Spatial Distribution of Inversion to Soil Fertility Factors in Guang'an Based on ETM+data
|School||Sichuan Agricultural University|
|Course||Utilization of Agricultural Resources|
|Keywords||ETM+ the soil types the land-used types soil fertility spatial distributioncharacteristics|
In order to save the cost to predict the soil fertility and space distribution features in the intensive and convenient landscape type complex areas, this article takes472analyzed soil samples by through Landsat5shot of the spectrum ETM+in2007remote sensing image, in waste of Guang’an area USES grid sampling technology at the same time, and apply ArcGIS9.3and ERDAS software, geometric correction, filter to eliminate the terrain and object surface texture shadows, difference and imaging process of remote sensing image caused in the noise, from the band like the grey value. Then it takes the SPSS13.0package of multiple regression analysis to set up organic carbon (SOC), total nitrogen (TN) and available nitrogen (AN) and potassium (AP) and bands pixel grey value, soil type (Gleyed paddy soils, Bleached paddy soils, Percogenic paddy soils, Hydromorphic paddy soils, Submergenic paddy soils, Acid purplish soils, Neutral purplish soils, Calcic purplish soils) and land uses (Dry land, Paddy field, Woodland, Grassland)between prediction equations. Choose the optimum model through the measurement values and between the average error of the model prediction, root mean square error and average relative error of the three test indexes, and predict the distribution features of the measured data by using ordinary kriging method. The following results are:The total amount of SOC is11.52±3.20g kg-1, TN is10.94±0.13g kg-1, AN is93.51±8.67mg kg-1, AP is5.96±1.09mg kg-1, AK is69.56±10.08mg kg-1.Establish ETM+remote sensing images and soil fertility factor regression models and select the optimum model is the results for soil types, land uses and5band pixel grey values or6band pixel grey values of the prediction model which can be established to predict the fertility. Prediction of SOC, AN and AK for5bands and soil type value a regression model combined best effect (P<0.01); TN prediction for the band6types of land use value and combined the best established regression model (P<0.01); AP is forecast5band and land use type of value, soil type value a regression model combined best effect (P<0.01), this shows that using ETM+remote sensing image figure can be quickly and easily.The higher SOC area mainly distributed in the northern of Guan’an, the lower area distribution in the eastern; Higher TN mainly distributed in the north and northeast of the area, lower TN is in large distribution east, south and southwest of this area. The higher AN mainly distributed in the northern of this area, lower of it distribute in the east and south. High AK area and AP mainly distribute in the middle of Guangyuan and low AP and AK distribute in the northern area distribution.