Font Size: a A A

The Research On Saline-alksli Soil Grading And Governance Of The Yellow River Delta Based On The RS And GIS

Posted on:2016-01-25Degree:MasterType:Thesis
Country:ChinaCandidate:B GuoFull Text:PDF
GTID:2283330470450514Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
For a long time, saline-alkali issue has always been global issues and problems.Saline-alkali soil is damage for crops which is destroying their survival environment,affecting their normal growth and even leads them to death. Therefore it will hinderthe normal development of agricultural production and the environment. According tothe FAO and UNESCO statistics, the total area of saline-alkali soil is approximately954million hm2, while in China is more than99million hm2. Saline-alkali hasbecome a significant factor restricting the development of China’s agriculturaleconomy.The problem of the saline-alkali soil has been especially outstanding and hasrestricted the development of the agriculture in Yellow River Delta. So it’s urgent tostrengthen governance. It is said" Know yourself and know your enemy, you will winevery war." The primary task of governance is to understand the distribution of thesaline-alkali soil and the salinization degree in the area.After field soil sampling in the Yellow River Delta and the experiments of thesamples, I obtained the soil salinity of soil samples in sampling points and generatedthe points of sampling points on ArcGIS10.1platform with the soil salinity of soilsamples.The same period remote sensing images which from Small Satellite Constellationfor Environment and Disaster Monitoring and Forecasting B Star (HJ-1B satellite)were advance processed and extracted the remote sensing image reflectance values ofeach band in the location of sampling points. By using SPSS platform, the calculationof Pearson correlation coefficient aimed at each band of reflectance values andcorresponding soil salinity were figured out and then got the standard deviation ofeach band. Finally the diagnostic coefficient was obtained which could able to showthe sensitivity of each band to the soil salinity. The results showed that the soil salinitysensitivity of Band1Band2and Band3which significantly higher than the Band4wasmore unified, these three bands can be used to quantitative remote sensing retrievaloperations for soil salinity.In order to setting up the relationship model between image reflectance valuesand soil salinity, the three bands: Band1Band and Band3which were regarded as independent variables were conducted a multiple linear regression analysis operationby using stepwise regression analysis method on the SPSS platform. The resultsshowed that Band2was excluded while Band1and Band3were Conducted aregression operation as retained independent variables and got significant operationalresults.BP Neural Network is widely used with its strong non-linear regression mappingcapability and outstanding self-learning and adaptive ability. By introducing the BPNeural Network into the remote sensing retrieval of soil salinity, we can takeadvantage of their own characteristics, which provided a new method and approachfor quantitative remote sensing inversion. In this paper, by using the structure of2-16-1, which is2nodes in the input layer,16nodes in the hide layer and2nodes inthe output layer, I built up a three-layer Neural Network model which concluded inputlayer hide layer and output layer. The network model training was operated whenmade the Band1and Band3as an input node and the Log-Sigmoid as the trainingfunction. After accuracy test which was to contrast the predictive power between BPNeural Network model and multiple linear regression models, the result was obtained,that in most cases the former was stronger than the latter. Thus it is a good way toconduct simulation for soil salinity in the Yellow River Delta by using neural networkwhich acquired by training.The results showed that the degree of saline-alkali soil salinization of the YellowRiver Delta is more serious, the proportion of native saline-alkali which wasrepresented as severe saline and salt pan is nearly70%. And their distribution is moreconcentrated in the Yellow River and the areas concentrated with large waters. Due tothe high water potential, large amount of water and other factors, the Groundwatertable in these areas is turning shallow, which is aggravating the degree of salinization.Finally, the causes of the Yellow River Delta saline-alkali were described andsome relevant proposals for governance were put forward in this paper. Although it isunable to prevent the formation of saline-alkali soli in the area by using artificialmeasures, however, it will be possible to relieve the degree of saline-alkali soilsalinization through scientific and rational ways, thereby reducing the constraints toagricultural development from saline-alkali.
Keywords/Search Tags:the Yellow River Delta, Saline-alkali grading, Remote Sensing Retrieval, BP Neural Network, Saline-alkali governance
PDF Full Text Request
Related items