Landscape Analysis And Quality Evaluation Of Laoshan Scenic Area Based On Forest Resource Monitoring System

Laoshan scenic area is the ecological barrier and biodiversity conservation area of Nanjing.Forests are the main body and ecological location is extremely important.Grasping the level of ecological landscapes and the quality of forest resources in Laoshan scenic area can effectively provide a strong theoretical basis for local governments to formulate major policy decisions and forestry bureaus to formulate forest management plans.It also responds to the government's construction of the“Eco Pukou”slogan and as an important link in“Green Nanjing”.This study took Nanjing Laoshan scenic area as the research object,combined the survey data of forest resources monitoring system with 3S technology,extracted the information of the Laoshan scenic area,and then used Fragstats4.2 to extract the landscape pattern indexes at three levels.Analyzes the landscape pattern dynamics of Laoshan Scenic area from 2001 to 2017.At the same time,the landscape pattern index is combined with the forest management inventory data to construct the quality evaluation system of forest resources in Laoshan scenic area based on the subcompartment scale and analyzes the quality of scenic forest resources from different aspects.The main research contents and conclusions are as follows:(1)Land type information extraction.The multi-scale segmentation of the Landsat image in the Laoshan scenic area is combined with the ROC parameters to select the optimal segmentation scale and the weight of the heterogeneity factor in each level.The theoretical optimal feature space is selected by the category distance matrix.Then,classifer the Landsat image at different maximum dimension and statistics the classification accuracy.The feature space corresponding to the highest classification accuracy is taken as the optimal feature space.According to the fuzzy statistical method,the optimal membership functions of different features at each category are determined,and a fuzzy multi-classifier model is constructed based on the nearest neighbor classifier and the fuzzy rule classifier.The research classifies the scenic area into five categories:woodland,agricultural land,construction land,unused land and water area and extracts land information in Laoshan scenic area.At the same time,the object-based C5.0 decision tree and pixel-based maximum likelihood method,SVM method and ANN method are used as comparison experiments.Among the object-oriented classification methods,the fuzzy classification method has the highest accuracy,the overall accuracy is 91.79%,the C5.0 decision tree method is the second,and the overall accuracy is 89.2%.The woodland information has higher extraction precision in the two methods.The fuzzy classification method is used to extract woodland information with a precision of 93.56%,the C5.0 decision tree method achieves90.31%.Based on the pixel classification method,the SVM method has the highest precision and the overall accuracy is 88.09%.The ANN method has an overall accuracy of 87.89%and the maximum likelihood method has an overall accuracy of 84.91%.At the same time,the object-oriented fuzzy classification and C5.0 decision tree classification can better solve the problem of“pepper-salt phenomenon”which is commonly found in pixel-based classification.To sum up,this study used fuzzy classification method to extract the land use information of Laoshan scenic area in 2001,2005,2009,2013,and 2017.(2)Analysis of scenic landscape patterns.At the landscape level,the landscape patterns of Laoshan scenic area in 2001,2005,2009,2013,and 2017 were analyzed using PAFRAC,COHESION,SHEI,and other landscape pattern indexes.With the construction and continuous reform of scenic eco-parks,the degree of landscape fragmentation in scenic areas has deepened.From 2001 to 2017,the number of patches has increased and reached 653 in 2017,and the patch density reached 5.9281;2017 The CONTAG index increased compared with 2001,reaching59.4094%.The connectivity of the dominant patch type increased;the SHDI value fluctuates narrowly and reaching a maximum value of 0.7255 in 2009.At the class level,LPI,PLAND,IJI,and other indexes were used to analyze the types of patch.Among them,woodland was the main type in the scenic area.LPI of woodland was much higher than other class,and the classes were clustered and dispersibility is more sensitive to disorganized human activities.IJI fluctuates greatly,but the overall trend is upward.With the mature construction of the scenic area,the overall development process is slowly entering an orderly state,and the classes tend to be balanced.distributed.(3)Quality evaluation of scenic forest resources.Combining the landscape indexes with the forest management inventory data to construct the quality evaluation system Laoshan scenic area forest resources.The system proceeding from the three principles of objectivity,real-time,and easy access,and includes ecological stability,forest health,forest landscape contains 6 indicators,and 16 elements based on the subcompartment scale.The AHP method is used to determine the weight of each index.The equidistant grouping method divides the Laoshan scenic area's forest quality evaluation criteria into five grades:excellent,good,medium,poor,and inferior.There are more middle quality subcompartments in Laoshan scenic area,accounting for 53.91%of the entire arbor forest,and few excellent quality and inferior quality subcompartments accounting for1.5%and 3.12%respectively.Among the different forest types,the area of special purpose forests is the largest,with 6708.77hm~2,among which 53.85%of the total area of special purpose forests is occupied by medium quality subcompartments;the smallest timber forest area is78.4hm~2,and 38.6%of the timber forests are inferior subcompartments equal to 13.8%of the scenic area.Among different types of mixed forests,the forest quality of coniferous forests is the worst,the forest structure is irrational,and the natural succession process is slow.The Laoshan scenic area needs to carry out targeted,effective and reasonable forest tending in time,actively adjusts the structure of the stand,and focus on the improvement of the tending and stressing measures for poor quality subcompartments and inferior quality subcompartments.