Scenario-based land cover change modeling and its implications for landscape pattern analysis in the Neka Watershed, Iran

摘要: Abstract Land cover changes and urbanization cause destruction of natural habitats threaten biodiversity. modeling is one the most important procedures for evaluating this trend. This study was performed with objective comparing multi-layer perceptron (MLP) artificial neural network logistic regression (LR) in predicting land change quantifying future landscape using metrics Neka River Basin, a small part eastern Hyrcanian forest, northern Iran. For purpose, first, analysis carried out satellite imagery, from 1987 to 2011. Then, transition potential conducted MLP LR 5 different scenarios. A Relative Operating Characteristic (ROC) detect degree correlation between variables transitions LR. In addition, accuracy rate assessing employed. prediction 2011 2017. The assessment model determined by actual map predicted Landscape indices 1987, 2001, 2006, 2011, 2017 were calculated analyzed Fragstats determine impact on fragmentation. result showed that during 1987–2001, agriculture main contributor increased built-up area. conversion orchard residential, 2001 2006. Forest regenerated agricultural lands, 2006 maximum minimum amounts Cramer's V obtained empirical likelihood variable distance river. Overall kappa best scenario based 88%. Furthermore, results major deforestation will occur surrounding forest areas residential development outskirts town Neka. Increased fragmentation continue 2017, more shape complexity be observed, Basin become diverse abundant. provide useful information Reducing Emission Deforestation Degradation (REDD) project.