1 Introduction

The El Niño-Southern Oscillation (ENSO) climate variability mode is a phenomenon of ocean–atmosphere interaction that occurs in the Tropical Pacific Ocean and is considered the primary driver of climate variability in various regions of the Earth (Trenberth 1997; Philander 1998; Adamson 2022). ENSO has two distinct phases: a warm phase known as El Niño and a cold phase known as La Niña (Berlato et al. 2005; Abreu et al. 2023).

Between 2015 and 2016, El Niño was one of the strongest in history (Varotsos et al. 2016; Burton et al. 2020) and was called a Mega El Niño due to the increase in extreme droughts in various regions of the world (Jimenez et al. 2018; Xie et al. 2022). Drought is a primary factor of wildfire and its increasing frequency projects a higher risk of fires, too. Understanding the dynamics of fires in arid areas, e.g., in Africa, is crucial for comprehending ecosystem properties and their impact on the global carbon (C) cycle (Wei et al. 2020). Furthermore, fires are significant in African ecosystems (Trollope and Trollope 1997; Bowman et al. 2009; Ribeiro et al. 2019), with the continent responsible for 70% of the global burned area (Russell-Smith et al. 2021) and 50% of C emissions related to fire (Giglio et al. 2013; Jiang et al. 2020). Arid lands cover 43% of the African continent and play a critical role in the C cycle and the livelihoods impacted by wildfires (Wei et al. 2021; Russell-Smith et al. 2021). Africa is known as the "Continent of Fire" and can be characterized by extensive biomass burning (Trollope and Trollope 1997; Strydom and Savage 2016). Fires represent crucial factors affecting ecosystem processes (Ribeiro et al. 2019), especially in grassland areas (Cao et al. 2015), followed by impacts on biogeochemical cycles, vegetation dynamics, and interference with Africa's climate conditions (Sedano et al. 2020; Russell-Smith et al. 2021).

Many factors contribute to fires in the African continent, such as population density, season (i.e., only 5% of wildfires occur in the rainy season), vegetation type (e.g., grasslands, shrublands, savannas, and forests), vegetation state (phenology; dry or wet; alive or dead), and landscape connectivity (hindering or facilitating spread), but the role of rainfall pattern stands out and determines fire variability (Archibald et al. 2009, 2011; Strydom and Savage 2016; Russell-Smith et al. 2021; Loudermilk et al. 2022).

According to satellite data and statistical models, Andela and Van der Werf (2014) found that changes in the rainfall patterns during the transition from El Niño to La Niña accounted for 51% of the increase in fires in Southern Africa. They demonstrated that rapidly changing demographic and socioeconomic situations have led to a significant increase in the area burned due to the conversion of savannas into agricultural lands. Burton et al. (2020) utilized the coupled fire-vegetation model JULES-INFERNO to investigate the impact of the 2015/16 Mega El Niño on fires in South America, Africa, and Asia, but the model did not accurately represent the increase in burned areas during El Niño conditions in Africa.

Numerous studies have already been conducted on fires in Africa from numerical modeling and the impact of rainfall (De Sales et al. 2019; Strydom and Savage 2016; Burton et al. 2020), followed by applied statistics (Andela and van der Werf 2014; Ribeiro et al. 2014, 2020) to the use of satellite products (Chen et al. 2017; Saha et al. 2019; Jiang et al. 2020; Sedano et al. 2020; Wan and Roy 2022) and risk analysis (Ansari et al. 2023a, b). However, only a few studies have exclusively addressed the influence of El Niño and its categories on fire risk based on a coupled ocean–atmosphere model along with a fire risk index, which represents a novel methodology applied to the African continent.

