Temporal And Spatial Variation And Prediction Of Ecological Carrying Capacity In The Mountain-River-Sea Transition Zone Based On Machine Learning And PLUS Model

The mountain-river-sea transition zone is a complex,differentiated and comprehensive area with prominent contradictions between man and land.It is a complex area with multiple ecosystems.In recent years,my country’s high-speed development of science and technology economy has greatly improved productivity,but problems such as population growth,extensive use of land resources,and ecological damage have become increasingly prominent.In the process of reform and opening up,economic development has neglected ecological protection,and the development of many regions has exceeded the ecological environment carrying capacity of the region.In the "Several Issues and Opinions on Establishing a Longterm Monitoring and Early Warning Mechanism for Resource and Environmental Carrying Capacity" issued by the Central Committee of the Communist Party of China,it is pointed out that it is necessary to strengthen the protection of natural ecosystems of mountains,rivers,lakes and seas,control ecological overload,and strictly control the ecological environment caused by the unreasonable use of land.question.At the same time,land use change is the most obvious manifestation of human activities,and land use change can change the structure,pattern and service functions of ecosystems.Therefore,it is necessary to start from the land use to realize the research on the ecological carrying capacity,which is of great significance to the protection of the ecosystem and the scientific and rational utilization.In this study,the karst-Beibu Gulf in southwestern Guangxi was used as the research area,and the evolution of ecological carrying capacity in the transitional zone between mountains,rivers and seas was analyzed based on land use change.Comparative analysis was carried out and the temporal and spatial changes of land use from 1990 to 2020 were explored.The temporal and spatial evolution and evolution trend of ecological carrying capacity from 1990 to 2020 were studied,and the driving mechanism was detected.Based on this,the simulation accuracy of the PLUS model was verified.And predict the ecological carrying capacity pattern of the region in 2030 and 2040,and provide theoretical reference and technical support for the high-quality ecological development and spatial optimization of the region.This study draws the following conclusions:(1)Three machine learning algorithms were selected for land use classification,and the overall accuracy of different algorithms and the accuracy of different land types were compared and analyzed to determine the optimal classification algorithm—random forest algorithm.The overall classification accuracy for 4 years was 80%.% or more,and coupling multi-source data and classification algorithm can ultimately improve the overall classification accuracy of land use and the classification accuracy of each land type compared with a single algorithm.(2)From 1990 to 2020,the area of arable land and grassland gradually decreased,and both land use dynamic degrees were negative.They were concentrated in Nanning City,Yulin City,and hilly and flat areas away from water sources,mainly occupied by construction land;the forest land area gradually recovered,The dynamic degrees are all positive values;the water area fluctuates up and down,and the area changes little;the area of construction land increases gradually,mainly occupying cultivated land and grassland.(3)By revising the yield factor and equilibrium factor,and improving the ecological footprint model,the average value of the ecological carrying capacity per unit area of the four phases from 1990 to 2020 is 4.53,which belongs to the medium and high carrying capacity level.Taking 2000 as the turning point,the trend of rising first and then falling,showing a karst The mountainous area presents a downward trend as a whole,with medium and low bearing capacities as the main ones;the watershed and its surrounding areas are dominated by low bearing capacities;the Beibu Gulf coastal zone shows a slight downward trend as a whole.The ecological carrying capacity of cultivated land and construction land is on the rise,the transfer-in amount is mostly provided by forest land,and the ecological carrying capacity of forest land is on a downward trend.The transfer-in and transfer-out amount is basically provided by the grassland.The ecological carrying capacity was basically stable with a slight decline.The three time periods from 1990 to 2020 were all dominated by the type of ecological carrying capacity that remained unchanged.The increase in the area of ecological carrying capacity in each city was less than the decrease,showing a downward trend as a whole.The cities of Nanning,Qinzhou and Chongzuo suffered the most damage.(4)Combined with the analysis of the three detectors in the geographic detector model,the influence of driving factors on the spatial distribution of ecological carrying capacity is not independent of each other,and changes in any factor may affect the spatial distribution of ecological carrying capacity.main driver.(5)The PLUS model constructed by random forest algorithm(RF)and cellular automata(CA),the calculated Kappa coefficient value is 0.736,the accuracy basically meets the requirements,and can be used to predict the temporal and spatial pattern of ecological carrying capacity in 2030 and 2040.The ecological carrying capacity per unit area in 2030 and 2040 is between 1.299 and 8.832,both of which are medium to high carrying capacity levels,showing a transitional trend of increasing carrying capacity from the southwest karst area to the southeast coast.From 2020 to 2040,the area of low bearing capacity and low bearing capacity will decrease,and the area of medium bearing capacity,medium high bearing capacity and high bearing capacity will increase.The ecological carrying capacity of forest land and construction land showed an upward trend,while cultivated land,water area and grassland showed a downward trend.