Geospatial data can support IFAD projects in analysing agricultural value chains and prioritizing investments. An important contribution is to monitor existing project interventions but also designing new ones based on future developments. Earth Observation (EO) for better-informed Decision Making (EODM) is a three year WCA grant that has supported different IFAD initiatives in Senegal (and in Cameroon and Mali). This is the work EODM carried out in Senegal together with the Programme d'Appui au Développement Agricole et à l'Entreprenariat Rural (PADAER) project and the Centre de Suivi Écologique (CSE), a national technical partner of IFAD. Activities are implemented in the PADAER area intervention. The work targeted the Communes of Makacolibantang (Tambacounda Region, Tambacounda Department), Bandafassi (Kédougou Region, Kédougou Department), and Ndorna (Kolda Region, Médina Yoro Foulah Department) in Senegal, which represent areas of high production for cereals in the country.
Geospatial data was collected from open sources such as OpenStreetmap (e.g. roads, villages, rivers). In addition the CSE provided its national datasets (e.g. boundaries of the communes, location of villages, land cover map and additional feeder road network) and conducted a ground survey to collect more local data (e.g. PADAER interventions) and verify the existing ones.
EODM also produced EO-based maps to support the value chain analysis, including:
EODM and CSE organized workshops to create capacity on the use of GIS and GPS for PADAER and its partners through participatory mapping techniques. Additionally to validate and integrate information for the value chain analysis, e.g. current capacity of storage facilities, statistics on crop production for each village, actual conditions of the roads, accuracy of the crop land map, etc.
Questions
The value chain processes identified are: production, collection, transport, storage, processing and sale of the production; also sales of inputs (fertilizers, pesticides, seeds, etc.). This exercise covers the processes from crop production to storage and transport of products as well as the delivery of inputs (especially during the rainy season).
1) Is the existing storage capacity of villages sufficient for the current and potential production and are the warehouses well positioned ? If not, where should new storage facilities be located?
2) Where should all-weather feeder roads be best located to channel the production towards storage infrastructures and markets and to deliver inputs to the farmers?
The figure below shows a map of the storage status (deficit: blue cross or surplus: green cross) for each VC. A black cross shows VC without warehouse facilities.
Looking forward
Apart from assessing the current situation, it is very important to analyse the potential needs for storage and feeder roads. Future scenarios are simulated (projected to 2030) which take into account population growth and, as a consequence, expansion of cropland and creation of new villages. All this determines a projection of the potential production and therefore an accrued need in terms of storage and road infrastructures. We have simulated this scenario using Agent Based Modelling (ABM) and participatory modelling techniques.
How can investments best be directed towards impact? ABM is a method for creating computer models with autonomous agents interacting in space, that can be utilized to understand complex problems in relatively data scarce environments. Based on ABM we have created an agricultural investment tool to assess where and how much to invest. By constructing a rule-based, digital representation of an agricultural system, users can test the effects of interventions in advance before enacting them in the real world. Users can test how different levels of investment impact economic output, inclusion in the supply chain, and exposure to environmental risks. They can further examine how to plan investments when prioritizing some of these factors more than others.
The projected cropland is modelled based on the currently cultivated land (for the major crops in the PADAER value chains, (i.e. riz de bas-fonds and mais) as identified from the satellite information. Expansion is modelled further based on: current population in each Commune, yearly population growth, size of the household holding for the 2 crops, and average attainable yield. A new village is created whenever the cropland expansion exceeds a maximum travelling distance from the farmers’ households to their fields.
Why is this important?
The exercise was regarded as very useful by the PADAER project stakeholders participating in the evaluation workshops. The participatory mapping and analysis make visible problems in the current infrastructures as well as opportunities for future investments. According to PADAER these will provide very valuable inputs in the design of the new phases of the project as well as in its M&E system. Capacity building efforts from EODM have supported PADAER staff and stakeholders who now can take a detailed and holistic stock of their local environment by way of advanced data analytics and satellites. The ambition is to scale this exercise to cover a much larger area in Senegal and at the same time develop the needed capacity at country level.
Sensor technologies such as those mounted on satellites can greatly inform resource management, development and systems planning. The EODM project is staking out paths for integrating satellite derived data in IFAD activities, driving transparent and evidence-based development of agriculture.
Do you think this is something for you?
