Assessment of Spatio-Temporal Trends in Rainfall Indices in Senegal: Validation of CMIP6 Models over the Historical Period and Projections Under Future Climate Scenarios

Abstract

Senegal, like many West African countries reliant on natural resources and agriculture, faces severe impacts from climate change. This study provides an analysis undertaken by the United States Agency for International Development (USAID) under the Senegal Water Resources Management Activity, investigating historical and projected rainfall extremes to assess potential risks to water resources under future climate scenarios. Using bias-corrected CMIP6 data validated against the Enhancing National Climate Services (ENACTS) dataset for 1985–2014, we assess model performance through time series analysis, spatial distribution, and Taylor diagrams. We examine changes across three time periods—1985–2013 (historical), 2021–2040 (near future), and 2041–2060 (distant future)—focusing on nine key rainfall indices relevant to agriculture and water security. The results indicate that CMIP6 models capture historical rainfall patterns well. The models MPI-ESM1-2-HR, MIROC-ES2L, MRI-ESM2-0, CanESM5, and GISS-E2-1-G show the best performance and are recommended for climate impact assessments. Spatial analysis reveals prolonged dry periods in the north and heavier rainfall in the south. Under SSP585, the near future shows an increase in consecutive dry days (CDDs) and a decline in extreme rainfall events in northern Senegal, whereas the distant future projects a reversal with intensified rainfall (Rx5day). The south shows contrasting patterns, with increasing rainfall intensities in the long term. These findings highlight shifts in rainfall regimes and underscore the urgency of integrating future climate scenarios into adaptation planning. This study recommends extending analysis to temperature extremes due to their implications for agriculture and public health.

1. Introduction

Climate change represents one of the most critical challenges of the 21st century, affecting key development sectors such as agriculture, water resources, health, and infrastructure across the globe. In Africa, where economies are heavily reliant on rain-fed agriculture, changes in rainfall patterns and extreme weather pose significant threats to livelihoods and ecosystems. According to the latest IPCC report, surface temperature increases in Africa have outpaced the global average, contributing to shifts in regional climate systems that influence the onset, intensity, and duration of rainfall, particularly in the West African Sahel [1].

In recent decades, rainfall extremes in West Africa have been characterized by increasing inter-annual variability, shifts in the onset and cessation of the rainy season, and the occurrence of both droughts and intense rainfall events [2,3]. This variability poses challenges for agriculture and water resource management, as well as for disaster preparedness and adaptation planning [4]. Sacré Régis et al. [5] used the CHIRPS dataset to investigate historical changes in precipitation extremes across West Africa, showing a heterogeneous pattern of both dry and wet extremes that complicates regional planning. Ajayi and Ilori [6] projected an increase in the intensity and frequency of heavy rainfall events in West Africa under future climate scenarios, suggesting a potential escalation of flood risks and water-related hazards.

In Senegal, several studies have highlighted the vulnerability of the country to rainfall extremes and their impacts. Sarr et al. [7] applied a changepoint analysis approach to examine trends in selected extreme precipitation indices in Senegal, revealing significant shifts in rainfall patterns across different agro-ecological zones. Faye [8] conducted a frequency analysis of drought events using the Gumbel method, demonstrating increased drought intensity and duration in recent decades. Faye et al. [9] used the Expert Team on Climate Change Detection and Indices (ETCCDI) indices to characterize rainfall extremes, confirming a marked increase in variability with implications for water availability, agriculture, and health. Nouaceur [10] emphasized the growing occurrence of urban floods due to a combination of climate change and land-use changes in the Sahelian zone, underscoring the need for improved monitoring and adaptation strategies.

Moreover, studies by Descroix et al. [11] and Panthou et al. [3] highlighted changes in the hydrological cycle of the Senegal River Valley and the Middle Niger Basin, linking them to broader climatic shifts and increased risks of extreme rainfall events. These studies stress the importance of integrating both observational analyses and climate model projections to better understand and manage the risks associated with rainfall extremes.

Given the increasing frequency and severity of rainfall extremes in Senegal and the broader West African region, it is crucial to assess both historical and future changes to inform national adaptation planning. However, few studies have combined high-resolution observational datasets with the latest generation of climate projections (CMIP6) to assess rainfall extremes in Senegal. Addressing this gap, this study employs bias-corrected CMIP6 projections validated against the Enhancing National Climate Services (ENACTS) dataset [12] over the 1985–2014 period. We analyze nine rainfall indices recommended by the ETCCDI to capture a comprehensive picture of rainfall extremes. In accordance with the definition by Lienou et al. [13], the indices used in this study are considered “moderate extremes indices,” primarily based on percentiles with defined thresholds to assess moderate extremes that typically occur a few times per year, rather than high-impact weather events. A detailed description of the indices and their calculation is available at http://etccdi.pacificclimate.org/list_27_indices.shtml (accessed on 10 June 2025) or in the work of Karl et al. [14] and Peterson et al. [15].

The objectives of this study are threefold: (1) to assess the performance of CMIP6 models in reproducing observed rainfall extremes in Senegal; (2) to investigate spatial and temporal trends in key rainfall indices under historical and future climate scenarios; and (3) to identify potential hotspots of change to support decision-making for adaptation and disaster risk reduction. We hypothesize that CMIP6 models, once bias-corrected, can adequately reproduce historical rainfall patterns and provide credible projections of future extremes. By comparing near-future (2021–2040) and distant-future (2041–2060) periods under different Shared Socioeconomic Pathways (SSPs), our study aims to provide decision-makers with actionable insights for building resilience in Senegal’s water, agriculture, and health sectors.

2. Data and Methods

2.1. Study Area

This study focuses on Senegal, located between 12 and 17° N and between 18 and 11° W (Figure 1) in the Sahelian zone of West Africa. The climate is Sudano-Sahelian, with a pronounced dry season and a single rainy season driven by the West African Monsoon and northward movement of the Inter-Tropical Convergence Zone (ITCZ), typically spanning June to September [13].

