Research On Forest Change In Changbai Mountain Using RS And GIS

Forest landscape change is one of the main problems of global environmental change and has an important impact on the global carbon cycle.After the founding of our country,the forest has undergone a large-scale deforestation and restoration process,especially after the reform and opening up,population,economic growth and policy implementation caused relatively dramatic changes in forest land in a relatively short period of time.Northeast Changbai Mountain is an important primary forest area in the country.At present,commercial logging has basically been stopped,and it is gradually moving towards artificial and natural restoration processes.From the previous largescale deforestation to "returning farmland to forest",the change of forest landscape pattern in Changbai Mountain has important representative value for studying the changes of forest landscape after reform and opening up in China,especially the forest landscape process in Northeast China.Remote sensing technology is widely used in global land change monitoring,providing spatio-temporal dynamic data of forest changes at various scales,and GIS technology is used for the analysis of landscape patterns and processes.In recent years,Forman's fragmentation theory is gradually being used to obtain the spatial process of forest change,to assist in the analysis of forest change processes and regional ecological status.The ANN-CA model combines "top-down" and "bottom-up" simulation methods can use different driving factors to simulate landscape changes more reasonably,which helps to analyze the effect of Forman landscape fragmentation theory applied to forest change simulation.In this paper,Landsat satellite image interpretation was used to obtain the five phases of land use data in the Changbai Mountains from 1977 to 2018.Based on the land use data,the spatial process of forest change was generated in a new way.Then,the forest landscape change trend and its driving mechanism were analyzed.In addition,the spatial process of forest change was used as the driving factor to input the ANNCA model for simulation and prediction the forest change in Changbai Mountain.The specific process is as follows:1)Obtained Landsat satellite images which can cover the entire area,and interpreted the five-phase land use data(1977,1990,2000,2010,2018);2)According to the Forman theory,analyzed the differences in the six forest spatial processes(including: perforation,fragmentation,degradation,disappearance,expansion and increase),determined the topological relationship between each process and multiple land-use categories,and introduce algorithms for generating spatial processes;3)Analyzed the forest change in Changbai Mountain and the driving mechanism behind it based on the land use data and the spatial process of the forest change;4)Established ANN-CA model based on BP neural network,then the forest spatial process and other driving factors were regarded as input variables of ANN-CA model.Hierarchical random sampling was used to extract the sample data for training the model,and the connection weights of the neural network were calibrated to simulate forest changes in 2010-2018.And the corrected weights and drivers were used to predict future forest changes;5)The simulation results ANN-CA-Forman(The process of forest spatial change was added)were compared with the results of the original ANN-CA model,and the effect of the forest change spatial process on the simulation of forest changes was analyzed.The results show that the total forest volume in the Changbai Mountains decreased during 1977-2000 and increased during 2000-2018.The loss of forests occurred throughout the research process,and most of them were converted into cultivated land and grassland.The development of agriculture and forestry is the main cause of forest loss.The restoration of part of the grassland to forest from 1990 to 2000 originated from the local forestry department and the management was strengthened after completing the system reform.After 2000,under the policy of returning farmland to forests,cultivated land was the main land category for forest restoration.The large number of grasslands was the result of regular logging and renewal of forests by commercial institutions after obtaining permission from the forestry department.The policies of different periods made the forest fragmentation of Changbai Mountain first intensify and then slow down.The fragmentation process was the major forest loss process from 1977 to 2000 and the degradation process dominated after 2000.The process of forest restoration was always centered on expansion,which occurred after 1990.The amount of forest recovery reached its maximum during 2000-2010 and slowed down after 2010.Forest loss and recovery basically reached a balance between 2010 and 2018,and the forest area increased slightly.Statistically,the accuracy of the forest simulation of ANN-CA-Forman reached 80.10%,which is 16.17% higher than the original ANN-CA model.In the spatial distribution,the ANN-CA-Forman model reduced the probability of erroneous plaques,which could simulate the changes in the details made the result closer to the real scene.The forecast of forest change in 2026 showed that the total forest would decrease by 256.5774 km2.The increase in the proportion of the fragmentation process indicated that the Changbai Mountains still had the potential risk of increasing fragmentation degree in the future.In general,the forest area of Changbai Mountain experienced a more obvious harvesting-restoration process from 1977 to 2018,and the forest area experienced a process of decrease-increase-minor increase.The degree of landscape fragmentation also gone from high to low.The ANN-CA-Forman model achieved better results than original ANN-CA-Forman model in simulating the forest change process.The prediction results also indicated that in accordance with the change trend of the forest after 2010,the forest area of Changbai Mountain forest area still had a reduced risk in the future.