1. Introduction: squalls and their relation to cumulonimbus storms Squalls are an integral part of the cumulonimbus lifecycle. The intense rainfall which accompanies cumulonimbus events forces a downward flow of air (the downdraught) which spreads horizontally outwards when it hits the surface. The winds associated with this horizontal outflow are the squalls which are the object of this study. The study region is well known as a zone of intense cumulonimbus storms and associated squalls; the severity of these events is in notable contrast to the relatively benign climate of the region at other times.
In order to understand and predict the occurrence of these squalls we need to break the problem into two components:
Firstly we need to deal with the likelihood of occurrence of cumulonimbus events, including the type of cumulonimbus expected;
Secondly, we need to determine the nature and severity of the squalls which may result from a cumulonimbus storm which occurs.
Random / statistical nature of cumulonimbus
Cumulonimbus storms are poorly predicted worldwide. This problem is primarily due to the inherently random nature of cumulonimbus convection. Cumulonimbus storms occur where there is conditional instability, and require a ‘trigger’ for each event to occur (see section 3). The characteristic cumulonimbus timescale is around 30-60 minutes and therefore events can develop very rapidly relative to desired forecasting lead times. This means that small errors in the description of those processes which initiate a cumulonimbus storm lead to explosive errors in the specification of the resulting storm. Furthermore, cumulonimbus cells breed new, daughter cells, so that errors accumulate and amplify. Accurate prediction of individual cumulonimbus events in the study area is therefore not possible over periods of more than a few hours. What we will aim for is a statistical description of cumulonimbus occurrence. We aim to predict the occurrence of cumulonimbus events within a spatial region broader than the events themselves, and then to predict the nature of the storms in such a region once they occur. This can be related to a probability of squalls influencing a fixed station within the broader spatial region.
Figure 1: Meteosat infrared satellite image of a field of convection in the study area (2 April 2004). Rough outline of forecasting strategy for tropical convection
Numerical weather prediction models are still not reliable for the prediction of cumulonimbus occurrence, whether the occurrence of individual systems or the existence of broader regions of cumulonimbus activity. Since cumulonimbus events are the key forecasting issue for the tropics, tropical forecasters need to use more empirical methods to predict the likelihood of cumulonimbus convection. The basic ingredients for such methods are:
Observations and numerical weather prediction output
Although the NWP models are not reliable in their representation of cumulonimbus, their short-term forecasts of winds and temperatures are good, particularly above the lowest levels of the atmosphere (e.g. Thorncroft et al. 2003). Forecasters use these model predictions, along with recent surface-based observations, to generate predictions of cumulonimbus occurrence ‘off-line’. Typically a forecaster will take vertical profiles of winds, temperatures and humidity from a weather prediction model, then use these to compute measures of cumulonimbus likelihood.
Empirical rules, stability indices
Given a large-scale environment, a number of rules and indices can be employed to predict the likelihood and nature of cumulonimbus events. These are reviewed in section N below. Although they are based on scientific reasoning, the direct application of these methods is empirically derived (e.g. certain thresholds may need to be determined). It is one aim of this project to test and calibrate these methods for the study region.
Some cumulonimbus systems are long-lived and propagate coherently. Observations of such systems in satellite data and ground observations can be used to extrapolate their subsequent behaviour for a short lead time, typically a few hours. These methods can be enhanced by knowledge of the environment into which a storm is moving (information perhaps obtained from a weather prediction model); this process is known as ‘nowcasting’.
Recent history of cumulonimbus
The recent history of cumulonimbus occurrence leads to effects which can either promote or hinder subsequent storms. cumulonimbus events, particularly over the oceans, lead to fluctuations in low level properties (temperature, humidity and winds) which can allow subsequent systems (including ‘daughter cells’) to develop more easily on timescales of an hour or so. Inversely, the effect of widespread cumulonimbus convection is ultimately, over a period of several hours, to stabilise the atmosphere, making subsequent events less likely. Balancing these opposing effects requires practical experience of forecasting in the study area.
The local geography of a given area influences cumulonimbus occurrence. For instance, over land cumulonimbus is preferentially developed over mountains. In the northern part of the study region, convective systems generated (or regenerated) over the Mali wetlands or the Guinea Highlands often propagate out over the Atlantic (e.g. Aspliden et al. 1976). More fundamentally, forecasters in a given region develop familiarity with local weather patterns and tools with which to analyse these patterns. In the WAM region for example, certain measures of the monsoon layer depth and northward extent of the Inter-tropical front (ITF) are used to assess the likelihood of convection over the continent.
