Projecting Climate Change Impacts on Benin’s Cereal Production by 2050: A SARIMA and PLS-SEM Analysis of FAO Data

Abstract

Globally, agriculture is facing significant challenges due to climate change, which is seriously affecting grain yields. This research aims to analyze the significant effect of climate change (temperature and rainfall) on cereal production in Benin. The choice of Benin is explained by its strong dependence on agriculture and its vulnerability to climatic variations. This study employed climate and agricultural data from FAO and ASECNA (1990–2020) to evaluate the impacts of climate change on cereal production. SARIMA time-series models were used for forecasting, while the PLS-SEM approach assessed the relationships between climate variables and cereal production. The findings reveal a rise in temperatures and a gradual decline in precipitation. Despite these challenges, the time-series analysis suggests that Beninese farmers are expanding cultivated areas, successfully increasing production levels, and improving yields. Projections to 2050 indicate an increase in areas and production for maize and rice, while sorghum shows a constant trend. However, even with these projections, it is recommended to explore, in more depth, the resilience strategies used by cereal producers to better understand their influence and refine the orientations of future agricultural policies.

1. Introduction

Climate change is one of the global issues we face today. It has diverse impacts and represents a major challenge to global food security [1]. In West Africa, a region that is particularly vulnerable due to its geographical and socio-economic characteristics, agriculture, an essential pillar of economic development, is seriously threatened [2]. The impact on food availability is considerable, particularly affecting farmers and the least advantaged populations, both in rural and urban areas [3].

Benin, a country heavily dependent on agricultural production, is not immune to the harmful negative effects of this increased climate variability [4]. In this country, the agricultural sector employs more than eighty percent of the population. In Benin, about 40% of the population contributes about 40% of the GDP; Benin generates nearly 88% of export revenues and has a predominance of cereal crops (notably maize, rice, and sorghum), which cover, on average, 49.5% of the cultivated land [5]. Regardless of the importance of crops, the yield balance is less and less positive, mainly due to climate change, as highlighted by Adjovi et al. [6]. Indeed, Benin has long been facing a contraction of the agricultural season, persistent deviations from usual climatic conditions, an increase in temperature, and a deterioration in the rainfall regime, thus transforming the landscape of agricultural production [7]. Studies conducted by Faye et al. [8] in Senegal on cereal crops also confirm that the decline in agricultural yields is mainly due to rising temperatures, which restricts production potential. For Ahossin et al. [9], grain producers mainly take crop losses and soil degradation as indicators in their assessment of climate change impacts.

This situation requires a rigorous evaluation of the effects of climate change—current and future—on cereal production, which is essential for food security and rural livelihoods. Previous research has not used predictive models that integrate long-term climate scenarios, although these tools are recognized as essential for anticipating future impacts on cereal production and developing appropriate adaptation strategies [10]. These models make it possible to identify the most critical climate variables, such as extreme temperatures and precipitation variability, and to assess their influence on agricultural yields.

Furthermore, many studies have examined the impact of climate change and producers’ adaptability techniques at different scales, both at the global and African levels, including at the national level [11,12,13,14,15,16,17,18,19,20,21,22,23]. Each of these studies agrees that a change in climatic conditions forces producers to seek effective adaptation methods or strategies. However, these studies have largely neglected the cereal sector, despite its importance, particularly for essential crops such as maize, sorghum, and rice. The few studies that have addressed this sector have focused on specific cereal crops, either maize or rice. Thus, this study aims to broaden the scope of assessment through analysis of the impacts of climate change on cereal production in Benin by focusing on the major cereal crops (corn, rice, and sorghum) that dominate this sector. By adopting an integrative approach combining climate and agricultural data, this study aims to provide a clear understanding of local climate and agricultural dynamics, while offering essential information to policy makers to guide adaptation strategies at national and regional levels.

2. Conceptual and Empirical Framework on Climate Change and Agricultural Policies in Benin

2.1. Concept of Climate Change or Climate Variability

According to the UNFCCC [24], climate change refers to a set of climatic alterations resulting from human activities that modify the composition of the Earth’s atmosphere, as well as natural fluctuations observed over similar time intervals. Similarly, the Intergovernmental Panel on Climate Change (IPCC) has defined climate change as any alteration in the climate over time, resulting either from natural variability or from human activities [8]. According to the latter, these changes are manifested by (i) an increase in temperature; (ii) changes in precipitation patterns and water availability; (iii) a rise in sea level and the salinization of ecosystems; and (iv) an increase in the frequent intensity of extreme events.

