Seismicity pattern of African regions from 1964–2022: b-value and energy mapping approach

Abstract

The African continent is one of the few tectonic plates where very few earthquakes have been reported. Sporadic poor data have jeopardized efforts to assess seismic hazards. In light of the tectonic setting, a review of the continent’s seismicity is crucial for predicting seismic hazards. For this study, seismicity data from USGS and ISC catalogs were collected from 1964–2022, in the latitude range of 40^oS-40^oN and longitude 20^oW-52^oE, from the surface to 700 km deep. Gutenberg’s parameters as well as the spatial variation of seismic energy were then analyzed as an assessment of Africa’s seismicity. The results show that NE Africa is at a higher stress level in terms of energy release than NW Africa. The Richter and Guttenberg’s constants of a = 5.61, b = 0.59; a = 6.55, b = 0.77; a = 6.44, b = 0.78; a = 7.41, b = 0.92; a = 5.94, b = 0.86; a = 5.97, b = 0.77 with the corresponding magnitude of completeness, mc, observed to be 3.7, 4.1, 4.4, 4.7, 4.8 and 4.6 were found for regions NW, NE, southern, eastern, western and central Africa, respectively. The distribution of focal depths vs. magnitude revealed that most of the seismicity is inherently shallow in its nature with an average focal depth of 21.22 km. In comparison, NW experiences relatively more profound events with an average depth of 32.19 km. A substantial rise in focal depth in North Africa is associated with the collision boundary. This information as a constraint on the tectonic plate near convergent boundary could be useful in plate modeling.

Key Policy Highlights

• The higher stress observed in the North Africa and East Africa continent is vital for policy makers for sustainable built environment

• The estimated minimum magnitude cut-off Mw 3.7 reveals that the recording facilities in African continent is now improving.

• An increased focal depth distribution of events in North Africa infers there is plate convergence and can be used to constrain African tectonic plate for plate modeling.

Keywords:

• Earthquake
• b-values
• energy mapping
• seismicity assessment
• African plate
• subduction zone

1. Tectonics of African continent

The seismic activity centers, such as rift zones, thrust and fold mountain belts, transform faults, and volcanic areas significantly influence Africa’s seismotectonics. The most noticeable regions of continuing tectonic deformation with significant seismic activity are the North African thrust and fold belt and the East African Rift system (Meghraoui and Pondrelli 2012; Misra and Mukherjee 2015). However, the most seismically active zones on the continent include the Cameroon Volcanic Line, the Congo Basin and the Plateau in Southern Africa (Meghraoui and IGCP-601 Working Group 2016).

The Atlas orogenic system, which includes the High Atlas, Tell-Atlas, Anti-Atlas, and Sahara Atlas, is where recent deformation and fault zones are concentrated. However, they are sparse to the east of Sidra (Great Sirt) Bay, where faults are represented pass through the Jebel El-Ahdar Mountains. This region of North Africa is part of the Alpine-Himalayan belt. According to estimates for the Neogene over the Middle and High Atlas Mountains, the Atlas deformation contributed significantly to the development of the plate boundary, accounting for 20–40% of the convergence between Africa and Europe (Gomez et al. 2000). Gibraltar-Sisly and the Atlas Mountains are notable collisional plate boundaries, which can be considered typical tectonic features along the Africa-Eurasia plate boundary. This area experienced moderate to large earthquakes, with the most notable event being the El Asnam Ms 7.3 earthquake on 10 October 1980, which was linked to a NW dipping, developing thrust fault (Philip and Meghraoui 1983; Yielding et al. 1989). It can be observed from Figure 1 that the fault system of NA where a complex fault system can be seen along the converging Africa-Eurasia plate boundary in the chain of Atlas Mountains.

Figure 1. North African fault system (Adapted from GEM Global Active Faults Database).

The Sub-Saharan African region’s most notable active continental rift system is the East African Rift System (EARS), which is also known as the East African Rift System. This divergent plate boundary, which separates the Nubian and Somalian plates, runs ∼ N-S across eastern Africa. The eastern and western branches of the EARS flank the Tanzanian craton, and the rift valleys form two significant lines. The Mozambique Channel has a third, southeast branch that finally connects to the Southwest Indian Ocean Ridge. The biggest seismicity recorded on the African continent was caused by the western branch of the EARS, which originated around 25 million years ago (Macgregor 2015).