2. West African climate – significance for squall formation The study region lies close to the west coast of continental Africa, and the occurrence of cumulonimbus squalls is very strongly related to the patterns of atmospheric temperatures, humidity and winds over the land as well as the ocean regions. In much of the study area, the midlevel winds have an easterly component and therefore cumulonimbus storms which are initiated over the land are steered by these winds over the ocean. Aspliden et al. (1976) showed that the large majority of intense cumulonimbus systems observed in the GATE ship array originated over land. In order to understand and predict squalls over the ocean regions, it is necessary to understand the cumulonimbus occurrence over Africa, and its downwind evolution.
Monsoon / ITCZ structure
The climate of the study region is dominated by the West African monsoon (WAM), which is, in turn, closely tied to the seasonal movement of the inter-tropical convergence zone (ITCZ). The ITCZ is the band of deep, cumulonimbus convection which lies close to the equator throughout the tropics, approximately following the zone of warmest SSTs. Since the summer monsoon brings the much-needed rainfall to west Africa, and is also related to the generation of hurricanes in the Atlantic, the northern summer season has been studied far more deeply than the winter season. However, the winter season remains important for squalls in the Gulf of Guinea, and the southern region of the study area in particular.
Figure 2: Map of region with summer monsoon patterns: monsoon trough (close to the ITF position), African easterly jet (AEJ), and low level monsoon layer winds. The zone of peak convective activity in the ITCZ lies around and to the south of the AEJ, although cumulonimbus events can occur well to the north and south of this. Summer
In the northern summer (between June and October), the monsoon trough (a band of low pressure) moves north over the southern Sahara, to around 23N, following the peak solar heating of the surface. The low pressure in the monsoon trough causes the low level winds to be drawn, as southwesterlies, from the Gulf of Guinea over the African continent, and therefore draws moisture over the land (Figure 2). In consequence, the supply of moisture in these monsoon winds allows the band of peak cumulonimbus activity within the ITCZ to move over land, to around 10-15N. In association with the low level SW monsoon winds, a midlevel easterly jet develops (the African Easterly Jet, or AEJ), with a peak of around 15 ms-1 at around 15N and 600-700 hPa altitude (around 3-5 km, or 10,000-15,000 feet). This easterly flow has some important impacts on squall occurrence:
The AEJ provides low level wind shear which organises the storms and can act to intensify the resulting squalls. The organised storms are often termed ‘squall lines’ (sometimes the more general term ‘mesoscale convective system’, or MCS, is preferred) and may have a north-south extent of several hundred kilometres. The significance of squall lines is that they persist for many hours (cases existing for more than 24 hours have been described), and propagate rapidly, at around 15 ms-1 ~ 30 kt ~ 15 degrees per day), so that a given squall line influences a broad swathe of the continent and the downstream Atlantic. Squall lines are discussed in more detail below.
It steers the cumulonimbus systems towards the west. Notably, in the northern part of the study region numerous squall lines (with return periods of around 4 days) move off the west coast of Africa over the ocean;
It is linked to the development of coherent, large-scale weather systems – African Easterly Waves – which organise the cumulonimbus systems, and in turn move westward over the Atlantic. The AEWs themselves do not produce strong winds within the study region, but they do strongly control the environmental temperature and humidity patterns, which control the likelihood and nature of cumulonimbuss when they do occur. Later in the summer, AEWs can act as precursors to hurricanes moving westwards across the Atlantic, but these almost all intensify to the west of the study region.
Although the ITCZ moves over land, along with the most intense squall lines, convective storms can occur in the Gulf of Guinea almost all year round. For instance, at Sao Tome only in the peak summer months (July – August) is convection suppressed.
Figure 3: Map of region with winter monsoon patterns: monsoon trough, low level winds, and midlevel winds. Peak cumulonimbus activity approximately follows the monsoon trough but events may occur to the north and south of this, especially in the east and over the Gulf of Guinea. Winter
In winter the peak solar heating moves into the southern hemisphere. Along the Guinea coast, the monsoon trough and the ITCZ remain tied to the coastal zone, and the peak cumulonimbus activity remains in the northern hemisphere or around the equator. Over southern and eastern Africa the monsoon trough and ITCZ move into the southern hemisphere. The ITF can on occasion move south of the Guinea coast, and the mid-level winds have a northward component above the monsoon layer, so that storms can be steered off land over the Gulf of Guinea.
Figure 4: vertical N-S section through monsoon layer based on Hamilton and Archbold (1945) and Parker et al (2004) which highlights the Saharan Air Layer (SAL) and different convective regions. cumulonimbus storms can be expected throughout the zone to the south of the Inter-tropical front (ITF). The African Easterly Jet (AEJ) is a summer monsoon feature. Synthesis
Hamilton and Archbold (1945) summarised the climatic zones which control the rainfall in the region, and this picture has been backed up by more recent studies. Intense cumulonimbus storms occur in the ITCZ region, to the south of the ITF. The ITF position fluctuates seasonally, between about 22N in summer and near the Guinea coast in winter. Further south, there is a more stable stratocumulus layer from which cumulonimbus events are not common (see also Aspliden et al. 1976).
