PS 16-100 - Vegetation recovery after man-made fire in the Korean protected demilitarized zone (DMZ)

Tuesday, August 9, 2016
ESA Exhibit Hall, Ft Lauderdale Convention Center
Seon-Mi Lee, Ecological Conservation, National Institute of Ecology, Korea, Republic of (South)
Background/Question/Methods

After the end of the Korean War in 1953, a military protected area was established completely dividing the Korean Peninsula (DMZ). In this area of 907 µx, military activities prohibit the entrance of visitors, which in turn has resulted in moderate ecological disturbance and good overall habitat preservation. Despite this fact, the area has not been totally exempt from man-made disturbances. Some examples are the installation of barbed wire defenses, cutting down of trees to build military roads, man-made fire to increase visibility around the DMZ area and so on. Among such disturbances, man-made fires are likely to most important environmental impacts, decimating the vegetation. The largest fires occurred in the year 2000 and again in 2005 in the eastern area of the DMZ. Following such fires, some areas were reforested with Pinus densiflora, but most were neglected due to the unconfirmed land-mines. We did vegetation survey though the Braun-Banquet method at; 1) the burned but naturally recovered areas dominated by Quercus mongolica and Q. variabilis, 2) the burned but reforested area with P. densiflora, and 3) an undisturbed area as a comparative control. The vegetation recovery was assessed by species composition, species rank-abundance, and changes in the Normalized Difference Vegetation Index (NDVI) from Landsat images in 1999, 2001, 2006, 2010, and 2015.

Results/Conclusions

The results of the Detrended Correspondence Analysis (DCA) revealed distinct positioning of plots throughout the Axis I; the undisturbed area (left), Quercus mongolica and Q. variabilis (center), the reforested Pinus densiflora (right). Species composition analyses showed a higher similarity between the two Q. mongolica and Q. variabilis community areas and the undisturbed area than to the P. densiflora reforested area. Findings from the species rank-abundance curve showed a higher number of species from the undisturbed area than all other plots (undisturbed area; 97 species, Q. mongolica; 38species,Q. variabilis; 40 species, P. densiflora; 19 species). The P. densiflora reforested area understory vegetation was dominated by Quercus species. As a consequence those areas will show a secondary succession from P. densiflora to Quercus-dominated forest in the near future. We conclude that passive recovery in the studied area is more effective than active reforestation. As expected from vegetation recovery areas after fire, the Normalized Difference Vegetation Index (NDVI) decreased after the large 2001 fire, followed by a gradual increase. This study further validates the importance of Landsat Imagery to assess and analyze vegetation recovery in the restricted access areas.