Forest Degradation Estimation Through Trend Analysis of Annual Time Series NDVI, NDMI and NDFI (2010–2020) Using Landsat Images

Abstract

Forest degradation plays an important role in greenhouse gas (GHG) emissions and climate change. Previous research has shown that more GHG has been emit-ted through forest degradation than deforestation. Therefore, its monitoring and estimation is important for strategy design to combat climate change. In this work, we intend to estimate forest degradation in Ayuquila River Basin, Mexico through vegetation trend analysis using annual time series vegetation indices (2010–2020) specifically, NDVI (normalized difference vegetation index), NDMI (normalized difference moisture index), and NDFI (normalized difference fraction index) derived from Landsat images. The vegetation trend analysis was carried out using a linear regression model and tested by Mann-Kendall for significance. Slope coefficient was used to indicate the vegetation trend: positive slope indi-cates vegetation regrowth and negative slope indicates vegetation degradation. For forest degradation, only significant trends with negative slope were analyzed (p < 0.05). To discard negative trends due to deforestation, a forest mask was ap-plied both at the beginning and at the end of the analysis. The accuracy assessment showed that the forest degradation estimation by time series NDVI obtained the highest overall accuracy of 81.33%, followed by NDMI with 73.33% and fi-nally NDFI with 72%.