Estimating Filed-Scale Soil Nutrient Levels in Shun Yi District, Beijing

摘要

The objective of this study is to estimate the field-scale soil nutrient levels (SNLs) of Shunyi District's croplands using 3S technologies. Based on the soil samples with GPS coordinates, a Landsat TM image was firstly utilized to identify cropland fields of the study area using support vector machine (SVM) classification module integrated with ENVI, and the overall classification accuracy reaches 87.68% and Kappa coefficient is 0.7996. Subsequently, according to the proposed classification criteria of SNLs by Beijing Soil and Fertilizer Work Station, four edaphic indicators including organic matter (OM), total nitrogen (TN), available phosphorus (AP) and available potassium (AK) were selected to estimate the filed-scale SNLs of Shunyi District. The result shows that four levels including very high, high, moderate and low are found except the very low level and they account for 0.4%, 42.3%, 53.3% and 4.0%, respectively. Furthermore, considering the spatial distribution of SNLs, it is better in the west, north and east than in the middle and south, while it is the worst in the middle.