The SPEEDY-HYCOM climate model offers several advantages over other climate models, such as its high flexibility and applicability to various environmental problems (Oliveira-Júnior et al. 2021; Gobbo et al. 2022). In many aspects, SPEEDY-HYCOM can match or even outperform state-of-the-art models in simulating climate means and variability, especially large-scale features (Kucharski et al. 2005; Kotsuki et al. 2022). The model is effective in investigating the sensitivity of simulated climate to changes in parameters within physical parameterization, which is crucial to understanding how different factors affect the climate (Held and Suarez 1978; Molteni 2003). Recently, the model has been used in studies on wildfire meteorology (Xu et al. 2016; Oliveira-Júnior et al. 2021).

Therefore, our research aimed to (i) investigate the potentiality of coupled SPEEDY-HYCOM model for wildfires and (ii) evaluate the influence of El Niño events on the dynamics of wildfires in Africa via the coupled SPEEDY-HYCOM model. The results shed light on vegetation burning in Africa and its complex interrelationships between fire, climate, and human activity. The findings from this research process provide valuable insights and point to several important considerations for sustainable fire management and the reduction of pollutant emissions.

2 Material and Methods

2.1 Study Area

Africa is a continent of about 30 million km2, accounting for 20.3% of the planet. It is the second most populous continent on Earth, represented by 54 countries (Fig. 1)—(Spinoni et al. 2021). The African continent is bounded by the Mediterranean Sea to the north (N), the Isthmus of Suez and the Red Sea to the northeast (NE), the Indian Ocean to the southeast (SW), and the Atlantic Ocean to the west (W). In addition to the main continent, Africa includes Madagascar and several archipelagos (Thomson 2022). Africa is the hottest continent on Earth, with approximately 60% of its land area consisting of arid lands and deserts (Griffiths 2013). According to the Köppen–Geiger climate classification map (Fig. 1), the tropical zone (rainforest and savanna) prevails between the N11° and S11° latitudes, but larger areas belong to the arid zones (deserts, semi-arid areas, and steppes) with increased risk of wildfires.

Fig. 1
figure 1

(Source: set of climate maps, spatial resolution of 1 km, topographically corrected 1980–2016, Beck et al. 2018)

African continent, with the Köppen–Geiger climate classification.

Full size image

2.2 SPEEDY-HYCOM Coupled Model

The Simplified Parametrization, primitivE-Equation DYnamics (SPEEDY) model is a spectral and hydrostatic atmospheric model, with eight layers (30, 100, 200, 300, 500, 700, 850, and 925 hPa—hectopascal) in sigma coordinates, with a horizontal truncation of T30, this corresponds to a horizontal resolution of approximately 3.95 degrees in both latitude and longitude (Gobo et al. 2022). For the implementation of the SPEEDY model, the following variables were added: long-wave radiation (LWR) and short-wave radiation (SWR), large-scale condensation, convection heat flux, latent heat flux, moisture, and momentum, as well as vertical diffusion processes (Held and Suarez 1978; Molteni 2003). SPEEDY offers computational advantages by having its processing twice as fast as other climate models of the same horizontal resolution. In this way, it allows for carrying out experiments on distinct scales, such as interannual, decennial, and even secular (Kotsuki et al. 2022). In addition to the advantages previously presented, the SPEEDY model reproduces climate system characteristics regardless of the region (tropical, middle, and high latitudes) (Kucharski et al. 2005).

The Hybrid Coordinate Ocean Model (HYCOM) is a model version of ocean circulation in hybrid coordinates from MICOM (Miami Isopycnic Coordinate Ocean Model) and was developed to simulate the global ocean circulation, as well as mesoscale and coastal processes, in addition to solving equations of momentum, mass, and energy balance in hybrid vertical coordinates (Bleck 2002). HYCOM uses the coordinates hybrid, as the ocean has columns of water that are not stratified, or unstable convection from International Comprehensive Ocean–Atmosphere Data Set. These coordinates allow a better transition from the mixed layer (ML) to greater ocean depth (Bleck 2002).

