leading to formation of cloud & precipitation .. if of course there is sufficient humidity. The opposite mechanism (Advection of Negative [relative] Vorticity, or NVA) will lead to a broadening of a column, allowing it to contract vertically, with the air subsiding - a non-developmental state.

Thickness **

Thickness (as used on many of these sites) is a useful tracer of warm and cold air in the lower 'half' of the troposphere (surface to about 5.5km), where a large part of the 'real weather' is to be found. Bulk movement of warm air (high values) and cold air (low values) can be followed by looping the fields. In particular, note carefully areas where the thermal contrast is increasing as these are regions of potential major baroclinic development.

However, be aware that working out things like maximum temperatures, snow-level etc., from this rather crude measure is fraught with difficulty and best avoided.
For more, you can follow the links in the FAQ entry for Thickness ... here

** NB: Some will know this quantity at 'Relative Topography', hence the 'Re Top' seen on sites using the German language.


More and more you will read / hear the term 'ensemble' in modern-day weather forecasting. Beyond about 3 days (and in some circumstances less than that), it is not very useful to treat a particular model frame as the answer to the forecast for that particular time-step, but much better to adopt the ensemble techniques as briefly described in the FAQ Glossary, the relevant entry for which you can find ... here. You might also like to read the entry on the 'Poor Man's Ensemble Technique' ... here which explains how YOU can effectively use the various models here described in this fashion.

Jet Streams

On the broad-scale, these are 'key' to deciding the overall complexion of the weather-type in any one geographical region in the mid-latitudes, i.e. roughly between 35 and 75 degrees N/S. The various entries in the FAQ & Glossary are useful starters ... for a basic definition, follow the link: here, and to delve even more deeply into the mysteries of upper air patterns etc., try this page .... here

Equivalent Potential Temperature ('Theta - E')

A useful quantity because it is calculated taking the humidity content of the air into account - and as most 'interesting' meteorology is to do with atmospheric moisture, Theta E (QE) is a better 'tracer' of air-mass property (at 850 hPa**) than temperature alone. It is a conservative property (doesn't change much) during both dry and moist adiabatic processes.

> The actual value of Theta-E can be useful: it can be used in algorithms to calculate daytime maximum temperature and snow-risk for example (as for Theta-W#). Remember though that this variable couples temperature AND absolute humidity: so high values show areas of warm, potentially more humid air (e.g. tropical maritime or modified tropical continental air-masses); lower values pick out colder, lower-humidity content air (e.g. polar maritime air-mass).
> Where there is a marked or 'sharp' discontinuity in values (best seen on colour-enhanced output), then some attempt at frontal placement can be made.
> The location of axes of plumes of high (orange/red) Theta-E, act as a "focus" for significant thundery activity: a plume of such air running northwards over NW Europe ahead of an upper trough coming eastwards may herald the initiation of a ' Spanish Plume' (q.v.); event.
> Although not strictly appropriate to these charts, it is pertinent to note here that the vertical distribution of Theta-E (and Theta-W) with height is a useful diagnostic of whether an air-mass possesses Potential (or Convective) Instability (q.v). If a column of air is lifted bodily, the temperature decrease varies from level to level, particularly because some layers become cloudy (and acquire the released latent heat) sooner than others. Therefore, certain layers can become unstable simply because they have been 'mechanically' lifted - these can be found by inspecting a thermodynamic diagram (or tabulated list of data), and finding layers [roughly between 900 and 400 hPa] where Theta-E (or Theta-W) decreases with height.
#[ in the UK service, the Wet-bulb potential temperature (Theta-W) is more usually used for these purposes, but is not readily available on external web sites - Theta E will do the same job though.]

** [ Beware of using these parameters at 850 hPa as tracers of air-mass change slavishly. Provided the air in the lowest 2 km of atmosphere is well-mixed, then they are indeed excellent for this purpose. However, in stagnant situations, or where there is marked 'de-coupling' between the quasi-frictionless flow above about 900 hPa and the surface, (or boundary-level) air, changes at 1500 m (i.e. 850 hPa) may not tell the whole story. ]

CAPE & Lifted Index

I've 'bracketed' these two parameters, because they are often considered together:

CAPE: it would be better if this parameter was always referred to by it's full expansion: "Convectively Available Potential Energy"; unfortunately, this doesn't happen, so we have the situation where newcomers to the world of 'severe convection' are bemused by all the talk of a quantity 'CAPE' which is not adequately defined. I'll have a go here:.....

