长三角平原区县域土壤磷素流失风险及其空间不确定性的快速评估  

Rapid assessment of soil phosphorus loss risk and its spatial uncertainty in county areas of the Yangtze River Delta Plain

作　　者：陈剑 瞿明凯 王燕[1] 万梦雪 黄标[1] 赵永存[1] CHEN Jian;QU Mingkai;WANG Yan;WAN Mengxue;HUANG Biao;ZHAO Yongcun(Key Laboratory of Soil Environment and Pollution Remediation,Institute of Soil Sciences,Chinese Academy of Science,Nanjing 210008,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院南京土壤研究所土壤环境与污染修复重点实验室,南京210008 [2]中国科学院大学,北京100049

出　　处：《生态学报》2019年第24期9131-9142,共12页Acta Ecologica Sinica

基　　金：国家自然科学基金项目（41771249）;国家科技支撑计划课题（2015BAD06B02-2）;南京土壤研究所“一三五”计划和领域前沿项目（ISSASIP1623）;中国科学院青年创新促进会（2018348）

摘　　要：磷素是水体富营养化的关键限制因子,其中从农田土壤中流失的磷往往是水体磷素的主要来源。然而,土壤磷素的流失风险不仅与土壤磷素水平直接相关,其他环境因子,如距受纳水体距离、磷肥施用量、地表径流潜力等也强烈影响其流失风险。同时,基于有限样本预测得到的流失风险必然具有一定的空间不确定性。以长三角典型县域金坛区为研究案例,首先结合多个环境因子构建快速磷指数(RPI)评估模型,再利用稳健地统计学方法识别土壤全磷的空间离群值,并利用序贯高斯模型(SGS)模拟土壤全磷可能的空间分布格局,最后将其多个可能的模拟结果及上述主要因子输入到RPI模型,用以快速评估土壤磷素流失风险及其空间不确定性。结果显示,金坛区土壤磷素流失的高风险区和土壤全磷高值区分布格局在研究区北部、中部具有一定的相似性,而在中西部的旱地区两者出现差异性。高风险区主要沿着河流呈现条带状及斑块状分布,较高及以上风险区(快速磷指数值大于0.93)的面积占金坛区面积的65.88%。概率阈值分别设定为0.50、0.75、0.85、0.95时,其超标面积占金坛区总面积分别达到16.71%、5.74%、2.84%、1.04%。引入多个相关环境因子并结合经稳健处理的SGS进行流失风险指数的空间模拟和不确定性评估,可以快速评估区域农田土壤磷素流失风险及不确定性,进而为区域土壤磷素调控提供必要的空间决策支持。Phosphorus is a key limiting factor for the eutrophication of water bodies.The phosphorus migrating from farmland soils is commonly recognized as the main source of phosphorus in water bodies.Therefore,a better understanding of the loss risk and its spatial uncertainty of soil phosphorus are important for the decision maker to regulate the loss of soil phosphorus at the local and regional scale.In this study,a total of 259 farmland topsoil samples(0—20 cm)was collected from the Jintan District,Changzhou City,Jiangsu Province,for the analysis of soil total phosphorus contents.Apart from soil total phosphorus,we introduced other environmental factors,such as the distance from receiving water body,phosphate fertilizer application rate,and surface runoff potential,into the assessment of the loss risk of soil phosphorus.Furthermore,the prediction of risk indices based on limited samples inevitably has certain spatial uncertainty.Here,a rapid phosphorus index(RPI)assessment model was firstly constructed based on the multiple environmental factors introduced above.And then,the robust geostatistical methods were applied for identifying possible spatial outliers of soil phosphorus data,and the sequential Gaussian simulation(SGS)model with the dataset being removed the spatial outliers was used to simulate the possible spatial distribution pattern of soil total phosphorus.Finally,multiple equiprobable realizations of soil total phosphorus and related environmental factors were input into the RPI model to evaluate the risk of soil phosphorus loss and represent associated spatial uncertainty.The results showed that the high-risk areas of soil phosphorus loss were mainly located in the eastern and middle parts of the Jintan District,which were similar with the high-value areas of soil total phosphorus.The dryland soil total phosphorus located in the mid-west of the study area was relatively low,while the loss risk of soil phosphorus was still at a high level.Such spatial distribution patterns of soil phosphorus loss risk we

关 键 词：土壤磷素 稳健地统计学 序贯高斯模拟 流失风险 关键源区 不确定性 

分 类 号：S153[农业科学—土壤学]