Moreover, these differences are characterized by a lasting transformation of the climate state that extends over an extended period, typically several decades or more [25]. Following these clarifications, this study builds on the idea that climate change involves long-term changes in the mean climate position, variability, and extremes, such as temperature and precipitation, in a specific region. Understanding this definition is crucial because it allows us to select the climate parameters that best characterize climate change and that deserve to be taken into consideration in this assessment.

2.2. Impact of Agricultural Policies and Climate Change Adaptation Strategies

Adaptation, conceptually rooted in evolutionary theory, explores the ability of living things and systems to develop or modify their behaviors to survive and thrive in changing environments [26]. This notion, discussed by researchers such as Smit and Wandel [27] and Lambert and Rezsöhazy [28], also applies to the response to climate variation, requiring adjustments to changes in temperature, precipitation, and soil conditions [26]. Initially explored in the context of disease outbreaks in specific environmental conditions, the relationships between the natural environment and living organisms have evolved to include strategies to reduce the sensitivity of humans and the environment to the effects of climate change [26]. Therefore, climate adaptation involves adjusting practices, policies, and infrastructure to respond to new climate realities while taking advantage of the opportunities they present.

The theory of adaptive behavior of smallholder farmers highlights the importance of adaptation in agricultural decisions, especially in response to climate change [29]. Farmers adjust their objectives and actions according to their changing situation, making strategic decisions to cope with climate impacts [29,30]. This notion highlights that adaptation is a key element of farm management in the face of climate challenges, with the aim of improving performance while minimizing potential risks. The agricultural adaptation strategies available to cope with climate change can be grouped into four main areas, such as financial control, government programs and insurance, agricultural production practices, and technological development [31]. These categories vary according to the type of agents involved and the level of involvement (local, national), highlighting the importance of taking adaptation into account in studies to avoid overestimating the effect of climate change on agricultural yields, especially those of cereals.

Given the urgent need to adapt to the consequences of climate change, in Benin, many measures have been implemented at the national level to strengthen the resilience and adaptive capacity of producers. In this context, the country established the Strategic Development Guidelines (OSD 2006–2011) during the period 2006–2016, which have been periodically updated to develop an effective agricultural policy, stimulate economic growth, and position Benin as a dynamic and competitive agricultural force. To achieve these objectives, Benin adopted a Strategic Plan for the Recovery of the Agricultural Sector (PSRSA) and put in place several supporting policies [32].

Among the policies implemented in the agricultural sector, there is a strategy for mobilizing financial resources, where the government and its technical and financial partners finance various adaptation interventions targeting producers [33]. Furthermore, the promotion and use of improved seeds or climate-resilient varieties, training on best agricultural practices, facilitating equitable access to agricultural credit, and encouraging agroforestry practices are integral parts of the agricultural policies developed by the State to help the farming population adapt to new climatic conditions [33]. Furthermore, these policies have been reinforced by the introduction of 50 digital solutions in the agricultural sector, classified into four main areas: advisory, training, and information services for producers (56%); marketing of agricultural products (16%); facilitating connections between actors in the agricultural value chain (14%); and services for monitoring crop activities and mapping farms (32%) [32]. Although some of these solutions are still in the evaluation or testing phase, Gbedomon et al. [32] point out that 54% of these digital approaches are already operational, enabling producers to better respond to climate change adaptation issues.

As a result of these policies and the PSRSA, Benin has experienced a revival of agricultural growth that has enabled it to be self-sufficient in the production of cereals, roots, and tubers [32]. This suggests that these interventions have enabled producers to minimize the influence of climate change on cereal production. These observations could be considered potential disruptive elements, influencing the expected impact of climatic parameters on cereal production indicators in this study.

3. Methodology

3.1. Choice of Study Framework

Benin was chosen as the study area for several reasons. First, agriculture is Benin’s main source of income, making the country particularly vulnerable to the effects of climate change [33]. This increased vulnerability is a key reason for focusing this study on Benin, as it allows for an in-depth assessment of the impact of climate change on a vital sector of the national economy. Second, by opting for national data, this study thus offers a complete representation of its territory. This allows for a comprehensive picture of the significant effect of environmental degradation on cereal production in various agroecological contexts.

