**3.1 Changes in agroclimatic indicators in the near (2020–2049) and far future (2070–2099)**

To study the impacts of climate change on the evolution of these indicators on each grid point of Senegal, the changes between the historical period (1985–2014) and the climate simulations SSP126, SSP245 and SSP585 for the near future (2020–2049) and the far future (2070–2099) are represented in **Figure 1**. The middle of the whisker diagram represents the spatial average over Senegal, and it is important to note that here the multi-model average is considered. An amplification of the changes in the evolution of the agroclimatic indicators is observed when moving from the near future to the distant future and from the sustainability scenario SSP126 to the fossil fuel development scenario, SSP585. In the near future, seasonal precipitation totals will be broadly the same or even slightly lower than what is observed in the historical model simulations. This slight decrease is more a consequence of a high occurrence of dry sequences and a shortening of the rainy season illustrated by a late start to the season. This evolution will become more pronounced in the distant future with a decrease in average over the country of −10% for SSP126, −20% for SSP245 and over −35% for SSP585.

By analyzing the distribution of precipitation over the season, it is clear that the trend observed over the period 2020–2049 is confirmed and amplified in the distant future. Indeed, a strong increase in dry sequences (DSs and DSxl) will be observed, which will increase from +15% for SSP 126 to +50% for SSP585. This situation will favor the risks of reseeding and post-flowering water stress. Although at the beginning of the rainy season (June to July), the occurrence of rainfall breaks of 8 to 14 days is a fairly frequent event, and the observation of a DSl or DSxl after the semi flowering is not something that farmers want. Indeed, even if young millet plants show a certain capacity to adapt to such pockets of drought, the same cannot be said for maize, sorghum or certain legumes (groundnuts, cowpeas, etc.). The water requirement of crops during the heading-maturity phase (or reproductive phase) is one of the main factors conditioning their final yield. In general, this period corresponds to the critical reproductive phase for non-photoperiodic crops with a cycle length of about 90 days. A late start of the agronomic season, from 10 days for SSP126 to about 20 days for SSSP585, will also be observed, whereas the end of the season seems to be less sensitive to the effects of climate change. Paradoxically, extreme rainfall events, notably R99 (mm) and R20 (days), will increase by about 10–15% and 20–25%, respectively, by 2050 before decreasing slightly in the distant future. These results are perfectly consistent with studies [31–33] demonstrating the installation of a new rainfall regime over the Sahel characterized by false starts, a shortening of the rainy season, an increased frequency of intense rainfall and an occurrence of dry sequences, which is directly associated with global warming.

#### **3.2 Spatial distribution of the change in agroclimatic indicators**

In order to better analyze the spatial distribution of these indicators, the average change of these indicators on each grid point is represented in **Figure 2**. This spatial distribution of changes on these agroclimatic indicators allows to measure the spatial heterogeneity of the impact of global warming.

#### **Figure 1.**

*Projected changes in the evolution of the 14 agroclimatic indicators. Changes are defined as the difference between the historical period (1985–2014) and the climate simulations SSP126 (top panel), SSP245 (bottom panel left) and SSP585 (bottom panel right) for the near future (2020–2049; boxplot in black) and the distant future (2070–2099; boxplot in red).*

#### **Figure 2.**

*Spatial distribution of changes in the 14 agroclimatic indicators in addition to the mean surface temperature (Tas). The changes are defined as the difference between the historical period (1985–2014) and the SSP126 climate simulations for the near future (2020–2049).*

Indeed, despite a general decline in cumulative seasonal rainfall on all pixels of the country, some areas of the country will be more impacted than others. The center of the country and even the Central East will suffer the largest decreases in terms of cumulative seasonal exceeding −20% followed by the south with rates between −15

*Evolution of Agroclimatic Indicators in Senegal Using CMIP6 Simulations DOI: http://dx.doi.org/10.5772/intechopen.109895*

#### **Figure 3.**

*Ranking of the four models used according to the 14 agroclimatic indicators. The sign (+) of the arrow indicates an increase in the near future of the indicator and the sign (−) a decrease. For example, M2, which is the model where warming is stronger, tends to capture later season starts.*

and − 10%. The north and the coastal area of the country will record the smallest decreases (less than −10%). Precipitation intensity indicators will increase in the coastal areas, the southeast and the north except for R99 which will increase only in the south of the country. The 8-to-15-day dry spells (DSl) will increase more in the south of the groundnut basin, while DSxl will increase almost throughout the country. Finally, the delays in the start of the season observed in **Figure 1** affect the groundnut basin (Fatick, Kaolack and Kaffrine), the center-east and the southeast of the country more, so the seasons are shorter in these areas of the country. This characterization of agroclimatic indicators allowed us to set up this ranking matrix of four global climate models (GCMs), namely CNRM-CM6-1, CNRM-ESM2-1, IPSL-CM6A-LR and MRI-ESM2-0 (see **Figure 3**).