When a parcel is lifted by whatever means from the surface (or another level), it will cool according to the basic laws of thermodynamics. The rate of cooling is dependent upon whether the parcel is 'dry' (i.e. unsaturated) or 'moist' (saturated). The rate of cooling is known as the Dry Adiabatic Lapse Rate (DALR) in the former case, and the Saturated Adiabatic Lapse Rate (SALR) in the latter. By inspecting the temperature of the parcel at any point in its upward travels, and comparing with the 'environment' through which it is travelling (best done on a thermodynamic diagram), it will either be cooler (denser) than the environment (stable), neutral (the same temperature), or warmer (less dense) than the environment (unstable). In this latter case (unstable/parcel warmer than environment), then the parcel (packet, bubble whatever), has Potential Energy, manifest as continued upward motion, which was initiated, and made Available due to Convective activity: C-onvectively A-vailable P-otential E-nergy - CAPE. [ For more on stability, lapse rates etc., see the main FAQ.]

The greater the excess in temperature (parcel warmer than environment), then the greater the energy available - and all other things being equal (e.g. no compensating descending currents due to broadscale dynamic effects), this should result in stronger upward motion, potentially more 'severe' conditions (downdraught gusts, heavy precipitation, larger hail, higher tornadic threat etc., etc.).

CAPE is simply the integration (sum) of the energy that a parcel would have throughout its vertical 'life', once convection is released. It is calculated on a thermodynamic diagram by assessing the area between the Environmental Lapse Rate (ELR/actual temperature trace) and the Parcel trace, provided the temperature of the latter exceeds the former. [ See a crude example HERE ]
The units of CAPE are J/kg (energy per unit mass of atmosphere), and some typical values are:
> ~ 150 - 300: heavy/intense shower-rainfall with/without thunderstorm - mainly 'slight' TS
> ~ up to 1000: 'moderate/severe' thunderstorm
> ~ up to 2500: 'severe/intense' thunderstorm.

(Note carefully: many so-called 'critical' values for CAPE & LI - see below, have been defined based on spring/summer US experience and may not be applicable to European/'all-season' storms - do NOT treat these figures as a 'threshold' or a 'requirement'!)
Some other figures for CAPE are outlined in the main FAQ.

Lifted Index (LI): In many respects, this can be viewed as a very crude measure of the above - i.e. CAPE. It is very easy to assess (probably why it was devised in the first place), and so has gained widespread currency in the operational / severe weather world. It is defined simply as:-
LI=T(500) - T(P);
where LI=Lifted Index (degC), T(500) is the temperature at 500 hPa as found from an actual radio-sonde ascent (or computer/model output) and T(P) is the temperature that a parcel would have, if encouraged to reach the 500hPa level (about 18000 ft or 5600m) by whatever means. [ See a crude example, based upon a Skew T, logP diagram: HERE ]

It will be readily seen that when the parcel is cooler (denser) than the environment (i.e. a 'stable' state), LI will be +ve; when the parcel is warmer (lighter) than same, (i.e. the 'unstable') state, LI will be -ve. The more 'negative' the number, the greater the excess of energy the parcel has at this point, hence it's crude likeness to CAPE (above).

Over the years, some rough figures have been developed which relate the value of LI to expected conditions: note that they must NOT be used slavishly in this respect:
LI +ve: stable, non-convective.
LI=0: neutral - non-severe convection may be possible.
LI -ve: unstable .... and the following are offered as a guide ....IF CONVECTION IS INITIATED!
LI -1 to -4: moderate thunderstorms/mainly small hail,
LI <-4: possible severe thunderstorms/large hail,
LI <-10: possible severe/intense thunderstorm, whirlwind phenomena (e.g. tornadoes), 'giant' hail etc. (see cautionary note above re: applicability to European conditions)

Precipitation type forecasts

(i.e. rain or snow phase)
These charts show precipitation intensity (Niederschlagsstaerke) and precipitation type (Niederschlagsform). For all meteorological NWP output, it is important not to "follow the dots" slavishly, that is, not to believe every twist and turn of the output at every time-step. This is particularly so for this product.