Furthermore, this study is particularly important considering the dominant cereal crops in the country: maize, rice, and sorghum. These crops are not only the most cultivated but they also constitute the staple food of a large part of the Beninese population [5]. By focusing research on these specific crops, this study aims to provide precise and relevant information on production dynamics in the face of climatic hazards. Thus, Benin presents an ideal study setting due to its economic dependence on agriculture, its vulnerability to climate change, and the strategic importance of its staple cereal crops. These combined factors make the study both relevant and crucial for better visibility of the situation at the country level. Figure 1 details the geographical scope of the study, encompassing the whole of Benin.

3.2. Data Used

Building on previous studies, climate data such as mean annual temperature and precipitation were obtained from ASECNA, over a 30-year period (1990–2020), in line with the methodologies of several other authors [8,35]. This approach is also in line with scientific recommendations, including those of the World Meteorological Organization (WMO), to analyze the evolution of climate parameters such as temperature and precipitation [36]. In addition to these two climate parameters, previous authors such as Faye et al. [8] and Grami and Ben Rejeb [35] have mainly focused on harvested area and mean annual cereal yield (kg/ha) to assess production variability in the cereal sector. Building on these studies, the data in this research encompass parameters such as yield per hectare, total area harvested annually, and total cereal production (including maize, rice, and sorghum) harvested annually. The inclusion of total production as an additional indicator is supported by the OECD [37], which identifies it as a key measure of the status of a crop in a given region. These available data are from the FAO website covering the period 1990–2020 (

Table 1. Summary table of variables used for the model.

Variables	Description	Sources	Year
Climate factors
Precipitation	Mean annual precipitation and mean monthly precipitation were collected (mm).	Association of Birth Control Services (ASECNA)	1990–2020
Temperature	The monthly average temperature was collected (°C).	Association of Birth Control Services (ASECNA)	1990–2020
Production indicators for corn, rice, and sorghum
Area harvested (ha)	Annual area harvested by crop	FAO	1990–2020
Yield (kg/ha)	Average annual yield per crop	FAO	1990–2020
Total production (t)	Total quantity of product harvested per crop	FAO	1990–2020

). The FAO platform was chosen for its accessibility and complete data availability for each parameter targeted in this study.

3.3. Analysis Method

The chronological evolution of the parameters was assessed using time series analysis methods. These, also called “time series”, show how the variables have changed over time and are the most appropriate approach when dealing with a longitudinal study as in this case [38]. To highlight the influence of climatic parameters on the three indicators (harvested area, yield, and total production) of the production of the three crops, structural equation modeling (SEM), in particular the partial least squares (PLS) approach, was used. Indeed, structural equation modeling (SEM) is the best choice compared to other data analysis techniques, such as factor analysis or regression analysis [39]. This preference is explained by its ability to construct latent variables, manage modeling errors, test theoretical hypotheses with empirical data, and evaluate complex models involving several dependent variables as in the present case [40]. Mathematically, PLS-SEM aims to create a linear model with the following general form:

$$Y=XB+E$$

(1)

Or

• Y is a matrix of n discovered by m answered as variables (harvested area, yield, and total production for each crop);
• X is a matrix of n observations by p predictor variables (design) (mean annual temperature and precipitation);
• B is a matrix of p-by-m regression coefficients, and E is described as the error term of the model, which is used in the same dimension as Y;
• The relationships between each latent variable and its manifest indicators represent the “external models” (Equation (2)), while the relationships between the latent variables are called the “internal models” (Equation (3)).

$$X_{kj}={\pi }_{kj} {\delta }_k+{ϵ}_{kj}$$

(2)

where ${\mathit{X}}_{\mathit{k}\mathit{j}}$ is the vector the same with the jth manifest variable of the latent variable ${\mathit{\delta }}_{\mathit{k}}$; ${\mathit{\pi }}_{\mathit{k}\mathit{j}}$ is a loading or structural coefficients associated with ${\mathit{X}}_{\mathit{k}\mathit{j}}$; and ${\mathit{ϵ}}_{\mathit{k}\mathit{j}}$ is an error term (measurement errors of the manifest variables).

In this case, it is about the relationships between the “Climate” parameter and its own measurement indicators (average annual temperature and precipitation), then between each crop (corn, rice, sorghum) and its production indicators (area, yield, and total production).

$${\delta }_k=\sum _{i={\delta }_i}^{{\delta }_k}{\beta }_{ki}{\delta }_i+{\epsilon }_k$$

(3)

where ${\mathit{\beta }}_{\mathit{k}\mathit{i}}$ symbolizes the structural value connected to the variables’ connection ${\mathit{\delta }}_{\mathit{k}}$ and ${\mathit{\delta }}_{\mathit{i}}$; ${\mathit{\epsilon }}_{\mathit{k}}$ is a term of error related to the inherent latent variable ${\mathit{\delta }}_{\mathit{k}}$. Here, these are the underlying variables’ direct correlations (climate) with the latent response variables (corn, rice, sorghum).

