Convection

Convection

Specificities of deep convection over Africa

Forecasting precipitation events associated with convection is a major challenge for forecasters, with potentially dramatic consequences for extreme events. Indeed, precipitation forecasts from current numerical models are not very reliable, either for global models parameterising these very fine-scale and highly random phenomena, or for high-resolution models attempting to explicitly resolve each thunderstorm cell.

Moreover, the difficulty is increased over the Sahel where convection has different properties from other tropical regions. Thus, we note:

  • Temporal variability of rainfall over the Sahel is very high from one day to the next, so the persistence of rainfall over Africa is very low.
  • Spatial variability is also very high, so that nearby stations can have rainfall totals that can double even on a seasonal scale.
  • The convective inhibition (CIN) is also very strong and the relief weak, so that the triggering of convection is difficult to predict.
  • At upper levels, Africa is under the influence of the subsident branch of the Walker zonal cell (ascending branch over India and Asia) which strongly dries the free atmosphere. Thus, outside the low layers in the presence of the monsoon flow, the atmosphere is very dry.
  • Despite these strong brakes on convective activity, it succeeds in developing, constituting a paradox, illustrated below.

Neelin et al. (2009) found a robust link in observations for tropical oceanic areas between rainfall rate (P) and Precipitable Water (PW). As illustrated in the Figure opposite, a power law is obtained above a critical threshold of PW, which increases with the mean tropospheric temperature ˆT.

The same analysis applied to the data observed over Senegal, is superimposed (red dots). As expected, this Sahelian zone appears much warmer and drier ( ˆT ≥ 273 °K, PW ≤ 60 mm) than the eastern Pacific zone. More surprisingly, convection seems to be more efficient over the Sahel, since the critical threshold in PW to hope precipitation is only 36 mm, compared to about 55 mm over the Pacific for equivalent temperatures, ˆT . However, analysis of the probability distributions of rainfall over the Sahel shows non-Gaussian distributions with high variability. For a PW of 50 mm, the rainfall rate is 5 mm day-1, but the majority of the rainfall is concentrated over less than 20% of the area with a much higher intensity, which presents a challenge for forecasting these rainfall events.

Observed average precipitation rate ⟨P⟩ (in mm h−1) as a function of precipitable water (PW, mm), for the eastern Pacific for 1 K bins of the vertically averaged tropospheric temperature ˆT . Daily precipitation rates ⟨P⟩ (mm day−1) observed by the Sénégal rain gauge network for the 2000–2011 period have been superposed with red points. Source: Fig. 3.5 of the Handbook.

One explanation for this paradox is the role played by the African easterlies in the organisation of convection. The objective of this chapter is to explain the unique characteristics of convection over the Sahel in order to understand how it works and to propose a forecasting method (WASA-F chapter 11 of the Handbook) based on the parameters that forecasting models handle better than precipitation.

Processes and factors controlling Convection

Convection depends on a multiplicity of factors at different scales. To understand and predict convection it is essential to consider all these factors which can be classified into 4 types (see diagram below).

  1. Updraugts : Convection is often only seen through the convective updraugt, considering mainly convective instability indices such as CAPE (convective available potential energy) and CIN (convective inhibition energy). Other parameters such as PW must be added.
  2. Downdraughts : As all ascending motion must be compensated in order to conserve mass, the capacity of the atmosphere to favour downdraughts must be considered. Moreover, the evaporation of rain within them generates cold pockets spreading over the surface, called “density currents” (DC), which can trigger new thunderstorm cells.
  3. The wind shear : in the environment helps to organise convection. This is the case of the East African Jet ( AEJ), which is essential to organise squall lines over Sahel.
  4. Other forcings : These factors external to convection are multiple and on different scales, large scales such as African Easterly Waves, surface properties, diurnal cycle of the boundary layer…
Schematic representation of the four types of processes and related indices that govern the convective activity. CAPE: convective available potential energy; CIN: convective inhibition energy; DCAPE: downdraught convective
available potential energy. Source : Figure 3.7 of the Handbook.