Figure 1. Administrative boundaries of Senegal showing the 14 regions, overlaid with a locator map (inset) indicating the country’s location in West Africa. The map highlights key administrative divisions including Saint-Louis, Louga, Matam, Kaffrine, Kédougou, Tambacounda, Sedhiou, Kolda, Ziguinchor, and others. Latitude and longitude lines are provided for spatial reference.

Annual precipitation shows a strong north–south gradient, ranging from around 250 mm/year in the arid north to 1450–1500 mm/year in the southern and southwestern regions [16,17].

Temperatures average around 27–28 °C annually (1960–1990 baseline), with dry-season highs around 35 °C; seasonal fluctuations result in a temperature of 25–27 °C during the dry season and 27–29 °C during the rainy season [18].

Senegal is recognized as one of the “highest hotspots” for climate change in the Sahel, with significant vulnerability to drought, flood, coastal erosion, and heat waves [19]. Between 2001 and 2020, flood frequency doubled relative to 1980–2000, and major drought events (e.g., affecting some 320,000 people in 2018) are becoming more frequent [20].

2.2. Validation Data

The Enhancing National Climate Services (ENACTS) dataset was used to assess the performance of the NEX-GDDP-CMIP6 bias-corrected precipitation data. Developed by the International Research Institute for Climate and Society (IRI), ENACTS provides high-resolution, station–satellite merged daily climate data for African countries. It delivers robust products tailored for decision-making [12]. Details of the ENACTS product are presented in Table 1, and the ENACTS network is shown in Figure 2.

Table 1. Description of the ENACTS rainfall product used for validation. This table summarizes key characteristics including data sources, product type (combination of station and satellite data), spatial and temporal resolution, and the period of coverage. The dataset is developed by ANACIM as part of the ENACTS initiative.

Dataset	Full Product Name	Period	Spatial Resolution	Temporal Resolution	Reference
ENACTS (ANACIM)	Enhancing National Climate Services	1985–2013	0.0375° (~4 km)	Daily	(Dinku et al., 2022) [12]

Figure 2. Geographic coverage of the ENACTS data network across Africa. Countries are categorized by the scope of implementation: national level (orange), regional level (purple), and both national and regional levels (peach). The ENACTS initiative aims to improve climate data availability and use for decision-making by integrating quality-controlled station data with satellite and reanalysis products. Source: International Research Institute for Climate and Society (IRI), Columbia University (ENACTS|columbia.edu).

The ENACTS dataset used in this study integrates quality-controlled rainfall observations from national meteorological services and satellite estimates. The station density across Senegal varies between 1 station per 1500 km² and 1 station per 2500 km², with higher densities in populated and agricultural regions. Quality control includes automated checks for gross errors, outliers, and temporal consistency, as well as manual verification against station metadata and regional climatology following the ENACTS guidelines [12]. These steps ensure high reliability of the rainfall estimates, which are further bias-corrected against the station data to enhance their accuracy for regional climate analyses.

2.3. NEX-GDDP-CMIP6 Bias-Corrected Datasets

This study utilizes fifteen (15) statistically downscaled CMIP6 datasets from the NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) project [21] to assess projected rainfall events in Senegal (see Table 2). The NEX-GDDP dataset provides bias-corrected climate projections at a high spatial resolution (≈0.25° × 0.25°) and includes daily variables such as temperature (mean, maximum, minimum), precipitation, humidity, wind speed, and wind direction.

Table 2. NEX-GDDP-CMIP6 models, institutions, and countries of origin.

Model Name	Institution	Country	Spatial Resolution (Latitude × Longitude)
ACCESS-CM2	Australian Community Climate and Earth System Simulator (ACCESS)—CM Version 2	Australia	1.9° × 1.3°
ACCESS-ESM1-5	Australian Community Climate and Earth System Simulator—Earth System Model Version 1.5	Australia	1.9° × 1.3°
BCC-CSM2-MR	Beijing Climate Center (BCC), China Meteorological Administration (CMA)	China	1.9° × 1.3°
CanESM5	Canadian Centre for Climate Modelling and Analysis	Canada	2.81° × 2.81°
CMCC-CM2-SR5	Euro-Mediterranean Center on Climate Change (CMCC)	Italy	1.1 × 1.1
GISS-ES2-1-G	NASA Goddard Institute for Space Studies (GISS)	United States	2° × 2.5°
IITM-ESM	Indian Institute of Tropical Meteorology (IITM)	India	2.81° × 2.81°
MIROC6	Japan Agency for Marine-Earth Science and Technology (JAMSTEC)	Japan	1.4° × 1.4°
MIROC-ES2L	Japan Agency for Marine-Earth Science and Technology (JAMSTEC)	Japan	2.8° × 1.9°
MPI-ESM1-2-HR	Max Planck Institute for Meteorology	Germany	0.9° × 0.9°
MRI-ESM2	Meteorological Research Institute (MRI)	Japan	2° × 2.5°
NESM3	Nanjing University of Information Science and Technology (NUIST)	China
NorESM2-LM	Norwegian Meteorological Institute—Low-Resolution Model	Norway	1.9 ° × 1.9°
NorESM2-MM	Norwegian Meteorological Institute—Medium-Resolution Model	Norway	0.9° × 1.3°
TaiESM	Research Center for Environmental Changes, Academia Sinica	Taiwan	1.4° × 1.4°

The statistical downscaling technique employed is the Bias-Correction Spatial Disaggregation (BCSD) method, initially developed by Wood et al. (Wood et al., 2002, 2004) and adapted in NEX-GDDP-CMIP6. BCSD corrects systematic biases in raw GCM outputs and enhances spatial resolution by combining quantile mapping and spatial interpolation, thus preserving monthly trends and variability before disaggregating data to finer grids [22,23]. Specifically, BCSD adjusts model outputs via empirical cumulative distribution function correction on daily values, then spatially disaggregates using high-resolution observations (e.g., GMFD) to achieve a ~25 km resolution.

The NEX-GDDP dataset covers 1960–2100: a historical period (1960–2014) and future under four SSPs (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5), making it ideal for sectoral impact assessments in agriculture, hydrology, and energy.

