Spatial variation in current and historical management of Arabica coffee across forests in its indigenous distribution
Data files
Nov 15, 2024 version files 55.12 KB
-
cluster_summary3.xlsx
15.70 KB
-
coma_current_v3.csv
3.38 KB
-
coma_current_v5.1.csv
2.91 KB
-
Figure_4d_location_buffer.csv
6.34 KB
-
Figure_5_Fig_S2.csv
485 B
-
README.md
3.63 KB
-
Table_1_cluster_summary.xlsx
22.66 KB
Abstract
To guide conservation of forest biodiversity in a broad sense we need to understand the landscape-level variation in current and historical management practices of agroforestry systems. We collected data on coffee management practices across a large forested landscape in Ethiopia within Arabica coffee’s indigenous distribution, using interviews in 11 villages and field surveys and interviews at 78 coffee sites in the forest.
There was a gradient in coffee density (higher), liana cover (lower) and canopy cover (lower) from sites with high management intensity to sites without management. Recently, management intensity has increased in the forest edges. Interviews suggest that substantial areas of currently unmanaged coffee forests are a legacy of reforestation of abandoned (semi-) open landscapes in the late 19th century.
Despite a dynamic history of coffee cultivation across these areas, the conservation of forest biodiversity including unmanaged populations of genetically diverse Arabica coffee should be a priority, given few such remaining areas in Ethiopia and elsewhere in the world.
Explanations to the Data files and R codes for analyses
Spatial variation in current and historical management of Arabica coffee across forests in its indigenous distribution
See methods section in the paper for complementary descriptions and explanations of column names.
R-code for analyses and some figures:
FigS1.R: Principal component analysis and dendrogram of the coffee management, (based on coma_current_v3.csv)
Figure2_Coffee management clusters and their characteristcs.R: Statistical code for results presented in Figure 2 a-h, (based on cluster_summary3.xlsx)
Table S3.R: Statistical code and data for results presented in Table S3
Table S5.R: Statistical code for results presented in Table S5, (based on coma_current_v5.1.csv)
Figure4_a-c.R: Statistical code for results presented in Table S5 and Figure 4abc, (based on coma_current_v5.1.csv)
Data and files used in the R-script:
**Table_1_cluster_summary.xlsx: (**data for result presented in Table 1)
column names (for three Clusters, i.e sheet 1-3): SiteID=Site ID, ClusterID= Cluster ID, Shade tree management, Digging, Herbicide, Enriching, Intercropping, Mulching, Planting coffee, Planting shade tree, Pruning, Removal of dried coffee, Removal of epiphytes, Removal of liana, Stumping coffee, Thinning of coffee stand, Under Canopy clearing, Slashing, Harvesting.
cluster_summary3.xlsx: (used in Figure2_Coffee management clusters and their characteristcs.R, data for result presented in Figure 2a-h)
column names (for Figure 2a-c): Clusters, Circumferenceofoldest(cm)= Thickness oldest coffee, Heightoldestcoffee(m)= Height oldest coffee, Circumferenceofthickest(cm)= Thickness of largest coffee.
Column names (for Figure d-h): Clusters, allVery few = very few coffee density, allFew= few coffee density, allMuch= much coffee density, allVery much= very much coffee density, Tree density_very few= Very few Tree density, Tree density_Few= Few Tree density, Tree density_Medium= Medium Tree density, Tree density_High/very high= High/very high Tree density, canopy_Open= Open canopy, canopy_Partially open= Partially open canopy, canopy_Closed= Closed canopy, Liana_None= None Liana cover, Liana_Little= Little liana cover, Liana_Much= Much liana cover, Liana_Very much= Very much liana cover, Easily walkable, Intermediate= intermediate to walk, Difficult to walk
coma_current_v5.1.csv: (used in Figure4_a-c.R, data for result presented in Table S5)
column names: SiteID= Site ID, ClusterID= Cluster ID, Distancetoedge= Distance to edge, Elevation, Distancetoroad= Distance to road
Figure 4d_location_buffer.csv: (used in GIS to map buffer of 500 m around each site, data for result presented in Figure 4d)
Column names: Subsite ID, ClusterID= Cluster ID (1= intesively managed sites,2= low managment sites,3= not managed sites, n/a= coffee sites managed previously but now abondended, nocoffee=above coffee growing area or no coffee in the site), latitude, longitude, altitude (m), Distance to edge (m), Distance to road (m)
coma_current_v3.csv: (used in FigS1.R, data for results presented in FigS1)
column names: SiteID= Site ID, Shade tree management, Digging, Herbicide, Enriching, Intercropping, Mulching, Planting coffee, Planting shade tree, Pruning, Removal of dried coffee, Removal of epiphytes, Removal of liana, Stumping coffee, Thinning of coffee stand, Under Canopy clearing, Slashing, Harvesting
Figure 5_Fig_S2.csv: (data for result presented in Figure 5 and Fig S2)
column names: Current not managed, Current managed
To investigate coffee management dynamics in forest landscapes, researchers selected 11 primary forest areas representing different kebeles within the coffee-growing altitudinal range. A mixed methods approach was used, combining key informant interviews, guided field walks, and field surveys to gather quantitative and qualitative data. Each main site involved interviews with two key informants, totaling 32 male farmers and knowledgeable elders, who provided insights into historical and current coffee management practices. Additionally, 24 other male informants participated in field walks to collect quantitative data on environmental variables and conduct further interviews.
During the walks, researchers visited preselected sub-sites known from prior inventory work and identified additional significant locations suggested by informants. They collected data on the size and estimated age of the oldest coffee shrubs, coffee and tree density, liana cover, canopy cover, and walkability. A semi-structured interview method allowed for open-ended questions to gain deeper insights. Interviews were conducted in Afan Oromo or Amharic and later translated into English for analysis. Data collection occurred over six weeks, gathering information from 78 sub-sites where coffee was present. This comprehensive approach facilitated a better understanding of coffee management practices within the forest landscape.