Role of SAL in squalls
The Saharan Air Layer (SAL) is a very dry layer of air which lies above the humid oceanic air and monsoon air, and plays an important role in squall occurrence. The dryness of the SAL is essential for the development of intense downdraughts, because it allows intense evaporative cooling. Furthermore, the thermal profile of the SAL tends to have a rapid decrease of temperature with height – conditions which tend to correspond to high conditional instability (high CAPE). The SAL is present above the monsoon layer for a few degrees of latitude south of the ITF. The SAL depth and southward extent are also strongly modulated in the northern summer by AEW passages (see below): the ridge of an AEW corresponds to a southward protrusion of the SAL (e.g. Karyampudi and Carlson 1988).
AEWs are the coherent weather systems which propagate across the continent towards the west in the northern summer. The best analyses of composite AEWs were performed by Reed et al. (1977) based on GATE data. AEW structures are usually defined in terms of the north-south winds at around 700 hPa (around 10,000 feet). At this level, where the mean winds are easterly, the northerly (N) and southerly (S) components of the wind are related to the troughs and ridges in the following way (from west to east):
- N – trough – S – ridge – N – trough – S – ridge – N – trough – S – ridge – N -
The waves have horizontal scales of some 3,000 km and although their winds are intrinsically light, they act to organise and modulate the cumulonimbus systems which are embedded within them. The coupling between cumulonimbus systems and AEWs is not well understood, and cumulonimbus events can occur in any phase of an AEW. However, the primary pattern of organisation of cumulonimbus by AEWs is now well established, and depends on geographical location (fig. 10 of Duvel 1990):
West African coast region north of 12N:
cumulonimbus cloud is maximised in the southerly winds ahead of an AEW ridge
Coastal region south of 12N:
cumulonimbus cloud is maximised around the wave trough.
Figure 5: Schematic E-W section through a mature squall line (adapted from Houze and Betts 1981) Squall line structure
Squall lines over west Africa have been the subject of many studies. They are characterised by a leading edge of intense cumulonimbus rainfall, followed by a broader region of lighter rain. The squalls appear with the gust front, or squall front, at the leading (western) edge of the system, and are associated with intense downdraughts driven by precipitation. Squall lines often propagate fast relative to their environmental winds, so that their leading edge is sharp, often with clear skies up to the arrival of the storm (Hamilton and Archbold 1945). The intense squalls last a few minutes, but may regenerate with additional downdraughts within the system.
The cloud structure of squall lines is very variable: some consist of near-constant cumulonimbus cells, while others are composed of intermittent, broken or weaker cumulonimbus cells. Intense squall lines may weaken, or break up, and may then subsequently regenerate.
To an individual observer, there is little difference between isolated cumulonimbus storms and squall lines – in practice their horizontal extent is smaller and their timescales shorter. Squalls may be experienced from cumulonimbus thunderstorms which do not pass overhead, and in such cases the squall winds are approximately from the direction of the storm (Hamilton and Archbold 1945).
Over land there is a strong diurnal cycle of heating of the surface, and this leads to a strong diurnal cycle of cumulonimbus occurrence. Cumulonimbus cells begin to develop in late morning, continue to grow in the afternoon, during which period they may merge and organise into large, coherent storms (MCSs or squall lines), and the peak cumulonimbus activity occurs in late evening. The cumulonimbus is preferentially generated in regions of coherent triggering – notably mountains and wetlands (Rowell and Milford 1993). There is some evidence that the diurnal cycle in these regions of triggering leads to a later diurnal peak in cumulonimbus activity downwind of the source region.
For the oceans, Duvel (1989) has shown that the diurnal cycle of intense cumulonimbus is relatively weak, but that close to the coast (in practice within 10 degrees, thus including much of the study region) the peak cumulonimbus activity is around local noon. More specifically, Duvel (1989) found a local maximum in the Gulf of Guinea region around 1200 local time, and further north, off the West African coast, around 1400 local time.
3. Basic cumulonimbus processes
Figure 6: schematic updraught – shows parcel rising from PBL – draw attention to T difference – include sketch profiles: CAPE and CIN Significance of buoyancy of cloud parcels.
Cumulonimbus clouds form as a result of ‘conditional instability’ – they are a form of convection, and are driven by the buoyancy of the warm air rising within the cloud. Most measures of the likelihood of cumulonimbus convection are based on estimating the difference in temperature between the air within the cumulonimbus cloud, Tp, and the temperature of its environment, Te: the warmer the cloud air relative to its environment, the higher the buoyancy, and the more energy which is available to the vertical circulations of air within the cloud. Upward and downward motion within cumulonimbus clouds is extremely strong (of the order of 10 ms-1) and these clouds act to transport atmospheric properties very rapidly in the vertical.