According to elastic rebound theory (Reid 1911), earthquakes are caused by the release of accumulated energy within fault raptures of rocks during sudden movements of tectonic plates. The distribution and mechanism of active faults, which are the source of significant seismicity, are crucial in seismicity studies. The map of active faults was recently compiled by Styron and Pagani (2020) under the project ‘The GEM Global Active Faults Database’. Figure 1 shows the mechanism of faulting for North Africa. It demonstrates that convergent northern Africa is subject to reverse faulting, and the formation of sizable thrust systems and orogenic belts (the Atlas and Betic/Rif chains) is primarily to responsible for the continent’s seismic activity. Diverging borders, however, show typical faulting (dextral and sinistral). It can be observed from Figure 2, that the East African rift system is now experiencing typical faulting in that it is a developing divergent tectonic plate boundary. Large earthquakes of various sizes have been recorded in historical and current times in the region around the triple intersection (Afar triangle). The area is dominated by normal faulting, according to the region’s surface geology and the focal mechanisms of earthquakes (e.g. Kebede and Kulhánek 1991; Ayele 2002), which is in line with current worldwide analyses of active faults by Styron and Pagani (2020).

Figure 2. The East African rift system with major active faults shown green (adapted from Chorowicz 2005).

The spatial distribution of faults within the African plate is covered in different literatures (e.g. Morley 1999; Skobelev et al. 2004; Macgregor 2015; Poggi et al. 2020).

2. Introduction

Earthquake risks have been significant in African countries, specifically along and near plate boundaries and away from plate boundaries in the intraplate regions. Even though the African plate is widely considered a stable seismicity region, the presence of potentially damaging earthquakes has occurred throughout history. For instance, the 1960 earthquake in Agadir, with a magnitude of 5.8 at a depth of 15 km, caused estimated fatalities of 12,000–1500 (Utsu 2002). This demonstrates that even moderate events can lead to significant loss of lives and properties. During the last few decades, several earthquakes caused significant losses in Africa (Ambraseys and Adams 1991; Fenton and Bommer 2006). Moreover, with the increasing development of urban areas in Africa, especially in sub-Saharan Africa with modern houses (Tusting et al. 2019), there is an increasing risk of major earthquakes.

Recent studies have shown that the African plate is home to many significant and damaging earthquake tremors (Table 1). In general, Africa has been considered as a stable continent. Nevertheless, a large portion of modern-day Africa exhibits a variety of deformations, from compression and Orogenesis in NW Africa to compression across the Congo Basin to continental rifting in conjunction with the East African Rift System to the transition from continental rifting to active oceanic spreading in Afar, the Red Sea, and the Gulf of Aden (Craig et al. 2011). This paper examines the seismicity of the African continent by dividing the whole continent into five regions depending on the proximity of the extensional boundary for West Africa, East Africa, and Southern Africa, the collisional boundary for Northern Africa, and the intraplate region for central Africa. The study area in this study is bounded by the Mediterranean Sea in the north, the Red Sea in the North East, the Indian Ocean in the East, and the Atlantic Ocean in the West. It is divided almost equally in half by the Equator.

Table 1. Some significant earthquakes observed in Africa in recent decades.

Region	Country	Time	Magnitude Mw	References
Bukavu	D.R. Congo	2008	6.0	Meghraoui and IGCP-601 Working Group (2016)
north-western Guinea	Guinea	1983	6.3	Irinyemi (2021)
Zemmouri-Boumerdes	Algeria	2003	6.8	Meghraoui and IGCP-601 Working Group (2016)
Machaze	Mozambique	2006	7.0	Fenton and Bommer (2006)
Juba	South Sudan	1990	7.1	Meghraoui and IGCP-601 Working Group (2016)
EL Asnam	Algeria	1980	7.3	Meghraoui and IGCP-601 Working Group (2016)

Here, the classification of what to be stated is based on tectonic forces and sub-lithospheric activities. For instance, North Africa is much affected by the Africa-Eurasia collision, whereas Eastern Africa is widely affected by the GEARS and the accompanying tensional stresses emanating from the elevated topography. Whilst some part of Africa (e.g. Congo) is affected by compressive stresses from ridge push from the Atlantic Ocean (Ayele 2002). On the other hand, West Africa is a relatively stable region with small far-field stresses from the ridge push in the Atlantic Ocean (Zoback 1992). In comparison, southern Africa is characterized by tensional stresses similar to Eastern Africa, with some shallow earthquakes from human-induced seismicity.