1. Convective updraughts

Wet convection is characterised by upward motions in clouds, resulting from conditional convective instability. The parcel theory, which neglects accelerations due to the pressure term and turbulent mixing with the environment, reduces the vertical acceleration of the air parcel to the buoyancy term B

where g s the gravitational acceleration, and Ɵ′v is the virtual potential temperature deviation from that of the environment Ɵv0 taking into account the mass of the water vapour expressed as a mixing ratio q’v0. Parcels with a warm anomaly Ɵ′v will be more buoyant than their surroundings and will be accelerated upwards. The weight of the liquid and solid hydrometeors (respective mixing ratios ql and qs) will reduce the lift of the air parcel (hydrometeor loading).

In the framework of this theory, the ascent of a parcel from the lowest layers (Figure opposite) follows the dry adiabatic to the base of the cloud zlcl , then the wet adiabatic to the top of the convection ztopbis. The evolution of the buoyancy B during this ascent defines 3 layers:

  • 0<z< zlfc (level of free convection) B<0 its integral corresponds to the CIN (Convective Inhibition energy), which is a barrier to convection.
  • zlfc<z< ztop Free convection layer: B>0 its vertical integral corresponds to the available convective potential energy (CAPE for Convective Available Potential energy).
  • ztop<z<ztopbis Overshoot zone: B<0.The parcel is braked by its now negative buoyancy B, and can theoretically reach the ztopbis defined such that the vertical integral of B is equal to CAPE.
Illustration of the parcel theory for an air mass lifted from the surface. Source : Figure 3.8 of the Handbook.

To analyse the factors driving convective updraughts (table below), it is necessary to consider three parameters: CAPE and PW are favourable and necessary but not sufficient, while CIN tends to suppress the development of convection. Note that the CIN barrier delays the onset of convection and allows more energy to be stored in the subcloud layer. This is favourable to the increase of CAPE and PW and to the triggering of more intense thunderstorms later in the day, or the following days. Although these three parameters are essential for understanding the occurrence and intensity of convective updraughts, they are insufficient on their own to forecast convective activity.

FavourableUnfavourable
CAPEYes – but strong duirnal cycle
PWYes – if PW>40 mm for West Africa
CINYes – increase with CAPE and PWYes – strong over Sahel – delays convection
EntrainmentYes – reduces CAPE especially if dry environment
Water loadingOui si cisaillementYes – reduces the buoyancy
Favorables ou unfavourable parameters for convective updraughts

2. Subsidences and Density Courants

Owing to mass conservation, the development of updraughts requires the occurrence of compensating
subsiding motion somewhere. The subsidence can occur at different scales.

  1. At the scales of the cloud and of the convective system :
    • Intense micro-scale downdraughts (with downward speeds > 15 m s –1, horizontal scales < 1 km), generating intense ‘microbursts’ when impacting the surface. These events are highly dangerous for aviation, and may be generated from very small‐scale clouds.
    • Convective dondraughts (1 to 15 m–1, over a few kilometers) are associated with convective precipitation and updraughts.
    • Mesoscale subsidences (a few tens of cm s–1, over 10-100 km), associated with trailing precipitation underneath stratiform anvils. The subsidence results in characteristic ‘onion or diamond’ shape soundings in post squall‐line regions as proposed by Zipser (1977) and frequently observed over West Africa.
  2. At mesoscale : In the vicinity of the convection, compensating subsidence is forced by gravity waves in the environment, allowing the atmosphere to redistribute heat around the convective source.
  3. At synoptic and planetary scales : Over extensive areas, and far from the convection area, large‐scale subsidence of a few centimetres per second occurs, such as within the subsiding branch of the Hadley cell. Radiation cooling is then a key process to equilibrate the adiabatic warming occurring within the large‐scale subsidence.

