A geospatial dataset of inland valleys in four zones in Benin, Sierra Leone and Mali

The dataset described in this data article represents four agricultural zones in West-Africa that are located in three countries: Benin, Mali and Sierra Leone. The dataset was created through a research collaboration between the Africa Rice Center (AfricaRice), Sierra Leone Agricultural Research Institute (SLARI) and the Institute for Rural Economy (IER). The dataset was compiled to investigate the potential for rice production in inland valleys of the three countries. The results of the investigation were published in Dossou-Yovo et al. (2017) and Djagba et al. (2018). The dataset describes the biophysical and socioeconomic conditions of 499 inland valleys in the four agricultural zones. In each inland valley data were collected through a focus group interview with a minimum of three farmers. In 499 interviews a total of 7496 farmers participated. The location of each inland valley was determined with handheld GPS devices. The geographic locations were used to extract additional parameters from digital maps on soils, elevation, population density, rainfall, flow accumulation, and distances to roads, market places, rice mills, chemical input stores, and settlements. The dataset contains 65 parameters in four themes (location, biophysical characteristics, socioeconomic characteristics, and inland valley land development and use). The GPS coordinates indicate the location of an inland valley, but they do not lead to the location of individual fields of farmers that were interviewed. The dataset is publicly shared as Supplementary data to this data article.


a b s t r a c t
The dataset described in this data article represents four agricultural zones in West-Africa that are located in three countries: Benin, Mali and Sierra Leone. The dataset was created through a research collaboration between the Africa Rice Center (AfricaRice), Sierra Leone Agricultural Research Institute (SLARI) and the Institute for Rural Economy (IER). The dataset was compiled to investigate the potential for rice production in inland valleys of the three countries. The results of the investigation were published in Dossou-Yovo et al. (2017) and Djagba et al. (2018). The dataset describes the biophysical and socioeconomic conditions of 499 inland valleys in the four agricultural zones. In each inland valley data were collected through a focus group interview with a minimum of three farmers. In 499 interviews a total of 7496 farmers participated. The location of each inland valley was determined with handheld GPS devices. The geographic locations were used to extract additional parameters from digital maps on soils, elevation, population density, rainfall, flow accumulation, and distances to roads, market places, rice mills, chemical input stores, and settlements. The dataset contains 65 parameters in four themes (location, biophysical characteristics, socioeconomic characteristics, and inland valley land development and use

Value of the data
A large multidisciplinary dataset comprising 499 inland valleys in three countries in West-Africa that cover location, biophysical characteristics, socioeconomic characteristics and inland valley exploitation.
The dataset can be deployed to analyze the potential for agricultural development, to characterize diverse inland valley landscapes, to perform environment impact assessments, to classify land use from satellite imagery, etc.
The dataset contributes to food security research and assessments in West-Africa and leads to further understanding of the diversity of agricultural systems and their potential to contribute to food production and income generation for the rural population.
The dataset was deployed to assess the diversity and importance of inland valley agricultural systems to a regional scale in Sierra Leone [1].
To expand regional coverage the data can be linked to similar surveys conducted in inland valleys in Niger state (Nigeria), entire Burkina Faso and southern Mali [3,4].

Data
The dataset contains biophysical and socioeconomic information on 499 inland valleys in four zones in Benin, Mali and Sierra Leone (see Fig. 1). The inland valleys are geolocated with latitude/ longitude coordinates. The parameters (Table 2), grouped in four themes (Table 1), were obtained from farmers' responses during focus group interviews conducted in each of the 499 inland valleys between 2013 and 2014. Additional parameters were extracted from digital maps using the location of the inland valleys. Table 2 outlines the variables collected, and their source whether from the interviews or secondary spatial data sources.
The dataset is provided in Microsoft Excel format and contains seven sheets. The first sheet (source) provides citation information and refers to this data article. The second sheet (variable explanation) outlines the variables. After that the sheet location provides the unique identifier of each surveyed inland valleys and the geographic coordinates expressed in longitude/latitude. The unique identifier can be linked to the variables stored in four sheets, one for each of the four zones, called Mali, Sierra Leone, Benin_Ouémé supérieur and Benin_Mono-Couffo (Fig. 1).

Experimental design, materials and methods
This section provides a summary of the steps taken to develop the geospatial dataset. [2] provides a full description of the methodology that was followed.
Data collection was implemented in two phases. In the first phase, 499 inland valleys were identified in four zones in Mali, Benin and Sierra Leone. These were 100, 149, 100 and 150 inland valleys in Mono and Couffo departments (Benin), Upper Ouémé catchment (Benin), Sikasso and   Kadiolo districts (Mali) and Bo and Kenema districts (Sierra Leone), respectively. These sites were visited by teams of trained surveyors equipped with a questionnaire and a GPS. Focus group interviews with at least three farmers operating in the inland valley were held and their responses were recorded. Focus groups existed of maximum 7496 farmers and on average 15 farmers participated in the focus group interviews. With the use of handheld GPS devices, the coordinates of the inland valleys were obtained. In a second phase, the locations of the inland valleys were imported into a Geographic Information System and their quality was checked. Spatial information available in the public domain were downloaded and imported in a GIS. These included maps of soil parameters, topology, rainfall, settlements, roads, population density, etc. Information for each inland valley was extracted using the location information of the sites and the values were added to the dataset of questionnaire responses and observations. Table 2 provides on overview of the 65 parameters in the dataset and their source (whether from the field surveys or public domain sources).