Due to the interaction between seismically active Africa-Eurasia plate boundaries and faults, North Africa is characterized by active seismicity on the continent. The region covers countries such as Tunisia, Mauritania, West Sahara, Morocco, Algeria, Libya and Egypt. Along the plate boundary between Africa and Eurasia, the NW is known for recent major thrusts and strike-slip earthquakes. However, owing to two different tectonic regimes, the collision between Africa and Eurasian plates and the rifting of the Red Sea, the north-eastern area, which includes Libya and Egypt, is characterized by uncommon big and moderate regular faulting earthquakes (Meghraoui and IGCP-601 Working Group 2016).

Although numerous significant historical and contemporary earthquakes have devastated the area, West Africa, which spans from Nigeria to Senegal, is primarily regarded as a stable section of Africa. A notable illustration of an instance of intraplate seismicity that took place on a stable West African craton is the Guinea earthquake of 22 December 1983, with a magnitude of Mw 6.4. The portions of the CA with moderate to low seismicity zones include Angola, Chad, the Congo basin, and the Cameroun Volcanic Line (CVL). One of the world’s most prominent and least researched intracontinental basins, the Congo Intracratonic Basin is bordered to the east and west, respectively, by the Cenozoic East African Rift System (EARS) and the Atlantic passive continental margin. The primary deformation in both cases was of the thrust type, which is unusual for the African Plate. A compressive stress environment might be produced by the mid-Atlantic Ridge and the EARS plate boundary pressures (Ayele 2002).

A continuous review of the continent’s seismicity is essential for predicting seismic hazards within the tectonic setting. Because large earthquakes usually have a significant return period (e.g. for example Guinea, West Africa, the return period could reach ∼100 years for magnitudes that exceed Mw 6.0; Irinyemi 2021). Seismic studies need to be updated so that they can be used as input for seismic hazard assessment.

Africa’s poorly-built infrastructures and the region’s vulnerability make it very pertinent to understand the potential seismicity in the region. Even moderate-sized earthquakes can prove that they can cause significant loss of lives and properties, so thorough mapping is critical. This work assesses the potential seismicity of the region. The Assessment of the seismicity of Africa in terms of Gutenberg’s parameter variation in the six regions of Africa for the period 1964–2022 as well as the investigation of the planar variation of seismic energy in light of seismicity assessment in the Continent are considered as the main objectives of the present study.

3. Seismicity data

3.1. Sources

Earthquake data of six regions of Africa and its adjoining regions were compiled in the range of 40⁰S–40⁰N Latitude and 20⁰W–52⁰E Longitude for magnitude range 3–7.5, depth 0–700 km, magnitude type m_b, M_s, M_w, M_L, and M_D. For the present study, 93,691 earthquake events were collected from both ISC Catalogue and USGS catalogues for 1964–2022.

Based on the USGS classification, data depths up to 700 km are used to incorporate shallow, intermediate, and deep earthquakes. Shallow events range from 0–70 km, intermediate ones from 70–300 km, and deep ones from 300–700 km. Classification of seismic provinces is based on active tectonics and present-day continental deformation (Meghraoui and IGCP-601 Working Group 2016). The geographical limits of the study area and the epicentral distribution of the earthquake events are shown in Figure 3(a) and Figure 3(b), respectively.

Figure 3. (a) The six divisions of the study area. (b) Epicentral distribution of the declustered earthquake events.

Figure 4. Distribution of earthquake events by depth. The distribution of intermediate to deep events is shown in the African plate’s northern part, while the remaining continent exhibits a shallow depth distribution.

The date of the earthquake, the epicentral coordinates, and the magnitude of the earthquake in different scales, such as moment magnitude (M_w), surface-wave magnitude (M_s), body-wave magnitude (m_b), duration magnitude (M_D), local magnitude (M_L), moment magnitude (M_w), depth, and time of events, are all included in the earthquake information collected for each event in the database. Figure 4 presents the distribution of earthquakes.