The measurement model was evaluated first, because unreliable or invalid constructs lead to unreliable relationships between constructs in the structural model [40]. Thus, the precision and reliability of the measurements were assessed by performing an assessment of internal consistency, discriminant value, and convergent accuracy (AVE) [41,42]. The structural model was then evaluated by assessing four collinearity criteria (VIF), the coefficient of determination (R2), and path coefficients [41].

Table 2. Indicator for evaluating measurement and structure models.

Criteria	Acceptance Thresholds	References
Measurement model
Path Coefficients	≥0.7	[39,41]
Cronbach’s Alpha (CA)	≥0.6
Composite Reliability (rhoC)	≥0.6
Average Variance Extracted (AVE)	≥0.5
Structural model
Variation in inflation factor (VIF)	<5	[39,41]
Coefficient of determination (R2)	≥0.5	[39,41]

summarizes the thresholds set for each of these criteria in previous studies.

Production indicator forecasts:

The forecasts were made using seasonal ARIMA (SARIMA) models. Indeed, the analysis of simple and partial autocorrelation functions allowed us to note the presence of seasonality in the production parameters. In such situations, seasonal ARIMA (SARIMA) models seem to be the best suited [43]. These SARIMA models allow us to take into account the trend, seasonality, and random components in the data, which makes them particularly suitable for modeling time series with a strong and well-defined seasonality [44]. Mathematically, the variable ($X_{T+h}$ ($h\ge 1$), can be approximated by [44] the following:

$${\widehat{X}}_T\left(h\right)=\sum _{i=1}^m{Z}_{T+h}^i{\widehat{\beta }}_i+\sum _{j=1}^n{S}_{T+h}^i{\widehat{\gamma }}_j$$

(4)

with $S_t=\sum _{j=1}^n{S}_{T+h}^i{\widehat{\gamma }}_j$, the seasonal component checking $E\left(S_t\right)=0$; $Z_t$ representing actual data, ${\widehat{X}}_T$ forecasts, and ${\widehat{\beta }}_i$ the coefficients estimated from the T observations.

Moreover, among the indicators commonly used to assess forecast performance, this study relied on the mean absolute deviation (MAE), a method that quantifies the average of the mean deviations between the predicted value and the actual observed value [45]. It is given by the following:

$$MAE=\sqrt{{\displaystyle \frac{\sum _{i=1}^N{(Y_i-{\widehat{Y}}_i)}^2}{N}}}$$

(5)

With ${\widehat{\mathit{Y}}}_{\mathit{i}}$ being the predicted value and ${\mathit{Y}}_{\mathit{i}}$ being the actual value for an observation i.

The coefficient of determination (R2) was also determined for each projection performed. To carry out each of the different analyses developed, R software version 4.3.0 was used.

4. Results

4.1. Impacts of Climate Change on Cereal Production and Projections for 2050

4.1.1. Dynamics of Precipitation and Average Annual Temperatures from 1990 to 2020

Analysis of maximum (a) and minimum (b) temperatures in Benin, segmented into three distinct seasons, reveals clear trends (Figure 2). Temperatures, both maximum and minimum, are naturally higher during the long dry season (November to March), with maximum temperatures reaching up to 38 °C in 2020, compared to only 36 °C in the 1990s. During the transition period and rainy season, temperatures are cooler, although the transition period records maximum temperatures of around 34 °C in 2020, compared to around 33 °C in previous years. Minimum temperatures show an increasing trend during the rainy season and the transition period but a decreasing trend during the dry season, recording minima around 19 °C in 2020, compared to over 21 °C between 1990 and 2015. The results obtained are consistent with the conclusions of the International Committee on Climate Change [46], as well as with other studies [47], which document a progressive increase in temperatures both in Africa and globally, a symptom of global warming.

The chronological analysis of precipitation (Figure 2c) in Benin, differentiated according to the seasons delimited by the climatic characteristics detailed by the DGEC [33] (extensive wet period, extensive dry period, and transition period), reveals irregular fluctuations but a general downward trend. The additional rainy period (June, July, September, October) and the transition period (April, May, August) show a significant decrease in precipitation, from 5000 mm between 2000 and 2010 to less than 3500 mm in 2020. During the transition months, precipitation also declines, from more than 2000 mm to less than 1500 mm in 2020. The dry season (November to March) follows a similar trajectory, with a decrease from 400 mm in the 1990s to only 50 mm in 2020. These trends illustrate the increasing instability of precipitation due to climate change, showing a serious obstacle to the sustainability of agriculture in Benin and Africa [7,48].

