Analysis On Landscape Pattern Process Graphic (TuPu) And Its Feature Inversion

Geo-informatic spatial and temporal graphic (TuPu) is a forward geo-science research filed about complex geo-science phenomena or processes summarized or abstracted by graphics thinking manner. Its main advantage lies in mining the inherent rule of landscape ecology pattern and process, exploring the evolvement rule of environment, among which, spatial pattern and evolvement process of environment factors is becoming one of the issues of environment research. Landscape pattern process informatic spatial and temporal graphic (TuPu) is human cognizance and highly abstract presentation to"complex, open, gigantic system". Landscape pattern process informatic TuPu is a kind of research means which establish a series of multi dimensional illustration and graphic thinking manner of landscape pattern to multi-sources remote sensed imageries by RS and GIS technologies. During the course of its generation, landscape index analysis and spatial statistics method play the important roles, then, all kinds of spatial and temporal analysis models are constructed through setting up all kinds of analysis parameters, the past landscape is reconstructed. Thus, the dynamic, virtual, multi dimensional landscape is shown and the future is predicted. Landscape pattern and its process directly participates in and responds global change, global change research is feasible through landscape pattern informatic spatial and temporal graphic (TuPu), which will provide policy decision foundation to global environmental change.Landscape pattern process informatic TuPu is used to research landscape pattern, process and reciprocity with the manner of geo-informatic TuPu, thus, the distribution of landscape pattern and its evolvement rule will be grasped indeed, such objectives as describing the present, reconstructing the past and predicting the future will be realized. The idea of landscape pattern process informatic TuPu is put forward based on geo-informatic TuPu, landscape ecology, geo-science modeling theory, also its research framework and technical methods are discussed. The measurement theory is introduced into landscape pattern informatic TuPu, the evolvement process and mechanism of research field are analyzed through extracting the informatic TuPu units in research field. The mathematical models of feature inversion of landscape pattern process informatic TuPu are constructed, then, the inversion and the prediction of landscape pattern process are made.The theory and methods of Landscape pattern informatic TuPu are explored through the case of Mount. Changbai landscape pattern and its process, including TuPu establishment and its 3d visualization, TuPu measurement analysis, inversion and prediction of TuPu features. The research work provides a new research method and vision in landscape ecology field in theory, enrichs geo-informatic TuPu thory and method, constructs the reseach system of landscape pattern process informatic TuPu, in practice, makes the inversion and prediction of landscape pattern process informatic TuPu to research field on the basis of TuPu measurement analysis.The main research contents and conclusions are showed as follows:1. Geo-informatic TuPu and landscape ecology are studied systematically, in theory, system info of geo-informatic TuPu is further concluded, summarized, the framework of landscape pattern process informatic TuPu is established, its theory and method are explored also.2. On the basis of analyzing Mount.Changbai ecologic environment thoroughly, classification and its index system of Mount.Changbai landscape pattern informatic TuPu are put forward, then, informatic TuPu is composed and presented with 3D visualization in research region. A series of Mount.Changbai TuPu are built, such as grade TuPu, direction TuPu, vegetation cover degree TuPu, conformation TuPu,rock distribution TuPu, 3D landscape pattern process informatic TuPu etc., thus, forming the foundation for measurement analysis and inversion of landscape pattern process.3.The measurement theory is introduced into landscape pattern process, a series of landscape pattern index (TA/CA, NP, PAFRAC, PD, D, SHDI, SHEI, CONTAG, CONHESION, IJI), spatial statistics, PCA, dynamic change model are used to analyze landscape pattern process in research region. At the same time, the index system and research framework, which are used to describe landscape pattern quantitatively, are built according to superincumbent cognizant process. The research production indicates that the landscape style in research region has changed distinctly, water body, tundra, conifer woodland, conifer and broadleaf mixed woodland, broadleaf woodland etc. decreased, while residential area, road, birch forest, grassland, infield, naked land etc. increased correspondingly in recent 30 years. During 1977-1985, water body, tundra, conifer woodland decreased. Of all the species listed before, the largest one is conifer woodland, which reduce 30%, while, there is the most rapid increase in road (dynamic degree 152%/a), the area of infield increases distinctly. During 1985-1999, residential area, road, birch forest, grassland, infield increased distinctly, while, conifer and broadleaf mixed woodland decreased distinctly, next, broadleaf woodland. (1) The change feature of parcel size. Parcel size influences the matter transference of ecological system tremendously, the average area of residential area parcel in 1977 is largest, and then road area, water body area, grassland area etc., while, the average area of residential area and road parcel in 1985 decreases slightly, showing their cracked trend, on the contrary, the average area of birch forest parcel rises notably, showing the former birch forest joining together. (2) The change feature of parcel margin. The border density of parcel reflects not only the predominance degree of parcel, but also the complex degree of parcel figure. The conifer woodland, conifer and broadleaf mixed woodland, broadleaf woodland is the predominance styles with large ED value, showing the long borderline in unit area and complex figure. (3) The change feature of parcel figure. By analyzing the PAFRAC, the value of residential area is the smallest, while, the largest one is road, indicating the simple parcel figure of and residential area parcel and linear distribution of road, the value of other landscape species belong to near 2, complex borderline among all kinds of landscape species. (4) Landscape in research region shows characteristic of vertical terrain in evidence, but its inside structure has changed notably, the fragmentation degree has increased, complexity has boosted up. With the rapid development of tour industry, jamming affects landscape change further. Many uncertainty factors, such as plant diseases and insect pests, fire, disastrous climate etc., bring on more complex landscape variety. The forming and evolvement mechanism of landscape pattern in research region has been understood. The author considers that natural factors compose the foundation of environment pattern and its transformation, which control the environment distribution pattern in large view, while social economic factors is only driving factor, which generate main driving power in short time.4. On the basis of the feature of geo-cell automatic model, the NN model is introduced into transformation regulation of geo-cell automatic. There are 16 layers for input, 11 layers for output in BP NN, the BP NN training is constructed for the purpose of model transformation regulation. Then, the regional landscape pattern inversion model (GeoCA-APSHV) is set up. TuPu in 1985 and 1999 are used to inverse and predict the regional landscape pattern process according to the feature of research region. By contrast with TuPu in 1999, the inversed TuPu in 1977 is considerable accuracy, reaches over 80% accurate degree, then, impersonal evaluation to model application is made. Finally the ecological strategy of sustainable development of landscape pattern in research region is put forward according to the systematic analysis.5. Uncertainty impenetrate all aspects of landscape pattern information TuPu, from data gathering, data processing, data analyzing, to reconstruction and prediction. Mainly representation includes uncertainty of DEM and remote sensed images quality, uncertainty of landscape style classification and subsequent process, uncertainty of reconstruction and prediction model of landscape pattern information TuPu, and its model evaluation etc.The research production in the paper is very important to environmental problem prevention and cure in Mount.ChangBai, it provides important reference information for governmental policy decision, at the same time, there is good use for reference meaning to other region.