The spatial extent of Thiès region is between 10° 44′ 46″and 10° 52′ 46″north latitude and 78° 39′ 11″and 78° 44′ 13″west longitude. Thiès region was once an agricultural country where agriculture, especially groundnuts and vegetables, became essential to Senegal’s economy. Regarding land area, it is one of the smallest regions in Senegal, at about 6669.6 km² or 3.35% of the total area of Senegal. As shown in Figure 1-b, it is bounded to the North by the Louga region, south by the Fatick region, west by the Atlantic Ocean and Dakar region (the capital of Senegal), and to the east by the Diourbel and Fatick regions. The Thiès region had 2,162,831 inhabitants in 2020, according to ANSD.

From the perspective of the agricultural situation, the main crop types are peanut, maize, millet, sorghum, and cowpea. According to the agricultural data collected in ANSD (accessed on 22 October 2022, at https://senegal.opendataforafrica.org/gallery?tag=DAPSA), the sown land area of these main crop types listed above represented about 266,668.24 hectors in 2020. In the same period, the agricultural production of these crops was around 25,3784.08 tons. From the point of view of spatial land use morphology, the remote sensing image analysis in 2020 shows that agricultural land use (48.4%) and grassland land (18.4%) represented the most significant land area dominant morphology, accounting for 66.8% of the total. Similarly, artificial surfaces represented 3.5% of the region’s land area. Then, in this region, the topography is flat except for the “Plateau of Thiès,” which culminates at 105 m of altitude. The temperatures are generally high, and the annual temperature cycle is complex. The maximum temperature is 33.2°. In addition, the interannual evolution of rainfall shows that the average rainfall was about 461.65 mm from 2000 to 2020, according to the data collected by the National Agency of Civil Aviation and Meteorology (ANACIM). The soil’s characteristics were ferruginous tropical sandy soils that are slightly leached (Tappan et al., 2004).

The results presented in this section differed from one period to another (Table 3; Figure 4). Between 2000 and 2005, agricultural land use decreased by about −303.01 km². During this period, forested areas (168.55 km²) were the most common land use type, resulting in the loss of agricultural land. Also, artificial surfaces accounted for 40.22 km² of this lost value. The situation remained similar in the second period (2005–2010) due to agricultural land losses of approximately −273.29 km²; grassland was the primary contributor to the total loss. Compared with the two first periods, the agricultural land transition diminished from 2010 to 2015. The net transition of agricultural land was −110.05 km² or a difference of 193.05 km² compared to 2000–2005. Hence, the agricultural land net transition was intense from 2000 to 2005. Similarly, 2015–2020 shows that agricultural land was gained among the other land use types, with 119.56 km². As shown in Table 3, except for one period, this study’s inter-period results revealed that the net transition of agricultural land was negative, whereas grassland gained the mean important flow.

The study period may produce noticeable results regarding agricultural land transition. This analysis shows the total net agricultural land transition was about −566.80 km². With 315.90 km², grassland represented the most significant flow. With 167.51 km², bare land came in second place. Artificial surfaces are the most common land use type that may threaten agricultural land, accounting for approximately 148.95 km². In summary, over the past 21 years, the most considerable change has been the substantial transition of agricultural land with a −1.667% temporal index, of which grassland has represented the most critical land use that caused this transition (Table 4).

