The Study On Spatial Variability Of Forest Soils Second And Trace Element

The forest can improve the human living environment, water resource conservation and prevent soil and water loss. Forest plants interact with forest soil. Forest soils provide nutrients for forest plants. Forest plants affect soil formation. For this reason, the second and trace element in the forest soil had been investigated. That has great significant for forest restoration and regeneration. This paper in the case of Yuncheng and Yunan district on Yunfu city of Guangdong province and used the artificial neural network model for predicting the spatial distribution of forest second and trace elements, including the Spatial distribution characteristics of content. Then reveal current environmental the spatial distribution of soil second and trace element in the study area.The main results were as follows:1. For forest soil second and trace elements, the highest content was exchangeable Ca,then total sulfur, the last was exchangeable Mg in the study area. In the trace elements, the highest levels were total Pb, the second was total Zn, the last was total Cd. Total Cd, Zn belong to the primary standard in the soil environment quality standard. From all the second and trace elements, they were both less than the atlas of soil environmental background value, except total Zn.2. By comparison, the value of Coefficient of Variation leads to the conclusion that all of the elements Coefficient of Variation in Yuncheng district was between 33.04% and 98.19%. All of these elements Coefficient of Variation in Yunan district was between 33.43% and 106.34%. They were at medium Variation, except exchangeable Ca. Exchangeable Ca at big Variation. Exchangeable Ca had the highest Coefficient of Variation in Yuncheng and Yunan district, 98.19%, 106.34% respectively. Total Cd Coefficient of Variation was 33.04% in Yuncheng district and was small than other elements in the study area.3. Choose the best spatial interpolation method. Through the comparison of root mean square and mean absolute error in modeling and verification points. The results suggest that the precision of universal Krige method and inverse distance weighted interpolation method was better. Further compared the significant correlation degree of actually measured value and predictive value. It shows that two methods actually measured value and predictive value both have a significant correlation. But the value which actually measured value and predictive value's deviation showed that universal Krige method deviation is small than other methods. But universal Krige method interpolation method can't predict other elements well in the six elements, except exchangeable Mg.4. From the select of the model, exchangeable Ca, exchangeable Mg, total S, total Cd, total Zn showed BP artificial neural network model was the best, through two models the significant of correlation degree, including actually measured value and predictive value. Total Pb shows universal Krige method interpolation method was best. Further research on its optimal model was still needed. From spatial prediction map view, the artificial neural network can better reflect obvious on terrain changes. Compared with universal Krige method interpolation method, the artificial neural network can be used widely. So in a word, the artificial neural network was best.5. The spatial distribution of second and trace elements in the soil present regularity. The distribution of total Cd, total Pb, total Zn was gradually decreasing from west to east the distribution trend of the space in the study area. The exchange of Ca in general from west to east gradually increased and then decreased and then increased. The exchange of Mg and total S in general from west to east gradually decreased and then increased. Exchange of Ca and Mg in the low-value region was an irregular strip from north to south from the central part of the entire study area. The second elements were influenced by Leaching and corrosion and lead to a large amount of nutrient loss. The trace elements were influenced by human activity, including a chemical plant, pesticides, car exhaust and so on.