| Forest disturbance events can have a strong impact on the distribution,structure and function of forest ecosystem.The scarcity of data on forest disturbance characteristics(size,severity,etc.)creates uncertainty in carbon estimation and bottlenecks in sustainable forest management.Although there are already global forest disturbance products from Hansen’s team,whether its methods and accuracy can be further improved at the regional scale,especially in the subtropical regions where disturbances are frequent and cloudy,remains to be explored.Therefore,with Hunan Province as the research area,this study intends to explore the most suitable spectral index combination for subtropical forest disturbance monitoring,so as to more accurately reveal the temporal and spatial characteristics of forest disturbance.Based on the LandTrendr algorithm,this study generated the forest disturbance data set of Hunan Province from 1991 to 2021,revealed the temporal and spatial dynamics and trend characteristics of forest disturbance,explored the change trend of two disturbance characteristics(patch size and severity)of forest disturbance events in Hunan Province in the past 30 years,and analyzed the impact of disturbance on the landscape index of prefecture-level cities.The main results of this study are as follows:(1)The study found that the superposition of the five spectral indices B5,B7,NBR,TCA and B3 can obtain the optimal interference monitoring results,and the overall accuracy reaches 86.39%,which is higher than the accuracy of Hansen products in the same region(82.89%).From 1991 to 2021,the total forest disturbance area of Hunan Province was 11103.25 km2,accounting for about 10.54%of the total forest area of Hunan Province.The change in forest disturbance area generally showed a trend of first increasing and then decreasing,with a peak around 2010.(2)The average forest disturbance rate of prefecture-level cities in Hunan Province increased in the first three periods(1991-1995,1996-2000 and 2001-2005),and reached its peak in the fourth period(2006-2010).The average interference rate of the last two periods(2011-2015 and 2016-2021)began to decrease gradually.The average disturbance rate varies greatly among prefecture-level cities.For example,Changsha City had a maximum disturbance rate of 0.95%(2006-2010),while Zhangjiajie City had a maximum disturbance rate of only 0.15%(2011-2015).(3)The two major disturbance characteristics of forest disturbance in the study area(disturbance patch size and severity)showed a gradual increase trend,and there were more and more large-scale and high-severity disturbance events.The spatial scope of the disturbance gradually expanded from scattered areas to large-scale province-wide.The median average disturbance severity and mean plaque size over the six time periods shifted to the right over time.(4)The ENN index(Euclidian nearest neighbor distance index)of forest disturbed patches in all cities showed the same general downward trend,but the ENN index value of all cities increased during 2016-2021,indicating that the occurrence of disturbance events is becoming more and more concentrated.The PERIM index(patch circumference index)of forest disturbed patches showed the same overall increasing trend among prefecture-level cities,indicating that the area of disturbance patches was becoming larger or more irregular.In this study,we explore the combination of spectral indices most suitable for monitoring subtropical forest disturbance,and reveal the spatiotemporal variation characteristics of forest disturbance in Hunan Province in the past 30 years.On the one hand,this study finds a method that is more suitable for characterizing subtropical forest disturbance,and on the other hand,it provides more accurate spatial and temporal information on the dynamic changes of forest disturbance for evaluating the impact of forest management and policy and near-real-time forest adaptive management. |
PDF Full Text Request