The spatial analyses of the net transition of agricultural land at an interval level, as shown in Figure 5, revealed several aspects. In fact, with a variation coefficient of −0.71%, agricultural land dropped in all the communes in the first period (2000–2005). The highest negative value of net conversion was recorded in the Mbayenne commune (−0.65%) and the least in the Noto G. Diama commune (−12.35%). In these communes, bare land gained about 6.14 km² in Noto G. Diama and 0.20 km² in Mbayenne. The coefficient variation was - 4.68 from 2005 to 2010, with Thieneba at −24.22% and Keur Moussa commune at 40.06% of the net transition of agricultural land. From 2005 to 2010, the average net transition of agricultural land was around −1.26 km², higher than the first period, which recorded −4.51 km². Then, the net transition of agricultural land from 2010 to 2015 was similar to the above description. This period records a coefficient of variation of about −10.23 with an average net transition of −1.73 km². These cases were −39.81% in the commune of Mont-Rolland and 37.20% in the commune of Koul. In Koul, bare land contributes approximately 41.03 km²; in Mont Rolland commune, forested areas dominate 81.06 km² for this transition. From 2015 to 2020, the net transition was 43.09% in Noto G. Diama, compared to −34.73% in Sandiara. The coefficient of variation was 25.03 from 2015 to 2020.

During the study period (2000–2020), the coefficient of variation was about −2.22 (Figure 6). The spatial distribution of the net transition of agricultural land was in Mont Rolland commune, at approximately 33.22%. In this amount, forested areas lost about −43.99 km², which appears to have been the more common land use type during the agricultural land transition. In the Sandiara commune, the spatial distribution was −41.73%. Bare land was the most critical land use category, contributing to the loss of agricultural land at approximately 55.52 km². With an average of −7.56%, the agricultural land transition remained unequally distributed in the Thiès region from 2000 to 2020. In sum, Figure 5B highlights the relatively substantial dispersion. In other words, they are not significantly correlated regarding agricultural land transitions between the communes. The R² coefficient was 0.3964 from 2000 to 2020. Consequently, it appears that the potential driving factors may be different.

The temporal distribution of agricultural land net transition was nearly identical for all communes between 2000 and 2005, with an average of −15.53% (Figure 7). The temporal evolution of the net transition of agricultural land was −13.55% in the Ndiass commune and −16.69% in the Darou Khoudouss commune. From 2005 to 2010, Touba Toul commune recorded the highest temporal index of 303.62%. Or, compared with the average period (26.55%), the lowest speed was localized in Marouane commune, approximately −14.17%. From 2010 to 2015, the average temporal index was around 15.95%. Hence, this average has many characteristics because about 140.09% of net conversion is noted in the Koul commune, compared to the lowest of −15.84% in the Mont-Rolland commune. In contrast to the previous period, the speed of 2015–2020 was high, with an average of 17.59%. The highest speed was related to Noto G. Diam (188.17%) and the lowest in Nguiene commune, at −15.22%. In short, the inter-period analysis of the time distribution of the agricultural land net transition showed that the speed varied from commune to commune.

The study period was better analyzed to understand the characteristics of agricultural land’s net transition. The average speed from 2000 to 2020 was approximately 0.072%. This average hid several disparities within the commune. Twenty of the thirty-one communes investigated presented a negative speed. This negative value was located in Koul (−0.35%). Or, at −4.33%, the Sandiara commune had the slowest speed. The Mont Rolland commune (24.84%) recorded the most critical speed among the communes, recording a gain. Or, the commune of Pekess presented less speed by 0.18%. In this amount, Mont-Rolland recorded a net transition of agricultural land of 56.00 km² and Sandiara of −60.41 km². As a result of this analysis, despite the study being conducted in the same area, the spatial and temporal repartition of the net transition of agricultural land in the Thiès region displayed several characteristics that could be attributed to various economic and social factors.

This section examines the intensity transition on agricultural land. In fact, from 2000 to 2005, the rate of artificial surface conversion was 2.50% in the Noto G. Diama commune. The conversion of agricultural land to artificial surfaces covered approximately 4,006 km² in this commune. Similarly, Cherif Lo (1.81%) and Diender Guedj (1.91%) achieved high results. The proportion of agricultural land in this net transition of artificial surfaces was 2.63 km² and 2.45 km², respectively. Alternatively, the net transition of artificial surfaces in Ngueniene commune was −0.28%, and the cause of the decrease in artificial surfaces was wetland by 1.26 km². Between 2005 and 2010, the Malicounda commune experienced an essential transition of 3.65%. Fandene commune occupied second place with 2.21%. Malicounda lost approximately 6.45 km² of agricultural land to the profile of construction land, while Fandene lost about 1.55 km². We noted that eight communes recorded a negative value during this period—for instance, Mont-Rolland commune −0.055%—and bare land contributed to this loss.