To select the models presented in Table 2, we considered the availability of daily rainfall variables for the entire study period (1985–2100) and ensured that each model covered the historical baseline period and the future scenarios of interest (SSP2-4.5 and SSP5-8.5). The choice of models also took into account their frequent use in climate impact studies over West Africa, aligning with standard practices in regional climate modeling [24]. References such as NASA NEX-GDDP documentation [21], and studies on CMIP6 model evaluation in the Sahel region (e.g., [6,20,24]) guided our selection process to ensure relevance to West African climate assessments.

2.4. Methods

This study covers the period 1985–2100, where 1985–2013 represents the reference period and 2015–2100 the future period under two Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5). A special focus was given to two horizons: 2021–2040 (near future) and 2041–2060 (distant future), aligning with the Emerging Senegal Plan and the country’s Sustainable Development Goals (SDGs), with two SSPs (SSP2-4.5 and SSP5-8.5) representing moderate mitigation and worst-case socioeconomic pathways, respectively [21].

The reference period is used to evaluate the performance of the bias-corrected CMIP6 models against the ENACTS dataset. For the validation part, we focus solely on the precipitation variable and analyze each individual NEX-GDDP-CMIP6 model, as well as their multi-model ensemble mean (hereafter ENSMEAN) along the reference observational data. For future projections, we rely exclusively on the ENSMEAN to minimize uncertainties and enhance reliability [22]. Future extreme changes are analyzed by comparing the multiannual mean values for the periods 2021–2040 and 2041–2060 against the reference period (1985–2013). Due to data availability, particularly for the ENACTS dataset, we used 1985–2013 as the historical period for the validation phase.

This study examines projected changes in extreme precipitation events in Senegal for two future climate periods under two SSP scenarios. Nine extreme rainfall indices (Table 3) from the Expert Team on Climate Change Detection and Indices (ETCCDI) [14,15] are employed. These indices are computed from daily precipitation data during the monsoon season (June to September). A detailed summary of these indices is provided in Table 3.

Table 3. Indices calculated for the rainy season, from June to September.

Acronym	Index Name	Description	Unit
CDDs	Consecutive dry days	Number of consecutive days with rainfall < 1 mm	Days
CWDs	Consecutive wet days	Number of consecutive days with rainfall ≥ 1 mm	Days
RR1	Number of rainy days	Number of days with rainfall ≥ 1 mm	Days
SDII	Daily rainfall index	Average rainfall per wet day (≥1 mm) per year	mm/day
Rx5day	Max 5-day precipitation	Maximum total rainfall over 5 consecutive days	mm
R95pTOT	Very wet days	Total rainfall from days exceeding the 95th percentile	mm
R10	Heavy precipitation days	Annual count of days with rainfall ≥ 10 mm	Days
R20	Very heavy precipitation days	Annual count of days with rainfall ≥ 20 mm	Days
PRCPTOT	Total annual precipitation	Total rainfall on wet days (≥1 mm) during the season	mm

To evaluate the performance of the CMIP6 models, each individual NEX-GDDP-CMIP6 model was assessed against the ENACTS dataset for the historical period (1985–2013), focusing on the precipitation variable. This approach allowed us to examine the spread and variability across different models. For analyzing future projections of the nine rainfall indices (Table 3), we employed the multi-model ensemble mean (ENSMEAN). This choice was made to avoid presenting an overwhelming number of figures (i.e., 9 indices × 15 models) and to provide a clearer, more robust synthesis of the results. Using the ENSMEAN also helps mitigate individual model biases and enhances the overall reliability of the projections.

3. Results

3.1. Evaluation of CMIP6 Models Against Observations

Figure 3a depicts the annual rainfall cycle from ENACTS observation and each of the CMIP6 models for the reference period 1985–2013. We can see that both ENACTS data and all CMIP6 models were able to capture the annual cycle characterized by a single rainy season from May to October and peak rainfall in August. However, we notice that most of the models used tend to overestimate the annual amount of rainfall.

Figure 3. Comparison of ENACTS data with CMIP6 models for rainfall patterns over the period 1985–2013. (a) Annual cycle of rainfall (in mm), showing the monthly distribution and peak rainfall during the core rainy season. (b) Inter-annual variability of annual rainfall totals (in mm), illustrating the year-to-year fluctuations captured by ENACTS and various CMIP6 models. The analysis highlights differences in model performance in reproducing observed seasonal and inter-annual rainfall dynamics.

When comparing inter-annual variability (Figure 3b), it is notable that rainfall exhibits a sawtooth pattern over the historical period 1985–2013. This variability highlights years with higher rainfall amounts alternating with years of lower rainfall. This dynamic may have significant implications for ecosystems, agriculture, and other aspects dependent on rainfall regimes [25]. A substantial discrepancy between the ENACTS reference data and the CMIP6 models is observed, particularly regarding inter-annual variability, as illustrated in Figure 3b. This disparity can be attributed to various factors such as differences in the representation of climate processes and errors in modeling.

3.2. Spatial Representation of Seasonal Rainfall Patterns

Figure 4 presents the spatial distribution of rainfall during the historical period from 1985 to 2013, based on ENACTS data, individual NEX-GDDP-CMIP6 models, and the multi-model ensemble mean (ENSMEAN). By examining the different panels in the figure, one can assess the similarities and differences in the spatial distribution of rainfall among the climate models.

Figure 4. Mean June–September (JJAS) rainfall (in mm/month) over Senegal for the historical period 1985–2013. (a) shows observations from the ENACTS dataset, while panels (b–k) display simulations from 15 individual CMIP6 models.

A particularly noticeable feature on these maps is the pronounced latitudinal gradient in rainfall, with a clear decrease in isohyets from south to north. The southern regions, influenced by the summer monsoon and other meteorological systems, generally receive higher levels of rainfall. In contrast, the northern regions of West Africa are typically characterized by arid or semi-arid climates, with significantly lower rainfall levels. The central regions fall between these two extremes, receiving intermediate amounts of rainfall [26].