Moreover, local studies have also highlighted the perceptions of farming communities towards these climate changes. For example, Djenontin [49], Dedjan [50], and Yegbemey et al. [51] reported similar observations among farmers, particularly in northern Benin, regarding changes in rainfall distribution and variations in rainfall during wet periods. These changes pose considerable challenges for agriculture, particularly for cereals, highlighting the need for appropriate adaptations to mitigate the effect of global warming on agricultural production in Benin [6].

4.1.2. Dynamics of Production Indicators (Harvested Area, Yield, Total Production) for Cereal Crops (Maize, Sorghum, and Rice) from 1990 to 2020

From 1990 to 2020, maize cultivation in Benin experienced significant growth in terms of harvested area, yield, and total production (Figure 3 and Figure 4). The harvested area nearly tripled, from 457,903 hectares in 1990 to 1,267,159 hectares in 2020. This considerable expansion in harvested land reflects growing demand and may also be linked to agricultural policies favoring production intensification, as noted by Dayou et al. [52]. Maize yield also increased, although more moderately, from 895.4 kg/ha to 1271.8 kg/ha. This improvement in yield is likely related to the use of improved seed varieties and modern agricultural practices. As a result, total maize production quadrupled from 409,994 tons in 1990 to 1,611,615.13 tons in 2020. This increase in maize production is attributed to the exponential population growth observed in both rural and urban areas, which has significantly increased the demand for food and nutrition [53]. However, although the general trend in maize yield is upward, the Figure 3 shows a considerable decline in the last five years. This instability in maize yield is attributed to several factors, including the lack of specific fertilizers, shortage of agricultural labor, and poor climatic conditions [54].

As for rice, production also experienced positive dynamics during the same period. The harvested area increased from 7836 hectares in 1990 to 104,586 hectares in 2020, reflecting significant efforts to expand the land dedicated to this crop. Occupying the second place among the most consumed cereals in Benin, this expansion highlights its leading role in the country’s food diversification [55]. Similarly, the yield almost tripled, from 1396.1 kg/ha to 3935.3 kg/ha. These increases are attributable to the implementation of improved rice varieties resistant to unfavorable climatic conditions [56].

However, although an overall increasing trend is observed between 1990 and 2020, there is considerable variability from year to year, as shown in Figure 3b. This corroborates the findings of Niang et al. [57], who demonstrated through trials conducted in the Glazoué commune (central Benin) that rice yields fell by 16.4% between 2010 and 2013 due to a decline in overall rainfall and a delay in the onset of rains [58]. Nevertheless, with the overall increasing trend in rice yields, total rice production has also grown exponentially, from 10,940 tons in 1990 to 411,578.14 tons in 2020. These results are impressive given the challenges posed by climate change and demonstrate the importance of this crop in Benin.

During some agricultural seasons, rice production represents more than 80% of total cereal production in some production areas of Benin [59]. This highlights the significant increase in production levels and the dynamism of Benin’s agricultural economy during this period. However, sorghum dynamics present a contrasting picture compared to other crops (Figure 3c). The harvested area slightly decreased from 135,528 hectares in 1990 to 134,693 hectares in 2020. This small reduction may be associated with farmers increasingly favoring more profitable crops such as maize and rice [55]. Indeed, it is recognized that the demand for sorghum is mainly concentrated in Borgou, while the demand for crops such as rice and maize is significantly higher in areas beyond this region [60].

Nevertheless, the yield and total production of sorghum showed a notable increase from 730.7 kg/ha to 1100.5 kg/ha and from 99,026 tons in 1990 to 148,235.93 tons in 2020. This improvement can also be attributed to the adoption of more resistant varieties.

Figure 3. Dynamics of yields and total production of maize, rice, and sorghum in Benin from 1990 to 2020. Source: FAO [61].

Figure 3. Dynamics of yields and total production of maize, rice, and sorghum in Benin from 1990 to 2020. Source: FAO [61].

Figure 4. Dynamics of harvested areas of corn, rice, and sorghum in Benin from 1990 to 2020. Source: FAO [61].

Figure 4. Dynamics of harvested areas of corn, rice, and sorghum in Benin from 1990 to 2020. Source: FAO [61].