As shown in Figure 8, from 2010 to 2015, the average intensity was 0.9%. The highest intensity of artificial surfaces was localized in Ngoudiane commune at 9.57%. This percentage was the highest for all the previous periods. Agricultural land (2.46 km²) and bare land (3.5 km²) represented the land use types that contributed to the significant value of construction land. The intensity was lowest in Taiba Ndiaye commune (−19%), whose bare land (3.05 km²) has caused this phenomenon. Between 2015 and 2020, the results in Malicounda (5.51%) showed diminished intensity compared to the previous period. In this commune, agricultural land losses add about 3.38 km² to the profile of construction land. Then, the Sindia commune occupied the second place at approximately 4.52%, and the loss of agricultural land to the artificial surfaces profile was more critical than in the Malicounda commune, with 5.97 km². Globally, the inter-period results show that the artificial surface intensity is moderately high. Even though the causes and effects of this change are different, we thought that the amount of agricultural land turned into artificial surfaces was significant.

The study period was a comprehensive pivot to understanding the global intensity of the net transition of construction land. The Malicounda commune records the highest value at 11.34%. In this commune, agricultural land loss to artificial surfaces was about 20.20 km². This loss from agricultural land to artificial surfaces was approximately 8.79 km² in Fandene and 11.99 km² in Sindia commune, with an intensity of 9.82% and 6.93%, respectively. Or, the Thilmakha commune (0.1%) records the lower net transition, and the loss of agricultural land to artificial surfaces was about 0.296 km². Globally, 31 communes were registered, with 15 having a percentage higher than 1%. Geographically, these communes were located southwest of and central to our study area (Figure 8E). Consequently, this study’s transition from agricultural land to artificial surfaces was significant. Therefore, the influence of bare land on the net transition of construction and agricultural land is not negligible and needs a comprehensive investigation.

A growing body of literature agrees that socio-economic development occurred during the agricultural land transition (Kanianska, 2016). Conversely, according to farmers’ perceptions, natural environmental factors drive the agricultural land transition. Firstly, as shown in Figure 9, if we focus on the “strongly agree” response, the socio-economic factors, namely, the population growth factors (35.38%), record the most important “strongly agree” response among the other driving factors. Weaknesses follow it in the land use policy factor (29.06%), considered a political driving factor. Regarding natural and climatic factors such as temperature and hydric erosion, they recorded 25.13% and 23.93%, respectively. Secondly, Table 5 showed that the relationship between soil salinization and deforestation was the most significant, with 0.971. In the same sense, wind erosion and soil salinization (0.944) occupied the second place. In that setting, the main driving factors that may facilitate the agricultural land transition in Thiès' region are the natural environment and biophysical factors. Alternatively, among socio-economic factors, one of the most interesting results was observed between the lack of agricultural investment and land policy (0.695). This context was followed by land policy and land prices (0.586). The lower value (0.246) was observed between urban expansion and population growth. Consequently, it seems that the manifestation of agricultural land transition is not the same between industrial and developing countries. In conclusion, farmers’ perceptions of socio-economic drivers of agricultural land transition in this region are insignificant compared to natural environmental drivers.

