Assessing Local Expert Data Quality for Forest Monitoring; A Case of Kafa, Ethiopia

Organisator Laboratory of Geo-information Science and Remote Sensing

wo 9 april 2014 09:30 tot 10:00

Locatie Gaia, building number 101
Droevendaalsesteeg 3
6708 PB Wageningen
+31 317 48 16 00
Zaal/kamer 1

by Elias B. Gebremeskel (Ethiopia)


Recent advancement in spatial data collection technologies dramatically increases the contribution of ordinary people to collect and disseminate geospatial data. At the same time, there is an increasing general agreement that community based forest monitoring can play a crucial role in producing and sharing information about the condition of forest resources in time and space. Despite the advantages of the community based monitoring, there are also a lot of doubts and concerns that existed in the scientific community related to the quality of the data. Therefore, this research is aiming to assess the quality of forest monitoring activity data sets, which is collected by local experts in Kafa Biosphere Reserve in Ethiopia. The research was conducted to test the quality of  local experts data for REDD+ mechanism to track the forest change and carbon emissions. In this research, we examines the quality of local experts data relative to the reference data sets of remotes sensing time series images of 2000 to 2012, GIS data sets, and ground based validation measurements. The main variables are date of forest disturbances, size of the forest disturbance, drivers information, location and coverage of forest disturbances. The spatial variables of the local experts data were assessed using the spatial data quality parameters whereas the temporal variables were compared through BFAST monitoring on Landsat time series images and Visual interpretations on high resolution images of Spot and Rapid Eye. The results show that the local experts can perform and produce quality data comparable to validation measurements by experts. We found a regression correlation value of 0.84 for area/size estimation and ~65% percentage of correctly classification accuracy of drivers information of forest disturbances. Furthermore, the result confirms that local experts have a short time delay in detecting forest disturbances compared to high resolution remote sensing time series data of Spot 5-Rapid Eye satellite images than of Landsat imagery. Based on this case study results, we suggests that the local expert data can enhance the quality of forest monitoring data of remote sensing particularly in detecting near real time forest disturbances.

Keywords: Community Based Forest Monitoring; Spatial Data Quality; Local Experts; REDD+; Ethiopia.