PATSPO project , CIFOR-ICRAF, Addis Ababa & Nairobi, December 2022


Introduction

What to Plant Where in Ethiopia allows users to select suitable tree species and their best matching seed sources at any planting location in Ethiopia. It further gives contact information to order the seed. The seed sources included in the tool are all from Ethiopia’s National Tree Seed Source Registry, which is developed with support from the PATSPO project.

The location of the planting site is selected in the 'Planting location' TAB.

Suitable species and available seed sources are provided independent from each other in this App (and actually more species are listed for seed sources).

- Suitable species are listed in the 'Suitable species' TAB.

- The TAB of 'Select species' chooses the species for the 'Source ranking', 'Source locations' and 'Source details' TABs.

A different set of environmental variables to calculate 'environmental distance' can be chosen in the 'Variables' TAB.

Details on the methodology used is given in the TAB of 'Methods'.

A simplified version of this App is available as a version for practioners .


Planting location

Select the planting location by clicking on the map below

Loading...

Selected planting location: coordinates and planting zone

Loading...

Suitable species

The table below lists species predicted to be suitable to the planting site in baseline and/or future climates.

Note that suitability maps are not available for all species with seed sources (see in the 'Methods' TAB).

Field 'habitat' indicates whether habitat is lost (< 4 GCMs out of a total of 9 predicting presence in the future scenario), uncertain (4-5 GCMs predicting presence), kept (> 5 GCMs predicting presence) or new (unsuitable habitat under baseline conditions and > 5 GCMs predicting presence).

Field 'GCMs' indicates the number of CMIP6 Global Climate Models predicting future presence at the planting location.

Loading...

Select species

Seed source ranking

Selected species (select species from the 'Select species' TAB)

Loading...

Sources are ranked by environmental distance. In case of ties, geographical distance (Distance.km) is used as tie-breaker.

Loading...

Seed source locations

Selected species (select species from the 'Select species' TAB)

Loading...

The map shows the location of the seed sources listed in the 'Source ranking' TAB.

Seed sources are labeled by their ranking. Click on the marker to see the Seed.ID and Location.

Loading...

Seed source details

The table below gives more details about a particular seed source among those seed sources available for the selected species. Sources are listed by their rankings, as documented in the 'Source ranking' TAB.


Loading...

Selected environmental variables

The selected variables are the variables used for calculating the environmental distances between the planting site and the seed sources.

The default selections correspond to environmental variables used for species suitability modelling in a climate change atlas.


Selected bioclimatic variables from those available from WorldClim

Selected bioclimatic variables from an expanded set

Selected soil and topographic variables

Methods

Planting zones

Planting zones correspond to the atlas of Potential Vegetation of Ethiopia , with following adjustments from the 2010 version:

- The vegetation type of transitional semi-evergreen bushland was added as documented here .

- The vegetation type of intermediate evergreen Afromontane forest was added as documented here .

- Boundaries for transitional rain forest were adjusted as documented here .

- Boundaries for Lake Tana were adjusted.

Description of the other vegetation types are available from the vegetationmap4Africa .

The seed-collection zones were created by resampling the vegetation map to the same resolution as the environmental raster data.

Note that the zonation corresponds to the baseline climate as we have not yet modelled shifts in the zones under future climate scenarios.

Environmental data

Nineteen bioclimatic variables were obtained from WorldClim 2.1 . Future climates for the 2050s (2041-2060) correspond to a low emissions scenario (SSP 1-2.6) and a high emissions scenario (3-7.0) of Shared Socioeconomic Pathways of CMIP6 . These climate values are median or extreme (see next section) values from those projected by the nine Global Climate Models initially listed in WorldClim 2.1. The resolution of environmental rasters was 2.5 arc-minutes, the highest resolution in WorldClim 2.1 for future projections.

The bioclimatic variables were expanded with the envirem package (version 2.2) and the ensemble.PET.season function of the BiodiversityR package (version 2.13-1) . The moisture index was calculated from the WorldClim BIO12 and the envirem annualPET.

Soil variables were obtained from SoilGrids250 , calculating the average afterwards from soil depths at 5-15, 15-30, 30-60 and 60-100 cm.

The topographic wetness and roughness indices were obtained from ENVIREM .

The default environmental variables are those that were used for species suitability modelling. In the version for practioners of this App, this default subset of variables is always used.

Extreme values for future climates

Extreme values for future climates corresponded to either maxima or minima from the projections by the nine Global Climate Models. The selection of the direction (towards maxima or towards minima) was guided by potential linkages to extreme drought or heat events.