According to the findings of this study, between 2000 and 2020, a diverse range of factors, including those related to the economy, politics, and nature, may significantly impact Thies region’s agricultural land morphology. An in-depth exploration of the dynamics and existing problems in farmland morphology is crucial to formulating targeted protection policies (Lyu et al., 2021). Therefore, understanding agricultural land use dynamics in transition is critical for ensuring national food security in developing countries like Senegal. The agricultural land resource is an indispensable production factor for national economic development and farmer households. Economic growth generally leads to increased demand for land and changes in land utilization patterns (Chen et al., 2020). This situation is a reality in Thiès because the results show that between 2000 and 2020, the net transition of artificial surfaces was about 148.95 km² and evolution of 210.7%. According to ANSD, between 1976 and 2019, the regional urbanization rate increased steadily from 29% to 51.7%, respectively. Then, Thiès region has served as a secondary region of Senegal to promote socio-economic development (Faye and Du, 2021). Promoting socio-economic development requires constructing new infrastructure, and the lack of comprehensive land policies may induce agricultural land transition. This context is essential because fast urbanization appears to be a factor that may induce the transition of agricultural land. Expansions of housing, transportation, industry, retail sales, schools, and other developments are driving farming off of land (Francis et al., 2012). The analysis demonstrates that the communes of Fandene (Department of Thies) and Malicounda (Department of Mbour) have the most intense agricultural land transition. Or, in the North, the agricultural land use seems to have been caused by other factors such as climatic according to the survey. Then, the rural exodus and other economic activities, such as tourism in the western part of the Thies region, partly explain this rapid increase in the urban population. The majority of the population and economic activity of the Thiès region is concentrated in these communes, which serve as the city centers. The migration percentage for employment problems represents 13.3% in the Thies region for ages 15–35. Accordingly, the migration led almost to urban spatial issues and agricultural land abandonment. So, as cities grow and spatially expand in the Thiès region, agricultural land is converted into residential land (Picard and Selod, 2020).

In addition, the Thiès region was one of the backbone regions of Senegal’s mining industry. Previous studies highlight that mining activity in the northern part of the region, such as Taiba Ndiaye commune, strongly impacts the population’s socio-economic life (Henri Marcel SECK, 2021), including agricultural land use. So, this situation can justify the rapid conversion of agricultural land use to bare land in this part of our study area. Regarding infrastructure, this area has been chosen to host significant structuring projects such as Blaise Diagne International Airport and the Special Integrated Economic Zone. In addition, there are the urban poles of Diamniadio and Lac Rose, the industrial zone of Diamniadio, and the motorway projects. In this situation, the demand for land became more and more significant. The survey results did not highlight this spatial pressure on land, which revealed the lowest relationship between urbanization and population growth at 0.246. Thus, there is uncertainty regarding the influence of urban expansion on agricultural land transition. As a result, the social survey results demonstrate that urbanization was most significant in the western part of our study area, like the Sindia commune. This region’s part of the west appears to have the highest urbanization because of its proximity to Dakar (the capital of Senegal). In 2017, the Dakar region accounted for 39.5% of Senegal’s economic units, whereas Thiès and Diourbel accounted for 11.5% and 9.9%, respectively (Ministry of Economy Finance and Planning, 2017). In short, based on the spatial and survey data, we thought that the link between rapid urbanization and population growth might cause the rapid change from farmland to no agricultural land in our study area.

This section’s policies and cultural driving factors are the dualities between state land use policies and traditional land use practices. Since the Industrial Revolution, the economic development of Western Europe and North America has been characterized by continuous urbanization accompanied by a gradual phasing-in of urban land property rights over time (Cai Yongyang, 2015). Conversely, in developing countries such as Senegal, urbanization has been accompanied by several conflicts. The issue is the competition between land users, weak land tenure, etc. For instance, the drastic loss of livelihoods, including agricultural land, in the peri-urban areas of Sebougou (Mali) is primarily related to the crisis in Mali’s land management system (Coulibaly and Li, 2020). So, the agricultural land transition process may be more complex in Senegal than in industrial countries. The weak land use plan may significantly affect the rapid agricultural land transition. For instance, the duality between customary and modern land policy led to many issues in Thies. A relevant study highlights that on the West African continent, the land is mainly accessed through an informal and customary channel (Durand Lasserve and Selod, 2013). So, converting agricultural land use to another no agricultural land activity became more significate. This assertion shows that land on Thies region remains subject to traditional practices.

