Knowledge base

The website meteo.pl deals primarily with short-term forecasts rather than global climatology, but we will try to answer briefly: The media report on the effects of significant climate warming - more intense tropical cyclones, heat waves, droughts or floods in various regions of the world - but we hardly ever learn about the causes behind these phenomena. Apart from tropical cyclones, which are fairly regular and located in a specific climate zone, it is the heat, droughts and floods that are caused by changes in, or characteristics of, atmospheric and oceanic circulation - the underlying causes of climate change. It should be mentioned that it is not only ocean currents that transport the energy absorbed by the Earth in the equatorial and subtropical zones towards the poles, the transfer of energy by latent heat (energy is absorbed on evaporation of water and given back when it condenses in a cloud) and the transfer by "normal" advection are equally important. When analysing climate change on the basis of observations made regularly over the past 150 years, it should be remembered that the geological epochs of cooling and warming of the Earth's climate - the Miocene, Pliocene and Pleistocene - have lasted for 23 million years, while the warming in the current Holocene epoch is only less than 12 000 years old. It is like comparing a year to its last 4.6 hours, and in this calculation 150 years of observation has a measure of 210 seconds. Studying climate is not just about checking the temperature, this element is the result of many processes - the circulation of the atmosphere and oceans, the phase changes of water, the formation of clouds and precipitation - as we constantly write about in our comments, showing by example how variable the atmosphere is.

Concepts of atmospheric convection are evolving as measurement techniques (satellite and radar and automated precipitation intensity measurements) develop. In the formal terminology contained in the Meteorological Dictionary, PTE/IMGW 2003 the notion of a supercell and, even more, of a mesoscale convective system does not appear. There are no professors of synoptic meteorology in Poland, storm hunters simply transfer concepts from the USA to our local ground. But remember that NY is at the latitude of Naples and the northern border of the USA is the southern edge of Poland. In addition, there are frequent inflows of air from the Gulf of Mexico, from disappearing tropical cyclones, with a significant water vapour content and thus enormous variability energy. The course of convection phenomena on the Great Plains has nothing in common with the course of convection over the Central European Lowlands; additionally, we lie in the shadow of the Alps, Sudetes and Carpathians. We began to introduce the concept of the supercell in our commentaries published from 1998 onwards, when it became apparent that the numerical system was forecasting "peacock eyes" of rainfall in excess of 30 mm/3h every summer. Going back to the definition - we consider a supercell to be an ensemble of multiple CB clouds, covered by a common anvil, with thunderstorms and heavy rainfall of course, but the ensemble does not produce its own circulation other than interacting in close proximity (20-30 km). A supercell is 80-200 km in diameter under central European conditions. A mesoscale convective system (in English MCS) produces its own longer-lived circulation in the form of a wave or a small low-level centre on a meridional front, or forms without a front in an air mass with significant variability energy and water vapour content. From numerous satellite observations and sometimes from media reports, such MCS in and around Europe can be observed south of the Alps, the southern Carpathians and over the Mediterranean and Black Sea coasts; no typical MCS's are observed over Poland.

The criterion for classifying a tropical cyclone as a hurricane is not the pressure value at its centre but the wind speed, which should exceed 32 m/s near the centre of the vortex, in which case the tropical cyclone is declared a category one hurricane. One and the other - pressure, or rather pressure gradient, and wind - have an undoubted relationship, but practical considerations suggest that wind speed should be used to define the category of hurricane along with the impact it has.

EQUIVALENT-POTENTIAL TEMPERATURE - represents the amount of heating of air due to condensation of all water vapour contained in it (at a given pressure), followed by adiabatic (without heat exchange with the surroundings) reduction of so heated air to the pressure of 1000 hPa. Te-p isotherms are depicted on a map of the 850 hPa area (altitude approx. 1.5 km above sea level), since the position of atmospheric fronts and the area occupied by Arctic, polar and tropical air can be seen more clearly. The 850 hPa surface map also shows air temperature isotherms, but does not indicate the areas occupied by individual air masses.