4.1.3. Evaluation of the Influence of Climatic Factors on Cereal Production Indicators Using the Partial Least Squares Method (PLS-SEM)

• Evaluation of the measurement model

The evaluation of measurement techniques was conducted to assess the reliability and validity of the measures. The product of this evaluation, using the PLS-SEM algorithm, is presented in

Table 3. Measurement model evaluation indicators.

Variables	Indicators	Coefficients	Cronbach Alpha (CA)	Composite Reliability (RhoC)	AVE	Number of Indicators Deleted
Climate	Precipitation (mm)	−0.578	−0.727	0.146	0.609	0
	Temperature (°C)	0.942
Corn	Area harvested (ha)	0.941	0.898	0.937	0.834	0
	Yield (kg/ha)	0.792
	Production (t)	0.994
Rice	Area harvested (ha)	0.974	0.959	0.974	0.925	0
	Yield (kg/ha)	0.928
	Production (t)	0.983
Sorghum	Yield (kg/ha)	0.437	0.518	0.476	0.392	2 (Area and Production)

. Several commonly used indicators were explored to analyze the PLS-SEM model. Among these indicators are internal reliability measures, including lambda coefficients, Cronbach’s alpha (CA), and composite reliability (rhoC). These measures aim to determine whether the indicators or latent variables are homogeneous, that is, whether each convergent latent variable adequately and meaningfully captures the response of the dependent variable.

Indicators with coefficients less than 0.7 (e.g., precipitation) were not eliminated if their deletion would not affect the composite reliability. However, indicators that had lower coefficients were deleted if their deletion would help the AVE. Based on the results of the measurement model evaluation, two indicators were specifically eliminated in the context of sorghum cultivation: harvested area (coefficient of −0.132) and total production (coefficient of 0.437). The results showed that external loads, CA, CR, and AVE were all within the acceptable thresholds. Table 3 presents the results of the measurement model evaluation.

• Evaluation of the structural model

The PLS model examines the importance of climate on production indicators of three major crops in Benin: maize, rice, and sorghum (Figure 5). Annual precipitation and annual mean temperature are used as climate indicators, and their impact on the harvested area, yield, and total production of each crop is analyzed.

For maize, the total coefficient of climate effect is estimated at 0.7659, indicating a positive and significant influence of climate on maize production (p-value < 0.0000). The coefficient of determination (R²) is 0.5866, showing that 58.66% of the variability in maize production can be explained by climatic variations. Specifically, while annual precipitation has a negative influence, annual mean temperature has a positive influence. This results in an expansion of the harvested areas, yield, and total production of maize in response to higher temperatures, despite variable precipitation. The deliberate use of maize in Benin’s agricultural development plan and exponential population growth, which increases food demand, further favor the growth of this crop [53].

For rice, the total coefficient of the climatic effect is 0.7594, also significant (p-value < 0.0000). The coefficient of determination (R²) indicates that 57.67% of the variability in rice production indicators can be explained by climatic variations. As for corn, the positive influence of the annual mean temperature on the three production indicators compensates for the negative effect of annual precipitation, favoring an increase in the yield and total production of rice. This trend is attributed to the use of better agricultural activities and the increase in food demand due to population growth.

For sorghum, the total effect of climate is estimated at 0.5349 but is statistically insignificant (p-value = 0.1100). The coefficient of determination (R²) shows that only 28.61% of the variability in sorghum production can be explained by climatic variations. This highlights the limited and statistically insignificant impact of climate on sorghum, suggesting that other factors may play a more critical role in determining its production.

Overall, the results show that climate, particularly mean annual temperature, has a significant and beneficial impact on maize and rice production. These crops benefit from higher temperatures despite fluctuating annual precipitation, which improves indicators such as yield and provides producers with incentives to expand their harvested areas to maintain high levels of total production. By contrast, the impact of climate on sorghum is less significant, indicating the potential influence of other factors.

Based on the demonstrated influence of climatic parameters on agricultural production in the literature [7,8,48], the results of this research indicate that the lack of significant influence on sorghum may be due to the increasing adoption of adaptation methods by cereal producers. Indeed, the implementation of adaptation methods, for example, the use of improved cultivars and resilient agricultural practices, helps to mitigate the potential effects of climatic variations [20]. These adaptation strategies probably already reduce the expected impacts of precipitation and temperature on yields and production [62,63].