Land tenure is closely related to culture and institutional values in society (Suryadi et al., 2021). In Senegal, Law N*64–46 of 17 June 1964, governs land use management in Senegal and stipulates that land belongs not to the state, territorial communities, or users but to the “Nation” (Niang, 2017). In other words, this law stipulates that most of the land in Senegal is national domain land that does not belong to the people who use it because the law has abolished customary rights. According to the National Agency for Demography and Statistics, approximately 88.61% of plots were not registered between 2017 and 2020. This situation shows that agricultural land remains under traditional dominance, and the mode of land acquisition is passed down from generation to generation. As a result, landholders can sell, transfer, or reuse it for other purposes without significant legal constraints. Hence, no registered land became an obstacle not only to agricultural development but also caused rapid urbanization (Faye, Du and Zhang, 2022). Accordingly, the continuous fragmentation of agricultural land from generation to generation without an adequate policy for its protection remains a reality. So, the dominant morphology changes in Senegal, particularly in the Thiès region, is heavily influenced by property rights and a lack of agricultural investment. In another work, the lack of land policies is often seen as the main factor in agrarian land transitions. Political and economic reforms are imperative for farming and related sectors. But today, it is difficult for developing countries like Senegal to initiate specific reforms. This situation arose because the politic of decentralization, for example, remains closely linked to the obligations of structural adjustment policies and is constrained by the logic of liberal reform (Boutinot, 2003). Then, according to FAOSTAT data, over the period 2000–2020, the average credit to agriculture was about US $48.82 million. This situation seems to indicate that external financial institutions partly finance Senegalese agriculture. As a result, the agricultural policy reforms, including Senegal’s land system, remain challenging and complex. In summary, this may imply that when the social development level is low due to the unclear property rights of cultivated land and underdeveloped agricultural technology, the social awareness of cultivated land protection is relatively weak, and cities are disorderly expanded (Lv et al., 2022).

This study’s main factors are natural factors such as rain variability, deforestation, and temperature. Rain variability has a significant impact on the evolution of sown land use. In fact, in Africa, particularly in the Sahel, after the rainy periods of the 1960s, many researchers noted rain anomalies in the early 1970s (Ambiente, 2016). This situation was pointed out, too, in Senegal. Tappan et al. (2004) revealed that rainfall in the Groundnut Basin ranged from 400 to 800 mm in the 1960s to 200–600 mm in the 1990s (Tappan et al., 2004). Then, according to the data collected by the National Agency for Civil Aviation and Meteorology (ANACIM), the average rainfall from 2000 to 2020 was 461.65 mm. Or, this average hid several disparities. For instance, in 2000, the average was 607.90 mm, compared with 317.90 mm in 2005. These analyses show a difference of 290 mm. The above discussion shows that rain remains a problem in our study area because Senegal’s agricultural production still depends entirely on the rainy season. This variability’s consequences are reflected directly in farm productivity and yield (FAYE Bonoua, 2016). In addition, low annual rainfall, frequent dry spells, and the shortening of the rainy season affect the vegetative cycle of crops (Faye Mbagnick, 2018). During the social survey, several farmers attested that their agricultural land was decreased, and the most critical factor was rainfall, weak agricultural investment, and so on. Therefore, weak agricultural productivity diminishes agricultural land area, directly inducing agricultural land abandonment and population migration. This is because agriculture is Senegal’s main activity (ANSD, 2020). The household, affected by the weak agricultural productivity, moves to a big city such as Dakar to look for new activities. This statement may explain the strong relationship between rainfall and lack of investment in agricultural land (0.893). However, agricultural production depends on rainfall (Isabelle et al., 2019) and the availability of outputs. In the same sense, the temperature data from the same period from 2000 to 2020 reveal that in 2000 the average temperature was 24.85°, and in 2020 it was 28.01°. So, the temperature was gradually rising. This augmentation of the temperature seems to have a negative impact on agricultural land use, as shown by the survey results (0.971).

