Data for simulating near-term climate change impacts on Kenyan tree cover using LRange
Data files
Mar 17, 2026 version files 177.41 KB
-
AboveGroundLiveBiomass_Responses.csv
27.22 KB
-
Climate_data.RDS
10.62 KB
-
Herb_Responses.csv
30.08 KB
-
Manuscript_Plots_Rscripts.R
5.73 KB
-
PrimaryProductivity_Responses.csv
29.82 KB
-
README.md
10.99 KB
-
Shrub_Responses.csv
29.78 KB
-
SimulationOutputProcessing_Rscript.R
5.02 KB
-
Tree_responses.csv
28.14 KB
Abstract
Data associated with our paper “Near-term climate change impacts on Kenyan tree cover” are archived here. In this work, using L-Range – an ecosystem model – and the most recent downscaled climate projections we explored the durability and wider impacts of ongoing afforestation efforts across Kenya. These data represent input layers for the ecosystem model L-Range which is a localized version of the global rangelands model GRange. LRange retains GRange’s model architecture. However, within L-Range inference is constrained to the spatial extent of Kenya by creating requisite input spatial layers that most accurately represent land use, land cover, soil, and vegetation conditions extant across the nation. Using this model we explored how existing and planned tree cover across Kenya would respond under multiple future climate scenarios. Our simulations indicate that, under all scenarios, tree cover across Kenya will remain stable or show increasing trends in the near term. This will be accompanied by increases in overall woody vegetation, driven by shrub cover expansion and the contraction of herb cover areas. These impacts will be particularly pronounced in areas dominated by savannas and deciduous tree cover. Alongside these changes in vegetative cover, simulations indicate declines in net primary productivity and aboveground live biomass. The repository also includes a link to an interactive web-based decision-support application hosting results from this worl is also included.
Dataset DOI: 10.5061/dryad.sj3tx96j0
Description of the data and file structure
Overview
Near-term changes in Kenyan tree cover and broader ecosystem impacts were explored by conducting simulations using L-Range, an ecosystem model. At monthly time steps L-Range simulates plant regeneration, primary production, decomposition, competitive interactions between herbaceous and woody vegetation, as well as the cycling of nitrogen and carbon, and flow of water through ecosystems. The goal was to understand how climate change and fire can influence afforestation outcomes across Kenya in the year 2050 assuming that the Kenyan government achieves 10% tree cover extent by 2030 (baseline conditions).
Extant conditions across Kenya are represented using gridded datasets representing soil texture, distribution of vegetation classes (herbs, shrubs, and trees), land cover, biomes, and monthly climate data. The input data are at a 10 km resolution and simulations were restricted to areas dominated by natural vegetation (i.e., urban areas, deserts, bare areas, and water-covered areas were excluded). To represent past and future climatic conditions four climate scenarios or Shared Socioeconomic Pathways (SSP) were considered. For each scenario (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5), bias-corrected and downscaled monthly minimum and maximum temperature (tmin and tmax) and total precipitation (pr) projections from 13 Global Circulation Models (GCM) were obtained for a historic period (1950-2014) as well as for a future time period (2015-2100) . These data were resampled from their native resolution of 50 km to match the resolution of our analysis (10 km). Due to size constraints the climate data are not included within the LRange.zip files, however we link here the code to access these data. The downloaded files should be placed within the appropriate SSP sub-folder within the Layers folder. Simulations were repeated using climate data from each of the 13 GCMs for each SSP scenario. We include here the summarized results from these simulations, the R code used to produce these summaries as well as R code used to generate plots presented in the paper.
Files and variables
LRange.zip (Software, Zenodo) includes all program and input files associated with installing L-Range and running L-Range simulations. Input files are placed in the Layers subfolder and include data associated with Soil Composition, Vegetation Cover, Land Use and Land Cover, and Fire maps. Spatial layers are placed within the Layers subfolder. Data sources, and descriptions are detailed below. All input layers are included as ASCII files and were derived by subsetting and rescaling global datasets to represent Kenya's geographic extent at a resolution of 100 sq.km.
Soil composition: Contains layers representing 6 soil variables (bulk density, soil carbon, clay, silt, gravel, and sand) from the top and sub soil layers. These layers were obtained from The Harmonized World Soil Database (Nachtergaele, 2009).
Vegetation Cover: Layers representing herb, shrub, evergreen tree and deciduous tree cover across Kenya. These data were obtained from the Global 1km Consensus Land Cover classification dataset (Tuanmu and Jetz, 2014). Two layers representing a successful afforestation scenario (evergreen_enhanced.asc and deciduous_enhanced.asc) are also included.
Fire maps: Contains monthly fire probabilities for the period between 2000 and 2020 (Chuvieco et al, 2018). Files were used to generate an average fire probability layer from which fire probabilities were extracted for each landscape unit.
Land Use and Land Cover: Layers representing the landscape unit and land use classes (ESA, 2017), a file indicating the latitude of each landscape unit (latitudes.asc) and a unique landscape unit identifier (Zone_Id.asc) are also included.
Parameter files: Files sociated with setting parameter values for each landscape unit for the historic and double fire scenarios are included (LU_RamanFoley_reclass_updated_fut_(nf/df).grg) in the Parms folder. The folder also includes a legend file for the land use classes (goge_fin.leg) and a file (Sim_Parm.grg) that is used to set up the simulations. A given climate-fire scenario (e.g. SSP1 no fire) is set up by updating the Sim_Parm.grg file to reflect the path to the appropriate climate scenario sub-folder and the correct landscape unit file.
