Prediction of selenium-rich characteristic crops production area based on neural network model:a case study of Zhangbei County,Hebei province
This paper is the result of geochemistry.[Objective]In recent years,selenium enrichment industry has developed rapidly,but because there is no significant positive correlation between selenium content in soil and selenium content in crop seeds,the planting regionalization of selenium-enriched agricultural products based on soil selenium content has certain uncertainties.Therefore,how to establish the selenium-rich model of agricultural products to predict the production area of selenium-rich agricultural products and make reasonable planning is an urgent scientific problem.[Methods]The crops and soils were collected from Zhangbei county of Hebei province.And pH,organic carbon,silica,aluminum trioxide,calcium oxide,magnesium oxide,ferric oxide,potassium oxide,sodium oxide,selenium,arsenic,mercury,cadmium,lead,copper,zinc,chromium,and nickel were measured.The influencing factors of selenium uptake in crops were studied,and the prediction model of crop selenium content based on neural network was established,and the yield area of selenium-rich crops was predicted,base on the results of measurement and the data of 1/250000 multiple geochemistry collected.[Results]There were only 2.31 percent of topsoil sites reached the standard of selenium-enriched,and they mainly distributed in the northern area of Sanhao Town.There are 6 naked oats which meet the standard of selenium-enriched,and the selenium-enriched rate is 16.67 percent.While there are 7 flaxseeds which meet the standard of selenium-enriched,and the selenium-enriched rate is 21.86 percent.There was a negative relationship between crops'bioconcentration factor for selenium with organic carbon,ferric oxide and aluminum trioxide,a positive relationship with pH in soil.The prediction model of selenium content in flax and naked oat seed was established by machine learning,and the results were reliable.The potential production areas of flax and naked oats were mainly distributed in Zhangbei,Xiaoertai,Tailugou,Youlouou and Haojiaying Town,etc,and the output areas were 2 486 km2 and 2 406 km2,respectively.[Conclusions]The area of selenium-rich crop producing area defined by the prediction model is much larger than the selenium-rich soil area,which is of great significance for the development of selenium-rich agriculture in Zhangbei County.
万方数据