The output is decidedly 'blocky', and the output resolution (i.e. what you actually see), is certainly rather course; a single 'block' appears to cover one or two medium-sized English counties. Whether the model resolution is equally course is not known: indeed, if orography is used (and it would be unusual if it isn't), then it must be very crude.

The output is in mm per hr (mm / h); this is fine for liquid precipitation (e.g. rain), but snow output (rosa / purple / schnee) is also shown as mm / h .. these aren't mm of snow on the ground: this is the rainfall equivalent. As a very rough conversion, multiply the intensity by 10, so 1.5mm / h rainfall-equivalent=> 1.5 cm / h. But remember, that not only is this a highly variable quantity, it assumes that the model has got a perfect representation of the lower-tropospheric temperature / humidity profile, and the precipitation intensity is correct, and the model 'knows' about the surface (snow covered, frozen etc.), and that all the snow settles!

In my view, it would be best to regard the 'snow' phase colours as scales of probability in any one situation: the deeper the shade of purple, the higher the probability of snow reaching the surface .. be very suspicious of the lower-decimal (roughly < 0.8 mm/h) figures for snow rate.

Freezing levels

(or 0°C isotherm levels)
On the Topkarten site, there is a section where you can choose from various fields with 3 hr time-steps (though over a limited domain); in particular, and most useful currently, the field labelled:- " 3h Niederschlag " will show precipitation amounts in millimetres. (I assume where the phase is snow then it is 'liquid water equivalent', though I can't see any legend to that effect), and also shown are contours of the height of the zero degC level above mean sea level (in metres): remember to make an allowance for your location - so if you are 200 m altitude, and the forecast is 500 m amsl, the zero deg.C level will be 300 m above you.

This latter shouldn't be taken as the 'snow-level' as such, because of course snow can descend (depending upon the relative humidity in the boundary layer & intensity) well below this level, but it is a useful guide & provided it doesn't change too much with time, it can be used to infer the 'usable' (for winter sport purposes) snow-line in hilly / mountain areas.

If the freezing level is less than ~300 m above your elevation (remember that the output heights are amsl - adjust for your location), then it is reasonable to assume that there is a >50% probability of PPN falling as snow inland. If it is <100m then there is a >=90% probability of snow. (However, in coastal areas with an onshore wind off a relatively warmer ocean, then these figures will not apply so well if at all). [ Ref: Boyden Met.Mag 1964 ]

The horizontal resolution of the GFS is approx. 40 km out to 180hr (~80 km beyond that), and the output resolution of these charts, according to the site information, is 50km, so for this display you are seeing near-full model resolution, as opposed to the main charts, which are 'thinned' to ~100 km. The GFS is still a crude model to use for 'local' work though, bearing in mind that modern mesoscale models are working down to the ~10km or better level.

The Omega Equation

There is an equation in atmospheric physics known as "the Omega equation": it allows synoptic-scale vertical motion to be accurately estimated by analysing temperature and wind fields above the boundary layer, leading to a value for vertical velocity (allocated the greek letter 'omega'), in pressure co-ordinates (dp/dt).
[ By the way, because pressure decreases with increasing height in the atmosphere, ascending air gives rise to a negative value: therefore on computer output, -ve values imply ascending air, +ve values descending air. ]

In it's broadest exposition, the Omega equation brings together the forcing due to vertical variation of absolute vorticity advection, the thermal advection terms (various) and components due to diabatic warming. These factors can be viewed subjectively (e.g. a PVA area engaging a focus of strong warm advection will lead to strong upward motion and vigorous surface development).


A numerical model though needs to have all this in terms of rigorous mathematical formulae (albeit with some approximations) that can be solved. On the way to calculating all this development, the models spew out all sorts of diagnostics that can be used by forecasters: one such are 'Q-vectors'. These are found by using a 'cut-down' form of the Omega equation developed by Hoskins, Draghici & Davies in 1978.

The vector "Q" is dependent upon the rate of change (with time) of the potential temperature gradient brought about solely by horizontal change in the wind-field (derived using model contour fields, geostrophic.)