To estimate the buoyancy of the cloudy air we need to estimate the in-cloud temperature, Tp(z). It turns out that the progress of the air up through a cumulonimbus cloud, from cloud base to cloud top, follows basic physical processes of energy conservation and saturation humidity, so we can make a good estimate of the in-cloud temperature if we know the temperature, humidity and pressure of the air which enters the cloud at its base. Standard equations can be solved to find the in-cloud temperature at any level, as the air ascends (Emanuel 1994).
Given the in-cloud temperature, we need to find its difference from the environmental profile in order to obtain the buoyancy. The temperature of the air in the cloud environment, Te, generally decreases with height, and varies according to location, and to the passage of weather systems, especially in the study region around West Africa. In order to estimate the buoyancy of air within cloud, we therefore need to know the temperature profile, Te(z), at the specific time and location where we might expect cumulonimbus clouds to occur.
Concept of convective inhibition
Observations of cumulonimbus clouds and their environmental conditions show that while the clouds achieve significant positive buoyancy through much of their depth, there is a layer near cloud base where the air is negatively buoyant – the in-cloud air is colder than its environment. This corresponds to a layer of convective inhibition, or CIN, which literally acts to inhibit the initial growth of cumulonimbus storms. The existence of CIN means that cumulonimbus storms are spatially intermittent – even in an environment of conditional instability we do not see cumulonimbus events at all locations nor on all occasions. Furthermore, some energy must be supplied to the low level air in order to lift it sufficiently to overcome CIN and allow a cumulonimbus storm to develop – this energy source is known as a trigger. Over land, triggering is caused by flow over hills and other land surface anomalies. Over the ocean, triggering is thought to be dominated by temperature and wind fluctuations in the boundary layer – significantly these fluctuations are strongly dependent on the recent history of antecedent cumulonimbus storms.
The inhibition of cumulonimbus in its early stages is also dependent on the dryness of the air just above cloud base. As small cumulus clouds evolve, they mix environmental air and can be killed-off if this environmental air is too dry. This process is particularly important in the study region if dry Saharan air moves above the location in question.
In summary, the development of a cumulonimbus storm involves:
A trigger, which lifts air through the CIN layer
Positive buoyancy of the cloud air
Cumulonimbus lifecycle: downdraughts
As a cumulonimbus cloud matures, over a period of perhaps 30 minutes, large precipitation particles form and begin to fall to earth. These particles have two effects on the air which act to reinforce each other:
The weight of the precipitation, falling into air from above, imparts a downward force on the air, and,
Evaporation of the precipitation chills the air, reducing its buoyancy and ultimately making it colder than its environment.
The combination of these two processes leads to the formation of a downdraught within the system. Air descends rapidly (again with speeds of order 10 ms-1), driven by its negative buoyancy and the weight of precipitation. When the downdraught reaches the ground it spreads out in the horizontal – it is this horizontal air motion which comprises the squall winds which are the subject of this study.
What makes an intense downdraught?
The intensity of downdraughts in a given storm is dominated by the characteristics of the mid-level air which is drawn into the cumulonimbus system, and then descends rapidly to earth. The evaporative effects are maximised if the mid-level air is dry – precipitation falling into dry air will evaporate strongly and chill the air very efficiently. Again, in the study area, the possibility of Saharan air moving above the ocean is significant. The precipitation intensity will also have a strong influence on the downdraught evolution: we expect this to be related to the intensity of the cumulonimbus updraught, although it is also affected by microphysical effects such as the availability of cloud condensation nuclei (CCNs). The microphysical effects are not well understood at present, and are not taken into account in operational prediction systems.
Downdraught -> squalls
The development of squalls near the surface when a downdraught reaches the ground is controlled by 2 processes. Close to the downdraught itself, a jet-like flow may occur if the downdraught is sufficiently intense. In this case, the momentum of the downdraught air is high, and the horizontal airflow is driven by high pressures in the downdraught core. Furthermore, if there is a large change in wind velocity between the downdraught origin level and the surface, the downdraught may carry this momentum with it. In the region of the AEJ, for example, this enhances the easterly momentum of air arriving at the surface.
Figure 7: Schematic of cold pool and gust front. The depth of the gravity current is typically around 500 – 1000 m and the head has a length which is typically of the same order – perhaps a few km. A jet-like flow may occur near the downdraught if the downward momentum is high enough. In any event, turbulence will be relatively high close to the downdraught. The ambient wind shear indicated favours an active gust front on the left side, as indicated. At a greater distance from the storm, and in less intense cases, the momentum of the flow is dissipated by turbulent drag and the horizontal winds are driven by the temperature differences between the cold pool of air near the surface, and the warmer environmental air. In such cases, the winds in the cold pool are well described by the gravity current model, which has been studied widely in observational, laboratory and numerical studies. A sketch of a gravity current is given in Fig. 7. The flow is composed of a head region, in which turbulence is maximised, and a ‘following flow’ which is more laminar. The region of the head may be a length of the order of a few times the current depth: for a typical cold pool this length may be a few kilometres. The speed of the gust front is approximated by
where h is the depth of the flow (around 500 m), Δθ is the temperature deficit in the gravity current and θ0 is the background temperature. For typical squall temperature deficits of around 3%, this implies a gust front speed of around 10-20 ms-1 as observed, and the following flow has a speed which is comparable with this (note that turbulent gusts within the flow may be higher than this – as in section 4 below). A gust front over the ocean may persist for around 30 km from an isolated cumulonimbus source (Forecasters’ reference book, 1993).
