Skip to main content

Historical and Forecast Land use Data for Ibadan, Nigeria 1984 - 2040

Cite this dataset

Eyelade, Damilola; Clarke, Keith; Ijagbone, Ighodalo (2022). Historical and Forecast Land use Data for Ibadan, Nigeria 1984 - 2040 [Dataset]. Dryad.


The SLEUTH model was implemented for Ibadan Nigeria, a city in West Africa with a mostly unregulated growth trajectory. The input physiographic and historical data for SLEUTH was compiled at spatial resolutions ranging between 15m – 90m. The temporal resolution of inputs was based on 35 years of historical data spanning the years 1984 to 2020. The urban (built-up commercial, residential, and industrial) land use increased almost four-fold from 3.0% of the total area in 1984 to 11.2% by 2020. The Forest extent was halved from 73.1% to 36.7% while agriculture and savanna extents have expanded over time.  Predictors of growth and the effect of tiling and spatiotemporal resolution on calibration and future growth forecasts with SLEUTH were determined. Transport and suburban edge spread were found to be the best predictors of growth. Model calibration metrics were found to generally improve with finer spatial resolution. Uncertainty reduced as the time interval between input land use data shortened but other accuracy metrics tended to improve as the time interval increased. An intermediate time interval and pixel resolution provided the best compromise of uncertainty and other accuracy metrics. Tiling the data allows a good fit within individual tiles but a reduction in pattern and trend recognition across land-use types. Year 2040 projections when optimized to allow the highest confidence and lowest uncertainty, suggest a continuation of rapid growth in suburban clusters and along transport routes. This collection of citywide data and growth forecasts for Ibadan can aid other researchers and planners interested in inputs for related projects and models in transport, water resources, planning, policy evaluation, and environmental management.


The land use and urban data were derived from Landsat satellite imagery. The overall accuracy (OA) of the classification for specific dates was 98.5% (1984), 80.7% (2000), 85.4% (2006), 87.5% (2013) and 94.5% (2020). Vector datasets for the period of study were derived from OpenStreetMap datasets and comparisons with historical satellite imagery.  All data have been reprojected to UTM Zone 31, WGS 1984 to allow for uniformity and measurements in metric units. Calibration was initially carried out using data spanning the entire study period. Subsequently, data up to 2013 was used to predict 2020 actual land use. Calibration validation was then carried out using pixel-based accuracy metrics. Finally, optimum calibration values were used for forecasting growth up to year 2040 at 30m resolution.

Usage notes

The files in this folder include :
1. An unzipped data folder named "Ibadan" with 6 subfolders described below:
    A. Ibadan_Physical_and_Transport_shapefiles: This folder contains input vector datasets including historical transport layers as well as areas of interest (AOI), rivers, waterbodies, Metropolitan Local Government Area (LGAs), Calibration and accuracy metrics for tiled data  (Ibadan_15m_metro and Ibadan_30m_metro).
    B. Ibadan_Physical_and_Landuse_Rasters: The folder includes all landuse and urban raster data at 15m resolution in GeoTIFF format for 1984 -2020. It also includes Exclusion, Hillshade, slope and DEM GeoTIFF rasters at 15m resolution.
    C. Ibadan_SLEUTH_GA_Inputs_and_Log_Files: This folder contains all input files in GIF format at 15 – 90m resolution,  scenario files for running SLEUTH as well as calibration log files
    D. Ibadan_SLEUTH_Ouput_Calibration_Validation_Image_and_Statistic_Files: The folder contains GIF image files from calibration validation at 15, 30, 60 and 90m resolutions as well as calibration validation statistics. 
    E. Ibadan_SLEUTH _30m_Forecasts_2020_to_2040: This folder contains tiled and untiled 2040 forecast results at 30m resolution.
    F. Figures_From_Paper: This folder contains the figures found in the companion paper

2.  A zipped file containing bash scripts for running preprocessing steps

3.  A zipped folder containing custom c coded for tiling SLEUTH input datasets

4. A README file with information about the provenance, overall structure, organization, and relationships between files 
   as well as links to download SLEUTH and other required stand alone software.