摘要当前在揭示区域土壤全氮（Soil total nitrogen，STN）变异性特征时，图斑连接依然是主要的研究方法。但STN数据究竟与何种图斑连接能更精确地反映区域STN的空间分布特征，在不同土壤类型区仍然存在不确定性。本研究以南方典型红壤丘陵区的江西省余江县为例，基于规则网格（1km×1km）在其中部地区采集的129个土壤样点，通过采样点STN含量值分别与所在采样栅格连接（PGr）、与土壤类型图斑连接（PSt）、与土地利用图斑连接（PLu）及与土壤—土地利用复合图斑（PStLu）四种不同的图斑连接方法获得区域STN空间分布特征，并通过67个验证样点的预测值和实测值对不同连接方法的不确定性进行对比分析。结果表明，PGr方法的预测精度最低，而分别考虑土壤类型和土地利用方式间差异的PSt和PLu方法的预测精度则较高，其中PLu方法的精度最高，优于PSt方法。而综合考虑土壤类型和土地利用方式间差异的PStLu方法精度在四种方法中最高。可见，在红壤区利用图斑连接方法进行STN空间预测和模拟时，PGr方法的不确定性最大，应避免使用。而充分考虑对STN有重要影响的土壤类型和土地利用方式，可大幅降低其结果的不确定性，其中同时考虑土壤类型和土地利用方式的方法不确定性最小。74641

该论文有图12幅，表2个，参考文献20篇。

毕业论文关键词：图斑连接法采样栅格土壤类型土地利用方式红壤丘陵区

Effect of Different Polygon-based Methods on Revealing STN Spatial Variation

Abstract Polygon-based method is still one of the widely applied for getting spatial variation characteristics of regional total nitrogen (STN) at present。 However, it is still unclear that linking STN data to what polygon can more accurately reveal the spatial distribution characteristics of STN by this method in each soil region。 In this study, 129 soil samples were collected with 1 km×1 km grids through the central regions of Yujiang Country in Jiangxi Province which is a representative area of hilly red soil region。 The spatial distribution contours of regional STN content was obtained by four different polygon-based methods: including linking STN content values to sampling grid which the sample belongs to (PGr), to corresponding soil type polygon (PSt) ,land-use polygon (PLu) and soil type-land use composite polygon (PStLu)。 The uncertainties of four methods were evaluated and compared with on the basis of 67 validation samples。 The results showed that the prediction accuracy of PGr method was lowest, while the accuracies of PSt and PLu methods, considering the difference of STN content among various soil types and land-use patterns were greatly improved。 Furthermore, PLu has the highest accuracy, better than PSt。 PStLu, both considering the difference of STN content among various soil types, has the highest accuracy compared with four polygon-based methods, Obviously, when STN spatial was predicted and simulated by polygon-based methods in the red soil region, the uncertainty of PGr method is the largest, and this method should avoid being used。 The methods of PLu and PSt giving full consideration with the effects of soil type and land use on STN content can greatly reduce the prediction uncertainties, and PLu is suggested the prior method。 

Key Words:polygon-based methodsampling gridsoil typeland use pattern hilly red soil region

目录

摘要Ⅰ

Abstract-Ⅱ

目录Ⅳ

图清单-Ⅴ

表清单-Ⅴ

变量注释表-Ⅵ

1 绪论-1

2 研究方法-2

2。1研究区概况-2

2。2土壤样品采集及实验室分析2