Maximum values were calculated for variables that mainly correspond to temperatures and PET values:

- BIO01 to BIO11 from WorldClim

- BIO15 from WorldClim

- maxTempColdest and maxTempWarmest from envirem

- various variables derived from monthly PET values calculated by envirem

- continentality, monthCountByTemp10 and thermicityIndex from envirem

Minimum values were calculated for variables that mainly correspond to precipitation and moisture index values:

- BIO12 to BIO14 and BIO16 to BIO19 from WorldClim

- aridityIndexThornthwaite, climaticMoistureIndex and embergerQ from envirem

- Moisture Index

- growingDegDays0 and growingDegDays5 from envirem

- ariditySeason from BiodiversityR

Species suitability

Species suitability in baseline and future climates were obtained from a climate change atlas for Africa . Ensemble suitability modelling for this atlas was done by ensemble suitability modelling with BiodiversityR via methods partially described in Kindt (2018) . The explanatory variables selected for suitability modelling via a process that limited their, Variance Inflation Factors are the default variables for calculating the environmental distances between the planting site and the seed sources.

Note that for 10 species, Albizia malacophylla, Cytisus proliferus, Corymbia citriodora (synonym: Eucalyptus citriodora), Cupressus lusitanica, Eucalyptus globulus, Eucalyptus grandis, Eucalyptus saligna, Millettia ferruginea, Moringa stenopetala and Pinus radiata, species suitability modelling results could not be obtained as a result of limitations in available presence observations in Africa. Therefore, no information is given on projected changes in habitat for these species. However, seed source details are provided for these species in the App.

Species listed for seed sources

Whereas suitability maps are not available for all species, all species with seed sources are listed for seed source rankings. The user can however opt to only show those species that are suitable at the planting location in the baseline or future climate.

Seed source ID

The 'Source.ID' field used in the App corresponds to the first characters of the source location, the first three characters of the genus name and the first three characters of the species name. For example, 'ENAR_Acaaby' was derived from the location of Enarj Enawga for Acacia abyssinica.

Ranking of seed sources

The ranking of seed sources was based on their distance in 'environmental space', calculated as the average Euclidean distance in a multivariate configuration obtained by the variables selected in the 'Variables' TAB. Euclidean distances were calculated after a principal components analysis performed via function rda with argument scale=TRUE of the vegan package (version 2.5-6) . The average distance was calculated by dividing the Euclidean distance by the number of variables, and therefore increasing the number of variables should not necessarily result in an increase of the environmental distance.

The geographical distance was obtained from the distGeo function of the geosphere package (version 1.5-10) . The geographical distance was not used for ranking seed sources, except as 'tie-breaker' in cases were there was a tie for the environmental distance.

Information on tree uses

The key references consulted included:

- A useful tree and shrubs manual developed for Ethiopia by RELMA-ICRAF

- A useful tree and shrubs manual developed for Kenya by RELMA-ICRAF for two species not listed for Ethiopia ( Moringa stenopetala and Pinus patula )

- Plant Resources of Tropical Africa (PROTA)

Details about the star rating system of PROTA are available here .

The use categories and subcategories correspond to classifications of uses in the summary tables of the useful tree manuals. A separate use category corresponds to information from PROTA.

Administrative data

Information on the Region, Zone and Woreda (District) of seed sources was obtained from the GADM database (version 3.6) . Information on the Kebele corresponded to details of the seed source, if recorded in the field.


Versions

Version v.2022.12.r

- Seed sources were selected as a subset from a Seed Sources Database (SSDB) created in 2022.

- Information on bioclimatic, soil and other environmental characteristics were collated for the selected seed sources.

- Species links to the Agroforestry Species Switchboard in the seed source details were standardized with the current names in the 2022 update of the Switchboard.

- Included a list of all species besides lists for use categories.

Version 1.0 (2021)

- First openly released version.

Credits

App developed by Roeland KINDT, Abrham ABIYU, Ian DAWSON, Lars GRAUDAL, Jens-Peter LILLESO, Yigardu MENGESHA, Soren MOESTRUP, Carsten NORGARD and Paulo VAN BREUGEL.

Planting zones correspond to the atlas of Potential Vegetation of Ethiopia .

Species suitabilities in baseline and future climates correspond to a climate change atlas for Africa .

This App is an output from the Provision of Adequate Tree Seed Portfolio in Ethiopia (PATSPO) project that is supported by Norway's International Climate and Forest Initiative through the Royal Norwegian Embassy in Ethiopia. Additional funding came from the CGIAR Research Program on Forests, Trees, and Agroforestry , supported by the CGIAR Trust Fund, and from Germany's Federal Ministry for Economic Cooperation and Development via support to the Transforming the Quality of Tree Planting partnership platform.


Suggested citation:

Kindt R, Abiyu A, Dawson IK, Graudal L, Lillesø J-PB, Mengesha Y, Moestrup S, Nørgard C, van Breugel P. 2022. What To Plant Where in Ethiopia. Version for researchers 2022.12.r. Nairobi and Addis Ababa: World Agroforestry. Accessed via https://www.worldagroforestry.org/output/what-to-plant-where-in-ethiopia on DD-MM-YYYY.