The parcel theory developed for an ascending flow to define the CAPE, can be applied for a subsiding flow as illustrated in the Figure opposite. However, the trajectory of the parcel requires strong assumptions corresponding to an extreme and theoretical case where the evaporation of precipitation is strong enough to reach the wet adiabatic while remaining at constant pressure, then to follow the wet adiabatic down to the surface.

The DCAPE (Downdraught Convective Available Potential Energy) index proposed by Emanuel (1994) is defined as the area between the trajectory of the parcel and the profile of its environment.

In this extreme case the parcel reaches the surface (blue arrow) with the temperature Ɵ′w (of the wet thermometer) at the parcel’s original level. If evaporation is not sufficient the trajectory will be intermediate (green arrow) between the wet adiabatic and the dry one (red arrow). There is therefore a high degree of uncertainty about the trajectory of the subsidences compared to that of the convective updraughts. This is an important feature that complicates the convection forecasting.

Illustration of the computation of the DCAPE index on a tephigram, for a subsiding parcel taken at 800 hPa. The blue arrow indicates the descent of air that reaches the ground completely saturated, the red arrow denotes dry adiabatic descent, and the green arrow represents the most common occurrence, lying between the moist and dry extremes. Therefore, there is considerable uncertainty in the temperature of the downdraught at ground level.Source : Figure 3.10 of the Handbook.

The analysis of the thermodynamic trajectory of the subsidence in the Figure above allows the identification of the favourable parameters, namely:

  • Middle layers (and above cloud base) as dry as possible (Ɵ’w minimum) to promote evaporation of precipitation cooling the subsiding parcel to compensate for its adiabatic heating
  • Sufficient precipitation
  • A subcloud profile as close as possible to the dry adiabatic, conditions reached in the afternoon.
  • Highest possible cloud base height

This analysis does not take into account other factors such as dynamically forced subsidence (via pressure forces) which can have positive buoyancies.

Density Current (DC)

The evaporation of precipitation within subsidences generates cold air masses called “cold pools” which spread out at the surface as density currents. A schematic view of a density current is shown in the figure opposite.

  • Its head is deeper – typically 1–2 km in depth in the Sahel – and characterised by a rotor circulation. Its propagation (10-20 ms-1 over Sahel) lifts ambient warmer air in the form of a roll cloud that can trigger new convective cells.
  • This discontinuity between the two air masses is sharp (~ 500 m), as confirmed by the DC passage signature commonly observed at the surface, namely :
    • sharp drop of temperature and of specific humidity (relative humidity may rise or fall, depending on the case),
    • rotation of winds, intense gusts, a pressure jump (~1 hPa) and the arrival of rain a little later.
  • Behind the head region, the tail of the cold pool is shallower, the winds slow down, and over a period of hours the low levels recover.
  • The upper part of the density current is turbulent with intense wind shear, especially in the head region.
Schematic view of a density current, and photo of the arrival of a density current lifting dust, a phenomenon referred to as haboob. Source : © CNRS Photothèque, F. Guichard et L. Kergoat, Mali, August 2004. Source: extract from figure 3.11 of the Handbook.

Over Africa cold pools can be intense (surface temperature drops larger than 10 °C) and deep (1–2 km), representing an important source of dust lifting over bare soil and allowing the visualisation of the cold pool (above picture). Cold pools are a key feature causing the triggering of new, secondary convective cells in situations of strong inhibition, as occur over the Sahel. This process of secondary triggering helps to organise convection, propagate it and enhance its duration.

3. Wind shear

The variation of the horizontal wind (i.e. shear) in the vertical, both in intensity and direction, is the third factor to consider in order to understand and forecast the convective activity. Early in the 1960s Ludlam
pointed to strong shear as a key feature for severe storms (Ludlam, 1980). A tilting updraught enables
precipitation to fall out of it – thus reducing the drag – and to feed a cold downdraft by rain evaporation. A tilted updraught and a strong downdraught may work side by side to produce intense and long‐lasting
convection.

