Post-fire Forest Recovery Study With Multi-source Remote Sensing

Forest is the main component of the whole terrestrial ecosystem. It provides not only eccential living sources for human beings, but important ecological functions for terrectrial surfaces on earth. In recent years, periodic outbreaks of forest fire have been observed that showed different fire characteristics and intensities in differet climates, land surfaces, and human activities. Forest fire would not only affect the ecological balance of natural ecosystems, but also cause serious degradation of landscape patterns and forest productivity. Also, the process of post-fire forest recovery is complicated, varying with fire intensities and human interferences. Therefore, quantitative information of forest biophysical characteristics and sptio-temporal patterns is of great importance to better understand post-fire forest recovery and to assist sustainable management. The forest fire prevention system based on optical satellite remote sensing and GIS technology can solve the disadvantages of the traditional investigation methods, but the study found that the vegetation index of forest canopy tends to be saturated when the canopy leaf area index reached 4.0, and the green vegetation layer in rapid recovery in 3-5 years after disaster in optical image which Is not conducive to the long-term monitoring of forest restoration. Active microwave remote sensing technology with the help of the radar signal is not affected by cloud effects and can penetrate the canopy to characteristic of wood layer. An empirical linear regression model was developed to estimate the amount of forest stock based on multi band multi polarization radar backscatter coefficient and forest volume in previous research.Based on the 1987 catastrophic fire in Greater Khingan Mountains, Northeast China, this study tests the capability of quantitative mapping of forest recovery with multi-source remote sensing. The study area is the TUQIANG Forest, Greater Hhingan Mountain Forest Bureau. The L band ALOS PALSAR image in 2010 and C band Radarsat-2 image in 2011 serve as the primary data source. A Landsat TM image, the 2010 forest inventory data and the 2015 field measurements are collected to establish the relationships between biophysical parameters(volumne) and SAR backscattering coefficients under different fire intensities. Primary findings the studies are as follows:(1) Adopting an index of normalized burn ratio, the fire intensity map a fire intensity map of the 1987 catastrophic fire is extracted from the Landsat TM image.(2) The correlation anlaysis shows that, in the same poliarization(HH), backscattering coefficients of the L band ALOS PALSAR image has higher correlations with forest volume than C band Radarsat-2. Consequently, the L band ALOS PALSAR image can better reflect the distribution of forest volume.(3) Among different polarizations(HH, HV and HH/HV) of the L band ALOS PALSAR image, the HV image performs slightly bettern than HH, and much bettern than HH/HV combination, in regards to their relationships with forest volume.(4) The L-HV ALOS PALSAR image better reflects the recovery of burned area in Greater Khingan Mountains at a saturation point with forest volume of 100m3/hm2.(5) By adopting the idea of space instead of time to assume that the forest volume before the fire and the forest volume of unburnt area approximate. From this perspective, severely burned area of forest also did not recover to the fire volume level before the fire after 23 years, forest restoration is the fastest in medium fire area, the light area almost reaches forest volume level before the fire. Generally speaking, Forest volume recovery is slow after the fire, and and has a strong negative correlation relationship with the fire intensity.(6) Topography(slope and aspect) slightly affects forest recovery in the study area, with the south-facing areas recovering better than other areas. There is no significant correlation between slope and forest recovery in the forest fire area of low intensity, and is significant correlation in high intensity area. It presents certain positive correlation between slope and forest recovery.