This in-depth analysis not only highlights the overall trends in the spatial distribution of rainfall but also enables a detailed comparison between observed data and projections from different climate models. This provides a solid foundation for understanding regional weather dynamics. It is important to note that the seasonal accumulation of precipitation is generally underestimated by the climate models in the north and overestimated in the south.

3.3. Statistical Assessment Using Taylor Diagram

Figure 5 presents a Taylor diagram comparing the bias-corrected outputs of 15 CMIP6 models with ENACTS climate reference data. The diagram simultaneously displays three precipitation statistics: centered root mean square error (RMSE), correlation coefficient (r), and normalized standard deviation (STD), providing a concise visual summary of model performance [27].

Figure 5. Taylor diagram summarizing the performance of 15 CMIP6 models in simulating rainfall over Senegal compared to ENACTS observations. The diagram displays the spatial correlation coefficient, normalized standard deviation, and centered root mean square error (RMSE) for each model. The black dot represents the multi-model ensemble mean (ENSMEAN), and each colored point corresponds to an individual model. Models closer to the reference point (1 on the x-axis, standard deviation = 1) exhibit better agreement with observations.

Along the correlation axis, most models show a correlation coefficient approaching 0.8, indicating a strong ability of the CMIP6 models to reproduce the reference dataset and supporting their use for future climate analyses. The concentric semicircles centered on the point (1,1) represent the RMSE, where closer proximity to this point denotes lower error and better agreement with observations. The normalized standard deviation axis highlights a spread among models, with values ranging from 1 to 1.25, reflecting challenges in accurately capturing rainfall variability over Senegal.

Figure 5 thus summarizes the relative performance of the 15 CMIP6 models in simulating rainfall compared to ENACTS observations. The black dot represents the multi-model ensemble mean (ENSMEAN), while each colored dot corresponds to an individual model. Models that lie closer to the reference point (correlation = 1; standard deviation = 1) exhibit stronger agreement with observed data.

Based on the Taylor diagrams presented in this study, the models that consistently exhibit the best performance—characterized by a high correlation (close to 0.9) and a standard deviation close to 1—are MPI-ESM1-2-HR, MIROC-ES2L, MRI-ESM2-0, CanESM5, and GISS-E2-1-G. These models consistently align closely with the ENACTS reference dataset, effectively capturing both the magnitude and variability of rainfall across Senegal. Their performance underscores their suitability for use in climate impact assessments in the region.

3.4. Historical and Projected Spatial Patterns of Rainfall Indices

Figure 6 illustrates the spatial distribution of key rainfall indices over Senegal for the historical period (1985–2013) during the JJAS season, providing a comprehensive overview of the baseline characteristics of rainfall extremes.

Figure 6. Spatial distribution of rainfall indices over Senegal for the historical period (1985–2013) during the JJAS (June–September) season. (a–h) represents a different rainfall-related index, including dry and wet spell durations, precipitation extremes, and total rainfall.

The CDD index (Figure 6a) highlights a pronounced north–south gradient, with longer dry spells in the northern regions (Saint-Louis, Louga, and Matam) and shorter dry periods in the southern regions (Ziguinchor, Kolda, and Sedhiou). This gradient reflects the influence of the West African Monsoon and the contrasting arid conditions in the Sahelian north.

The CWD index (Figure 6b) shows an inverse spatial pattern compared to the CDD index, with higher wet spell frequencies in the southern regions and lower frequencies in the north. This pattern aligns with the more humid conditions in southern Senegal and the transitional rainfall zone in the central regions.

The RR1 index (Figure 6c) confirms this latitudinal gradient, with the highest number of rainy days in the southern regions, indicating the south’s dominance in seasonal precipitation.

The SDII index (Figure 6d) exhibits higher daily rainfall intensities in the southern regions and lower intensities in the north, reinforcing the south’s susceptibility to more intense rainfall events.

The Rx5day (Figure 6e) and R95pTOT (Figure 6f) indices both display high values in the southern and central regions, indicating areas prone to short duration but intense rainfall events that may cause localized flooding.

The R10mm (Figure 6g) and R20mm (Figure 6h) indices also reflect this north–south contrast, with higher frequencies in the southern regions and lower frequencies in the north. This further emphasizes the south’s vulnerability to intense rainfall events, with potential implications for agriculture and water management.

Collectively, these indices demonstrate a clear latitudinal gradient in rainfall extremes across Senegal, highlighting the interplay between the West African Monsoon and the regional climate systems. These baseline conditions provide essential context for interpreting future changes in rainfall extremes and assessing the potential impacts of climate change on water resources, agriculture, and socio economic activities across the country.

To investigate the temporal evolution of rainfall extremes across Senegal, spatial trends in precipitation intensity and frequency indices were assessed for the near-future (2021–2040) and distant-future (2041–2060) periods under the moderate-emission scenario SSP245. Figure 7 and Figure 8 present the spatial distribution of these indices during the JJAS season, providing critical insights into potential shifts in rainfall patterns across the country.

Figure 7. Projected spatial distribution of rainfall indices over Senegal for the near-future period (2021–2040) during the JJAS (June–September) season under the CMIP6 SSP245 scenario. (a–h) The panels represent different rainfall indices, including dry and wet spell duration, precipitation extremes, and total rainfall.

Figure 8. Projected spatial distribution of rainfall indices over Senegal for the near-future period (2041–2060) during the JJAS (June–September) season under the CMIP6 SSP245 scenario. (a–h) The panels represent different rainfall indices, including dry and wet spell duration, precipitation extremes, and total rainfall.

In the near-future period (2021–2040), substantial spatial heterogeneity emerges across multiple indices. As depicted in Figure 7, the consecutive dry day (CDD) index (Figure 7a) projects an intensification of dry spells across the northern regions, with values exceeding 15 days, while southwestern areas exhibit shorter dry spells. Conversely, the consecutive wet day (CWD) index (Figure 7b) indicates higher wet spell frequencies in the south and a marked decline in the north, reinforcing the latitudinal gradient of precipitation regimes.