2.3 SPEEDY-HYCOM Model Simulations

The first-year simulation began in January 1960 due to the initial spin-up to stabilize the model and not bring noise to the simulation process. Restoring the boundary conditions is useful when seeking balance for oceanic models, in which the temperature and surface salinity must follow the data from model input. Similarly, they are also used in models of atmospheric conditions when sea surface temperature (SST, °C) is prescribed. For such SST conditions, we used the Extended Reconstructed Sea Surface Temperature (ERSST), which is a global monthly analysis of SST data derived from the International Comprehensive dataset Ocean–Atmosphere Dataset (ICOADS) (Huang et al. 2020).

In the context of heat fluxes, a plausible argument can be made that these fluxes are proportional to the deviations. This concept aligns with the formulation of heat flow described in Fick's 1st law. In physical terms, for example, an SST warm anomaly leads to increased heat flow from ocean to the atmosphere. It, therefore, leads to ocean cooling, which tends to restore the previous equilibrium condition. To initialize the SPEEDY-HYCOM model, we utilized initial conditions such as sea surface temperature climatology (SST, °C), temperature at 2 m above ground (°C), precipitation (mm.day−1), evaporation (mm.day−1), and albedo (%). All of these data are derived from the NCEP/NCAR reanalysis (Kalnay et al. 1996) and covers the period from 1979 to 2015.

2.4 El Niño Classification

El Niño events in the Eastern Tropical Pacific were classified via SST anomaly based on a predefined climatology between 1961 and 2020 for the entire time series of SST, obtaining linear anomalies. The series time of the SST anomalies was created for the region between 5° N and 5° S and 110° W and 170° W. Events were defined in three consecutive periods of three months, superimposed on values above 0.55 °C, being called El Niño events. The events were defined in three consecutive three-month periods, overlaid on values above 0.55 °C, and were referred to as El Niño events. The event thresholds were classified as follows: (i) weak, with an SST anomaly between 0.55 °C and 0.95 °C; (ii) moderate, with an SST anomaly between 0.96 °C and 1.45 °C; and (iii) strong, with an SST anomaly exceeding 1.45 °C (Table 1). The classification was based on simulations of the coupled ocean–atmosphere model SPEEDY-HYCOM.

Table 1 Classification of El Niño extremes via SST anomaly simulated by the SPEEDY-HYCOM model. El Niño types are Weak (W), Moderate (M), and Strong (S)
Full size table

2.5 Meteorological Fire Danger Index (MFDI)

The MFDI was used in the study to assess fire hazards in Africa. The MDFI represents how predisposed vegetation should be burned, as the longer the time without rain (summer weather and/or prolonged drought), the greater the risk of vegetation burning (Camia et al. 1999; Vitolo et al. 2020). The calculation of the MFDI consists of information on vegetation cover and values daily maximum air temperature (Tmax, ºC), minimum relative humidity (RH min, %), and accumulated rainfall (P, mm) of the site.

The Drought Day Index (DD) is calculated daily based on the product of 11 precipitation factors taking into account accumulated precipitation over 11 previous periods (days) (Eq. 1 and Table 2)

$${\text{DD}} = \mathop \prod \limits_{i = 1}^{11} \exp \left( { - \beta_{i} ,P_{i} } \right),$$
(1)

where \({\beta }_{i}\) is the decay factor, and \({P}_{i}\) represents the accumulated precipitation in millimeters (mm).

Table 2 Ranges and values of the decay factors for the 11 periods that integrate the Drought Day (DD) index
Full size table

The Base Danger (BD) was estimated for each day by combining DD for that day with a sinusoidal curve representing the effects of phenology for different types of vegetation, as per Eq. (2)

$${\text{BD}} = 0.45\left( {1 + \sin \left( {\min \left( {A \times {\text{DD}},180^\circ } \right) - 90^\circ } \right)} \right).$$
(2)

In this context, the argument of the sinusoidal function is measured in degrees, and the coefficient values for ‘A’ are 2.4 for closed savanna and closed forest, and 3.0 for savanna and open forest. The rationale for using BD is that, for the same DD value, the fire danger is higher.