In this blog we only scratched the surface of what is possible using geospatial data for value chain analysis. However, we hope we were able to show you the importance of this approach for this application and many more. If you would like to receive more information on the approach itself, and how to apply it to other areas and questions, do not hesitate to contact us.
 |
| Location of the three areas targeted |
Geospatial data was collected from open sources such as OpenStreetmap (e.g. roads, villages, rivers). In addition the CSE provided its national datasets (e.g. boundaries of the communes, location of villages, land cover map and additional feeder road network) and conducted a ground survey to collect more local data (e.g. PADAER interventions) and verify the existing ones.
EODM also produced EO-based maps to support the value chain analysis, including:
- A flood risk map that was combined with the existing road infrastructure to assess roads for potential flooding and erosion, and also used for identifying potential bas-fonds areas (irrigated lowlands). The figure below gives an example of flood risk map for Bandafassi.
- Baseline maps including accurate classification of cropland and forest land based on satellite information to identify: the current production areas and protected areas where cropland expansion should not occur.
 |
| Flood risk map for Bandafassi |
This land cover information was used to generate baseline maps for all Communes together with, for instance, their administrative limits, the location of each village (further categorized as village centre and village satellite), the network of roads and trails, as well as PADAER and other partners’ interventions. These include storage facilities, irrigation schemes (bas-fonds aménagés) and all-weather feeder roads (pistes de désenclavement).
 |
| Baseline map of Makacolibantang |
EODM and CSE organized workshops to create capacity on the use of GIS and GPS for PADAER and its partners through participatory mapping techniques. Additionally to validate and integrate information for the value chain analysis, e.g. current capacity of storage facilities, statistics on crop production for each village, actual conditions of the roads, accuracy of the crop land map, etc.
Questions
The value chain processes identified are: production, collection, transport, storage, processing and sale of the production; also sales of inputs (fertilizers, pesticides, seeds, etc.). This exercise covers the processes from crop production to storage and transport of products as well as the delivery of inputs (especially during the rainy season).
Based on these processes we can derive a number of questions for identifying and prioritizing project interventions and investments that can be addressed through spatial analyses based on geospatial data.
1) Is the existing storage capacity of villages sufficient for the current and potential production and are the warehouses well positioned ? If not, where should new storage facilities be located?
2) Where should all-weather feeder roads be best located to channel the production towards storage infrastructures and markets and to deliver inputs to the farmers?
Methodology
To answer the first question we classified the villages in each Commune as village centre (VC) or village satellite (VS) based on their role in both the collection of products and the delivery of inputs. A VS functions as first collection point of the products and has limited warehouse capacity. Products are then transported to a VC by truck where there are larger warehouse facilities. Farmers are organized through the Organisation Paysannes (OPs). Each OP has an unique village location and for each village PADAER collects statistics of the current areas planted and production for each crop in the relevant value chains. Totals can be calculated for each VC and geo-located. The marketable production in each VC is then matched against the current storage capacity of PADAER or other partners. Each VC is finally classified in terms of deficit or surplus. This information can now be geo-located and the two used for planning and evaluating investment needed: each VC should have sufficient storage capacity to cover its needs.
When addressing the second question we should start from the consideration that most VC are currently not served by roads that are easily accessed by trucks. In Figure 4 these are shown in orange, while roads in white are trails which in practice are used by trucks but with difficulties, especially during the rainy season. In principle all VC should be connected with such all-weather feeder roads. New feeder roads are prioritized for investment based on the size of the production areas they serve and the number of VC they connect. This is also taking into account the possible risk of being inundated in the rainy season, which should be minimized.
Outcomes
Current situation
The table below shows how current marketable production and storage capacity are matched to assess the current status of each VC and then for the whole Commune. Villages highlighted in red have a storage gap (indicating that a storage must be built), while those highlighted in green have a surplus, always with reference to the current situation. The exact needs in terms of new storage are calculated. Overall, Bandafassi has a gap of around 419 tonnes.
To answer the first question we classified the villages in each Commune as village centre (VC) or village satellite (VS) based on their role in both the collection of products and the delivery of inputs. A VS functions as first collection point of the products and has limited warehouse capacity. Products are then transported to a VC by truck where there are larger warehouse facilities. Farmers are organized through the Organisation Paysannes (OPs). Each OP has an unique village location and for each village PADAER collects statistics of the current areas planted and production for each crop in the relevant value chains. Totals can be calculated for each VC and geo-located. The marketable production in each VC is then matched against the current storage capacity of PADAER or other partners. Each VC is finally classified in terms of deficit or surplus. This information can now be geo-located and the two used for planning and evaluating investment needed: each VC should have sufficient storage capacity to cover its needs.