The frequency of rainy days (RR1; Figure 7c) remains highest in the southern and southeastern regions, highlighting the concentration of rainfall events in these zones. The daily rainfall intensity (SDII; Figure 7d) shows elevated values in the southern regions (exceeding 11 mm/day) and lower values in the north, indicating persistent spatial disparities in rainfall intensities. The maximum 5-day precipitation accumulation (Rx5day; Figure 7e) and the very wet day index (R95pTOT; Figure 7f) both exhibit higher intensities in the southern regions, underscoring the potential for extreme rainfall events.

The frequency of heavy rainfall events, represented by R10mm (Figure 7g) and R20mm (Figure 7h), is also projected to remain higher in the south, indicating a heightened likelihood of extreme rainfall episodes that could exacerbate flood risks. Collectively, these projections for the near future suggest a continuation of the historical north–south gradient in rainfall extremes, with the south remaining wetter and more vulnerable to high-intensity rainfall events.

For the distant-future period (2041–2060), Figure 8 reveals a further intensification of these spatial patterns. The CDD index (Figure 8a) projects persistent and extended dry spells in the northern regions, with values surpassing 18 days, while southwestern regions continue to exhibit shorter dry periods. The CWD index (Figure 8b) maintains higher wet spell frequencies in the south, though with an overall decline in the central and northern regions.

The RR1 index (Figure 8c) projects consistently high frequencies of rainy days in the southern and southeastern regions, while the SDII index (Figure 8d) shows elevated daily rainfall intensities persisting in the southern zones. The Rx5day (Figure 8e) and R95pTOT (Figure 8f) indices indicate substantial intensification of extreme rainfall events across the western and central regions, highlighting an increased risk of flooding.

Additionally, the R10mm and R20mm indices (Figure 8g,h) continue to show higher frequencies of heavy rainfall days in the south, reinforcing the vulnerability of these areas to intense precipitation events under future climate scenarios.

Overall, the SSP245 scenario projections for both the near- and distant-future periods indicate a continued north–south gradient in rainfall extremes, with the north characterized by increasing dry spells and the south by persistent and intensifying extreme rainfall events.

Figure 9 presents the projected percentage changes in rainfall indices for the near-future period (2021–2040). The CDD index (Figure 9a) indicates a general reduction in dry days in the central and southern regions, with percentage changes reaching up to –40%, suggesting shorter dry spells. In contrast, modest increases are observed in some localized northern areas. The CWD index (Figure 9b) shows a mixed pattern, with slight increases in the north and decreases in parts of the south and center, implying shifts in the duration of wet spells that may impact seasonal rainfall regimes.

Figure 9. Spatial distribution of changes (%) in rainfall indices over Senegal for the near-future period (2021–2040) relative to the historical baseline (1995–2014), under the CMIP6 SSP245 scenario during the JJAS season. Panels illustrate projected changes in (a) CDDs, (b) CWDs, (c) RR1, (d) SDII, (e) Rx5day, (f) R95pTOT, (g) R10mm, and (h) R20mm.

The RR1 index (Figure 9c) projects substantial increases in the number of rainy days, particularly in the north where changes exceed +60%. This suggests a northward extension of wet day frequencies, consistent with a possible northward shift of the monsoon. The SDII index (Figure 9d) shows modest increases (up to +20%) in daily rainfall intensity in the north and northwest, indicating that rainfall events may become more intense in these regions.

The Rx5day index (Figure 9e) shows a mixed but mostly positive signal, with increases up to +40% in parts of the west and northwest, suggesting an increased risk of heavy rainfall events over short durations that could lead to flash floods. The R95pTOT index (Figure 9f) also shows increases in the north and center (up to +30%), highlighting a growing contribution of very wet days to total rainfall.

The R10mm (Figure 9g) and R20mm (Figure 9h) indices both exhibit increases across most regions, particularly in the north and center. R20mm shows the most pronounced increase in the northwest, with changes exceeding +100%, indicating that heavy rainfall events may become significantly more frequent in these areas.

Figure 10 extends the analysis to the distant-future period (2041–2060). The CDD index (Figure 10a) shows a continuation of the declining trend in dry days across much of the south and central regions (up to –40%), while some areas in the north show modest increases, reflecting an intensification of the north–south gradient in dry spells. The CWD index (Figure 10b) suggests further reductions in the south and center, indicating that wet spells may become shorter in the southern regions.

Figure 10. Spatial distribution of changes (%) in rainfall indices over Senegal for the near-future period (2041–2060) relative to the historical baseline (1995–2014), under the CMIP6 SSP245 scenario during the JJAS season. Panels illustrate projected changes in (a) CDDs, (b) CWDs, (c) RR1, (d) SDII, (e) Rx5day, (f) R95pTOT, (g) R10mm, and (h) R20mm.

The RR1 index (Figure 10c) reveals continued increases in the number of rainy days in the northern regions (up to +60%), suggesting a persistent northward shift in rainfall frequency. The SDII index (Figure 10d) shows consistent increases in daily rainfall intensity across much of the country, especially in the north and northwest, with changes reaching +20%.

The Rx5day index (Figure 10e) projects significant increases in the western and central regions (up to +40%), suggesting a continued trend towards more extreme rainfall events. The R95pTOT index (Figure 10f) shows widespread positive changes (up to +40%) across much of Senegal, reinforcing the role of very wet days in shaping seasonal rainfall totals.

The R10mm (Figure 10g) and R20mm (Figure 10h) indices show consistent increases, particularly in the north and northwest. R20mm shows the strongest signals in the northwest and center, with changes exceeding +100%, indicating a substantial intensification of heavy rainfall events.

Overall, these projections indicate that under SSP245, Senegal will likely experience an increase in the frequency and intensity of rainfall extremes, especially in the northern regions.

3.5. Temporal Trends in Rainfall Indices Under SSP585 Scenario

To assess the temporal evolution of rainfall extremes across Senegal, spatial trends of precipitation intensity and frequency indices were computed for two future periods—2021–2040 (near future) and 2041–2060 (distant future)—under the high-emission SSP585 scenario. These trends, depicted in Figure 9 and Figure 10, provide critical insights into the projected shifts in rainfall characteristics on the national scale.

