Transition Potential Modeling of Land-Cover based on Similarity Weighted Instance-based Learning Procedure and Its Implication in the REDD Project Design Document

Abstract

Reducing Emissions from Deforestation and Forest Degradation (REDD) is a climate change mitigation strategy employedto reduce the intensity of deforestation and GHGS emissions. In recent decades, drastic land use changes in Mazandaran province caused a substantial reduction in the amount of Hyrcanian forests. The present research based on objectives of REDD projects paid to identify of forest cover changes in a range of Marzan Abad and Kojour Districts in the Mazandaran province using of Landsat satellite images from 1984, 2000 and 2014. In this study, for the first time in Iran, using similarity weighted instance-based learning (SimWeight) approach, forest cover changes modeling was performed, and for validation, statistics of relative operating characteristic (ROC), ratio of hits/false alarms and figure of merit was applied. Finally, using voluntary carbon standard (VCS) methodology CO2 emissions for the 30 next years (until 2044) was calculated. The results showed that forest cover decreased about 4008 hectares and 3635 hectares during 1984-2000 and 2000-2014. The validation results indicated that ROC equal to 0.95, the figure of merit equal to 26 percent and the ratio of hit/false alarms equal to 82 percent reflects high accuracy of the model. Eventually, REDD project's implementation results designated that under the baseline scenario about 705336 tCO2e will release into the atmosphere over the 30 next years that REDD project can prevent the release of 491697.91 tCO2e. With respect to increasing deforestation in Hyrcanian forests and their important role in the mitigation of climate change, using the methodology offered can be estimated and predicted land cover changes and the impact of REDD projects on reducing GHGS emissions, and the REDD results can be used to complete the Project Design