The effects of relative humidity and air temperature were also taken into account for each day through the so-called Humidity Factor (HF) and the Temperature Factor (TF), defined in Eqs. (3) and (4)

$${\text{HF}} = - 0.006 \times {\text{RH}} + 1.3$$
(3)
$${\text{TF}} = 0.02 \times T + 0.4,$$
(4)

where HF is the Humidity Factor, RH is the Relative Humidity, TF is the Temperature Factor, and T is Surface Air Temperature. Fire danger increases (decreases) when relative humidity is lower (higher) than 40%. Temperature has a linear relation with the risk increasing when the temperature exceeds 30 °C.

For each day, the value of MFDI is obtained by multiplying the Base Danger by the Moisture and Temperature Factors, as per Eq. (5) and Table 3

$${\text{MFDI}} = {\text{BD}} \times {\text{TF}} \times {\text{HF}}.$$
(5)
Table 3 Classes of meteorological fire danger and respective ranges of the Meteorological Fire Danger (MFDI)
Full size table

The methodology used in the study is summarized in Fig. 2.

Fig. 2
figure 2

The flowchart of the methodology applied in the study

Full size image

3 Results

3.1 Fire Risk Modeling

By simulating the MFDI in lagged months, according to the Weak ENSO category (Fig. 3), we identified the variability of the maximum risk (high and very high) in the time lags. Based on lag-0, which represents the maximum peak period in ENSO (Fig. 3a), the maximum risk of MFDI identified was concentrated in the equatorial belt and in Southern Africa with a higher risk of fire in the following countries: Congo, Gabon, and Madagascar. Up to 3 months before the peak of an ENSO Weak (Fig. 3b), there was a decrease in the moderate risk area and an expansion of the very high-risk area. The very high risk expanded to West (W) African countries, thus, delimiting a wide and homogeneous range of forest fires. Furthermore, the region around the South of Southern Africa practically became less sensitive in terms of MFDI. Six months before the peak of Weak ENSO (Fig. 3c), there was an expansion of maximum risk to the North (N) of the continent and again no risk to the S of Southern Africa. For 9 months before the peak of ENSO (Fig. 3d), a condition of stabilization occurred in the region where there was a maximum risk, concentrated in the equatorial belt of the continent and expansion of moderate areas and weak for S southern Africa.

Fig. 3
figure 3

Simulations with the SPEEDY model for a control period (1961–2021), for weak ENSO regarding the spatial distribution of MFDI and its categories

Full size image

Regarding the Moderate ENSO category (Fig. 4), there was a similarity between the previously noted fire risk in the Weak ENSO category (Fig. 3), mainly in relation to spatial variability between time lags where there was a region with a predominance of maximum risk, mainly located in the equatorial belt and migrating to the N, mainly 6 months before the peak of the Moderate ENSO category (Fig. 4c).

Fig. 4
figure 4

Simulations with the SPEEDY model for a control period, for Moderate ENSO regarding the spatial distribution of MFDI and its categories

Full size image

Concerning the other two ENSO events, it is important to highlight the important state of the main ENSO category (Fig. 5). The region of East (E) Africa, specifically between Kenya and Somalia had been identified as having the maximum risk according to the MFDI, in all conditions (lag -0, -3, -6, and -9 months; Fig. 5). Extensive undergrowth is prevalent, particularly in the S, Southwest (SW), and E of Africa.

Fig. 5
figure 5

Simulations with the SPEEDY model for a control period, for Strong ENSO regarding the spatial distribution of MFDI and its categories

Full size image

3.2 Influences of Precipitation and Evaporation

Regarding the El Niño Weak lag-0, there was no change in evaporation throughout the continent. In contrast, at lag-3, a positive evaporation anomaly was observed in Central Africa, spanning from the coast to the interior of the continent, with particular emphasis on the countries of Guinea, Ivory Coast, Ghana, and Nigeria. This positive anomaly was followed by negative evaporation rates in Southern Africa, primarily along the border between Zambia, Zimbabwe, Namibia, and Botswana (Fig. 6).