The Figure opposite illustrates the impact of the shear on a DC.

  • In the case of no shear (a), the circulation is symmetric at the convective scale (negligible Coriolis term), and the DC (blue) generated by the convection spreads in all directions, taking a circular shape. The forced updrafts are distributed along the leading edge of the DC, and are therefore weak (dotted red circle).
  • In contrast, in the case of linear shear (b) (here easterly shear, as over West Africa), mid‐level dry air feeding the density current transports momentum and the density current spreading is no longer symmetric. It is faster and deeper on its downshear (here western) edge, where the convergence is concentrated and stronger allowing the triggering of new convective cells (in red).
Vue schématique (en coupe verticale, et de dessus respectivement en haut et en bas) de l’impact du cisaillement sur le courant de densité (en bleu) et son efficacité à déclencher de nouvelles cellules convectives (en rouge) : (à gauche) en l’absence de cisaillement et (à droite) en présence d’un cisaillement uni-directionnel. Les profils de vent de l’environnement sont relatifs à la propagation du CD. Source : Figure 3.12a et b du Handbook.

The general rule is thus: in the case of shear, the density current tends to trigger new convective cells on its downshear side.

Vertical motion in a sheared flow induces non‐hydrostatic pressure deviations (~1 hPa). In simple terms, a convective updraught (downdraught) behaves as a ‘soft obstacle’. A horizontal dipole of pressure opposes the ambient horizontal wind change and reduces the momentum difference between the draught and its environment. Such fluctuations of pressure at different levels generate vertical accelerations that favour new updraughts on the downshear side. Over West Africa the shear is moderate compared with mid‐latitudes, but stronger than over the tropical oceans, contributing to differences in convective activity between these tropical regions.

For the sake of completeness, the rotation of the shear should also be considered, in order to understand specific organised convection such as supercells, as developed by Weisman and Klemp (1986). As shear rotation is weak over West Africa, supercells are uncommon, explaining why tornadoes have not been recorded in West Africa.

4. Other Forcings

External forcings represent the last type of processes that govern convection and are an important source of predictability for convection over West Africa.

  • The forcing by the synoptic flow and in particular those associated with the East African Waves (AEWs) are essential to predict convection and are developed in Chapitre 2 : Systèmes synoptiques. Thus the conceptual scheme of the African Easterly Waves and Jet (AEW and AEJ) – Misva (aeris-data.fr) allows to locate the preferential zones of development of the mesoscale convective systems (MCS) within the waves.
  • The other external forcings to be considered are small-scale and are briefly presented below. Section 3.1.4.1 (pages 108-113) of Chapter 3 of the Handbook details these forcings that affect the life cycle of convection and more particularly its initiation.
    • Orographic Forcing : Plays an important role in triggering convection over the terrain above 600 m, more by thermal than dynamic effect. The most favourable massifs are the Jos Plateau, Darfur, Aïr, the Cameroon mountains and Fouta Djalon.
    • Coastal Circulations : MCS triggering frequently occurs in coastal areas, due to breezes resulting from the temperature contrast between the ocean and the continent.
    • Surface-Atmosphere Coupling : Surface heterogeneities (temperature, soil moisture, vegetation) generate breezes in the boundary layer whose convergence at the level of contrast zones favours the triggering of convection.
    • Convergence Lines : Mesoscale circulations generate near-surface convergence lines that favour the triggering of thunderstorm cells. They can have multiple causes (residue of a density current… and the above forcings), but the difficulty is the lack of observations to detect them. The VIS satellite imagery makes it possible in certain cases to visualise the cumulus clouds forced by these convergence lines.
    • Regeneration of Storms : Typically, the dissipation of convection generates heterogeneities of temperature, humidity and circulation in the boundary layer, which the next day favours the outbreak of thunderstorms.
    • Gravity Waves : The stable stratification of the troposphere allows the occurrence of gravity waves triggered by convective vertical motions. In Africa, the powerful and deep cumulonimbus clouds are important sources of gravity waves propagating over long distances and playing a role in the triggering of new convective cells. The absence of routine observations of these waves (clear-sky radar) makes it impossible to detect them, but their role cannot be neglected.
    • Diurnal Cycle : The strong cycle of the continental surface and the boundary layer – especially in the dry zone – forces the initiation of convection in the early afternoon.