The role of wind shear
Wind shear is known to have a strong control on the structure and evolution of cumulonimbus systems: some change in the wind vector with height is necessary for long-lived storms to occur. The wind shear acts to give some tilt to the cumulonimbus system, so that the downdraught and updraught can be decoupled. Without wind shear, the precipitation falls into the updraught and kills off subsequent storm development; with shear the precipitation can fall through relatively dry midlevel air and the updraught processes can continue. The gust front is sharper and more intense on the upwind side relative to the storm and is also enhanced on the downshear side (Parker 1996 and references therein; see figure 7). Secondary, daughter storms are often generated at the gust front.
4. Squall internal structure
Gust fronts, microbursts and downbursts
It is a natural tendency of cumulonimbus storms to produce a gust front, where the cold pool of air which has been chilled by precipitation meets the warmer ambient air. In extreme cases, when the downward momentum of the downdraught air is sufficiently strong, a jet-like flow may occur near the surface and more severe microbursts or downbursts may occur. Characteristic features of these, from observations in the USA, include:
Size less than 4 km
Calculated vertical velocities 25-25 ms-1
Observed horizontal velocities up to 30 ms-1
Tremendous low-level wind-shear
(Forecasters’ reference book, 1993)
These structures are widely studied worldwide –the fluid dynamics of these flows is the same regardless of geographical location, given the ambient wind conditions and surface state (land/ocean).
Ideas from gravity current dynamics
We have a picture of the flow regimes emanating from the source. Some laboratory studies have shown the details of the velocity structures within gravity currents, from which we can begin to estimate squall structures. In particular these studies can tell us the turbulent intensities (extreme gusts) in relation to the bulk properties of the system. Kneller et al. (1997) found that mean flow velocities within a gravity current were up to 30% greater than the speed of the gravity current head (corresponding to the gust front) and that instantaneous velocities (corresponding to squall winds) were up to 50% greater. In terms of height, these peak velocities were found within the middle of the flow: for cumulonimbus cold pools this corresponds to a depth of around 250-500 m.
Turbulence intensities were greatest within the head, as has been observed by numerous authors. The turbulence is mainly generated in the shear layer at the top of the gravity current head, but occurs throughout the depth.
It is known that gravity current behaviour is conditioned by the winds and stability of the environmental profile. A very stable layer in the low levels of the environment can modify the gust front – on occasions multiple gust fronts, or ‘bores’ may occur, through interaction with waves on the stable layer. Vertical shear of the wind in the environment of the gust front will tend to enhance the depth of the flow (if the shear is directed downstream, from the perspective of the gust front).
In summarising these ideas in the context of cumulonimbus cold pools, the passage of the gravity-current involves the sudden arrival of the head (the gust front or squall front), followed by strong winds and turbulence. The turbulence is maximised in the head region of the flow, which has a length of a few km and therefore a timescale of a few minutes. Maximum winds and turbulence occur above the surface, in the interior of the gravity current flow. These features all correspond well with observed squalls (e.g. Sommeria and Testud 1984; Chong et al. 1987).
Wind profiles within the cold pool flow
The head of a gravity current is highly anisotropic, and involves strong turbulence with horizontal lengthscales similar to the depth of the current (or order 1 km). As a result, we cannot hope to apply simple averaging or profile methods to explain the wind profiles in this part of the flow. We can make estimates of peak winds based on observed and laboratory cases, which indicate peak gusts of perhaps 50% above the mean speed of the gravity current.
Behind the head of the current, the flow is less turbulent and the horizontal lengthscales may be shorter. Here, there are some standard profile methods which could be employed to explain the vertical profiles of winds and turbulence in the lower part of the flow (up to, say, 40 m). These methods are based on modifications to the logarithmic wind profile with height, to account for the effects of turbulent fluxes of heat and momentum. In this region, away from intense cumulonimbus activity, we also observe that the winds are horizontally homogeneous. This means that statistical measures of winds and turbulence can be translated from temporal to horizontal
Few studies in the region have described the profiles of squall winds in detail. Hamilton and Archbold (1945) described the characteristics of squalls over Nigeria. Peak squalls occur just before, and during the peak cumulonimbus rainfall, lasting perhaps half an hour (about 50 km), and moderate after this. More detailed case studies have been made during the COPT81 experiment. For instance Sommeria and Testud (1984) describe the evolution of wind profiles in the lower 100 m on the arrival of a gust front at Korhogo, Ivory Coast (around 6W, 9N). Over this depth the squall windspeeds, at the arrival of the gust front, are almost constant with depth. This reinforces the observation from gravity current studies that in the head / gust front region of the cold pool, the gusts are deep, and cannot be studied with equilibrium-based profile methods.