4.2. Prediction

4.2.1. Autocorrelation Function (ACF)

Figure 6 presents the autocorrelation function (ACF) of the different production indicators of the three cereal crops. The analysis of this compiled Figure allows us to assess the seasonality of the data for each crop (maize, rice, and sorghum). For each indicator and for the three crops, a periodic pattern of decreasing peaks occurring on an annual scale is observed. The interval between two consecutive peaks or two consecutive decreases corresponds to the duration of the seasonality of the data.

Thus, the presence of decreasing annual peaks in the ACF graphs suggests annual seasonality. This observation validates the choice of the SARIMA method for forecasting.

4.2.2. Forecasts of Area, Yield, and Total Production of Cereal Crops

Projections of harvested areas, yields, and total production of cereals in Benin show mixed trends until 2050 (Figure 7). For maize, a gradual increase in harvested areas is projected to reach 2 million hectares by 2050, with a potential total production reaching 2.5 million tons. However, yield appears to stabilize around 1200 kg/ha. This growth in area and production is likely due to favorable agricultural policies and increasing food demand driven by population growth. The stability of yield, despite the increase in other indicators, can be attributed to persistent climatic constraints and challenges related to the advancement of agricultural techniques. These results contrast with predictions made a decade ago, which anticipated a possible 20% decline in major cereal production in sub-Saharan Africa by 2050 due to rising temperatures [64]. This suggests that in the Beninese context, adaptation methods—such as mulching, the use of organic fertilizers, innovative variables, fertilizers, pesticides, and agroforestry—are helping producers to better manage the situation [20].

For rice, the forecast indicates a significant increase in all three indicators. The harvested area is expected to reach 200,000 hectares, the yield is expected to exceed 6600 kg/ha, and total production is expected to exceed 600,000 tons by 2050. These optimistic projections contrast with other forecasts, which suggest that yields from rainfed agriculture (all crops combined) could halve, with an increase in average temperature between 2020 and 2050 [65].

For sorghum, forecasts show stability on the three indicators: around 150,000 hectares for the harvested areas, 11,000 kg/ha for yield, and 150,000 tons for total production. This stability may be due to the natural resilience of sorghum to climatic variations but could also reflect a stagnation in efforts to improve the productivity of this crop.

Overall, these projections need to be considered in the context of the potential impacts of climate change on agriculture in sub-Saharan Africa. The United Nations Environment Programme (UNEP) Africa Adaptation Gap report indicates that warming of 2 °C by 2050 could reduce overall agricultural incomes in sub-Saharan Africa by 10%, with higher temperature increases potentially reducing yields by 15–20% [65]. For maize, the report predicts a 30–40% decline in yields in West Africa, with rising temperatures also likely to affect the quality of agricultural products [65]. These projections highlight the urgent need to scale up adaptation efforts to strengthen the resilience of agricultural systems to impending climate challenges. This becomes even more critical when considering the projections of Yegbemey et al. [51], who estimate a potential reduction of 13 to 15% in average annual rainfall by 2100, particularly in northern Benin. Such a decrease could seriously affect cereal crops in all agro-ecological zones of the country [51].

5. Discussion

The results of this research show significant trends in climate dynamics and cereal production in Benin between 1990 and 2020. The observed increases in maximum and minimum temperatures across seasons highlight an acceleration of global warming, corroborating the results of local studies such as those of Dossa and Miassi [66] and Issa et al. [48]. The decline in rainfall during critical crop growth periods highlights the challenges of rain-fed agriculture in Benin, which remains the main cultivation method. These climatic changes exacerbate the vulnerability of cereal crops, particularly maize, rice, and sorghum, directly impacting their yields and total production [67]. Farmers’ perceptions of these changes, as reported by Djenontin [49] and Yegbemey et al. [51], reinforce the importance of strategic adaptations to mitigate negative impacts on agriculture.

Regarding cereal crop dynamics, the results indicate a notable increase in harvested areas and yields for maize and rice, despite the instability of climatic conditions. This increase is attributed to the extension of cultivated land, the use of better seed varieties, and modern agricultural practices [52,56]. However, sorghum production shows a slight decrease in harvested areas, offset by an increase in yields and total production, probably due to the adoption of more resilient varieties [55]. The research highlights the importance of crop diversification and the adoption of adapted technologies to improve the adaptability of agriculture to global warming. Initiatives to improve irrigation systems and introduce more resilient crop varieties are crucial to ensure food security in Benin, as highlighted by Faye et al. [8] and Iwédiga et al. [68].

