Font Size: a A A

Early Warning Research Of Oasis Soil Salizization Of The Lakeside In Bosten Lake

Posted on:2016-02-21Degree:MasterType:Thesis
Country:ChinaCandidate:D LiangFull Text:PDF
GTID:2283330479475677Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
In the last 20 years, soil salinization seriously restricted the sustainable development of oasis agriculture.With the mechanism of soil salinization gradual deepening and monitoring technology gradually improving,the eatly warning research of oasis soil salizization became one of the priorities in soil salinization research.At present, there were some scholars doing some research on disasters and land degradation by using the theory of early warning.The early warning model based on BP neural network achieved good results.The lakeside oasis in Bosten Lake was the study area. The main data sources included field survey data and Landsat satellite data in 2000, 2010, and 2013.The remote sensing images of three different periods were analyzed by spectral and spatial features.Based on the mechanism of soil salinization, using the theory of early warning, the warning indicators of soil salinization, the indicators of warning and the indicators of warning signs were selected, which were considered from human factor and natural factor.Through the Matlab software, the soil salinization early warning model based on BP neural network was constructed and used to discuss the situations of soil salinization.The conclusions were as follows:1.From 2000 to 2010, water body area reduced 591318 m2. The area of natural vegetation increased 4077750 m2. The area of unused land increased 183569 m2. The area which water changed to natural vegetation was 11372 m2. The area which water changed to unused land area was 2829067 m2. It occupied 99.6% of the total. The change area of unused land was 2645498 m2. Among it, the area which changed to natural vegetation occupied 53.1% of the total.2.The main content of soil salt was SO42- and K++Na+.Total salt content was from0.36 to 23.16g/kg. The main types of soil salinization were chloride sulfate type and sulfate type. The content of silt was the highest in the research area. The content of silt decreased from South and north to center gradually. Showing the spatial distribution in the east to the west, the silt content of the West Lake lakeside zone was higher than which were in the northeast and the center. Groundwater salinity changed in the range of 497.95~667.66mg/L, showing the trend of gradually increasing from north and south to the center. The groundwater salinity slightly was higher in western region than it in eastern region. The depth of groundwater was shallow. The deepest underground water level was-3.13 m, the most shallow underground water level was-0.30 m. The majority area was located in the-1~-2m, showing the trend ofgradually increasing from west to east. In 2013, the land use types of the four types of land were : Unused land > farmland > Natural Vegetation > water body. Water body area accounted for 12.1% of the total area. The total area of cultivated land was 31.6%.The total area of the natural vegetation was 23.3%. Unused land accounted for 33.1%of the total area.3.Based on the mechanism of soil salinization, combining with the actual situation in the study area, this study selected the mineralization degree of groundwater, groundwater depth, soil clay content percent, soil silt sand percent content, soil sand percent content, Wetness, NDVI, NDWI and land using code, a total of nine pre alarming index. Through the Matlab software, the BP neural network early warning model is constructed. And then,the model was used to simulate the status of soil salinization. Testing this early warning model, the regression fit was 0.9884 and the forecast accuracy was 0.6932. The prediction effect was relatively good.4.The study area was basically light, which was the highest in light and medium.Mild warning area accounted for 46.9% of the total area. Moderate warning area accounted for the total area of 37.8%. Severe warning area accounted for 15.3% of the total area. With regard to the spatial distribution, the northwestern and southern regions was a mild warning area. At the boundary of Kaidu River into the Bosten Lake, to the East, south two directions, salinization warning level gradually increased,until the dam. Severe warning was in the region where was from the lake entrance, to the pumping station 3~5 km.
Keywords/Search Tags:Soil salinization, Soil salt, Warning Indicators, BP neural network, Lakeside oasis in Bosten Lake
PDF Full Text Request
Related items