A number of storms passed over the HAPEX-Sahel array of surface instruments (13-14N, 2-3W) during August 1992, and the wind and thermodynamic data are available for analysis of the spatial coherence of the 10-minute averaged winds.
5. Prediction tools In accordance with the constraints posed by the fundamentally unpredictable nature of cumulonimbus events (section 1), we have a prediction strategy which involves predicting the number, intensity and characteristics of events which may occur in a given sample region. To do this, we will:
Compute instability and inhibition measures (e.g. CAPE, K-index, CIN etc.) and use them to assess the likelihood, characteristics and likely severity of cumulonimbus events (section 5.1)
Compute measures of expected downdraught intensity (DCAPE, temperature deficit) (section 5.2)
As discussed in section 3 above, most measures of the likelihood of cumulonimbus storms are based on an estimate of the buoyancy of cloudy air relative to its environment. Since the in-cloud air originates at low levels, all of these measures combine low level (e.g. 850 hPa) air properties with mid-level temperatures. At the outset it should be recognised that errors in these parameters are usually related to the errors in the low level air properties because (a) low level air properties change more rapidly, and over shorter spatial scales, than those at mid-levels (especially over land); and (b) low level measures include both temperature and buoyancy effects, whose errors may combine.
a. CAPE – Convective Available Potential Energy CAPE is the most detailed measure of cumulonimbus energetics: it is proportional to the integral of the cloud buoyancy from cloud base to cloud top, and therefore incorporates the negatively buoyant CIN layer. Since it is a measure of energy, it has units of J kg-1. For a cumulonimbus cloud, it is a necessary condition that CAPE be positive, so that there is sufficient energy in a rising cloud column to maintain the circulation. Observed values in thunderstorm environments often may exceed 1,000 J kg-1, and in extreme cases may exceed 5,000 J kg-1.
b. The Lifted Index can be calculated by subtracting the temperature of a parcel of air lifted from the surface to a midlevel height (defined to be a pressure of 500 hPa) from the existing in-situ temperature at that height. If the observed 500 hPa temperature is colder than the lifted air parcel then the lifted index is negative and the atmosphere is unstable to vertical air motions, like those found in cumulus and cumulonimbus clouds. The lifted index numbers provide an indication of the likely thunderstorm severity, rather than likelihood of occurrence.
0 to –2
-3 to –5
Table 5.1 Lifted Index – Thunderstorm Severity
c. The K-index is useful in determining the probability of occurrence of a thunderstorm. The K index is computed as:
K Index = (T850 - T500)+ TD850 + (TD700 - T700)
Where T850, T700 and T500 are the 850 hPa, 700 hPa and 500 hPa temperatures, respectively, and TD850 and TD700 are the dew point temperatures at 850 hPa and 700 hPa. All temperatures are in °C.
The first term on the right hand side is a measure of the lapse rate between the 850 and 500 hPa layer and indicates less stable air when the difference is a large positive number. The second term represents the amount of moisture in the lower layers of the atmosphere. As the dew point at 850 hPa increases, more latent heat is released when the air is lifted beyond saturation, and therefore the in-cloud temperature increases. The last term represents the (negative of) dryness of the 700 hPa layer. A large difference between the dew point temperature and temperature at 700 hPa indicates that the air at that level is dry and will tend to suppress the early development of cumulonimbus.
THUNDERSTORM PROBABILITY (%)
Table 5.2 K-Index - Thunderstorm Probability d. The Showalter Index (SI) is used to determine the stability of the lower part of the troposphere. An air parcel is lifted from an initial position at 850 hPa where localized low level influences are greatly reduced (unlike the Lifted Index, which takes a parcel from the surface). The air parcel is lifted to 500 hPa and the environmental temperature is then subtracted from the parcel temperature to obtain the value of the Showalter index. The risk of severe weather activity is defined as follows:
RISK OF SEVERE WEATHER ACTIVITY
No significant activity
1 < SI < 3
Showers possible with other source of lift
-2 < SI < 1
Thunderstorms possible (generally weak)
-3 < SI < -2
Thunderstorms more probable (possibly strong)
-6 < SI < -4
Strong or severe thunderstorms possible
SI < -6
Any thunderstorms likely to be strong or severe
Table 5.3 Showalter Index - Risk of Severe Weather Activity e. The Total Totals index is a simple index derived from the temperature lapse rate between 850 hPa and 500 hPa and moisture content at 850 hPa. The index is defined as follows and the risk of severe weather activity according to this index is defined in Table 5.4.
