Spatial estimation of soil total nitrogen using GIS:a case study in Xingguo county,Jiangxi Province

Spatial distributions of total nitrogen(TN) in soils are difficult to estimate because of their inherent spatial variabilities and lack of sufficient sample data.Soil-landscape model,based on parent material and topographic factors,was generated through applying GIS spatial analysis technique and a digital elevation model(DEM) in large areas.Based on 151 sample data,spatial distribution characteristic of TN was studied.Correlations between TN and topographic factors were analyzed and a regression model was established to predict TN content through linear regression analysis.The results for surface soils(0~20cm) showed that the average content of TN was 1.06g/kg in topsoil,with TN content between 0.5g/kg and 1.0 g/kg occupying the largest area(1580km~(2),49.3% of the total area) and that over 2.0g/kg the smallest(127km~(2),3.9% of the total area)in spatial distribution.Also,soils derived from sandstone and shale had the lowest average TN content(0.88 g/kg soil),soils developed on phyllite showed the highest average TN content(1.35 g/kg soil).Aspect among topographic variables had an important effect on TN content.The correlations of TN content with parent rocks,elevation and aspect were positively significant(P0.05),slope exposure had the highest correlation with TN content(r=0.62).The correlation of TN content with slope was insignificant(r=-0.03).In the soils of the whole study area,the step regression analysis of the TN content,topographic factors and parent rocks showed that elevation,slope exposure and parent rocks were the best factors for predicting the TN content in topsoil(P0.05).The multiple linear regression model is: TN=0.095+0.005×parent rocks+0.001×elevation+0.115×slope exposure classificationn=113R~(2)= 0.637 The spatial distribution of TN content could be predicted by using a multiple linear regression model and DEM(with a 30m×30m grid).