•There is some evidence regarding the effects of climate shocks on long-term mobility, but understanding of short-term mobility responses is limited.
•This limitation is due to challenges in capturing these movements in traditional data sources, such as surveys and census, particularly in developing countries.
• Mobile phone data provides cost-effective alternatives, provides valuable insight into short-term mobility, and allows for an assessment of how rainfall conditions affect temporary movements in Senegal.
Senegal, like many countries in sub-Saharan Africa, exhibits a high level of human mobility. A population primarily engaged in agricultural activities has been migrating to classify their livelihoods for centuries (Adriansen, HK (2008), Lalou and Delaunay, 2017). Agriculture's production and revenues are heavily dependent on rainfall and are vulnerable to climate shocks. As a result, migration is often adopted as a strategy to address fluctuations in climate, climate and environmental conditions. This is adopted by post-post mail in response to diversifying revenue streams or in response to harmful impacts (De France et al., 2023).
This study explores the relationship between human mobility and environmental stressors, with a particular focus on temporary internalization. This is often understood due to data limitations. Understanding the scale of internal climate movements is important to the country, so it is planned and prepared.
Mobile phone data offers greater advantages over traditional survey-based devices for measuring temporary internalization
Use mobile phone data to infer temporary migration patterns
Previous studies on transition rely primarily on research data and focus on long-term movements. However, finer mobility patterns such as seasonal, circular, or temporary movement clearly outweigh long-term movements (Coffey et al. 2015; Bryan et al. 2014). However, capturing these movements using traditional research methods is extremely difficult. In this regard, mobile phone data has emerged as a promising alternative to studying more nuanced human movements, particularly in the context of development (Blumenstock 2012; Demissie et al. 2019).
The study uses call detail records (CDRs) from millions of Sonatel (Orange Senegal) customers for three years (2013-2015) (1). CDRs are traditionally collected by mobile network operators for billing purposes, but contain more information than they can effectively capture population movements. The spatial density of antennas, combined with frequent telephone use, usually results in very thin data and is a unique resource for studying human mobility.
Mobile phone data offers greater advantages over traditional survey-based devices for measuring temporary internalization. These data avoid recall bias, a common problem, especially in short-term events research. Additionally, they allow short-term mobility patterns to be measured at a fraction of the costs of all traditional data collection, both nationally and over the years. In sub-Saharan Africa where resources are limited, mobile phone data can provide valuable insight into the temporary internal migration dynamics. Conversely, unlike traditional survey data, the sample of mobile phone users is not random and may affect the representation of the total population. Nevertheless, this study shows that selection bias is minimal and that the data are likely to represent a significant portion of the adult population, both rural and urban.
Chi et al. Migration events have been identified in CDRs using clustering methods based on recent research. (2020). The user-level migration trajectories are then aggregated to generate a spatially detailed measure of temporary migration (see Blanchard and Rubrichi, 2024).
Explaining the temporary transition patterns in Senegal
Mobile phone data provides new insights into the size, timing, duration and direction of Senegal's temporary travel flows at unprecedented levels of spatial and temporal resolution. Several important findings emerge from this analysis.
First, the data show that temporary migration is common in Senegal, with an estimated 3.6 million immigration events occurring over at least 20 days in 2015. We consider that multiple events may involve the same individuals who participate in at least 20 days of immigration events per year. In contrast, long-term mobility was estimated at around 2%, highlighting the importance of short-term mobility.
Second, the median and mean duration of temporary migration episodes in Senegal are estimated to be 38 and 50 days, respectively. The duration variation is relatively modest, with over 70% of transition episodes lasting less than two months.
Third, Senegal's mobile currents are widely distributed, forming a wide network of connections between rural and urban areas (Figure 1). Major cities like Dakar, Touba, Ziguinchor, Thiès and Kaolack not only send many migrants, but also attract temporary immigrants from across the country, bringing 23% of the total flow in 2015 with Dakar alone.
Figure 1. Movement between rural and urban zones flows in 2015.
Finally, the temporal distribution of short-term movements reveals clear seasonal patterns. All three years examined have seen a marked increase in the migrant population, beginning in June and peaking in August to September. In particular, the number of immigrants has been driven primarily by migration from rural areas, more than doubled from early June to late September 2013. This trend indicates systematic labor migration from rural to urban areas during the wet season (June to October), suggesting a strategy for income diversification. This finding challenges the perception that temporary migration is limited to urban movements from rural areas and is consistent with recent local observations in Senegal (Delaunay et al., 2016). Furthermore, significant rural-to-rural movements have occurred during the rainy season, consistent with increasing agricultural activity, with rural immigrant inventory still rising since October, perhaps due to harvesting activities.
Figure 2. Temporary immigrant stocks for 2013-2015 are from the Origin Destination Zone.
Assessing the effect of rainy season conditions on temporary migration
Mobile phone data shows consistent temporary movement patterns over the years, but these patterns can be influenced by rainy season conditions. On the one hand, insufficient rainfall conditions at locations of private homes may serve as a push factor to promote migration to diversify income sources (Findley 1994; Henry et al. 2004; Delaunay et al. , 2016). On the other hand, it can increase liquidity constraints and complicate the transition process. Furthermore, unfavourable rainfall conditions at potential destinations can reduce their appeal. To quantify these effects, transition estimates of origin from cell phone data are combined with satellite-based rainfall measurements from 2013 to 2015. The regression method evaluates how rainy season conditions of both origin and destination affect movement numbers and employs half-moon measurements to accurately capture the timing of the effects throughout the agricultural year.
The analysis revealed several important findings.
Effects of malignant rainfall conditions in the rainy season: poor rainfall in the origin of the rainy season significantly reduces migration to both rural and urban areas during the harvest season, leading to an increase in migration to other rural areas during the offseason. During the offseason, rural-to-rural migration in response to poorer rainfall is stronger and more developed livestock sectors in poorer areas. Poor households appear to be more dependent on migration as a coping strategy due to financial constraints and limited local employment opportunities. Naturally mobile idyllic households tend to raise resources and move more during periods of inadequate rainfall, seeking water from their regular homes. The consequences of poor rainfall in subsequent seasons: dry conditions lead to spatial redistribution from rural to rural during subsequent offseasons. Detrimental Effects on Movement: Overall, dry conditions have a negative effect on both the ability to move during harvest season and the expected returns from Movement. The simulation exercise revealed a drought scenario with a 17% reduction in temporary migration during the harvest season in 2014 compared to a neutral scenario in which all locations receive average precipitation levels.
Conclusion
Short-term internal migration is essential for economic decision-making in developing countries. However, it remains difficult to measure it accurately. This study shows the possibility of mobile phone data to shed light on temporary movement patterns in Senegal, particularly depending on the conditions of the rainy season.
The study analyzes how rainfall in both hometown and destination affects transition decisions, revealing that dry conditions can reduce harvest season migration while leading to rural relocation during the offseason. In particular, there is little evidence of significant migration of urban migration from rural areas following disadvantages, which contradicts the general narrative.
The findings highlight the need for customized policies and response options based on local susceptibility to rainfall conditions. They also suggest that interventions to support production and consumption during harmful climatic conditions need to be carefully timed to be effective.
Mobile network operators can play an important role in informing such policies through data. Orange's Flux Vision Solution serves as a useful tool to replicate such results, contributing to Orange's ambition to become a player of economic, social and environmental resilience.
(1) Mobile phone data was provided by Sonatel (Senegal) as part of the D4D Challenge and Opal project.