Q-vectors are not a measureable quantity like the actual wind or the temperature - they are more akin to the 'thermal wind', with which they are in fact allied.
They are derived, as indicated above, by manipulation of the Omega equation. In brief, the concept of the Q-vector is a mathematical representation of the atmosphere's attempt to restore the thermal field once it has been disturbed from its 'equilibrium' state of both hydrostatic (1) and geostrophic (2) balance. This 'restoring force' is deemed to be via ageostrophic motions, which are most important in determining where development is occurring.

(1): hydrostatic balance: the 'struggle' between the pressure gradient acting upwards, due to decreasing pressure with height, and the force of earth's gravity acting in the opposite direction - the hydrostatic equation represents this process.
(2): geostrophic balance: the 'struggle' between the pressure gradient (all planes, though usually taken as horizontal) and the deflection due to the earth's rotation - the equations of motion represent these.]

Where Q-vectors are convergent (i.e. negative divergence), the forcing is normally associated with tropospheric ascent (or low-level development / falling pressure.
Where Q-vectors are divergent, the forcing is normally associated with tropospheric descent (or low-level decay / rising pressure ).

Also, where Q-vectors are directed from cold to warm air, (at 850 hPa) this implies frontogenesis (a tightening-up of the thermal field); the opposite implies frontolysis (or frontal decay).

On various charts available via the Westwind site:-

FQ :=Forcing due to Q-vector convergence / divergence.

FQ: +ve divergence==> descent / non-developmental
FQ: -ve divergence==> ascent / developmental (or convergence)
The higher the number (+ve or -ve), the deeper the colour and the greater the forcing term. These data are usually shown in association with the 500 hPa field, roughly at the level of non-divergence (or maximum vertical velocity).


:=the component of Q 'normal' (i.e. at right-angles) to the isotherms and for most work, the isotherms used are those at 850 hPa; for mountainous terrain though, 700 hPa should be used.


FQn: strongly -ve in cold air=> frontogenetic


FQn: strongly +ve in warm air=> frontogenetic


Best used to find areas of 'strong blue' (positive) immediately adjacent to 'strong red' (negative); this implies synoptic-scale processes in place enhancing frontal development / ascent etc (or frontogenesis).



Example of CAPE & LI

 Example CAPE & LI on SkewT, LogP  On this example (of a Skew T, log P diagram), CAPE is the area enclosed by: X - A - TOP - B - Y - X (shaded green); Lifted Index is given by the difference (observing the sign) between the temperature (B-A): see explanation below:....
 A  The point (and associated temperature) where a saturated parcel given this particular ascent crosses the 500 hPa isobar=T(P)
 B  The actual temperature, T(500), from the radio-sonde ascent in this particular situation. The difference B-A [ or T(500) - T(P) gives the Lifted Index.
 X  The temperature at the surface (SFC) which initiated this particular parcel's ascent. From X to LFC, ascent-cooling will be at the Dry Adiabatic Lapse Rate (DALR).
 Y  The lowest 100 m or so of the ascent as given in the radio-sonde ascent; this will become heavily modified due to afternoon heating.
 LFC  The 'Level of Free Convection', above which the parcel, rising from the surface, will be saturated, and cool at the Saturated Adiabatic Lapse Rate (SALR), in this case from LFC to TOP.
 TOP  The point at which the temperature of the parcel=temperature of the environment, and buoyancy is neutral; the parcel theoretically ceases to rise of it's own volition - but given enough positive energy below this point, tops may (& often do) 'overshoot'. (also known as the 'Equilibrium Level')


ThetaE, ThetaW & derived parameters

The relationship between Theta E, Theta W and theoretical maximum temperature and snow probabilities (latter based on Bradbury, 1970).