Figure 5. Seismicity distribution of Africa continent in G-R. The raw data were compiled from the ISC and USGS catalogues and then declustered.

3.2. Conversion of different magnitudes into M_w

The ISC catalog consists of heterogeneous magnitude scales reported from different agencies, thus necessitating homogenization with regression equations. In this study, empirical relations were used to convert from body wave magnitude (mb) both in USGS and ISC, from surface-wave magnitude (Ms) local magnitude (ML) and Duration magnitude (MD) to corresponding moment magnitude (Mw).

For conversion of different magnitudes to Mw (Scordilis 2006), for global data including Africa; (1) $$M_w=0.85(\pm 0.04)m_b+1.03(\pm 0.23);3.5\le m_b\le 6.2$$(1)

Then the relation in Equation (1)(1) $$M_w=0.85(\pm 0.04)m_b+1.03(\pm 0.23);3.5\le m_b\le 6.2$$(1) can be used to estimate the moment magnitude (M_w).

Conversion $M_s-M_w$ (2) $$M_W=0.67(\pm 0.05)M_s+2.07(\pm 0.03);3.0\le M_S\le 6.1$$(2) (3) $$M_W=0.99(\pm 0.02)M_s+0.08(\pm 0.13);6.2\le M_S\le 8.2$$(3)

Conversion $M_L-M_w$ (4) $$M_W=0.722M_L+0.743;M_L>2,\mathrm{ }\text{for}\mathrm{ }\text{intraplate}\mathrm{ }\text{domains}$$(4)

Conversion of $M_D-M_w$ according to Akkar and Bommer (2010) (5) $$M_w=0.764\left(\pm 0.04\right)M_D+1.379\left(\pm 0.2\right) ;3.7\le M_D\le 6.0$$(5)

Conversion from $M_{s(\mathit{ISC})}-m_{b(\mathit{ISC})}$ to M_w according to Weatherill et al. (2016) (6) $$M_w=\{\begin{array}{l}0.616M_S+2.369;M_S\le 6\\ 0.994M_S+0.1;M_S>6\\ 1.084m_b-0.142;m_b<6.5\end{array}$$(6)

Conversion from $M_{s(\mathit{NEIC})}-m_{b(\mathit{NEIC})}$ to Mw, according to Weatherill et al. (2016) (7) $$M_w=\{\begin{array}{c}0.723M_S+1.798;M_S<6.5\\ 1.005M_S-0.026;M_S>6.5\\ 1.159m_b-0.659;m_b>6.5\end{array}$$(7)

3.3. Declustering of catalogue

Prior to computing seismicity, removing dependent events from the catalog are the first step in assessing earthquake hazards. Therefore, only shocks that trigger independent behaviour should be considered in the analysis. This can be achieved by removing dependent events (foreshocks and aftershocks) from the updated catalog by specifying a window both in the space and time domain.

Finding mainshocks and aftershocks is easy using windowing methods. The following shocks are classified as aftershocks for each earthquake in the catalog of magnitude M if they happen within a specific time window, t(M), and a certain distance, d(M). When the strongest earthquake happens later in the series, foreshocks are classified similarly to aftershocks, i.e. as aftershocks. The size of the most significant shock in a series determines how the time-space windows are reset.

Even though there are several window sizes in the computation of declustering of catalogs, Gardner and Knopoff (1974) method is a widely used method owing to its ease of applicability. According to Gardner and Knopoff (1974), the window sizes in the space and time domain are shown below (Equation (9)(9) $$d\left(km\right)={10}^{0.1238\mathrm{*}{M}_w+0.983\mathrm{ }}t\left(\mathit{days}\right)=\{\begin{array}{c}{10}^{0.032\mathrm{*}{M}_w+2.7389},\mathit{\text{ if }}{M}_w\ge 6.5\\ {10}^{0.5409\mathrm{*}{M}_w-0.547},\mathit{\text{ if }}{M}_w<6.5 \end{array}$$(9) ). (9) $$d\left(km\right)={10}^{0.1238\mathrm{*}{M}_w+0.983\mathrm{ }}t\left(\mathit{days}\right)=\{\begin{array}{c}{10}^{0.032\mathrm{*}{M}_w+2.7389},\mathit{\text{ if }}{M}_w\ge 6.5\\ {10}^{0.5409\mathrm{*}{M}_w-0.547},\mathit{\text{ if }}{M}_w<6.5 \end{array}$$(9)

Hence, in this study, ZMAP (Wiemer 2001) with the choice of Gardner and Knopoff (1974) algorithm is used to decluster our catalog.