The PLS-SEM model confirms the significant influence of climatic factors, particularly the annual mean temperature, on cereal production in Benin. Although the observed positive impact may not correspond to initial expectations, it should be noted that studies conducted in various regions and for different crops have reported both positive and negative effects on agricultural production [46]. This highlights the complexity of the interactions between climate and agricultural production, which depend essentially on local contexts and agricultural practices.

The model further indicates that climate, particularly temperature, has a positive impact on maize and rice production, with high coefficients of determination for these crops. According to Edoun and Mongbo [53], the evolution of production levels of these cereals is mainly determined by population growth and by agricultural policies favorable to the expansion of maize and rice cultivation. By contrast, the impact of climate on sorghum is less significant, suggesting that this crop is either more resilient to climatic variations or influenced by other factors, such as local food preferences [60,69].

Projections to 2050 show contrasting trends for cereal crops in Benin. For maize, forecasts suggest a gradual increase in harvested areas and total production, despite a stabilization of yields that could be attributed to persistent climatic constraints [64]. Other studies attribute this trend to population growth, with maize remaining the most consumed food in Benin [5].

For rice, the projections are more optimistic, with notable increases in all indicators, driven by rising food demand and the adoption of improved agricultural practices. However, these projections contrast with studies predicting a decline in cereal yields of up to 40% by 2050 in West Africa due to global warming [46]. These contrasting results arise from differences in scale, as projections for West Africa often group cases together where producers adapt with those where they do not, as highlighted in related reports [56]. Nevertheless, the results are consistent with the projections of Jalloh et al. [70], who predict an increase in rice production in West Africa over the same period. Finally, projections for sorghum show stability across indicators, reflecting its natural resilience to climate variation. These results explain the crucial importance of the adaptation and adoption of resilient agricultural practices to ensure food security in Benin in a context of imminent climatic challenges [51,55].

These findings highlight the need to strengthen agricultural policies aimed at adapting to climate change. It is crucial to promote initiatives that encourage innovative practices, such as crop diversification, improved irrigation systems, and the dissemination of climate-resilient crop varieties. Additionally, fostering the adoption of climate-smart agricultural practices through targeted incentives can reduce reliance on rain-fed agriculture. Combining these measures with enhanced access to tailored training and financing programs for farmers can sustainably bolster food security in Benin, addressing the growing challenges posed by climate change.

The analysis approaches employed in this study are robust and provide highly relevant insights into the relationships between climatic factors and cereal production. Structural Equation Models (SEMs), in particular, are among the most widely used tools in the literature for assessing relationships between multiple variables. Their adoption in this study underscores a methodologically sound approach to exploring complex dynamics. However, as with any analytical framework, SEMs are not without limitations. A key concern lies in their capacity to establish causal relationships, as highlighted by ongoing debates in the literature [71,72]. While SEMs offer significant value in modeling intricate variable interactions, their application to non-experimental observational data can sometimes lead to ambiguous or misleading interpretations [72]. This is a limitation intrinsic to the method rather than a reflection of the study itself.

Additionally, this study does not incorporate certain critical factors, such as political interventions targeting producers, socio-demographic and economic conditions, or the adoption levels of innovative technologies, all of which could substantially influence cereal production. Addressing these gaps calls for further research that integrates a broader range of determinants and adopts a more comprehensive analytical framework. Such an approach would enable a deeper understanding of the complex interplay between climatic, socio-demographic, and economic factors, offering a more nuanced perspective on cereal production in Benin while enhancing the robustness and policy relevance of future findings.

6. Conclusions

This study demonstrates the significant impact of climatic factors, such as temperature and precipitation, on exacerbating the vulnerability of cereal crops like maize, rice, and sorghum in Benin, even with the gradual adoption of improved agricultural practices. The findings underscore the urgency of intensifying policies that promote climate-adapted agricultural practices to enhance the resilience of farming systems. By combining technological innovation with efficient resource management, this study provides a strong foundation for countering the adverse effects of climate shocks.

To build on these findings, future research could explore the complex interactions between climatic, socio-demographic, and economic factors influencing cereal production. Special attention should also be given to evaluating the real effectiveness of agroecological and climate-smart practices adopted by producers, considering the diversity of agricultural systems and local socio-economic constraints. A more robust analytical approach—combining qualitative and quantitative data and integrating these dimensions—should be developed to propose more reliable production and forecasting models. Such an approach would enable a more comprehensive assessment, refining recommendations for adaptive agricultural policies and improving preparedness for climate challenges, thereby ensuring long-term food security.