Forest Landscape Structure And Characteristic Analysis Of LouFan County Based On GIS And ANN

This study was conducted the application of the theory and method of Forest Landscape Ecology, basing on data of the Loufan county forest resource and the supporting of GIS to finish quantitative analysis of the forest landscape pattern in Loufan County. Then, the BP artificial network model technology was introducing and taking Loufan County of forest landscape pattern in elevation, residents land area and the distance to the rivers and lakes of the distance of three ecological factors as input variables to construct the diversity index in study area and the fractal dimension of the BP neural network model. This method preliminarily explicated whether the artificial neural network model (ANN) was applied to the study of forest landscape ecology. Therefore, some reasonable suggestions were put forward on the basis of the forest landscape in Loufan County Ecological Construction. The results showed that:1. Lou fan County’s forest area was79470.55hectares, accounting for61.64%of the total area of territorial resources.19.49%of Arbores,8.31%of woodland;41.05%of forshrub land;6.90%occupied forest land area non-mature afforestion land occupied forest land area;0.83%of not stumpage forest and23.41%of suitable land for forest were included.2. In forest of Loufan County, mature and Young Plantation accounted for69%, in which mature Plantation was62%, followed by Young forest, over mature forest, near mature forest. Arbores stock volume accounted80.4%total stumpage volume. Forest resources were mainly concentrated in the mature forest, accounted for59%.3. Forest land area in LouFan was79470.55hectares, forest coverage was12%, LouFan afforestation trees was29.95%, which was below the national average and the entire province.4. The majority lands were Barren Mountain and land and scrublands in Loufan Xian.They accounted for60.17%of the total area. It was obvious that barren mountain were occupied more areas than the scrublands.The other10.95%included pure forest, economical forest, mingled forest and opening forest etc.5. Mixed forests, pure forest, artificial reforestation not into forest, sparse woodlot, economical forest landscape plaques were scattered in LouFanXian forest landscape, constituting the forest landscape heterogeneity in this area.6. The whole LouFanXian’s forest landscape diversity level was low, mainly controlled by a few landscape elements. It showed that the forest landscape type was relatively single. The study area evenness index was relative high, the landscape elements type distribution was evenly between plaques.7. Based on the analysis of the structure of LouFanXian forestry land landscape, using Elevation H, settlement from the river and lake J and number distance L as Input vectors and fractal dimension F and diversity index S as Output vectors estimate3:S:1multi-input and single-output BP neural network model. The results showed that the fractal dimension of the BP neural network model fitting accuracy was96.41%, and the diversity index the BP neural network model fitting accuracy was89.72%. It was convergence well.8. Under the known factors such as residential land, altitude, distance from rivers and lakes and others, we can use the BP neural network model to simulate the diversity index and fractal dimension, provides a new method for the future study of landscape ecology index analysis.