When addressing the second question we should start from the consideration that most VC are currently not served by roads that are easily accessed by trucks. In Figure 4 these are shown in orange, while roads in white are trails which in practice are used by trucks but with difficulties, especially during the rainy season. In principle all VC should be connected with such all-weather feeder roads. New feeder roads are prioritized for investment based on the size of the production areas they serve and the number of VC they connect. This is also taking into account the possible risk of being inundated in the rainy season, which should be minimized.
Outcomes
Current situation
The table below shows how current marketable production and storage capacity are matched to assess the current status of each VC and then for the whole Commune. Villages highlighted in red have a storage gap (indicating that a storage must be built), while those highlighted in green have a surplus, always with reference to the current situation. The exact needs in terms of new storage are calculated. Overall, Bandafassi has a gap of around 419 tonnes.
 |
| Storage needs in the current situation (Bandafassi) |
 |
| Map of storage status for each village centre |
The outcomes above indicate in which villages new storage facilities should be built and their location.
In the image below the factors considered for prioritizing the construction of new feeder roads are shown. The network of existing minor roads and trails is used for locating new feeder roads. Each portion of this network is analysed separately on the basis of the number of VC which can be connected; the marketable production which can be channelled; and the points which are at risk of flooding.
 |
| Prioritizing the realization of feeder roads |
Apart from assessing the current situation, it is very important to analyse the potential needs for storage and feeder roads. Future scenarios are simulated (projected to 2030) which take into account population growth and, as a consequence, expansion of cropland and creation of new villages. All this determines a projection of the potential production and therefore an accrued need in terms of storage and road infrastructures. We have simulated this scenario using Agent Based Modelling (ABM) and participatory modelling techniques.
How can investments best be directed towards impact? ABM is a method for creating computer models with autonomous agents interacting in space, that can be utilized to understand complex problems in relatively data scarce environments. Based on ABM we have created an agricultural investment tool to assess where and how much to invest. By constructing a rule-based, digital representation of an agricultural system, users can test the effects of interventions in advance before enacting them in the real world. Users can test how different levels of investment impact economic output, inclusion in the supply chain, and exposure to environmental risks. They can further examine how to plan investments when prioritizing some of these factors more than others.
The projected cropland is modelled based on the currently cultivated land (for the major crops in the PADAER value chains, (i.e. riz de bas-fonds and mais) as identified from the satellite information. Expansion is modelled further based on: current population in each Commune, yearly population growth, size of the household holding for the 2 crops, and average attainable yield. A new village is created whenever the cropland expansion exceeds a maximum travelling distance from the farmers’ households to their fields.
More scenarios can be elaborated based on different assumptions on the available level of investment. A scenario with constrained resources (scenario no. 1) can be developed, which can cover only a limited number of village storage and road access needs. Or otherwise the level of investment can be raised to cover all needs (scenario no. 2). The two, as well as intermediate scenarios, can help prioritizing the investments or defining the total level of investment required to cover all needs in a specific Commune.
Outcomes are compared in the following figure. The increased investments in the second scenario simulation sees more villages connected, more production reached by all-weather roads and more flood-prone roads upgraded.
 |
| Comparing potential production scenarios with different investment levels |
Why is this important?
The exercise was regarded as very useful by the PADAER project stakeholders participating in the evaluation workshops. The participatory mapping and analysis make visible problems in the current infrastructures as well as opportunities for future investments. According to PADAER these will provide very valuable inputs in the design of the new phases of the project as well as in its M&E system. Capacity building efforts from EODM have supported PADAER staff and stakeholders who now can take a detailed and holistic stock of their local environment by way of advanced data analytics and satellites. The ambition is to scale this exercise to cover a much larger area in Senegal and at the same time develop the needed capacity at country level.
Sensor technologies such as those mounted on satellites can greatly inform resource management, development and systems planning. The EODM project is staking out paths for integrating satellite derived data in IFAD activities, driving transparent and evidence-based development of agriculture.
Do you think this is something for you?
In this blog we only scratched the surface of what is possible using geospatial data for value chain analysis. However, we hope we were able to show you the importance of this approach for this application and many more. If you would like to receive more information on the approach itself, and how to apply it to other areas and questions, do not hesitate to contact us.
Who we are?
Earth Observation for better-informed Decision Making (EODM) is a three year WCA grant that has supported different IFAD initiatives in Senegal (but also Cameroon and Mali).
![]()
![]()
![]()
![]()
![]()
Earth Observation for better-informed Decision Making (EODM) is a three year WCA grant that has supported different IFAD initiatives in Senegal (but also Cameroon and Mali).