3.4. Catalogue completeness

In seismic risk analysis and prediction, the Gutenberg-Richter b-value is an essential parameter. However, sometimes the FMD could be erroneous and lead to bias due to the incompleteness of the instrumental catalogue. This could happen because some magnitudes may not be felt and remain undetected by the sensors, thus affecting the reliability of the empirical relation. Therefore, it is necessary to set the local cut-off point, M₀, in such a way as to maximize the reliability of the G-R relation. Furthermore, the Mc value depends on many general factors, such as the number and quality of the sensors, their geometrical configuration, and the environmental noise level. Hence, accurate computation of the b-value is crucial to the effectiveness of the method.

In this study, a maximum curvature (MAXC) method proposed by Wiemer and Wyss (2000) was used, where the magnitude of completeness is obtained by maximizing the best-fit equation of the frequency magnitude relation. The estimated completeness magnitude of the catalogue is 4.4 which is clearly shown from Figure 5. The G-R parameters were computed in the maximum likelihood method (Aki 1965) with consideration of cut-off magnitude.

Figure 6. Temporal distribution of cum. no. of events.

3.5. Review of G-R parameters

Here, the review of some representative G-R parameters with the main findings for the six regions of Africa is presented for different time ranges.

Gutenberg’s constants, a = 5.75 and b = 1.23, were discovered by Alabi et al. (2012) in their investigation of the seismicity of southern Africa. Their findings showed that southern Africa is less seismic. The presence of a moderate earthquake, however, could be problematic in the future. The likelihood of significant earthquakes occurring in this area shortly was low. However, a significant earthquake could happen due to the uncertainties surrounding earthquake projections, particularly in intraplate locations.

Kadiri and Kijko (2021) studied the seismicity of West Africa in G-R relations. According to their finding, West Africa lies in the tectonic seismicity with a value of 4.1 and the corresponding b-values of 0.77. On the other hand, Irinyemi (2021) calculated the G-R parameters to determine hazard levels in west Africa and found that the estimated hazard levels were high in the Palaeozoic area of Guinea with a uniform b-value of 0.70 ± 0.12 and a = 4.2.

From 2013 up to 2016, Asefa and Ayele (2021) studied the space-time distribution of earthquakes in the East African Rift System (EARS). The average b- and a-values of their study events over this period were 1.01 and 6.5, respectively, according to the frequency-magnitude distribution (FMD) of those events. The b-value result suggests a region under relatively high stress, with earthquakes of tectonic origin predominating. While Lamessa et al. (2019), in their study, for the latitude range of (4^oN-20^oN) and longitude (34^oE-48^oE) for the period 1960 to 2016, found a value of 6.9 and b = 1.03. Moreover, Poggi et al. (2017) analysed the seismicity parameters of the Gulf of Aden (Near the Horn of Africa) for the time range of 1904–2013 using the MLE approach and found b = 1.02, a = 5.38.

From October 1997 to December 2013, Ali (2016) examined the seismicity of Egypt and surrounding regions using data from the Egyptian National Seismological Network (ENSN) that was limited to latitude (22°N-34°N) and longitude (25°E-36°E). His b-values ranged between 0.50 and 1.19. The Levant Aqaba with b value= 0.96 and the Cyprian Arc with b value= 1.19 have the highest b values, whereas the Suez-Cairo-Alexandria fault with b-value= 0.812 and the active Kalabsha fault with b value= 0.554 have the lowest b values. At the same time, Poggi et al. (2020) looked into the b-values as part of a study project on hazard assessment in northwestern Africa (High/Middle Atlas). As per them, the Tell Atlas’ b-value was 0.98, whereas the Middle Atlas’ was 1.1 for the G-R relation. The G-R parameters for some representative areas of Africa are presented in Table 2.

