Using Landsat satellite imagery for assessment and monitoring of long-term forest cover changes in Dak Nong province, Vietnam

Abstract

Forests are essential in regulating climate and protecting land resources from natural disasters. In Vietnam's Dak Nong province, forest cover has changed significantly between 1989 and 2021. This study applies remote sensing and geographic information systems (GIS) approaches to detect negative changes in forest cover as well as other land cover types. The maximum likelihood classification tool was used to classify Landsat images for the years 1989, 2001, 2011, and 2021, with post-classification accuracy evaluated through kappa coefficient statistics. The potential to based classification on Normalized Difference Vegetation Index (NDVI) and Soil Adjusted Vegetation Index (SAVI) to detect changes in forest cover compared with supervised classification was also evaluated. The land use and land cover change detection results show that the forest area decreased from 77.54% of the study area in 1989 to 33.97% in 2021, with a total forest loss of 2,953.48 km 2 and only 117.12 km 2 of newly planted forest during this period. Broadly, forest cover in the area has been severely reduced, often due to indiscriminate logging and expansion of agricultural land on the forest edge.

Assessing forest changes in Yok Don National Park and surrounding areas, Dak Lak province, Vietnam

“…An error matrix test was conducted to derive the percentage of the user's accuracy (UA) and producer's accuracy (PA). Then, the overall accuracy (OA) and kappa coefficient were calculated according to equations ( 3) and ( 4) (Thien and Phuong, 2023).…”

“…A visual representation of the classification methodology is provided in Figure2. The LULC classes included forest (natural vegetation, forestry, parks, and individual trees), agricultural land (cultivated outfields, homestead garden fields, and small scattered plots of grazing lands), water (river, wetlands, lakes, ponds, and reservoirs), bare soil (unused lands, empty lands, open space, fallow lands, earth/sand fillings, bare soil), and buildings (residential, commercial and industrial services, and transportation network)(Thien and Phuong, 2023).…”