The data contained in the tables correspond to the alerts generated by the Terra-i Ethiopia system. Each alert generated by the system is equivalent to approximately 6.25 ha.
Click on the following links to download the data
National level dataset:
Table with statistics at national level (csv)
Other administrative levels:
Table with statistics at Regionals level (csv)
Table with statistics at Zones level (csv)
Table with statistics at Woredas level (csv)
Table with statistics at Protected area level (csv)
To generate these data, the data were first projected onto a UTM projection (datum WGS84).For the case of the Terra-i raster, the pixel size was set to a value of 250 m.
For a correct interpretation of the data the following indications should be taken into account:
● Terra-i is a system that detects forest changes. Thus, for a proper interpretation of forest loss (deforestation), changes should be queried using the eco-region layers representing forests. Another option is to download the geo-referenced data, and then perform the analysis based on the extracted data using reference layers by the user;
● If a MODIS pixel is detected by Terra-i it is assumed to be completely changed even if this alteration was only partial over the total area of the pixel. This can usually lead to an overestimation of the true value;
● The spatial resolution of Terra-i (~250 m) does not allow observation of small-scale events. The Terra-i alerts presented in the tables and reports should therefore be used as an indicator of the rate trend for multi-temporal analysis every 16 days or annually. However, it should never be considered as the exact figure of change in the analysed area as high resolution data must be considered for this. Terra-i is a good tool to provide alerts of coverage change events to support decision making.
DATASET: The data distributed here is in CSV format for different administrative level. These data are derived from the Terra-i output at 250m spatial resolution, in decimal degrees and datum WGS84. It is derived from the USGS/NASA MODIS data. CIAT processed this data to provide habitat change maps. The detections were made using algorithms described by Reymondin et al. (2012).
Additional data sources:
● Land use Land cover map of Ethiopia for the year 2003 (Source: RCMRD)
● National forest and non-forest map: generated by EFD for the year 2020.
DISTRIBUTION: Users are prohibited from any commercial, non-free resale, or redistribution without explicit written permission from CIAT. Users should acknowledge CIAT as the source used in the creation of any reports, publications, new data sets, derived products, or services resulting from the use of this data set. CIAT also request reprints of any publications and notification of any redistributing efforts. For commercial access to the data, send requests to Louis Reymondin (email@example.com).
NO WARRANTY OR LIABILITY: CIAT provides this data without any warranty of any kind whatsoever, either express or implied, including warranties of merchantability and fitness for a particular purpose. CIAT shall not be liable for incidental, consequential, or special damages arising out of the use of any data.
ACKNOWLEDGEMENT AND CITATION: We kindly ask any users to cite this data in any published material produced using this data, and if possible link web pages to the CIAT-Terra-i website (www.terra-i.org).
REFERENCE: Louis Reymondin, Andrew Jarvis, Andres Perez-Uribe, Jerry Touval, Karolina Argote, Julien Rebetez, Edward Guevara, Mark Mulligan (2012), A methodology for near real-time monitoring of habitat change at continental scales using MODIS-NDVI and TRMM. Submitted Remote Sensing of Environment.
The georeferenced data is available for download in *.asc format, which can be used in GIS software such as QGIS, ArcGIS. The data is available from Jan 01, 2004, to currently, updated every 16 days when new satellite images are available. The annual raster accumulates the alert by year and the detection of vegetation and loss forest every 16 days raster present per year the loss accumulative every 16 days in Ethiopia. The start and end date of the period are shown in the file name.
Limitation: These data represent a beta version of our detection method using Sentinel-1. Therefore, it needs calibration and validation.
The georeferenced data is available for download in *.shp format, which can be used in GIS software such as QGIS, ArcGIS. The data is available from Jan 01, 2020 to 2022, updated every 12 days when new satellite images are available. Each shape file presents an update with all the forest loss alerts in REDD PFM sites within the Bench Maji Zone in southwest Ethiopia. that are detected within the time period. The start and end date of the period are shown in the file name.
This near real time forest change detection communication and visualization tool is developed for Ethiopia through the collaboration between the Ethiopia forest development office, UNDP, CIFOR-ICRAF, and Alliance Bioversity-CIAT, as part of the REDD+ Investment Plan Phase I project.
The general objective of the project is to provide decision makers, land use planners and local forest rangers, researchers, and the general public access to near real-time data on forest change (loss) at a national level, and to pilot a refined Terra-i system at sub-national levels.
Terra-i is a collaboration between the Alliance Bioversity & CIAT (Alliance Bioversity & CIAT, based in Colombia), The Nature Conservancy (TNC, world environmental organization), Forests, Trees and Agroforestry Program(FTA), la Escuela de Negocios e Ingeniería (HEIG-VD, con sede en Suiza) y Kings College de Londres (KCL, con sede en el Reino Unido).
Terra-i has at its core a tool that detects land cover changes resulting mainly from human activities in near real time, producing updates every 16 days. It currently runs for the whole of Latin America and all the tropics.
The system is based on the premise that natural vegetation follows a predictable pattern of changes in greenness from one date to another caused by site-specific land and weather conditions over the same period. The so-called neural network is "trained" to understand the normal pattern of changes in vegetation greenness for a site in relation to terrain and rainfall. Sites where the vegetation greenness suddenly changes beyond the normal limits of the predicted values with respect to the actual satellite image values are then marked as areas of change. These processes are run on several servers and is updated with new images every 16 days and with a resolution of 250 m.
Ethiopian forest development, Development fund Norway, EthioWetlands, UNDP, CIFOR-ICRAF