TT = T850 + TD850 - 2T500
Scattered heavy / few severe thunderstorms and isolated tornadoes
Scattered to numerous heavy / few to scattered severe thunderstorm / few tornadoes
Numerous heavy / scattered severe thunderstorms and scattered tornadoes
Table 5.4 Total Totals Index - Risk of Severe Weather Activity 5.1.2 Measures of inhibition Convective inhibition - CIN - is defined, like CAPE, as an integral of cloud buoyancy, but restricting the integral to the negatively buoyant layer in the lower region of the cloud. CIN is then the energy needed to lift a mass of air to the point where it becomes buoyant, and can then rise freely. High values of CIN mean that cumulonimbus events may be suppressed entirely, even if there is positive CAPE.
Dryness of the layer just above the boundary layer (say 700 hPa, as in the K-index definition) will also tend to suppress the early stages of cumulonimbus development.
5.1.3 Role of shear / organisation The wind shear in the lower levels of the atmosphere is known to have a strong control on the organisation of the resulting cumulonimbus systems. In particular, some wind shear is necessary if long-lived, organised systems are to occur. Weisman and Klemp (1982) have attempted to describe this by using a Bulk Richardson number. The Bulk Richardson Number(BRN) is a ratio of CAPE and wind shear squared (Moncrieff and Green 1972). It must be used with other thunderstorm parameters when making thunderstorm decisions: it is a discriminator, not a predictor. BRN is usually a decent indicator of convective storm type within given environments. It incorporates buoyant energy (CAPE) and the vertical shear of the horizontal wind, both of which are critical factors in determining storm development, evolution, and organization.
BRN = CAPE / [0.5 (U2)]
where U is a measure of the vertical wind shear, typically in the 0-6 km layer AGL. High values indicate unstable and/or weakly sheared environments in which multicellular systems may be likely; low values indicate weak instability and/or strong vertical shear, which which highly organised cumulonimbus, such as supercells, may occur.
5.2 Predicting squall characteristics of cumulonimbus In attempting to predict the likely intensity of downdraughts and squalls, we use measures of the temperature of air which is brought down from midlevels to a layer close to the surface. This kind of estimate is less well-defined than measures of the updraught intensity (such as CAPE) since we do not know in advance the origin level of the downdraught, and we do not know the amount of water available for evaporation. With this in mind we use measures which define a priori the origin level of downdraught air, and then compute the maximum possible cooling which could occur in this air as it descends to the surface.
5.2.1 The downdraught CAPE (DCAPE) is computed, in a similar manner to CAPE, as an integral of the negative buoyancy of air. This is performed by assuming that the air is first cooled to saturation, then descends to the surface with its state of saturation being maintained by evaporation throughout the process. This can be regarded as a measure of the maximum possible energy available to descending air (although it does not take into account the weight of precipitation).
DCAPE can be computed for any level in the environmental profile – in practice we choose a level or range of levels appropriate to the source for downdraught air which is in the mid-troposphere.
5.2.2 The temperature deficit of the squall air can be estimated by the same method used to compute DCAPE (that is, by assuming the air is saturated when it reaches the surface); again, this should be regarded as the maximum possible temperature deficit – air which is subsaturated when it reaches the surface will be warmer, and its negative buoyancy will be less extreme. The Met Office Forecasters’ Handbook (1993) recommends an origin level for the downdraught air as the level at which the wet bulb temperature crosses 0C, and gives forecast peak squalls corresponding to a given temperature deficit.
5.2.3 Wind shear U, is typically computed over the same range of height as used for DCAPE or the downdraught temperature deficit. High wind shear may correspond to strong gusts due to the transport of momentum and vorticity to low levels, in the downdraught.
7 Summary and conclusions
Squall characteristics are determined by the characteristics of the cumulonimbus systems in which they are embedded, and these are in turn determined by the larger-scale properties of the atmosphere. Individual cumulonimbus systems cannot generally be predicted in a deterministic sense (except over extremely short timescales of an hour or so – nowcasting): we need to predict the statistics and probabilities of cumulonimbus occurrence, and to use knowledge of the larger-scale atmosphere to predict the characteristics of the storms which do occur.
Squalls are dynamically equivalent to gravity currents and therefore consist of:
A head region, following the gust front for a few minutes, in which winds and turbulence are strong.
A ‘following flow’ in which the winds are more horizontally homogeneous. Increased wind strengths may last for several hours.
The study region is well known for intense squall systems due to the presence of high instability in the atmospheric profiles, as well as dry mid-level air and strong vertical shear of the winds. The latter two properties are favourable for the development of strong squalls, as have been observed. In analysing and predicting these systems we can use:
measures of atmospheric stability (e.g. CAPE) which assess the energy available to storms;
measures of inhibition (e.g. CIN, and dryness at low- to mid-levels);
measures of storm organisation (e.g. wind shear and BRN); and
measures of potential downdraught intensity (e.g. DCAPE and temperature deficit).