Desertification is land degradation in drylands resulting from various factors, including climatic variations and human activities (European Union, 2011). In other words, land use and cover changes, such as deforestation, affect the climate system and land-atmosphere interactions (Sy et al., 2017). In that context, agricultural land transition and deforestation will be inextricably linked. This assertion was supported by our research, which discovered that the net change of agricultural land and the net transition of forested areas decreased during the study period. The social survey results indirectly show a high relationship between deforestation and wind erosion (0.943). This relationship induces subsequent socio-economic consequences, such as agricultural productivity. Relevant studies show that rural economic development (Liu et al., 2014) and urbanization (Rondhi et al., 2018) are the primary forces driving the transition from farmland to non-agricultural land use. As revealed by the social survey, the relationship between rainfall variability and deforestation was strong (0.897) from then on. Furthermore, according to survey results, many smallholders believe the garden could be an excellent solution for dumping this matter in the event of less rain. In contrast, the lack of financial means is critical for obtaining the solution. In addition, to support their families, many sell agricultural land or choose migration to diversify their income. So, this practice was seen as facilitating agricultural land transition and fragmentation.

Natural factors like rain runoff and bio-physical drivers like hydric erosion are closely linked. Hence, the consequences of the decrease in rainfall are reflected in Senegal by the degradation of the natural environment. Then, drought leads to the degradation of the vegetation cover, the soils being subjected to erosion and runoff, and the accentuation of acidification and salinization (Ndong, 1995). Hence, Senegal’s forests have decreased from 4.4% in 1965 to 2.6% in 2000 (Tappan et al., 2004). Similarly, our results show that forested areas have dropped by −41.01 km². In that setting, the forested areas in Senegal continue to be reduced. Consequently, forest degradation may affect agricultural land quality because previous studies have highlighted that soil organic carbon is crucial in regulating soil quality functions and ecosystem services (Amoakwah et al., 2022).

Multiple physical, chemical, and biological degradations increasingly threaten the soil. Soil erosion affects 56% of the land surface worldwide (Van Oost et al., 2007). Specifically, soils in Africa affected by moderate to severe water erosion cover more than 12 million hectares, or 18.5% of the total national territory (Faroukh Tsouli A, 2017). Then, the Groundnut Basin of Senegal (including our study area) is confronted with chemical and physical-biological degradation, which has become more intense. Thus, the soils are impoverished, restructured, and chemically exhausted by wind, water erosion, and recurrent droughts (Reseau Agro-Innov, 2017). The survey results show that the relationship between soil salinization and hydric erosion was about 0.690. This situation signifies that biophysical driving forces like wind and water erosion are some of the main factors that may facilitate the agricultural land transition in the Thiès region.

Consequently, agricultural land degradation was occurring in Senegal’s Thiès region. This continuous degradation may impact agricultural productivity because agricultural soil erosion is thought to perturb the global carbon cycle (Van Oost et al., 2007). Hence, the deterioration of forested areas significantly impacts agricultural land quality. According to the above background, the agricultural land transition is an incentive-driven process (Rondhi et al., 2018). According to the social survey data and the summary in Figure 10, biophysical factors appear to be one of Thies region’s most important driving factors that induced agricultural land transition. However, the study clarifies that social, economic, and natural factors are all tied together in the agricultural land change in our study area.

This discussion has shown that the Thiès region’s agricultural land faces several driving factors, particularly natural factors, which have been identified as the main factors. From then on, the first recommendation regarding agricultural land protection is implementing a comprehensive land use reform policy. This context exists because implementing clear rules between users is the first key to protecting it. So, according to the complexity of the land management system in Senegal, reforming land tenure must cooperate with its complexity rather than attempt to substitute customary land practices. From an economic point of view, we think that increasing agricultural production, creating new jobs, and raising income may stop agricultural land from changing because, according to survey results and interviews, agricultural land changes rapidly when abandoned or sold.