Table 2. Review different b-value studies for the six regions of Africa.

Area		Region	a-value	b-value	References
Western Rift		CA	–	0.773	Mavonga and Durrheim (2009)
South Sudan		EA	3.579	0.92	Uluka et al. (2020)
Egypt		NEA	–	0.5–1.19	Ali (2016)
High/Middle Atlas		NA	3.92–4.45	1.10	Poggi et al. (2020)
Rif/Tell Atlas		NA	3.72–4.47	0.98	Poggi et al. (2020)
EARS		EA	6.5	1.01	Asefa and Ayele (2021)
Southern-Ethiopian Rift		EA	6.9	1.03	Lamessa et al. (2019)
South Africa		SA	5.75	1.23	Alabi et al. (2012)
Gulf of Aden		EA	5.38	1.02	Poggi et al. (2017)
Mweru–South Katanga		CA	4.05	0.99	Poggi et al. (2017)
Western Africa		WA	4.1	1.01	Kadiri & Kijko (2021)
Guinea		WA	4.2	1.03	Irinyemi (2021)

3.6. Final catalogue

After removing duplicate events, 93,691 events were eligible for the final screening, which were then declustered with the Gardner and Knopoff (1974) algorithm. The remaining events were then filtered with a minimum cut-off value, and the final catalog with 13,019 events in moment magnitude is selected for analysis. The temporal variation of the cumulative number of events in the catalogue is shown in Figure 6. From the graph, the slope increases for the years 2004–2010, indicating the seismicity rate is higher in this time interval.

Figure 7. Magnitude versus time distribution of the study area.

It can be noted form Figure 7 that a dense distribution of large-magnitude earthquake events occurred in this time interval.

4. Seismicity parameters

The relationship between the size distribution of earthquakes and magnitude was discovered by Gutenberg and Richter (1944) and takes the form of the power law (Equation (10)(10) $$\hspace{0.17em}\log \hspace{0.17em}N(m)=a-bm$$(10) ). (10) $$\hspace{0.17em}\log \hspace{0.17em}N(m)=a-bm$$(10) where N is the number of earthquakes with a magnitude greater than or equal to m, and a and b are constants that describe the frequency-magnitude distribution (FMD).

Parameter a describes the productivity of the volume or quantity of earthquakes within the study area, and b the slope of the spatial mapping of frequency-magnitude distribution (FMD), describes the relative size distribution of events. The value of b is generally 1 or close to 1 for seismicity of tectonic origin (Müller & Jokat, 2000). A value of b greater than 1 indicates predominantly small earthquakes, while a smaller value of b indicates the dominance of more significant earthquakes concerning smaller events. Low-magnitude earthquake frequency declines, especially in seismically active regions, signify stress build-up. Lower ‘b’ readings before the great earthquakes reflect this (Gadkari and Mukherjee 2023). Moreover, high b-values in the volcanic compartments could explain the high seismicity in the region (Sanchez et al. 1995). In contrast, the regions with low b − values imply the considerable differential stress of the Earth’s crust, thereby pointing toward the end of the seismic cycle (Schorlemmer et al. 2005). Researchers (e.g. Utsu, 1965; Scholz 1968) pointed out in their studies that a relation could exist between b-values and stress and/or strain for a given locality.

Due to enormous implications in geodynamics and hazard assessment, correct estimation of Mc and Gutenberg’s constants is crucial in the GR relation. Generally, there are two approaches to computing the b values: least square fit and the maximum likelihood methods. However, the least square method is subject to bias because of the assumption that all the data point carries the same weight and the residuals are Gaussian-distributed, which cannot be justified at higher magnitudes (Naylor et al. 2010).

To avoid such inconveniences, the maximum likelihood method (MLE) proposed by Aki (1965) was used (Equation (11)(11) $$\frac{\hspace{0.17em}\log \hspace{0.17em}e}{M_{av}-\left(M_c-0.1/2\right)}\approx \frac{0.434}{M_{av}-M_c+0.05}$$(11) ). (11) $$\frac{\hspace{0.17em}\log \hspace{0.17em}e}{M_{av}-\left(M_c-0.1/2\right)}\approx \frac{0.434}{M_{av}-M_c+0.05}$$(11) where M_av is the mean value of the sample, M_c is the threshold above which the FMD is considered complete, and 0.1 is the bin size.

