Spatio-temporal Dynamics of Forest Cover in Mid-Upstream Basins of Taedong River in DPR Korea Using Landsat Time Series.

Deforestation has become a critical concern that brings about changes in the global ecological environment. The forest inventory is the first step toward the adoption of the necessary decisions at the national and regional level for the protection of the ecological environment and the sustainable development and use of forest resources. The present study analyzed the forest cover change dynamics in the mid-upstream area of the Taedong River in the Democratic People's Republic of Korea during 1990–2020 by applying remote sensing technology. For the analysis, annual image datasets layered with spectral index bands fitted to vertices of the normalized burned ratio segment curve were created by implementing LandTrendr algorithm in the Google Earth Engine platform. The sample dataset used to conduct random forest classification was collected by sampling from pixels in which ecological change events were not observed during the study period. The overall accuracy was 95.12 ± 0.14% in all years. The results showed that over the past 20 years (1990–2010), the forest area of the study area decreased by 652.5 km², accounting for 13.6% of the total forest area in the starting year (1990). The annual rate of deforestation in the study area between 1990 and 2000 was 1.043% year⁻¹, which was much higher than the global deforestation rate (0.184% year⁻¹) during the same period. For the period from 2000 to 2010, the deforestation rate was 0.423% year⁻¹, 3.4 times the global deforestation rate (0.125% year⁻¹). During these two time periods, deforestation in the study area was substantial. However, from 2010 to 2020, the forest area showed a tendency to increase, especially in the midstream region where it increased significantly. The average growth rate of forest area in the midstream region was much higher than that in the upstream region, where there was no significant change in forest area between 2010 and 2020. The current forest area of the research area was estimated at 92% of the forest area in 1990 (4783.9 km²). [ABSTRACT FROM AUTHOR]