Study On Socioeconomic Force To Evolvement Of Soil Heavy Metal And F,Sb In Chengdu Plain

In recent 20 years, the high-speed development of the urbanization of Chengdu plain had a profound effect on the soil circumference, which should not be ignored. This article focused on the study of the qualitative and quantitative effect on the accumulation of heavy-metal in soil, which was caused by the development of the urbanization of Chengdu plain, while attempting to find a correct unite-point from the two issuese—the urbanization and the accumulation, each of which belongs to a respective domain of science —the social science and the natural science. This was significant not only for the harmonious development of both the high-speed economic growth and the advisable use and protection of soil resources, but also for the accomplishment of the continuable development of both society and economy in urban and rural regions.Applying to data intensive technology of factor analysis, Index of socioeconomic factors which were constructed overall index, indirect index and direct index were classified and extracted. These index had effects on heavy metals, fluorine and arsenic. There were 10 in 1982 and 20 in 2002. Furthermore, in terms of scores of principal divisor 14 districts and counties in the centre of Chengdu plain were classified by taking advantage of clustering analysis. Through sequencing of grey degree of association between heavy metals in soil and factors (in 1982, in 2002), most highest and higher factors connected to degree of association with heavy metals were chosen in every region and computed major factors which can affect a variety of heavy metal content. The differences in different regions and overall influences of major factors on heavy metals, fluorine and arsenic were discussed.Factor analyses and region clustering analyses showed that, according to primary index after factor analyses, index which may affect heavy metal pollution were described as six major factors in 2002, including major socioeconomic production value, input of indirect pollution, and agricultural environment in soil, rural conditions, developmental rate and agricultural husbandry. Three major factors in 1982 about socioeconomic development, rural conditions and developmental rate were also obtained. Therefore, the ideal description of index obtained from major factors and analytic results proved rationality of frame which consists of overall, indirect and direct drive index. Because of complicated conditions on economic development and ecological environment in study area, this study area (in 1982, 2002) were classified three through region clustering analyses. The first class contained Guanghan, Jingyang, Xindu, Shuangliu, Jingtang, The second class contained Longquanyi, Dujiangyan, Chongzhou, Pengzhou. The third class contained Wengjiang, Qionglai, Xinjing in 2002. The first class contained Shuangliu, Qionglai, Chongzhou, Xinjing, Pengzhou, Pujiang, Dujiangyan and Jinyang, The second class contained Longquanyi and Jingtang; the third class contained Xindu, Guanghan, Pixian andWenjiangin 1982.Overall influences of major factors on heavy metals, F and As indicated that, the major influence factors of heavy metal content was single in 1982. However, the degree of impact of these factors in 2002 were socioeconomic level factors and development rate ones, which influenced content of As, Cr, Pb, Cu and so on. Meanwhile, the content of Hg, Cd and F was influenced by the factors of socioeconomic development, indirect pollution input and agricultural husbandry. The content of Cr, As Pb and Cu was influenced by the factors of indirect pollution input and soil agricultural environment fundamentally, the content of As, Pb and Cu had correlated to the factor of socioeconomic rate.Influence of different factors in different area (space) on accumulation of heavy metals showed that, because the whole socioeconomic development was in the first stage in Chengdu Plain at the beginning of 1980s, the major factors which influenced first section and third section were socioeconomic development level, and ones which influenced second section was the factor of socioeconomic rate. However, there was an obvious change in 2002. As far as first section which had better economics, agricultures and urbanization was concerned, the major factors were indirect pollution input, major socioeconomic production value, soil agricultural environment and so on. As far as second section which had better economic development and ecological environment was concerned, the major factors were indirect pollution input and rural conditions. The accumulation of heavy metals in the third section was lower than the whole study area. The third section which had unbalanced level of socioeconomic development and better ecological environment was influenced slightly by indirect pollution input and socioeconomic rate.There were 11 index acquired by the superior capability and momentum of BP network. These index were non-agricultural population proportion, average GDP, growth rate of GDP, industrial output value above a certain scale, passenger-kilometers by year, freight ton-kilometers by year, average cultivated area, yield of grains per hectare, multiple crop index, output value of agriculture per hectare and discharge of waste water by industrial enterprises. Three layers of BP network were established. There were 11 imports, 1 export and 1 hidden layer. Activation function in hidden layer presented as "S" and was liner function in export layer. Expectation error was 0.02. The most training frequencies were 8000. Initial value of learning rate was 0.01.Compared target vector of training samples with simulated export value and verified fitting extent of network model after network training, the simulated values in a variety of heavy metals were near to measured values .Meanwhile, average fitting errors of model were 0.4% below and fitting precision were 99.6% above, which verified fitting extent of BP network was very high. Compared BP network with the method of traditional regression, the former had higher precisionabout prediction of heavy metal content in soil. The extent of average precision was between 78% and 83%, while the extent of average precision of regression was only between 68% and 74%. Therefore, model of BP network had superiority.Predicted values of influenced factors in 2005 and in 2010 were regarded as input of network and merged with the former samples then retraining it. Through this work, the weight values of network were updated to acquire predicted values of Xindu in 2005 and in 2010. Predicted values for reference were acquired after integrative manage.