Processed Data for Cassava field perimeters survey in Uganda and Côte d'Ivoire, 2018

Cassava surveys were conducted to gather information on cassava cultivation in Côte d’Ivoire and Uganda during four weeks between February and January 2018. A total of 69 locations in Côte d’Ivoire and 96 locations in Uganda were included in the surveys. To facilitate the measurement of the survey locations, a predefined fishnet grid of 100 x 100 square meters was established using the ArcGIS Collector app. The selection of survey locations was randomised, with approximately 15-20 kilometre intervals along major motorable roads in each country.At each selected sampling location, the survey team assessed an area of approximately 200 x 200 square meters, divided into four predefined quadrants measuring 100 meters by 100 meters each. The team recorded various aspects related to cassava fields within the study area, including the perimeter of all cassava fields, the size of smaller cassava patches, and the number of individual plants outside of the main field patch. Additional attributes of the fields and patches were documented, such as whether cassava was intercropped, the age of the cassava plants, and the density of each field (categorized as high, medium, or low density). In the case of intercropped fields, the other crops present were also noted.To ensure complete coverage of the surveyed area, the surveyors had the option to activate a tracking function in the ArcGIS Collector app. This function automatically marked the route followed by the survey team on the screen. In certain areas where access was challenging or safety concerns arose, such as suburban regions, only one or two 100 x 100 square meter quadrants were selected for practical reasons.The data collected during the surveys were exported and saved as polygons and points representing the surveyed locations. These data underwent post-processing to determine the proportion of the study area covered by cassava fields. The area of cassava fields was calculated based on the field and patch perimeters, while for individual plants, a 0.5-meter radius was assumed around each plant.<b>dat_civ.csv </b>and <b>dat_uga.csv. </b>These are the processed survey data which merges, aligns and tidies the various different data files from the original survey (https://doi.org/10.6084/m9.figshare.23657391.v1). Column names are as in the original files with the additional columns listed below,tot_area_ind_plants: Total area consisting of individual cassava plantstot_ind_plants: Total number of individual cassava plants1story_buildings: Number of 1 story buildings2story_buildings: Number of 2 story buildingstot_buildings: Total number of buildingstot_cassava_area: Total area in cassava productiontot_monoculture_area: Total area of cassava monoculture productiontot_intercrop_area: Total area of intercropped cassava productionno_fields: Total number of fieldstot_cassava_area_w: Total area in cassava production weighted according to formula in methodstot_monoculture_area_w: Total area in cassava monoculture production weighted according to formula in methodstot_intercrop_area_w: Total area of intercropped cassava production weighted according to formula in methodscass_area_presence: Binary variable indicating whether sample location had any cassava productionThe total area in cassava production at each survey location A_C was calculated as A_C= (∑_(i=1)^M α_i +∑_(j=1)^N β_j +∑_(k=1)^K γ_k ) / δwhere α_i is the area of a cassava monoculture field and M is the total number of monoculture fields at the survey location, β_j is the area of a cassava intercropped field and N is the total number of intercropped fields at the survey location and γ_k is the area of an individual cassava plant and K is the total number of individual plants at the survey location. δ is the total area of the survey location. A secondary measure of total cassava production was calculated to incorporate i) a lower density of cassava production in intercropped fields (calculated as a weight of 0.75) and ii) the qualitative assessment of cassava density within each field or patch. Specifically, weights ω_(i,j) were assigned according the assignment below. All other fields with no specific density recording were given a weight of 1.Very High=1.75High=1.5Regular=0.75Sparse=0.5Very sparse=0.25<br><br><b>UGA_buff_pointstats.csv</b> and <b>CIV_buff_pointstats.csv</b> contain the buffer summaries calculated aligned to each sample location. Column names are as below;Centre_Lon Longitude of sample locationCentre_Lat Latitude of sample locationmean_Prod.2000 Mean, from a 2km buffer, of modelled Cassava production from CassavaMapmean_Prod.5000 Mean, from a 5km buffer, of modelled Cassava production from CassavaMapmean_Prod.10000 Mean, from a 10km buffer, of modelled Cassava production from CassavaMapmean_Harv.2000 Mean, from a 2km buffer, of modelled Cassava Harvest area from CassavaMapmean_Harv.5000mean_Harv.10000mean_lspop.2000 Mean population, from a 2km buffer, as obtained from LandScan 2014mean_lspop.5000mean_lspop.10000mean_settle.2000 Mean settlement density, from a 2km buffer, as obtained from a binary mask of WorldPop 2018mean_settle.5000mean_settle.10000The above are then repeated (with the exception of the settlement layer) for each statistical summary standard deviation (sd), minimum (min), lower quartile (lowerQ), median (median), upper quartile (upperQ), maximum (max).<br>