In Figure 11 related to the near-future period (2021–2040), several indices exhibit significant changes in their spatial patterns. As presented in Figure 9, the maximum number of consecutive wet days (CWDs; Figure 9b), mean daily rainfall intensity (SDII; Figure 9d), maximum 5-day precipitation accumulation (Rx5day; Figure 9e), and the frequency of heavy rainfall events (R10mm and R20mm; Figure 9g,h) demonstrate decreasing trends across large portions of northern Senegal. For example, SDII is projected to decline by approximately 15%, while CWDs may decrease by up to 50% in certain northern areas.

Figure 11. Projected spatial distribution of rainfall indices over Senegal for the near-future period (2021–2040) during the JJAS (June–September) season under the CMIP6 SSP585 scenario. (a–h) The panels represent different rainfall indices, including dry and wet spell duration, precipitation extremes, and total rainfall.

This paradoxical outcome—where total seasonal rainfall may remain stable or even increase despite a reduction in rainy days—underscores a key dynamic of climate change: the simultaneous intensification of dry spells and extreme rainfall events. The analysis of the consecutive dry day (CDD) index (Figure 9a) reveals that, while dry days are projected to decrease by up to 30% in the north, certain southwestern areas may experience increases of up to 30%, with the notable exception of the Kédougou, Kolda, Sedhiou, and Ziguinchor regions. These regional variations hold significant implications for water resource management and the planning of agricultural activities. Furthermore, the total number of rainy days (RR1; Figure 9c) is projected to decline across most of the country, including parts of the south, whereas some northern regions may experience increases of up to 40%.

In Figure 12, projections for the distant-future period (2041–2060) reveal the emergence of new spatial dynamics. As depicted in Figure 10, declining trends in CWDs, RR1, SDII, and R10mm are projected to intensify in the southern and central regions, indicating shorter wet spells and fewer rainy days (Figure 12a–h). Conversely, increases in these indices are projected in the northern regions, suggesting a northward shift in rainfall regimes under the high-emission scenario. Notably, the Rx5day, R95pTOT, and R20mm indices exhibit widespread positive trends, particularly across the western and central regions, implying that very intense rainfall events may become more frequent, even as average daily or seasonal rainfall totals decline. The R20mm index, in particular, shows marked increases across most of the country, except along the western coastal zone.

Figure 12. Projected spatial distribution of rainfall indices over Senegal for the near-future period (2041–2060) during the JJAS (June–September) season under the CMIP6 SSP585 scenario. (a–h) The panels represent different rainfall indices, including dry and wet spell duration, precipitation extremes, and total rainfall.

These contrasting patterns suggest a potential intensification of hydroclimatic extremes, characterized by more erratic rainfall distributions, extended dry periods, and a higher incidence of concentrated heavy rainfall events. Such changes could pose significant challenges to agriculture, water resource management, and the resilience of local communities and ecosystems.

In summary, the analysis of CMIP6 multi-model ensemble projections under the SSP585 scenario indicates an overarching drying trend across Senegal during the JJAS season in the coming decades. The distant-future period (2041–2060) is particularly concerning, as it exhibits the most pronounced reductions in seasonal rainfall accompanied by the greatest increases in the frequency and intensity of extreme precipitation events.

Figure 13 presents the projected percentage changes in rainfall indices for the near-future period (2021–2040) relative to the historical baseline (1995–2014) under the high-emission scenario SSP585. Figure 13a shows significant increases in the CDD index across the southern and southeastern regions, with values exceeding +60% in some areas, indicating an intensification of dry spells. In contrast, slight decreases in dry days are projected in localized areas in the northwest.

Figure 13. Spatial distribution of changes (%) in rainfall indices over Senegal for the near- future period (2021–2040) relative to the historical baseline (1995–2014), under the CMIP6 SSP585 scenario during the JJAS season. Panels illustrate projected changes in (a) CDDs, (b) CWDs, (c) RR1, (d) SDII, (e) Rx5day, (f) R95pTOT, (g) R10mm, and (h) R20mm.

Figure 13b displays a general decrease in the CWD index in the southern and central parts of Senegal, with declines reaching up to -40%, indicating shorter wet spell durations. However, localized increases are observed in some northern regions, reflecting potential regional variability. Figure 13c highlights substantial increases in the RR1 index in the northern and central regions, with changes reaching +60%, suggesting a northward shift in the frequency of rainy days, consistent with a potential shift in the West African M onsoon system.

Figure 13d reveals modest increases in the SDII index in the north and northwest, with changes reaching +20%, indicating a potential intensification of daily rainfall events in these areas. Figure 13e,f show moderate to significant increases in extreme rainfall events, with Rx5day showing increases up to +40% in the northwest and central regions, and R95pTOT displaying increases up to +30% in similar areas.

Figure 13g,h depict changes in the frequency of heavy rainfall events (R10mm and R20mm). These indices show relatively weaker and more variable signals, with some localized increases in the north and center but generally a neutral to slightly positive trend across most regions.

Figure 14 extends this analysis to the distant-future period (2041–2060). Figure 14a shows a generalized increase in dry days in the southern and southeastern regions, with values reaching +60% to +80% in some areas, indicating a further intensification of dry spells. Some reductions in dry days are observed in the northwest.

Figure 14. Spatial distribution of changes (%) in rainfall indices over Senegal for the distant- future period (2041–2060) relative to the historical baseline (1995–2014), under the CMIP6 SSP585 scenario during the JJAS season. Panels illustrate projected changes in (a) CDDs, (b) CWDs, (c) RR1, (d) SDII, (e) Rx5day, (f) R95pTOT, (g) R10mm, and (h) R20mm.

Figure 14b displays pronounced decreases in the CWD index across the south and center, with values reaching -60%, indicating shorter wet spell durations and highlighting a potential reduction in continuous rainy periods. Figure 14c shows sustained increases in the RR1 index in the northern regions, reinforcing the trend of more frequent rainfall events in traditionally drier areas.

Figure 14d presents consistent increases in the SDII index across the north and northwest, indicating continued potential for more intense rainfall events. Figure 14e,f project substantial increases in Rx5day and R95pTOT across the western and central regions (up to +40%), underscoring the increased risk of short-duration, intense rainfall events that could exacerbate flood hazards.