Organisation of Deep Convection

To discuss the organisation of convection, we use the concept of a “convective cell”, seen as an ascending region of a few to ten kilometres in size, extending vertically over a significant part of the troposphere. Each rising cell is associated with an area of precipitation identifiable by radar. Three types of convection organisation can thus be distinguished.

1. Different Types of Organisation

1.a The Single-Cell Strom (diagram opposite)

  • only one “convective cell”
  • short duration (30-50 min)
  • 3 stages :
    1. Formation of an ascending warm bubble Cu → congestus
    2. Mature: glaciation → deep Cb with precipitation → fall, evaporation, initialisation of subsidences and a DC
    3. Dissipation: subsidences at low and mid-levels, DC spreading over the surface, residual ascent at upper levels feeding an anvil
  • Typical Environment : weak shear → upright updraught → precipitation falls within the updraught and cuts off the flow that feeds the updraught
The Single-Cell Storm. Source : Chapitre 3 of the Handbook.
  • Short‐lived storms are common over West Africa, particularly in regions far to the south of the AEJ core in summer, where the wind shear is weak.
  • Not associated with severe weather, except occasionally microbursts.
  • Propagation at about the mean wind speed of the convective layer.

1.b The Multicell Storm (diagram opposite)

Definition: ensemble of shorter‐lived single‐cell storms at different stages of their life cycle that can persist over several hours up to a day or more.

Key Ingredient = moderated to strong wind shear in one direction (a linear hodograph).

The schematic scenario is the following.

  1. A growing ‘blue’ cell begins to produce precipitation falling out (due to shear) of the updraught and initiates a downdraught
  2. A new “red” cell is triggered downshear.
  3. About 20 mn later, the “blue” cell reaches its mature stage with an intense subsidence and a DC spreading above the surface downshear.
  4. This later reinforces the “red” cell, and favours the triggering of a new “green” cell along the CD leading edge.
  5. The dissipation of the first “blue” cell occurs 20 mn later, helping to fuel the CD that forces the development of the “red” cell,
  6. which then reaches its mature stage and will interact with the new “green” cell following the same scenario as the “red” cell that preceded it.
  7. The loop is thus closed. The multicell storm will be able to maintain itself through these interactions between the cells and achieve a long life span.
The Multicell Storm. Source : Chapitre 3 of the Handbook.

The propagation of the multicell storm (equivalent to a group velocity) results from the combination of the velocity of each cell (mean wind in the convective layer) and the discrete speed of initiation of the new cells. Thus the speed of the convective system as a whole may differ in speed and direction from that of the mean wind.

Multicell storms are common over the Sahel, where the shear below the AEJ provides favourable conditions. Their structure and climatology are presented in the following sections.

They may be associated with severe weather, in terms of heavy rainfall and squalls

1.c The Supercell Storm

The supercell storm is potentially the most dangerous of the convective storm types. It may produce hail,
tornadoes and strong winds; however, it has not been commonly recorded in West Africa, so it is only briefly mentioned here. The supercell consists of a single, powerful, quasi‐steady rotating updraught.

Apart from the required convective instability (CAPE), the key favourable ingredient for supercells is the vertical wind shear vector that needs both to be strong and to rotate at low levels. With this configuration, the supercell propagates differently from the mean wind toward the curvature of the
hodograph at low levels (concave side). Owing to the lack of rotation of the shear, supercells are seldom observed over West Africa. Nevertheless, in some regions, such as in the lee of mountains or on the flank of squall lines, rotation of the shear can occur, so that such storms are not excluded, even if much less intense than over the US Great Plains.