Fig. 6
figure 6

Evaporation simulations (mm) in the Weak ENSO category

Full size image

A similar situation was identified in lag-6 in relation to the previously highlighted areas in lag-3, with the intensification of evaporation anomalies. An exception was lag-9 in the El Niño Weak category, with negative anomalies of evaporation in central Africa and positive evaporation anomalies between Namibia and Botswana.

In the context of El Niño events, a comparison between the Moderate and Weak categories revealed relevant differences at lag-0. Specifically, a considerable decrease in evaporation was observed, with an emphasis on negative anomalies occurring across the entire of Central Africa (Fig. 7). The border region shared by Zambia, Zimbabwe, Namibia, and Botswana exhibited the most substantial positive anomalies in evaporation. This contrast becomes particularly evident during lag periods of 3 to 6 months after the El Niño Weak conditions. Concerning the lag-3, there was a reduction in the magnitude of positive anomalies and a shift to negative effects on evaporation across the continent. This change also corresponds to variations in evaporative patterns compared to the conditions observed during El Niño Moderate at lag-0. Notably, these patterns persist through the lag-6 period. Moreover, the negative anomalies of evaporation of lag-9 can be found in the central areas of Africa without the occurrence of positive anomalies of evaporation as in the Weak El Niño category.

Fig. 7
figure 7

Evaporation simulations (mm) moderate ENSO category

Full size image

The El Niño Strong category lag-0 indicated a higher occurrence of negative evaporation anomalies compared to the other categories, again highlighting Central Africa (Fig. 8). In lag-3 of El Niño Strong, there was a significant decrease in evaporation, unlike the other categories of El Niño. In the El Niño Strong lag-6 positive anomalies of evaporation in Central Africa and negative evaporation anomalies between Angola and the Congo. In lag-9, negative evaporation anomalies were observed off the coast of Africa, particularly affecting regions such as Gambia, Guinea-Bissau, and Mali. Conversely, positive anomalies occurred in Tanzania.

Fig. 8
figure 8

Evaporation simulations (mm) strong ENSO category

Full size image

There was no change concerning lag-0 and lag-3 of the El Niño Weak in precipitation across the continent except Sudan (Fig. 9). Regarding the lag-6 of El Niño peak, the largest positive precipitation anomalies occurred in all of Central Africa. There were minor precipitation records in Central Africa with emphasis on Gambia, Senegal, and mainly Chad and Sudan (Fig. 9).

Fig. 9
figure 9

Precipitation simulations (mm) weak ENSO category

Full size image

With the El Niño Moderate lag-0, the lowest precipitation rates across Central Africa as opposed to El Niño lag-0 Weak (Fig. 10). The higher precipitation records can be found obviously in the southern part of Africa, particularly in Madagascar. In lag-3, however, there were lower records of precipitation in some African countries, e.g., Mauritania, Niger, Sudan, Somalia, and the border between Angola and Congo. Regarding lag-6, the highest precipitation rates occurred in all of Central Africa, except in Ethiopia, unlike southern Africa with the lowest rainfall rates, opposite situation of lag-0. Highlight for lag-9 with the lowest rates of precipitation, mainly in Chad and Sudan, similar to El Niño category weak.

Fig. 10
figure 10

Precipitation simulations (mm) moderate ENSO category

Full size image

There was a significant increase in the smallest records of rainfall in El Niño Strong category for lag-0 as opposed to lag-0 in previous years, followed by a decrease in positive precipitation anomalies in Southern Africa (Fig. 11). It is worth highlighting the similarity between El Niño Strong and El Niño Weak with lags of 3 and 6 months, and the exception was the increase in anomalies of negative rainfall in southern Africa, mainly in Angola and the border between Zambia, Namibia, and Botswana. Furthermore, lag-9 was not similar to the lags of the previous year, it indicated a negative precipitation anomaly in Senegal and Mauritania and a positive precipitation anomaly, only in Tanzania.