This diagram combining temperature changes, and precipitation characteristics relative to the 1985–2014 history allow us to determine, in terms of trend, whether the models are warm/dry/extreme/early-onset, warm/moist/extreme/early-onset, cool/moist/moderate/early-onset, cool/dry/moderate/early-onset for the near-future SSP126 (2020–2049). For example, temperatures are expected to increase in this period by +0.98°C with IPSL-CM6A-LR (M3), +1.07°C with CNRM-CM6-1 (M1), +1.09°C with MRI-ESM2-0 (M4) and + 1.1°C with CNRM-ESM2-1 (M2). With regard to precipitation, the changes are variable: With 3/4 models predicting a decrease in precipitation on average over Senegal ranging from −1 to −13%, the largest decrease is recorded by the warmest model, namely CNRM-ESM2-1 (M2). While the second warmest model MRI-ESM2-0 (M4) predicts an increase in precipitation explained mainly by a high occurrence of events such as R95, WD, R10 and longer seasons compared to the historical period. Finally, the IPSL-CM6A-LR, which is the model with less warming compared to the other three models, predicts a climate with an increase in extreme rainfall events such as R99, SDII, R20 and all categories of rainfall breaks (DSs, DSm, DSl and DSxl). This approach could allow for sensitivity testing with agronomic impact models. In addition, it will allow us to better understand the response of crops to different degrees of warming and extreme climate in the different agroclimatic indicators defined (**Table 2**).

### **4. Discussions and conclusions**

This assessment of the evolution of 14 agroclimatic indicators shows that the effects of climate change that will affect all of Africa are already visible at the local level. Knowledge about future climate trends and their impacts has increased in recent years in Africa. According to the IPCC, monsoon rainfall is expected to increase over the central Sahel and decrease over the extreme western Sahel (Senegal area). The monsoon season is expected to have a delayed start. However, at a more localized scale, the question of the reliability of this information arises. The ability of global models to represent the Sahelian climate faces difficulties related to the limited availability of quality climate data and the many uncertainties intrinsic to the physics of the models. In order to better understand the impact of climate change on key indicators for the agricultural sector, four corrected global models are used to characterize their evolution in the near (2020–2049) and distant (2070–2099) future.

The results corroborate with the conclusions of the latest IPCC report that predict a decrease in rainfall in the western Sahel, agricultural droughts and extreme rainfall events. Indeed, we show that climate change will lead to an almost constant decrease for the different socioeconomic scenarios, of the seasonal cumulative rainfall in the near future below −10% for the ssp126 scenario and about −40% in the distant future for the ssp585 scenario. This decrease in rainfall is rather explained by an increase in rainfall breaks combined with a shortening of the rainy season as illustrated by the results. The frequency of the different categories of dry sequences will tend to increase to reach +20% for the DSl. This trend will persist in the distant future and could exceed +50% with the ssp585 scenario. The shortening of the rainy season will be mainly explained by a delay in the start of the rainy season which could be 1 month compared to the average observed over the 1985–2014 history. Furthermore, despite a decrease in precipitation, extreme rainfall events will increase in intensity (R99) and frequency (R20) in the near future for all scenarios before decreasing in the distant future. This increase in the frequency of heavy rainfall corroborates the results of Taylor et al. [34], Panthou et al. [35] and Chagnaud et al., [36] who attribute this intensification of rainfall to global warming, which particularly affects temperatures in the Sahara. Thus, a warmer Sahara intensifies convection in Sahelian storms through increased wind shear and changes in the Saharan air layer [37]. We also analyzed the strong spatial disparities behind these future trends. The results indicate that the central part of the country and the Central East will experience the largest decreases in terms of seasonal accumulation. Precipitation intensity indicators will increase most strongly in the coastal areas, the southeast and the north. The increase in dry sequences of the DSl and DSxl type and the delays in the start of the season in the southern part of the groundnut basin constitute a source of vulnerability for the agricultural sector, with significant consequences for food security, diseases, farm capital, loss of livestock, etc. [38]. In conclusion, our work could be the starting point for simulations of yields, biomass and several relevant indicators for the agricultural sector in order to quantify the impact of climate change on agriculture in Senegal. These studies could be carried out on various spatial scales (field, commune, department or region). This will allow for the exploration of adaptation strategies such as scaling up investments in irrigation, fertilization or pesticides and accelerating the adoption of climate-proof agricultural technologies such as drought-resistant crop varieties and collaboration between farmers and breeders for the mass use of organic fertilization.