1. The relationship for day maximum temperature was developed specifically for SE England. In mid-winter in particular, at latitudes further north, the expected maxima will be lower. Also note that the figures are not necessarily based on a plentiful data-set over the whole range of expectations; for example, at the very high and very low end of the scale, the numbers of events used to find the theoretical 'base' maxima will have been much smaller than for the 'middle', or most-likely area of the data-set. With experience, you may indeed find your own figures are better than these, so keep a record of what happens in each occasion.
2. The figures will not apply over a full snow-cover, as insolation gives a different heating response over such surfaces.
3. The techniques use the data from 850 hPa (or mbar), which is very roughly at 1500m (or 5000 ft) amsl. Under conditions of strong anticyclonic subsidence, this level may not represent the air mass in the lowest 50 hPa of the atmosphere. Cold-undercutting (and indeed warming over sea surfaces) should be allowed for. Also, in cases of very high pressure (over 1035 hPa / mbar), the 850 hPa / mbar level will be even higher up, and will represent the near-surface even less.
4. Once you have found the basic working figure for maximum temperature, 'correct' for cloud cover using the table given. Note that basic figures & corrections are only given to the nearest 1degC and the results should not be used slavishly - arguments involving the odd degree between one situation and another are pointless given the approximations involved & the accuracy of the raw input data (whether model or ascent-based).
5. In the case of the snow risk, remember it is for mean sea level. You need to adjust for altitude using the adjusting table given.

Method of use:
(a): Find expected ThetaE (Equivalent Potential Temperature) over the area in question at around the middle of the day - from latest NWP output, or from radio-sonde ascents.
(b): Against the appropriate value of ThetaE in the table below, look up the value for that month - interpolating / adjusting as necessary.
(c): Look at the list of corrections given underneath and choose appropriate value: apply.
(d): For snow probabilities, find the base (i.e. msl) figure from the extreme right-hand column.
(e): Work out the altitude of the site in question, and using the table of adjustments, find the % risk of snow for that altitude. (IMPORTANT NOTE: In my experience, these figures are too pessimistic and should be used as a very rough guide only. I find that you need persistent, moderate or heavy precipitation before these figures are relevant. What I do is use the 50% equivalent isopleth at msl ... 3degC ThetaW / 16 degC ThetaE ... as a 'tracer' for 'snow-worthy' air at lower levels, then use other means to decide on the precise risk.)

 850 thetaE
 850 thetaW
 Jan.  Feb.  Mar.  Apr.  May  Jun.  Jul.  Aug.  Sep.  Oct.  Nov.  Dec.  Snow
Prob. (at msl)
 54  18  19  20  21  23  25  26  26  25  24  22  20  19  
 46  16  17  19  19  21  23  25  25  23  23  21  19  17  
 40  14  16  17  18  20  22  23  23  22  21  19  17  16  
 36  12  15  15  17  19  21  21  21  21  19  17  15  15  -
 32  11  13  15  15  18  20  21  21  20  19  17  15  14  -
 30  10  13  14  15  17  19  20  20  19  18  16  14  13  -
 26  8  11  13  13  15  17  19  19  17  17  15  13  11  1%
 22  6  9  10  11  13  15  16  16  15  14  12  10  9  14%
 18  4  8  9  10  12  14  15  15  14  13  11  9  8  34%
 14  2  6  7  8  10  12  13  13  12  11  9  7  6  78%
 10  0  4  5  6  8  10  11  11  10  9  7  5  4  96%
 8  -1  3  4  5  7  9  10  11  10  8  6  4  3  100%
 6  -3  1  3  4  6  8  9  10  9  7  5  3  2  100%
 4  -4  -1  1  1  4  6  7  7  6  5  3  1  1  100%
 2  -5  -3  -1  0  2  4  5  5  4  3  1  0  -1  100%
 0  -6  -5  -3  -1  0  2  3  3  2  1  -1  -2  -3  

Corrections to be applied (for the maximum temperature figures):

A day of heavy overcast with precipitation: apply correction to above of at least -2 to -3degC, and for air-masses laden with cloud and having high low-level humidity (spring / summer), then the correction may be as high as -5degC.
A day of bright, virtually uninterrupted sunshine (but no snow cover) apply + 1degC. On days of strong sunshine over a dry ground (in a relatively warm air-mass for the season), then correction is at least +2degC, and may be up to +3degC.

Use the following table to adjust the % snow risk (at mean sea level) in the above to the required elevation:

 Elevation (in metres)  % PROB at msl >>>>  20%  40%  60%  80%
 50m    28%  50%  68%  84%
 100m    39%  59%  73%  89%
 150m    50%  68%  78%  93%
 200m    59%  73%  83%  98%
 250m    66%  78%  88%  100%
 300m    72%  83%  92%  100%
 350m    77%  88%  98%  100%