Finally, Figure 14g,h show continued trends of increasing heavy rainfall events, with R10mm and R20mm showing significant increases, particularly in the north and northwest. The R20mm index (Figure 14h) shows the most pronounced increase, exceeding +100% in some areas, highlighting the potential for more frequent heavy rainfall events.

Overall, Figure 13 and Figure 14 under the SSP585 scenario point to an intensification of rainfall extremes in Senegal, particularly in the northern regions. This shift suggests a potential reconfiguration of rainfall regimes, with implications for water resource management, agriculture, and disaster risk management.

3.6. Temporal Trends in Rainfall Index Anomalies

Figure 15 provides a comprehensive overview of the temporal evolution of rainfall indices anomalies over Senegal for the future projections (2015–2100) under the SSP585 scenario. Overall, the time series indicates a pronounced declining trend across most indices, suggesting a systematic shift in rainfall characteristics throughout the century.

Figure 15. Time series of anomalies (%) in rainfall indices over Senegal from 1995 to 2100, showing both historical values (in blue) and future projections under the SSP585 scenario (in red). The indices displayed are (a) PRCPTOT, (b) SDII, (c) RR1, (d) Rx5day, (e) R95p, (f) R99p, (g) R10mm, (h) R20mm, (i) CWDs, and (j) CDDs. Anomalies are computed relative to the 1995–2014 baseline.

The PRCPTOT index (Figure 15a) exhibits a persistent decline in total rainfall, with an average decreasing trend of approximately −0.23% per year, indicating a progressive reduction in annual precipitation totals over time. This downward trend is echoed in the SDII index (Figure 15b), which also shows a consistent negative trend of about −0.20% per year, suggesting a decrease in daily rainfall intensity that could imply less frequent intense rainfall events.

The RR1 index (Figure 15c), representing the number of wet days, displays a marked downward trajectory with an approximate trend of −0.43% per year. This result aligns with the reduction in wet spells and an increase in dry conditions, as also reflected in the declining trends of the CWD index (Figure 15i) at −0.09% per year and the increasing trend in the CDD index (Figure 15j) at +0.26% per year. These opposing signals highlight the potential intensification of drought conditions in the region.

Extreme rainfall events also exhibit a declining tendency. The Rx5day index (Figure 15d) shows a negative trend of approximately −0.23% per year, indicating a decline in the most intense 5-day rainfall events. Similarly, the R95p (Figure 15e) and R99p (Figure 15f) indices reveal decreasing trends (−0.46% and −0.61% per year, respectively), underscoring a potential reduction in the contribution of very wet and extremely wet days to the seasonal rainfall totals.

The R10mm and R20mm indices (Figure 15g,h) both show declining trends (−0.80% and −1.13% per year, respectively), reinforcing the notion of a decline in the frequency of moderate and heavy rainfall events under the SSP585 scenario.

Collectively, the projected trends presented in Figure 15 suggest a systematic drying of Senegal’s rainfall regime, characterized by declining total precipitation, fewer wet days, shorter wet spells, and a reduction in rainfall intensity and extremes. This evolving climate pattern underscores the increasing risks of drought and its associated impacts on agriculture, water resources, and socioeconomic systems, emphasizing the need for robust adaptation strategies tailored to the regional context.

3.7. Uncertainties in Rainfall Modeling in West Africa

This section addresses the uncertainties associated with climate model projections of rainfall in Senegal. Using outputs from CMIP6 models, this study analyzes inter-annual rainfall trends under the high-emission socioeconomic pathway SSP585, providing insight into the range and direction of potential future climate scenarios. Two projection horizons are considered: the near future (2021–2040) and the distant future (2041–2060).

Figure 16 illustrates the inter-annual trends in annual rainfall over Senegal under the CMIP6 SSP585 scenario, capturing both the near future (2021–2040) and the distant future (2041–2060).

Figure 16. Inter-annual trends in annual rainfall over Senegal under the CMIP6 SSP585 scenario for (a) the near future (2021–2040) and (b) the distant future (2041–2060). Each colored line represents a linear trend from an individual climate model, with dashed lines indicating trends that are not statistically significant (p ≥ 0.05) and solid lines indicating significant trends (p < 0.05) based on a two-sided t-test. In Figure (a), 40% of the models project an increase in rainfall, while 60% show a decrease. In Figure (b), only 13% of models project an increase, while 87% project a decrease, highlighting a broad consensus toward drying conditions in future projections.

In Figure 16a, corresponding to the near-future period, individual climate models show substantial inter-model variability in projected rainfall trends. Approximately 40% of the models exhibit positive trends in annual rainfall, suggesting potential increases in precipitation in some scenarios. However, the majority (60%) project negative trends, with some models showing marked declines in annual rainfall amounts. The predominance of negative trends implies that although some uncertainty remains, there is an emerging tendency toward drier conditions during this period.

In Figure 16b, representing the distant-future period, the ensemble consensus shifts more decisively toward declining annual rainfall. Here, 87% of the models project negative trends, reinforcing the signal of drying conditions. Only 13% of the models project increases in annual rainfall. The consistency and strength of these downward trends suggest a more robust signal of drying across Senegal in the second half of the century under the SSP585 scenario.

The dashed lines in both panels denote trends that are not statistically significant (p ≥ 0.05), while the solid lines represent statistically significant trends (p < 0.05) based on a two-sided t-test. Notably, the increased prevalence of negative trends in the distant future indicates a growing agreement among models that future climate conditions are likely to favor a reduction in annual rainfall, with important implications for water resource management, agriculture, and climate adaptation planning in the region.

Overall, Figure 16 highlights the evolution of model consensus on rainfall trends in Senegal, revealing a transition from mixed signals in the near future to a clearer indication of drying conditions in the distant future under a high-emissions scenario. This finding underscores the need for proactive adaptation strategies to manage the potential impacts of declining rainfall on livelihoods and ecosystems in Senegal.

