Application Of Biodegradable Polymer In Fields,Soil Quality Assessment And Yield Prediction Model Research

Slow/controlled release fertilizers are used to solve problems of nutrients loss and environmental pollutions caused by conventional fertilizers.However,the existing research results show that the current slow/controlled release fertilizers have problems such as higher prices and single nutrition elements,or envelope rupture of coated fertilizers that still cause nutrients bursts.In this paper,we selected a series of multi-nutrition-containing biodegradation polymer materials developed by Research Center for Engineering Technology of Polymeric Composites of Shanxi Province as the slow/controlled release fertilizers and studied their effects on physical properties and chemical properties of soil and corn physiological traits,yield,and quality of crops through the corn field experiment.The soil quality assessment model and the crop yield prediction model were established to lay the foundation for efficient design of multi-nutrition-containing biodegradation polymer materials and to reduce the cost of research of fertilizers.The specific findings are as follows:The multi-nutrition-containing biodegradable polymer materials have a significant influence on physical and chemical properties of soil.Application of the biodegradable polymer materials,as a kind of organic matter,have increased soil organic carbon content,subsequently decreased soil specific weight,increased microbial activity,thereby increasing soil urease activity and favoring for plant growth.Since the fertilization is beneficial to the growth of plants,the water absorption and evaporation of the crops are increased.And thus the fertilization reduces soil moisture content.Because there are a large number of ions and organic substances that can be used as buffers in the soil,the pH changes in each treatment are not significant in a short-term fertilization strategy.The changes of nitrogen phosphorus and potassium contents in soil showed that the biodegradable polymer slow/controlled release fertilizer has sustained nutrients release performance.Application of multi-nutrition-containing biodegradable polymer materials increased the yield and quality of corn.Among them,crop yield of boron-containing biodegradable polymer slow/controlled release fertilizer(BRF)and other-micronutrient-containing biodegradable polymer slow/controlled release fertilizer(WRF)was optimal,indicating that the growth of corn is sensitive to boron and other microelements(Fe,Mn,Cu and Zn).Application of biodegradable polymer materials can significantly increase the content of nitrogen,phosphorus and potassium,total amylum,amylose,soluble sugar and grain of corn,and have obvious effects on improving the taste and nutritions of fruits.The application of biodegradable polymer increased corn plant height,dry matter accumulation,and chlorophyll content of the plants,indicating that the sustained release of nutrients and the input of organic matter made the crops have good nutrient conditions in the later growth stage and achieved the aim of increasing yield and fertilizer using efficiency.Cluster analysis was performed using crop yield and soil qualities to evaluate the effect of application of biodegradable polymer materials on soil quality.Soil fertility quality was classified into three categories using crop yield,corresponding to low-yield soil(no fertilized soil),middle-yield soils(sulfur-containing biodegradable polymer slow/controlled release fertilizer(SRF)treated soil)and high-yield soils(multi-nutrition-containing biodegradable polymer materials except for SRF and organic-inorganic compound fertilizers(OM+INF)treated soils).The results show that the use of centroids and Euclidean distance clustering results in serious loss of soil information by directly clustering soil indicators,and the classification is not conducive to the evaluation of soil quality.The use of intergroup links and Euclidean distance clustering can be used to classify the soil into three classes: high-quality soil(BRF and OM+INF treated soils),medium-quality soil(WRF,Polyvinyl alcohol inorganic fertilizer compound fertilizers(PVA+INF),silicon-containing biodegradable polymer slow/controlled release matrial(GRF),NPK-containing biodegradable polymer slow/controlled release matrial(PRF)and SRF treated soil),and low-quality soil(no fertilized soil).After the analysis of principal components of soil indicators,the main components of soil indicators were used for clustering.The results showed that the main component clusters reflected the health quality of the soil: application of BRF can improve soil health quality.The application of OM+INF PRF,WRF and no fertilizer have no significant effect on the deterioration or destruction of soil.The application of SRF,GRF,and PVA+INF caused damage to the health of the soil.Multivariate linear regression and artificial neural networks were used to predict the crop yield using data of biodegradable polymer slow/controlled release materials and soil moisture content in field experiment,respectively.The results showed that the prediction accuracy of neural network is significantly higher than that of multiple linear regression.Using Bayesian regularization training function to reduce neural network overfitting improves the model prediction accuracy.After adjusting the input variables and model parameters,the optimal neural network prediction model was found.The yield prediction results of all fertilizer treatments basically conformed to the regularity of real results except for that of no fertilizer treatment and sulfur-containing biodegradable polymer slow/controlled release fertilizer treatment because of nutrient deficiency and cumulative sulfur poisoning.From the perspective of optimizing fertilizer design and guiding agricultural practices,this model has met the requirements.