The employment of these measures may require some empirical tuning for the study region in question.
In the study region there are strong seasonal variations in the atmosphere which control the cumulonimbus occurrence. These are modulated also by the passage of weather systems (especially African Easterly Waves in the northern summer) and by the diurnal cycle of heating of the land. Thus prediction and analysis of squalls in the region requires an analysis of daily data, and an appreciation of seasonal and subseasonal fluctuations.
Aspliden, C. I., Y. Tourre and J. B. Sabine, 1976. Some climatological aspects of West African disturbance lines during GATE. MWR, 104, 1029-1035.
Buckle, C., 1996, Weather and climate in Africa, Addison Wesley Longman Ltd, 312pp.
Chong, M., P. Amayenc, G. Scialom, J. Testud, 1987: A tropical Squall Line Observed during the COPT 81 Experiment in West Africa. Part I: Kinematic Structure Inferred from Dual-Doppler Radar Data, MWR, 115, 670-694
Duvel, J. P., 1989. Convection over tropical Africa and the Atlantic ocean during northern summer. Part I: Interannual and diurnal variations. MWR, 117, 2782-2799.
Duvel, J. P., 1990. Convection over tropical Africa and the Atlantic ocean during northern summer. Part II: Modulation by easterly waves. MWR, 118, 1855-1868.
Eldridge, R.H., 1957, ‘A synoptic study of West African disturbance lines’, Q. J. R. Meteorol. Soc., 83, 303-314.
Emanuel, K.A., 1994, Atmospheric Convection, Oxford University Press, 580pp
Forecasters’ Reference Book, 1993, Meteorological Office, Bracknell. Hamilton, R.A. and Archbold, J.W., 1945, ‘Meteorology of Nigeria and adjacent territory’, Q. J. R. Meteorol. Soc., 71, 231-264. Houze, R.A. and Betts A.K., 1981, ‘Convection in GATE’, Rev. Geophys. And Space Phys., 19, 541-576
Karyampudi, V.M. and Carlson, T.N., 1988, ‘Analysis and numerical simulations of the Saharan air layer and its effect on easterly wave disturbances’, J. Atmos. Sci.,45, 3102-3136
Kneller, B.C, S.J. Bennett and W.D. McCaffrey, 1997, ‘Velocity and turbulence structure of density currents and internal solitary waves: potential sediment transport and the formation of wave ripples in deep water’, Sedimentary Geology,112, 219-234
Moncreiff, M.W. and J.S.A.Green, 1972, ‘The propagation and transfer properties of steady convective overturning in shear’, Q. J. R. Meteorol. Soc., 98, 336-352
Parker, D. J., 1996, 'Cold pools in shear', Q. J. R. Meteorol. Soc., 122, 1655-1674
Parker, D. J., 2002, ‘The response of CAPE and CIN to tropospheric thermal variations’, Q. J. R. Meteorol. Soc., 128, 119-130.
Parker, D.J., C.D. Thorncroft, R.R. Burton, and A. Diongue-Niang, 2004, ‘Analysis of the African Easterly Jet using aircraft observations from the JET2000 experiment’, Submitted to Q. J. R. Meteorol. Soc.
Parsons, D.P., Yoneyama, K. and Redelsperger, J-L., 2000, ‘The evolution of the tropical western Pacific atmosphere-ocean system following the arrival of a dry intrusion’, Q.J.R.Meteorol.Soc.,126, 517-548.
Reed, R.J., Norquist, D.C. and Recker, E.E., 1977, ‘The structure and properties of African wave disturbances as observed during Phase III of GATE’, Mon. Wea. Rev.,105, 317-333.
Rowell, D.P., and J. Milford, On the generation of African squall lines, J. Climate, 6, 1181-1193, 1993.
Sommeria, G. and J. Testud, 1984. COPT81: A field experiment designed for the study of dynamics and electrical activity of deep convection in continental tropical regions. BAMS, 65, 4-10.
Thorncroft, C. D., D. J. Parker, R. R. Burton, M. Diop, J. H. Ayers, H. Barjat, S. Devereau, A. Diongue, R. Dumelow, D. R. Kindred, N. M. Price, M. Saloum, C. M. Taylor, and A. M. Tompkins, 2003, ‘The JET2000 Project: Aircraft Observations of the African Easterly Jet and African Easterly Waves’, Bull. Am. Met. Soc., 84, 337-351.
Weisman, M.L. and Klemp, J.B., 1982, ‘The dependence of numerically simulated convective storms on vertical wind shear and buoyancy’, Mon. Wea. Rev.,110, 504-520.20>15>