Monitoring Regional Forest Disturbance By Remote Sensing

Forest ecosystems, which are major parts of the terrestrial biosphere, play an important role in terrestrial carbon cycling and storage. However, the accuracy of regional forest carbon-flux estimation is greatly influenced by the lack of forest disturbance data. Thanks to large coverage, comprehensive and dynamic monitoring, remote sensing is becoming a useful tool for disturbance monitoring. How to the use a long time-series data to extract disturbance information is one of the significant problems for current study. In a case study of Wuning County in Southern China, this research adopted14Landsat TM/ETM+Images from1986to2011to study the time-series forest disturbance and to provide input parameters or regional verification for regional and global carbon budget study.Based on the spectral characteristics of forest subjected to disturbance and the sensitivity and anti-jamming capability of different forest detection methods, this paper analyzed the detection capability of different methods (NDVI, NDMI, NBR, IFZ and DI). The results show that all of the five detection methods can effectively distinguish disturbance information by using the remote sensing images of the year when disturbance occurred, however, DI and IFZ method are still have the ability to distinguish disturbance after three years, their Kappa coefficients are0.78and0.68respectively. In addition, the response of DI and NBR index are the most susceptible in all five indexes, the absolute difference reaches to8.61and6.74, respectively. Moreover, IFZ, NDMI and NBR have the ability to resist atmospheric interference. Therefore, we can say that DI and IFZ method are better for remote sensing monitoring of forest disturbance in Wuning County.In this study, we analyzed the forest disturbance-monitoring technology suitable for forests in Southern China based on a time-series trajectory analysis method. We selected Wuning County, where forest disturbances are frequent, as our research area. Reforestation efforts have also increased in this area. The method we used not only identifies repeated disturbances but also monitors forest recovery. In comparison with field observation, the DI time-series trajectory analysis method can effectively extract the forest disturbance information in Wuning County, and the Kappa coefficient of our disturbance products reaches to0.80with an overall accuracy of89.7%, showing a great potential of the technique for forest disturbance monitoring.By comparing the different spectral bands of Landsat, B5shows a great potential for separating the burned area from logged area, the difference between the two reflectance reaches to0.106. Furthermore, field surveys indicate that the region growth algorithm can be used to separate the burned area from logged area precisely. Through the analysis of the time-space disturbance characteristics of forest disturbance, we found that Wuning County suffered from the most dramatic disturbance in the1990s, especially in1992, and the trend of disturbance was going down from1986to1998. In contract, forest recovery was going upward from1992to2011. The most of the disturbance occurred in low-elevation areas where near roads and water and mainly caused by human activities.