A Study On Analysis And Prediction Of The Temporal And Spatial Variation Of Wetland Landscape Patterns In Dongting Lake

Wetland is the ecological landscape with the richest biodiversity and is one of the most important human survival environment on the Earth. Wetland has a variety of ecological functions, praised as the "Nephridium of the Earth"; Wetland also is the ecosystem subjected to the most serious damage by human activities in the modern history. The biodiversity of wetland is dwindling and the function is degenerating. In this context, analysis and forecast the temporal and spatial variation of wetland landscape pattern is of great significance for conservation and sustainable use of wetland resources.The study area of this paper lies in the Dongting Lake Wetland Reserve. In the support of RS and GIS technical, this paper takes the Landsat-TM images in 1989 and 1996 and the CBERS-CCD image in 2008 as data sources, then carries out analysis and forecast study for the temporal and spatial variation of wetland landscape pattern, sums up the causes which lead to wetland landscape change and and puts forward the countermeasures for wetland resource protection of the study area. The main study results of this paper are as follows:(1)This paper adopts optimum index factor method to select the optimize bands of remote sensing images, the results is that Landsat-TM image and CBERS-CCD image separately adopts 7,4,1-band combination and 5,4,1-band combination, which will make the images get the maximum amount of information and the true color, be in favor of image interpretation.(2)According to the attribute difference of the wetland resources, the wetlands of the study area are divided into 4 categories and 15 sub-categories, such as river wetland, lakes wetland, marsh wetlands and artificial wetlands; According to the operability of extracting wetland landscape information from the remote sensing images, the wetlands of the study area are divided into 8 kinds of wetland landscape types, such as lakes, rivers, reservoir-ponds, grass-bottomlands, mud-sand-bottomlands, the reed, the paddy fields, woodland. Compared the two classification schemes, the second one is more conducive to the extraction of the wetland landscape information from remote sensing images.(3)This article adopts "division-regional and division-level" information extraction method, that is, in different landscape regions and different levels of classification, using different information extraction methods, such as Normalized Difference Water Index, Normalized Difference Vegetation Index, single-band threshold method, supervised classification and artificial division, to extract various types of wetland landscape information from remote sensing images. The accuracys of information extraction from remote sensing images in three periods are above 85%, and are in line with the research requirements.(4)This paper adopts the landscape pattern indexs and the dynamic models to quantitatively analyse the spatial and temporal variation characters of the wetland landscape pattern of the study area. The results show that, from 1989 to 2008, the total area of the wetlands in study area is reducing, wetlands landscape patterns have become complicated, the landscape diversity of the whole wetland landscape is gradually increasing, and the landscape heterogeneity is gradually enhancing, the landscape dominance is gradually reducing, the degree of the whole landscape fragmentation is increasing.(5)Based on the analysis of landscape pattern variation, we can see that the wetland landscape of the study area is changed because of climate change, mud sedimentation, farming and aquaculture around the lake, replacing farmland into lakes, artificial cultivating reed and planting trees at the beach, and so on.(6)This paper adopts the markov model to respectively predict the trend of wetland landscape changes of the study area, we can see that, if keep on the trend from 1989 to 2008, when the wetland landscape pattern of the study area gets to a stable state, the area of the water-landscape such as lakes and rivers will be much reduced, and the area of woodland-bottomlands and grass-bottomlands will be obviously increased.(7)Based on the variation characters and variation trend of the landscape pattern in the study area, thia paper puts forward some measures to protect the wetland, such as taking different protection measures to different landscape zones and different landscape types, strengthening supervision and management, strengthening the monitoring and the study to the wetlands, strengthening publicity and education, and so on.