Study On Regional Landscape Ecology Assessment Based On Wavelet Neural Network

Landscape is a scale which contains geographical region and is included in ecosystem, many land use problems couldn't get reasonable explanation without being based on landscape scale. However, with the development of the modern industry, the material exchange between human activities and living environment is developing with astonishing speed. On the one hand, existing over-consumption of land, forest, energy, freshwater and other natural resources, especially to the development and utilization of renewable resources as animals and plants, which have been difficult to restorative regeneration, and also lead to the destruction of biodiversity and the extinct of valuable animals and plants. On the other hand, the increasing waste excreted to the environment causes the worsening of the eco-environment continuously, threats the survival of human and other livings seriously. Based on the above aspects, many ecological problems produced, mainly presented exhaustion of natural resources, severe water and soil erosion, spread of land desertification, degeneration of grassland, reduction of the forests ,disordered exploitation and excavation, indiscriminate killing of wildlife, serious destruction of its ecological environment and biological diversity decline. These problems have been become the restriction of the development of social economy and the survival of human beings, therefore, the protection of landscape eco-environment was on the agenda gradually as the constant increasing of the severity of ecological problems.The paper based on the summary the foundation of landscape ecology and neural network introduces wavelet neural network (WNN) to the field of landscape ecology assessment, an analysis of landscape assessment principle, process and methods based on WNN is presented. Forms relatively complete theory and method system for WNN model of landscape ecology. Concretely speaking, the paper makes a study on the following aspects:(1) Based on the concept and contents analysis of landscape ecology assessment, the landscape ecology assessment is deposed into quality assessment of landscape ecology part, vulnerability assessment of landscape ecology part and landscape ecological risk assessment part. And summarizes the methods of landscape ecology assessment, which are following: field survey, expert judgment, statistical analysis, RS and GIS mapping, mathematical model methods and so on.(2) The paper supports a comprehensive assessment index model based on WNN landscape ecology assessment, and then, proposes the principle and process of the model. The landscape ecology assessment based on WNN is as following in process:â‘ he present condition investigation of study area,â‘¡constructing assess index system of assessment region,â‘¢single analysis and assessment,â‘£evaluate comprehensively with WNN model.(3) According to the connotation of landscape ecological status, constructing the landscape ecology assessment index system based on WNN, which include landscape productivity, degree of landscape threatened and landscape stability to describe the productivity, elasticity and pressure facing of landscape ecological system.(4) The paper also proposes WNN model of landscape ecology assessment, main process includes constructing assess index system, calculating comprehensive risk index, comparing with assessment standard, etc. For the good accuracy and precision, the model result and that ANN model simulated could be analytically compared. However, the landscape ecological system is a system with complex influencing factors, during the calculating of comprehensive risk index, besides quantitative model, also needs qualitative analysis reasoning.(5)This paper will take Jiangyan City as an example, and study sustainable land-use landscape ecology assessment by using landscape ecology assessment WNN model. The results show that study region can be compartmentalized into five grades, high sustainable region, the area of which accounts the whole area 12.64%, less-high sustainable region, accounting 6.70%, medium sustainable region, accounting 35.04%, ordinary sustainable region, accounting 39.17%, inferior sustainable region, accounting 6.45%,and propose management measures against different grades. The study results also show, calculating assessment sustainable land-use wavelet neural network elaboration model which is proposed in this paper has fast, accurate and better-suitable such advantages.(6)This paper takes Taizhou domination area as study object. Through application of WNN model, start landscape ecology assessment of city edge region. The study shows, this model will most eliminate the effects of human factors, consider various factors of affecting landscape ecology, and make full use of statistical data information about characteristic parameters of landscape ecology. Through high nonlinear mapping, seek the implied intrinsic mechanisms between landscape ecology synthetic assessment and characteristic factors. So objectivity of obtained data can be strengthened and give concrete synthetic assessment results, which can provide important scientific basis of further landscape assessment management and construction.(7)This paper not only proposes measures of protecting landscape ecology environment against different grades, but also proposes corresponding policy and suggestions against main landscape ecology problems found during study. That include,â‘ putting rational distribution of landscape;â‘¡strengthening landscape diversity protection;â‘¢accelerating the construction of regional landscape corridor;â‘£Maximizing the role of green space system such as farmland.