Thus, forecasters are recommended to carefully examine hodographs to predict the storm type, propagation… and to evaluate risks.

2. The squall-line Conceptual Model

The fast-moving squall line (SL) is the predominant type of mesoscale convective system (MCS) over West Africa. It is a multicell, self‐maintained system having three main interacting components (Figure below).

1. A curved line of powerful cumulonimbus occupies the convective part, a few tens of kilometres wide, that can extend for several hundred up to a thousand kilometres in length. The convective part is associated with a strong updraught extending over the whole troposphere, overshooting into the lower stratosphere as
detected by cold IR satellite brightness temperature. Owing to the large latent heat release by convective cells, a low pressure (~1 hPa) (L in Figure (b)) develops at mid‐levels (~3 km), and intense precipitation is formed, feeding intense convective downdraughts. Owing to the elevated 0 °C isotherm (~4.5 km) over
West Africa, a large part of the precipitation is liquid and falls rapidly (~7 m s−1) underneath the convective part.

2. A large amount of solid (ice) hydrometeors are also formed in the deep tropical troposphere (~16 km). Falling slower (~1–2 m s−1), they feed a thick and widespread anvil called the stratiform region characterised by weak mesoscale ascent (a few tens of cm s−1) in the anvil and a weak mesoscale subsidence (a few tens of cm s−1) below the sloped, elevated anvil cloud base. Owing to this sloped anvil cloud base and to the mid‐level low pressure in the convective part, an inward horizontal pressure gradient develops. This favours the formation of a rear inflow that reinforces the AEJ behind the squall line and brings in drier air below the stratiform anvil, as
illustrated by Figure (b), with reinforcement of easterlies of up to 40 kt around the 700 hPa level.

a) Three‐dimensional schematic view of a fast‐moving squall line with the airflow circulation, the cold air mass forming the density current (DC) in blue shading, and some trajectories of hydrometeors. (b) Two‐dimensional conceptual
model of the squall line. (c) Its box representation. Source : Figure 3.14 of the Handbook.

3. The mesoscale subsidence is forced by the rear inflow and maintained by both the rain evaporation and the melting of ice particles, contributing to feed the ‘density current DC’, corresponding to the third key component of the squall line. The deep head of the DC propagates downshear, where lifting of ambient air over this deep head helps to overcome the CIN barrier and trigger new convective cells in the convective part.

3. Climatology and Classification of MCS

A climatology of West African MCSs over a 25‐year period (1986–2010) from Meteosat has been established using the algorithm described in Fiolleau and Roca (2013). The speed Vp and duration D parameters are used to describe MCSs in four classes :

  1. The class C1 [D < 9 h ; Vp < 10 m s–1] corresponds to small (between 5000 km² and 10 000 km²) numerous diurnal and slow‐moving MCSs.
  2. The class C2 [D > 9 h ; Vp < 10 m s–1] includes long‐lived slow‐moving MCSs. They are bigger (up to 20000 km2) than those in the C1 class.
  3. The class C3 [D < 9 h ; Vp > 10 m s–1] corresponds to diurnal and fast‐moving systems dissipating in the evening.
  4. The class C4 [D > 9 h ; Vp > 10 m s–1] corresponds to fast‐moving long‐lived squall‐line type systems. They are less numerous but are the largest (typically 30000 km2).

The climatological spatial distribution of cloud cover in JJAS for the four SCM classes is shown in the figure opposite. The coverage of these upper tropospheric clouds (expressed in h/month) is indicative of convective rainfall.