SimulationOutputProcessing_Rscript.R: R script with code used to process raw outputs from L-Range simulations. The code summarizes outputs from each simulation to generate results tables with summarized changes in tree cover, herb cover, shrub cover, aboveground live biomass, and mean monthly net primary productivity. % change in cover are measured by comparing cover characteristics in 2050 relative to the baseline year of 2030 such that positive values represent gains and negative values represent losses in cover.
Tree_responses.csv: Simulated tree cover changes from L-Range simulations from different SSP scenarios for different areas of Kenya are summarized.
Herb_Responses.csv: Simulated herb cover changes from L-Range simulations from different SSP scenarios for different areas of Kenya are summarized.
Shrub_Responses.csv: Simulated shrub cover changes from L-Range simulations from different SSP scenarios for different areas of Kenya are summarized.
AboveGroundLiveBiomass_Responses.csv: Simulated aboveground live biomass changes from L-Range simulations from different SSP scenarios for different areas of Kenya are summarized.
PrimaryProductivity_Responses.csv: Simulated mean monthly net primary productivity changes from L-Range simulations from different SSP scenarios for different areas of Kenya are summarized.
Results are organized under the same column headers in each of the 5 results files mentioned above:
- Year: 2050 - Mid-century climate change and disturbance impacts on afforestation efforts
- Areas: Distinct areas within Kenya within which afforestation-related outcomes were measured names, evergreen forested areas (Evergreen), deciduous forested areas (Deciduous), savanna dominated areas (Savanna), all areas with at least 10% tree cover (All), and areas that were newly afforested in 2030 (New).
- Scenario: The specific climate scenario that was considered (SSP1-RCP2.6, SSP2-RCP4.5, SSP3-RCP7.0,SSP5-RCP8.5)
- Fire: The fire disturbance scenario either Historic (past trends prevail) or Double (fire probability is doubled)
- GCM: 13 Global Circulation Model climate projections that were considered within each SSP scenario.
- value: % change in the variable of interest in 2050 relative to baseline values in 2030.
Manuscript_Plots_Rscripts.R: An R script with code used to generate vegetation and climate plots presented in the paper. The code can be used to call in any of the five .csv files included here.
The summarized simulation results from each SSP scenario and the plots were used to understand how climate change and fire scenarios will influence afforestation outcomes in Kenya in 2050.
Code/software
Simulations were conducted using L-Range a localized version of GRange: https://l-range.com/l-range.html.
The website hosts detailed tutorials on running simulations using L-Range. The adapted version of the model used in our simulations is included here (LRange.zip).
Simulation results were summarized using R (ver.4.4.0; R Core Team, 2024).
Simulation results can also be viewed using an interactive web-based decision support application that was produced in collaboration with the Kenya Forestry Research Institute and the Stockholm Environment Institute, Africa: Kenya Afforestation Decision Support Tool
Access information
Temperature and precipitation projections for the historic period and future climate scenarios were obtained from the Climate Impact Lab Global Downscaled Projections for Climate Impacts Research (CIL GDPCIR; Gergel et al., 2023) project. These data were accessed using code hosted here https://planetarycomputer.microsoft.com/dataset/cil-gdpcir-cc0#Ensemble-example
The downscaled gridded data are too large to include here. We, however, include summaries for the variables used in the study (Climate_data.RDS). The file includes year-wise ensembled means (mean across all GCMs) for mean monthly minimum temperature (tmin), maximum temperature (tmax), and annual precipitation (precipitation) for the Historic (1950-2015) and the four SSP scenarios (2015-2050). The change_clim object shows the percent deviations in temperature and precipitation in the future time period relative to mean conditions in the historic time period projected for each climate scenario (SSP).
Data were derived from the following sources:
-
Chuvieco, E., Lizundia-Loiola, J., Pettinari, M.L., Ramo, R., Padilla, M., Tansey, K., et al. (2018). Generation and analysis of a new global burned area product based on MODIS 250 m reflectance bands and thermal anomalies. Earth System Science Data, 10(4), 2015-2031. https://doi.org/10.5194/essd-10-2015-2018
-
ESA. (2017). Land Cover CCI Product User Guide Version 2. Tch. Rep. (tech. rep.). ESA. maps. elie.ucl.ac.be/CCI/viewer/download/ESACCI-LC-Ph2-PUGv2%7B%5C %7D2.0.pdf.
-
Nachtergaele, F., van Velthuizen, H. & Verelst, L. (2009). Harmonized World Soil Database Version 1.1. Rome, FAO and Laxenburg, Austria, IIASA. https://www.fao.org/3/aq361e/aq361e.pdf
-
Tuanmu, M.N., & Jetz, W. (2014). A global 1‐km consensus land‐cover product for biodiversity and ecosystem modelling. Global Ecology and Biogeography, 23(9), 1031-1045. https://doi.org/10.1111/geb.12182
-
Gergel, D. R., Malevich, S. B., McCusker, K. E., Tenezakis, E., Delgado, M. T., Fish, M. A., and Kopp, R. E. (2024). Global Downscaled Projections for Climate Impacts Research (GDPCIR): preserving quantile trends for modeling future climate impacts, Geoscientific Model Development, 17, 191–227, https://doi.org/10.5194/gmd-17-191-2024.