The constant of GR relation in Equation (10)(10) $$\hspace{0.17em}\log \hspace{0.17em}N(m)=a-bm$$(10) , a value is determined from the relation: (12) $$a=\mathit{logN}\left(M_c\right)+b{M}_C\mathrm{ }$$(12)

In this work, b-value, the slope of the G-R relation, and the constant ‘a’ are determined using an MLE method in log scale for absolute magnitudes only.

5. Energy mapping

Energy mapping is a critical parameter in analysing stress induced during earthquake tremors in the tectonic plates and the process of corresponding seismic hazard mapping therein. That is, it could be a measure of the potential for damage to man-made structures.

Earthquakes exceeding magnitude M_w 3.0 are used for energy mapping since most energy budgets dissipated are from major and great earthquakes. In the analysis, the relation suggested for stable continental regions such as the African plate by Kanamori (1977) is used (Equation (13)(13) $$\hspace{0.17em}\log \hspace{0.17em}{M}_0=1.5M_w+9.1$$(13) ) (13) $$\hspace{0.17em}\log \hspace{0.17em}{M}_0=1.5M_w+9.1$$(13)

To convert the moment magnitude into the radiated seismic energy, Equation (14)(14) $$E_s=0.5*{10}^{-4}{M}_0$$(14) introduced by Kanamori (1977) was used, where the radiated seismic energy is linearly related to the seismic moment, $M_0$_. (14) $$E_s=0.5*{10}^{-4}{M}_0$$(14) where Es is the energy radiated in Joules and $M_O$ is the seismic moment in Nm. Hence, from Equation (13)(13) $$\hspace{0.17em}\log \hspace{0.17em}{M}_0=1.5M_w+9.1$$(13) & Equation (14)(14) $$E_s=0.5*{10}^{-4}{M}_0$$(14) , energy radiated in Joules can be solved as in Equation (15)(15) $$E_s={10}^{(1.5M_w+4.8)}$$(15) (15) $$E_s={10}^{(1.5M_w+4.8)}$$(15)

The resulting sum of energy for the regions can be used to represent the stress released and accumulation of stress due to instrumental seismicity within the region from 1964–2022 time.

The intensity of the energy release of African continent is shown in Table 3. To enhance the comparison, energy intensity variation is also provided in graphic representation in Figure 8 for each regions.

Figure 8. The relative proportion of energy intensity in J/km² for the six regions of Africa.

Table 3. Intensity of seismic energy release from 1964–2022 in the African continent and adjoining regions. It is illustrated in the graph form in Figure 8 for better understanding.

Region	Energy released (J) × 10^16	Area (Km^2) × 10^6	Energy (J/km^2) × 10^9
NE Africa	2.79	2.87	9.74
E. Africa	3.26	5.19	6.29
NW Africa	3.26	6.34	5.14
C. Africa	0.99	6.97	1.42
S. Africa	1.41	11.23	1.25
W. Africa	0.32	6.56	0.48

6. Results & discussions

Figure 3(b) shows the seismicity pattern of Africa, and the adjacent area follows the triple junction branching into the E and SE parts of Africa, NE Africa, and the Gulf of Eden, with the centre emanating from the Afar region of Ethiopia. It follows the Great East Africa Rift System (GEARS). Apart from most earthquakes, epicenters look clustered in N Africa.

The graph in Figure 5 shows the seismicity of the African continent. As shown in the figure, the G-R constants for the continent are a = 7.25, b = 0.77, and Mc = 4.4. For the six tectonic provinces viz, NW Africa, NE Africa, S Africa, E Africa, W Africa, and C. Africa are a = 5.612, b = 0.59; a = 6.55, b = 0.77; a = 6.44, b = 0.78; a = 7.408, b = 0.92; a = 5.94, b = 0.86; a = 5.969, b = 0.77, respectively. The magnitude of completeness, Mc, was observed to be 3.7, 4.1, 4.4, 4.7, 4.8, and 4.6 for NWA, NEA, SA, EA, WA, and CA, respectively.