The differences observed between CMIP6 models arise primarily from variations in spatial resolution, which affect the representation of local-scale convective processes and topographic features critical to rainfall generation in West Africa. Additionally, discrepancies stem from differing parameterization schemes, especially those governing cloud microphysics and land–atmosphere interactions, which influence rainfall intensity and distribution. These factors contribute to model-specific sensitivities and uncertainties, reinforcing the need for ensemble approaches and downscaling techniques to enhance the reliability of rainfall projections in Senegal.

4. Discussion

The spatial and temporal analysis of rainfall indices conducted in this study highlights marked heterogeneity in rainfall patterns across Senegal, with substantial regional disparities. All indices—except for the maximum number of consecutive dry days (CDDs)—consistently show higher values in the southern part of the country, followed by the central and then the northern regions. These results align with previous research carried out in Senegal by [28], confirming the well-known south-to-north rainfall gradient. This gradient is primarily driven by the influence of the West African Monsoon, which enhances rainfall in the south, while the north is more heavily influenced by arid atmospheric circulation regimes.

Projections for the distant-future period (2041–2060) under the SSP585 scenario indicate a likely intensification of drought conditions in the north, coupled with an increase in extreme rainfall events in the south. These findings support earlier conclusions by [25], who projected a long-term reduction in rainfall linked to a weakening of rainfall intensity. This trend is particularly evident in the projected decline in the number of very wet days. At the same time, several studies including [29] have documented a rising frequency of intense rainfall events and a reduction in the duration of wet spells across the Sahel region, further confirming the trend towards more erratic and extreme rainfall.

For the near-future period (2021–2040), however, model projections are more divergent. Some indices suggest an increase in wet conditions in the north—reflected in fewer dry days (CDDs) and more frequent rainy days (CWDs, RR1)—while a simultaneous decline in these indicators is projected for the southern regions. This spatial and temporal variability is consistent with the findings of [30], who reported global shifts in rainfall patterns, including an increase in the frequency of heavy rainfall events and a concurrent decline in the duration of prolonged dry spells.

Importantly, the observed changes in rainfall indices appear to follow a latitudinal progression, with early signals emerging in the northern (Sahelian) zone, gradually extending to the central (Sudano-Sahelian) region, and finally affecting the southern (Sudanese) parts of Senegal. This pattern is corroborated by the work of [7], who identified a similar spatial evolution of climate anomalies across the Sahel-Sudanese transect [30].

The relationships between rainfall variability in West Africa, inter- and intra-annual oscillations, and regional atmospheric circulation are intrinsically linked. Inter-annual variability is strongly influenced by large-scale climate drivers such as the El Niño–Southern Oscillation (ENSO) [31], the Atlantic Niño [32], and the Madden–Julian Oscillation (MJO) [33]. These drivers modulate the strength and position of the West African Monsoon (WAM) system, which, in turn, shapes the onset, duration, and intensity of the rainy season across Senegal. Understanding these atmospheric dynamics is essential for interpreting the complex patterns of wet and dry spells in the region and for improving the accuracy of seasonal forecasts and climate projections.

Despite these insights, the analysis also reveals substantial uncertainties among CMIP6 models, particularly in their projections of inter-annual rainfall trends. The discrepancies between models highlight the influence of model-specific sensitivities, differences in the representation of physical processes, and the complex interactions of regional climate dynamics. These uncertainties underscore the need for cautious interpretation of the results and emphasize the importance of using ensemble means and multiple scenarios in climate impact assessments.

Further research is essential to improving our understanding of the drivers of rainfall variability in West Africa. This includes advancing knowledge of inter-annual and intra-annual oscillations, regional atmospheric circulation patterns, and their representation in CMIP6 models. As recommended by [34], future work should aim to enhance the representation of local climatic processes and feedback mechanisms to reduce uncertainties in model projections. Additionally, investigating the sources of bias in CMIP6 and comparing them with CMIP5—as highlighted by [35,36]—could help refine future climate projections and strengthen the robustness of scenario-based planning for the region.

5. Conclusions

This study assessed historical and projected rainfall variability across Senegal using nine standardized rainfall indices from the Expert Team on Climate Change Detection and Indices (ETCCDI). These indices capture rainfall intensity (PRCPTOT, R95pTOT, Rx5day, SDII, RR1), frequency (R10mm, R20mm), and duration (CDDs, CWDs), offering insights into climate-related risks to water resources and agriculture.

The analysis reveals significant spatial and temporal contrasts in rainfall dynamics. Northern Senegal shows a trend toward drier conditions in the near-future period (2021–2040) with a 3% increase in dry days and a 5% decrease in extreme rainfall (Rx5day). However, in the distant future (2041–2060), projections suggest a reversal, with a 5% decrease in dry days and a 15% increase in Rx5day, pointing to more intense but potentially less frequent rainfall.

In southern Senegal, near-future projections indicate a 13% increase in dry days, a marginal rise in CWDs (1%), and a 3% decrease in SDII, suggesting a drying trend and reduced daily rainfall intensity. By contrast, distant-future projections highlight a 24% increase in rainy days, a 6% decrease in CWDs, and a substantial 22% rise in Rx5day, in dicating increased intensity of extreme rainfall events.

Overall, these findings underscore a north–south gradient of increasing climate extremes, with potential implications for water availability, agriculture, and disaster risk management. The CMIP6 models provide useful guidance but exhibit uncertainties, particularly in simulating regional rainfall, highlighting the need for ensemble ap proaches, improved downscaling, and integration of local knowledge.

Future work should also consider temperature extremes, as they are critical for agriculture and public health, and expand the range of climate models and scenarios to enhance projection robustness.

The results of this study can inform the design of evidence-based adaptation strategies that account for regional rainfall variability and future climate risks. Key steps include integrating rainfall projections into agricultural planning to enhance crop resilience, supporting water resource management strategies that reflect changes in rainfall extremes, and aligning adaptation measures with national policy frameworks such as the Emerging Senegal Plan and the SDGs. Collaborative efforts with local communities, policymakers, and development partners are essential to ensuring that adaptation strategies are context-specific, feasible, and sustainable.