The low is a vortex with a vertical axis, the isobars have a cyclonic curvature, the air movement is counterclockwise, the pressure is lowest in the centre of the vortex. The magnitude of the pressure is not the most important; for practical reasons the cyclonic curvature of the isobars causing the wind to converge and forcing the air to move upwards is the most important. As a result, clouds form, then precipitation. A similar situation occurs in the case of a high, which is also a vortex with a vertical axis; its other features contradict those of a low: anticyclonic curvature of isobars, clockwise air movement. Instead of wind convergence in the area covered by the low, there is wind divergence and descending movements. These lead to the formation of a subsidence inversion (a layer in which the temperature increases with height; when there are no descending movements the air temperature decreases with height). Inversion, more precisely temperature inversion, has an important influence on the vertical development of cumulus clouds.

JET STREAM - a very strong airflow occurring near the tropopause. It is most often seen on weather maps at 300 hPa (about 9 km). It is characterised by a large wind speed gradient (decreasing wind speed per unit distance from the p.s. axis). The wind speed gradient is a measure of the vorticity that is transported to the lower layers of the atmosphere becoming a starter low on the atmospheric front. If there is no generation of a low, then the isobars take on a cyclonic curvature or change their previous curvature towards a cyclonic curvature. The changes described above take place in the area to the left of the p.s. axis, in the area to the right of the p.s. axis the anticyclonic vorticity is transported. VORTICITY - is defined as the rotation of a speed vector. The term made a career when in the 1940s a theory of the development of baric systems was formulated; the change in vorticity was taken as a measure of the development of a low or a high, rather than a mere drop or rise in pressure.

ADVECTION - is the horizontal movement of air resulting in changes at a given point in e.g. temperature, humidity and other indicators describing the weather so far.

The front wave precedes the formation stage of the low. The closing of the first isobar around the top of the wave is a sign of the formation of a young low. As it develops it becomes a mature low when the occlusion of atmospheric fronts begins. The low is a vortex. For a vortex to form from a front wave, it must be fed by cyclonic vorticity. It may come from a heat source; a typical area for such a source is the Mediterranean Sea in the cold season, or from a jet stream - in the latter case, the season is irrelevant.

Following your comments on the numerical weather forecasts, I have realised that when Arctic winds blow, it will be very cold, and when polar maritime air arrives, it will also be cold, but not as cold, and wet. Until now, I thought that the Arctic lay in the polar regions (or vice versa), or in any case that it was the same direction. Meanwhile, it turns out that Arctic winds are colder than polar winds, and that snow from polar directions is wet, while that from the Arctic is dry. Where is this meteorological Arctic then, and where are these meteorological polar regions? Because geographically speaking I suppose they are in the same place? In meteorology, the circumpolar area north of the 70th parallel is taken to be the Arctic area. The air over this area, which is covered with snow and ice (melting snow and ice in summer), takes on certain characteristics that distinguish it from air formed in other regions and is called Arctic air. (AA) The second area over which an air mass is formed that is radically different from the previous one is the area of subtropical highs (the Azores High in the Atlantic). The northern boundary of this area is near the 45th degree of latitude. The air residing in the Azores High area is warm and humid and is called tropical air (TA). When it reaches us, we experience it as mild air; no other air mass has this characteristic. The remaining part of the Atlantic, located north of the 45th parallel, in meteorological nomenclature is the polar area; here the polar air (PA) is formed. It is intermediate between warm and humid tropical air and cold and low vapour Arctic air. We therefore have three air masses; two fronts are sufficient to separate them. The Arctic front separates the AA from the PA, the polar front separates the PA from the TA.

The pressure field determines the direction and speed of the inflow of air masses at all levels of the atmosphere. Air masses, forming over warm or cold, dry or humid areas, bring a particular type of weather. All this is further complicated by processes on weather fronts and the impact of local conditions. Local conditions have a decisive influence on the type of weather and the course of atmospheric processes when there are no or negligible changes in temperature and humidity caused by the influx of a new air mass. And although there are no "pure" local or advective cases in weather, at least these two factors compete with each other; the type of weather determines the one whose contribution is greater.