Fig. 11
figure 11

Precipitation simulations (mm) strong ENSO category

Full size image

4 Discussions

The results obtained from MFDI confirm the previous findings of Andela and van der Werf (2014), which reveal that the trend of burning areas is directly related to ENSO, increasing by 51%. Strydom and Savage (2016) assessed the spatiotemporal distribution of fires in South Africa from 2003 to 2013 using a dataset of MODIS-derived Active Fire Hotspots and found that the eastern and northeastern regions were more susceptible to fires, in contrast to the MFDI, which identified only the northeastern sector. In recent decades, the areas affected by wildfires in South Africa have increased significantly, as indicated by studies by Joshua and Sukumar (2021) and Carter et al. (2021).

It is noteworthy that such countries are densely covered by forest (Ribeiro et al. 2019; Burton et al. 2020; Jiang et al. 2020) and, recently, it was verified via geo-spatialization of the MODIS product MOD14/MYD14 Fire and Thermal Anomalies in the Congo Basin and Forests Mountain ranges on the Northeast (NE) coast of Madagascar (Wan and Roy 2022). The moderate risk category can be found in Southern Africa, mainly in the Kalahari region. Saha et al. (2019) evaluated the impact of fires in the albedo of the terrestrial surface, since the generalized presence of fires in Africa has implications for the regional climate and the hydrological cycle, mainly in the Kalahari region, southern Africa. Previously, Joshi and Sukumar (2021) pointed out that the socio-anthropogenic conditions in Southern Africa exhibit a high positive sensitivity of the burned area to temperature, in contrast to northern Africa, which shows a high negative sensitivity, similar to the results obtained in this study using the MFDI index.

These conditions are directly associated with the variability of the rainy season in the African continent, primarily in the equatorial belt (Le Barbé et al. 2002; Liebmann et al. 2014; Andela and van der Werf 2014; Uele et al. 2017; Ayanlade et al. 2018), as also identified in the study through precipitation simulations based on the ENOS categories. It is worth noting that previous studies have linked the seasonal variability of precipitation to the influence of ENOS (Andela and van der Werf 2014; Uele et al. 2017; Ayanlade et al. 2018). Studies based on orbital products, for example, Jiang et al. (2020), identified long-term fire trends based on burned areas from the MODIS product (MOD13C1) and fire emissions from the Global Fire Emissions Database (GFED4s) in Central Africa. Joshi and Sukumar (2021), using Vegetation Type Fractions (MOD12Q1/MCD12C1) and Gross Primary Productivity (GPP) (MOD17A1) products, indicated an increase in the Central and Southern regions of Africa. The use of MDFI and the SPEEDY-HYCOM model showed similar results to previous studies for both mentioned regions.

This increased risk is directly linked to the soil moisture conditions, characterized by a high deficit in moisture. Consequently, it contributes significantly to the expansion of the MFDI risk area (Rojas et al. 2011; Dong-Yun et al. 2013; Cao et al. 2015; Burton et al. 2020). Our simulations on the precipitation and evaporation, since ecosystems largely depend on their distribution space–time in Africa (Konapala et al. 2020), with time lags (0, 3, 6, and 9; i.e., with a 0 to 9-month lag) for the above-mentioned El Niño categories.

According to De Sales (2019), the rainfall deficit is linked to evapotranspiration, resulting from a decrease in the net surface radiation balance, followed by a reduction in precipitation, especially in metropolitan areas with high population density in Southern Africa. Burton et al. (2020) also suggested a potential intensification in the frequency and extent of forest fires due to the projected population increase along with global warming in Africa. These factors are expected to amplify the impacts of forest fires based on the seasonal variability of rainfall and evaporation in the region, aligning with the findings in this study. The scientific and practical importance of understanding fires and their implications for the current and future climate of Southern Africa is underscored, with results similar to those in the study, except for the population factor.