  1. C1 systems are the more numerous but, with C3, have the weakest contribution to total rainfall. They are localised in the vicinity of elevated terrain (Fouta Djalon, Jos Plateau, Cameroon Mountains and Central Africa).
  2. In contrast, C2 systems have an important contribution to rainfall, especially in the Atlantic ITCZ, at the coast with maxima close to Conakry and the Niger Delta, and Central Africa.
  3. The contribution of C3 systems is also significant, especially in the Sahel band and over Central Africa.
  4. Fast‐moving and long‐lived C4 systems are concentrated in the Sahel band and strongly collocated with the AEJ, where they explain most of the rainfall. They correspond to fast‐moving squall lines.
Spatial distribution of cloud cover JJAS climatology of MCSs over 25 years of Meteosat performed using the Fiolleau and Roca (2013) tracking algorithm for the four types of MCSs: classifications C1, C2, C3 and C4. Resolution is 1° square and the unit is hours per month. Heavy isolines represent surface altitude above 600 and 1200 m. Source : Figure 3.16 of the Handbook.

Owing to the conditional character of the convective instability and to the strong convective inhibition (large CIN) over West Africa, especially over the Sahel, external forcings are new convection.

The climatology of the triggering location of MCSs (Figure opposite) sheds light on the mechanisms involved. First of all, as short-lived convective systems are more numerous than long-lived (and better organised) ones, the number of triggers for C1 and C3 systems is greater (maximum of ~ 700) than for C2 and C4 systems (maximum ~ 300).

The main types of forcing highlighted by this climatology are :

As for the previous Figure for the location of the triggering. Unit is the total number of triggering events over the 25‐year period of the climatology. Source : Figure 3.17 of the Handbook.
  • Orographic Forcing : Plays an important role in triggering convection over the terrain above 600 m, more by thermal than dynamic effect. The most favourable massifs are the Jos Plateau, Darfur, Aïr, the Cameroon mountains and Fouta Djalon.
  • Coastal Circulations : The coasts from Guinea to Liberia and Nigeria are conducive to the outbreak of CSM classes C1, C2 and C3.
  • The combination of these two coastal forcing zones, with the orographic forcing of Fouta Djalon and Cameroon mountains (as analysed by Vondou et al., 2010) leads to the triggering of numerous SCMs of all types (except class C4 for southern Nigeria and Cameroon), explaining the cloud cover and precipitation maxima in these two regions
  • Surface–Atmosphere Coupling : Although less frequent, triggering also occurs on flat surfaces with elevations below 600 m over the western Sahel, such as over Mali for MCSs of classes C3 and C4. To explain this type of triggering, Taylor et al. (2011) analysed the initialisation of more than 3000 MCSs using satellite data and found a clear preference for the initialisation of the first deep convective cells over areas of high soil moisture gradients resulting from the previous days’ rainfall. The figure opposite provides a conceptual model of this coupling mechanism.
Schematic depicting the impact of soil‐moisture
heterogeneity on convective initiation. Idealized soil‐moisture induced flows (blue arrows) under light synoptic winds (black arrow) create an ascent region (large red arrow) where the shallow, strong current opposes the mean wind. The preferred
location for convective initiation coincides with the ascent region induced by the heating gradient at the downwind edge of the dry patch. Additional convergence over the dry patch is provided by a
deep, weaker current at its upwind edge and cross‐wind gradients in soil moisture. Source : Figure 3.19 of the Handbook.

Adapted Products

1. Observations

Given the high spatio-temporal variability of convection, satellite observation, and in particular Meteosat with its high frequency (10 min), is the most suitable tool for observing convective activity, while being aware of the limitations of remote sensing measurements depending on the channels used. It is therefore advisable to refer to Chapter 9 “Remote Sensing” of the Handbook to master its use. The figure below illustrates the ability of the IR channel to track a MCS and the triggering of a new MCS by its density current.

Sequence of IR images illustrating the triggering of an MCS by a cold pool generated by a previous MCS on 27 September. Yellow arrows indicate the limits of the arc‐shaped leading edge of the cold pool as detected by the low clouds. Source: Extract from the Figure 3.27 of the Handbook.

Automatic algorithms for the detection and tracking of convective cells (MCS Tracking) have been developed and facilitate the analysis of satellite imagery to determine the life cycle, characteristics (intensity, depth, size…) and trajectory of these convective elements. For this purpose we recommend the use of the EUMETSAT operational product ‘Nowcasting’ SAF (NWCSAF/EUMETSAT).