In principle, yes, although not always. For example, UM model forecasts starting at 00 and 12 o'clock use a wider set of initial conditions than forecasts at 06 and 18 o'clock, so they tend to be more reliable. The second reason, more general — the model needs, especially in summer, during the period of intense convective phenomena, the so-called "run-up". This means that the first hours of the forecast may not yet show heavy rainfall or thunderstorms. It is then worth checking the forecast calculated earlier.

This is usually because the forecasts from the global models, prepared outside ICM, on which we run our calculations, have changed. The second reason is that the more distant the forecast, the more uncertain and less reliable it is. Which is not to say that there are no exceptions to the rule.

The mechanism for rainfall in the second half of the night and thunderstorms in the morning involves radiative cooling of the upper cloud layer. Rain in the second half of the night occurs when an unstable equilibrium develops in the upper cloud layer. It takes several hours for a cloud to develop into a precipitation cloud and rain begins to fall from the cloud formed under such conditions in the morning. A thunderstorm with some thunder spectacularly ends the convection in the middle troposphere initiated under the cover of darkness.

As we wrote earlier, the horizontal distance between adjacent nodes (grid step) of the UM model is 4 kilometres. The relationship between model spatial step and resolution is such that the smallest perturbation that can be described by the model has a length of two grid steps. Thus, for the UM model with a grid step of 4 km, all phenomena smaller than 8 km are sub-grid in nature, i.e. they cannot be resolved explicitly. The only way to determine their possible occurrence is through parameterisation, i.e. statistical linkage between large-scale variables and sub-grid processes. Hence the difficulty in accurately describing the occurrence of a local storm in the 4 km model.

The warm season is characterised by the development of daytime cumulus clouds with occasional rainfall. In the evening and at night the clouds disappear and the next day the cycle repeats itself. The horizontal extent of precipitation from these clouds ranges from a few hundred metres to a dozen-or-so kilometres. Sometimes they are larger when the clouds merge into larger assemblages. The UM model has a computational grid step of 4 km, so the precipitation field could be smaller than the grid step, or be comparable to the grid step. Reliability of precipitation forecasts tends to increase in the cool season, where clouds form due to slipping over the frontal area. These are stratus clouds and the width of the precipitation zone reaches up to 300 km, its length much more. Zones of precipitation from such clouds do not have "holes", as is the case with precipitation from cumulus clouds.

Forecasting rainfall is one of the more difficult tasks of numerical models. Sometimes the forecast error is due to specific local phenomena (see the local cloudburst question) and sometimes, for example, to erroneous boundary conditions fed into our models. To determine the probability of precipitation, we recommend area maps (so-called detailed maps), which allow you to see whether, for example, you are in the middle of a large area with precipitation or on the edge of it.

Meteorological models for local areas, such as Poland or Europe, require initial atmospheric state and boundary data. The source of these data are global models, i.e. models that calculate the weather forecast over the entire globe. For the UM model, data is provided by the global UM model calculated at MetOffice in the UK. In contrast, the WRF models are fed by the GFS model calculated at NCEP in the USA. The UM and WRF models are not fed with direct observational data. Observational data comes from meteorological services carrying out measurements in their own states, as well as collecting data from satellites and the sea. So, they are implicitly included in the process of weather forecast generation in our models, through the mechanism of their assimilation in global models.

The most obvious difference between the models, at first glance, is the grid step (this parameter is mentioned when deciding which model to choose: UM 1.5 km, UM 4 km, WRF 3.4 km). For this reason, the 1.5 km UM can include many fine details that are not mapped on the sparser WRF model grid. The difference in the number of calculation levels, UM 70 versus WRF 48, is important for the diagnosis and prediction of the vertical structure of atmospheric temperature and humidity. A larger number of computational levels results in better forecasts of cloud types, cloud thickness, vertical extent and a better starting point for precipitation forecasts. Each model slightly differently parameterises (approximates) various physical phenomena such as radiation transfer, cloud and precipitation formation, etc. The models also have different substrate characteristics (different grid and different substrate classes) and are fed with data from different global models, which sometimes affects the outcome of the final forecast.