Uele et al. (2017) reported negative anomalies of precipitation, as well as Sedano et al. (2020) used multitemporal Sentinel-2 imagery to monitor and quantify forest degradation extent, both in Southern Africa. Previously, Andela and van der Werf (2014) found that the changes in precipitation were driven by ENSO and contributed to the increase in wildfires over southern Africa. Burton et al. (2020) used the coupled fire-vegetation model (JULES-INFERNO) to assess the impact of El Niño 2015/16 on fire dynamics in Africa and the results indicated that there was an increase in area burned under El Niño conditions, mainly concerning the precipitation in southern Africa, being drier compared to the rest of the mainland, corroborating the result obtained for lag-6, regardless of El Niño category. Chen et al. (2017) found that the reductions in precipitation and terrestrial water storage have increased emissions from pantropical forest fires during and after El Niño compared to La Niña, being confirmed in the study using the simulations of precipitation and evaporation variables, mainly in Africa Central and Southern.

The SPEEDY-HYCOM model enabled the observation of the effects on Southern and Central Africa concerning the expansion of fire risk categories through MDFI. These changes can pose a threat to various biomes and directly impact the populations of the countries in both regions mentioned above. The results are supported by the findings of Rohat et al. (2019) who evaluated socioeconomic and climatic scenarios in the Central Africa region, which will be drastically affected by climate change, especially in terms of exposure to extreme temperatures resulting from anthropogenic actions and urbanization processes. Heat waves and prolonged droughts occurring in the African continent intensify the risk of fire as indicated in recent studies by Richardson et al. (2022) and Senande-Rivera et al. (2022).

5 Conclusions

The MFDI simulated by the coupled model SPEEDY-HYCOM for three categories of El Niño (Weak, Moderate, and Strong) in lagged months indicated that there are marked differences between the categories regarding the danger of fire in Africa. The El Niño Weak presents variability between maximum risk as well as time lags mainly in Southern African countries, where sustainable fire management should be encouraged, followed by emissions reduction and real-time monitoring of fires during El Niño events.

El Niño Moderate is similar to El Niño Weak, the exceptions are lag-6 and the highest occurrence in the equatorial zone of Africa, which in turn deserves attention by authorities and managers regarding environmental factors in relation to the occurrence of fire during the performance of both the El Niño categories. El Niño Strong occurs in East Africa, with maximum fire risk, regardless of time lags, such a condition is necessary for mitigation actions, as the category implies maximum danger of fire in Africa. Precipitation and evaporation simulations via the SPEEDY-HYCOM model point out that the El Niño categories act with intensity and duration different in Africa. However, Central Africa and especially Africa Austral deserves all the attention in relation to El Niño, with emphasis on the lags 0 and 6 (evaporation) and 0, 6, and 9 (precipitation). In summary, the SPEEDY-HYCOM is perfectly applicable using the classic and operational fire hazard index, as well as in the evaluation of the variability of the climatic conditions and their spatial distribution in Africa during the warm phase of ENSO.

This study utilizes the SPEEDY-HYCOM climatic model to assess the Meteorological Fire Danger Index (MFDI) and highlights distinct advantages compared to other methods for detecting wildfire risk in vegetated areas. The accuracy and predictive capabilities are significantly enhanced with the use of this model, providing a more effective approach to wildfire management and prevention in vegetated environments. These findings have significant implications for the safety and conservation of affected areas.

Initially, the study emphasized the need to identify the factor that exerts the greatest influence on the circumstances that promote the occurrence of forest fires in environments affected by El Niño. Understanding this critical component is fundamental for formulating effective strategies for prevention and mitigation. Furthermore, a more in-depth examination of these findings is essential to assess the socio-environmental consequences caused by forest fires resulting from El Niño events. These consequences have the potential to impact not only biodiversity and ecosystem health but also local communities, the economy, and air quality. It is also necessary to investigate La Niña events in the African continent using not only coupled ocean–atmosphere models but also satellite-derived data on burned areas, land use, and census data, with a particular focus on population density.