Forest Disturbance Monitoring And Trend Analysis Based On Time Series Of Landsat Data

Vegetation dynamic change monitoring by time series has always been the key and hot issues of remote sensing of environment.Forest is an important component of the terrestrial ecosystem,which dynamic change affects global terrestrial carbon cycle and carbon accumulation.It is an important significance to the terrestrial carbon cycle and climate change research by using time series data to monitor forest dynamic change.Due to the impact of the natural disasters and human activities,forest vegetation suffered frequent damage.In this paper,Changting County of Fujian Province as the research area,the research was on forest disturbance monitoring of 2000?2014 in Changting county,vegetation trend analysis and vegetation restoration of disturbance areas with Landsat time series data.The main research contents and results are as follows:(1)The preprocessing of Landsat time series data.Due to the impact of atmosphere,cloud and mist in the acquisition process,the preprocessing were radiation calibration,atmospheric correction,cloud and its shadow masking and filling the masking null value in before using the images.(2)Forest disturbance monitoring.Based on the theory of time series trajectory analysis method(LandTrendr),we monitored forest disturbance,and extracted key information of Changting county from 2000 to 2014.The change trajectory of LandTrendr is determined by the data,and we obtained vertices information from the change trajectory to extract key information of forest disturbance.The amount of forest disturbance in Changting county showed an increasing trend from 2000 to 2014.In 2008,forest disturbance reached the maximum capacity of 90km,followed by 2004 and 2009.The results are verified based on field survey,the visual interpretation and pixel of interference.The result of field survey was consistent with the monitoring result,the result of visual interpretation was consistent with the time of monitoring result by LandTrendr.Based on pixel scale accuracy verification to monitor forest disturbance showed that the overall accuracy for 87.74%and kappa coefficient to 0.79,showing higher accuracy of forest disturbance monitoring.(3)Vegetation trend analysis.Analysis of vegetation time series trend in Changting County by Sen+Mann-Kendall analysis,the result showed that Changting County forest area mainly showed no significant trend from 2000 to 2014,1.49%of the region showed significant degradation trend,which mainly caused by the forest disturbance,and 0.63%of the region was significantly increased,which distribution around non-forest area in central of Changting Country.For the region of forest disturbance,we analyzed the trend of pre-disturbance and post disturbance of the region showed no significance trend,the time of vegetation restoration and the state of vegetation recovery.Before the forest disturbance the most region showed an increasing trend with intensity of 0?0.05,after the forest disturbance the vegetation gradually recovered,6.03%of vegetation change intensity was 0?0.1.Most region of the disturbance was recovered very well within 6 years.Vegetation recovered 40%?80%within 4 years after the forest disturbance,and 0.6%of the regional vegetation recovered faster,which growth state was better than the pre-disturbance state within 4 years.