The MISVA website proposes to monitor convective activity :

2. Forcast

For the short term (0-6h) only nowcasting tools using observations (satellites, radar, surface…) can help the forecaster to establish a forecast. Chapter 6 “Nowcasting” of the Handbook presents the main approaches used with many illustrations.

Beyond that, numerical prediction of precipitation is poor, even within a few days, and the use of forecast ensemble processing gives only slightly better results than climatology. This is why the WASA/F method proposed by MISVA is to rely on variables that are better predicted, such as PW Paramètre Eau Précipitable – Misva (aeris-data.fr) among othet parameters.

The forecasting method WASA – ANASYG – Misva (aeris-data.fr) is detailed in the Chapter Chapitre 11: Cartes synthétiques d’analyse et prévision sur l’Afrique de l’Ouest : WASA/ of the Handbook.

It consists in examining the analyses and forecasts of 9 objects characterising the meteorological situation and then the forecaster deduces the expected convective activity and its type of organisation, based on the conceptual models linking convection to these 9 objects and on his experience. It is therefore advisable to examine the variables below, which are available daily on the MISVA website.

Références

Handbook

  • Chapter 3        Deep Convection          pages 90-129
  • Section 11.12       WASA-F Method: Convection          pages 444-451
  • Section 6.2.3 Nowcasting: shear and microbursts pages 234-245
  • Section 9.1.4.4.4 Convection RGB: pages 337-338
  • Section 9.1.4.4.5 Day microphysics RGB: pages 338-341

Illustrations et cas d’études

  • Handbook 
    • Section 3.2.1 Life Cycle of Convection on 12–16 August 2012 (CS02): pages 115-119
    • Section 3.2.2 A Cold Pool Case Study, 27 September 2014 (CS14): pages 119-121
    • Figure 5.21 Convective Dust Storms (Haboob):  pages 199
  • Site bilingue handbook case study (umr-cnrm.fr)
    • CS01: 1-10 Aug. 2012 – Train of AEWs with Breaking – Part 3. Convective activity over 10 days
    • CS02: 13-16 Aug. 2012 – An Archetype of AEW – Part 3. Convective activity over 10 days

Articles

Emanuel KA 1994. Atmospheric Convection. Oxford University Press: New York.

Fiolleau T, Roca R. 2013. Composite life cycle of tropical mesoscale convective systems from geostationary and low Earth orbit satellite observations : method and sampling considerations. Q. J. R. Meteorol. Soc. 139: 941-953. doi : 10.1002/qj.2174

Lafore J‐P, Moncrieff MW. 1989. A numerical investigation of the organization and interaction of the convective and stratiform regions of tropical squall lines. J. Atmos. Sci. 46: 521–544.

Ludlam FH. 1980. Clouds and Storms: The Behavior and Effect of Water in the Atmosphere. Pennsylvania State University Press: University Park, PA ; 405 pp.

Neelin JD, Peters O, Hales K. 2009. The transition to strong convection. J. Atmos. Sci. 66: 2367-2384. doi: 10.1175/2009JAS2962.1.

Taylor CM, Gounou A, Guichard F, et al. 2011. Frequency of Sahelian storm initiation enhanced over mesoscale soil-moisture patterns. Nat. Geosci. 4(7): 430-433. doi :10.1038/ngeo1173.

Weisman ML, Klemp JB. 1986. Characteristics of Isolated Storms. In: Mesoscale meteorology and forecasting, PS Ray (ed.). American Meteorological Society ; pp. 331-358.

Zipser EJ. 1977. Mesoscale and convective-scale downdrafts as distinct components of squall-line structure. Mont. Weather Rev. 105: 1568-1589. doi: http://dx.doi. org/10.1175/1520-0493(1977)105<1568:MACDAD>2.0.CO;2.

Search