It can be observed from Figure 8 that NE Africa dominates the energy release budget per unit area in the African continent with 40%, while E. Africa and NW Africa constitute 26% and 21%, respectively. The regions such as western, central and southern Africa total up 13% of the energy reservoir. Though Africa is generally considered a stable continent in seismic hazard events, northern Africa and its proxy (eastern Africa) can be regarded as regions of active seismicity.

This study also highlights the presence of collisional earthquakes near Eurasia-Africa plate boundaries due to subducting slabs into the mantle. A close observation of the distribution of earthquakes in depth (Figure 4) reveals a significant amount of intermediate and deep earthquakes in the N. Africa, which is attributed to the subduction zone (Hellenic), and the collision of Africa with the Eurasia plate. In essence, it is possible to extract information about the tectonic setting of North Africa and the extent to which the cooling slab deeps. A critical observation from the absence of intermediate to deep earthquakes in the EA and SA is that these areas are home to events of volcanic origin and tectonics of extensional basins. Due to the anomalous temperature in, particularly in EA, the crust is thinner, and the seismogenic depths are relatively shallow.

Due to the sporadic nature of earthquakes in the intraplate region lying at a remote distance away from the plate boundaries, coupled with insufficient amount of data in seismic recording networks, the study requires viable solutions such as numerical methods to reinforce the efforts being done in seismic hazard assessment programs economically with less effort.

7. Conclusions

Compared to other parts of the African continent, seismic stress release per area is comparatively higher in the northern region. More likely, this was caused by the seismicity caused by collisions between plates in the Eurasia-Africa boundary zone. Furthermore, it may be inferred from the greater b-value that the population of lesser events is higher in Northern Africa. With higher a-values of 7.408 for eastern Africa and 6.55 for NE Africa, respectively, the seismicity rate is higher. While the b-value for the NW region is lower in terms of the seismicity parameter, NW Africa will likely experience substantially higher stresses because of the complicated fault systems in the Atlantic range. Notably, the b-value of 0.92 and the high seismicity rate, or a-value of 7.408, demonstrating that eastern Africa is the region with active seismicity. With a b-value of roughly 1.0, Southern Africa is still significant. Despite being far from plate boundaries, the East African Rift System’s spreading rate (Pule & Saunders, 2009), the region’s elevated topography (Craig et al. 2011), and human-induced activities like mineral mining (Foulger et al. 2018) may have substantially influenced southern Africa’s seismicity.

As far as the epicentral distribution is considered, the pattern follows the margins of both collisional boundaries in the north, the combination of collisional-transform boundaries in NE Africa, and the extensional boundaries forming a tri-branching pattern in the Afar region of Ethiopia extending to Malawi in the far southern African region.

The regions average focal depths range from 11.53 km to 32.19 km, with South Africa having the lowest average and northwest Africa having the highest, with NE Africa coming in second. N. Africa experiences a disproportionately high number of moderate and profound earthquakes compared to the rest of Africa (Figure 4). This might have been connected to subduction zone collisional boundaries. Instead, with average values of 21.22 km focal depths, the seismicity in the African lithosphere is primarily shallow in depth. In actuality, southern Africa’s average focal depth is substantially lower than the global average. Of course, human-induced activities like mineral mining could have contributed to southern Africa’s extremely shallow nature of seismicity.

The study revealed significant findings about the tectonic setting of North Africa, most notably the influence zone of collisional boundaries near the Africa-Eurasia plate boundaries. In addition, the information extracted related to the depth of intermediate to deep earthquakes (Figure 4) is essential to constrain the model’s boundary condition for developing a mechanistic model of the African tectonic plate.

Acknowledgements

Authors are grateful for detailed review comments in two rounds provided by Prof. Soumyajit Mukherjee (IIT Bombay) and Dr. Mery Biswas (Presidency University, Kolkata) which were helpful for enhancing the quality of the manuscript.

Disclosure statement

The authors report no conflict of interest.

Data availability statement

The seismicity data used in the study are compiled from the USGS website (https://earthquake.usgs.gov/earthquakes/) accessed in 2022 GC and the ISC website (http://www.isc.ac.uk/) accessed the same year. In addition, derived data supporting the findings of this study are available from the corresponding author [Kavitha B] on request.