利用地形和遥感数据预测土壤养分空间分布

在GIS支持下,选择地形因子和遥感植被指数,建立土壤养分空间分布预测模型,应用回归克里格（Kriging）方法,预测吉林省农安县土壤养分（有机质和全氮）的空间分布。结果表明,11个环境因子中,相对高程、坡度、地形起伏度、坡度变率、归一化植被指数（NDVI）与土壤有机质和全氮含量均具有显著的相关性。地面粗糙度和地形湿度指数与有机质具有显著相关性,而与全氮的相关性不显著。相对高程、坡度、地面粗糙度、河流动能指数以及NDVI在土壤养分的多元回归预测模型中贡献较大,是预测土壤养分空间分布的最优因子。有机质和全氮在研究区的空间分布格局呈现由东南向西北逐渐减少的趋势,这种分布格局受地形和植被的综合作用,同时与土壤类型密不可分。精度检验结果表明,回归克里格方法能够提高土壤养分空间分布预测精度,是一种有效的空间分布插值方法。     

基于多元线性回归模型的土壤养分空间预测——以陕西省蓝田县农耕区为例

以陕西省蓝田县农耕区为例,基于土壤类型、地形、植被、人类活动等土壤养分影响因子,采用多元线性回归预测方法,对土壤养分分布进行空间预测。通过数据标准化和分类统计均值法将所有影响因子统一为相对度量值,根据影响因子与土壤养分含量的相关性显著程度,筛选合适的因子构建回归模型,最后使用随机抽取的验证样本集,计算了预测结果误差。结果表明:多元线性回归预测结果虽然与Kriging法预测结果宏观上具有基本一致的空间分布趋势,但是它克服了传统插值方法的不合理突变问题,更加精细地描述了土壤养分空间分布特征;其平均误差和均方根误差小于Kriging插值法,可作为集成多元影响因子对土壤养分空间分布预测的有效方法。分析发现,研究区内洪积土、水稻土、塿土、立茬土等土壤类型分布区,以及地势较低、坡度在0~8°之间、坡向为东坡、东南坡,等水热组合条件良好的区域养分含量较高。该研究可以为土壤资源合理利用、土地的科学管理以及测土施肥技术提供重要依据。     

广义线性地统计模型在典型亚热带丘陵区数字土壤制图中的应用

数字土壤制图(DSM)在当前的应用越来越广泛,其方法主要包括环境相关模型、空间自相关模型,以及这些模型的混合模型。理论上,混合模型相对单一模型具有明显的优势。广义线性地统计模型(GLGM)也是一种混合模型,相对于最常用的混合模型——回归克里格(RK),又具有能加入随机效应来解决土壤变异的非平稳性等优势。然而,GLGM因计算繁琐等缺点,在国内外应用较少。本文以广西南宁高峰林场内一小面积(3.03km2)丘陵为研究区,以14个地形因子为预测变量,使用GLMM及其与普通克里格(OK)相结合的模型(GLGM),对土壤有机碳(SOC)、pH、黏粒和阳离子交换量(CEC)的空间分布进行预测,并与常用的多元线性回归(MLR)、地理加权回归(GWR)、回归森林(RF)、普通克里格(OK)、RK和广义可加模型(GAM)进行比较。结果表明：GLMM在预测黏粒上准确度较高;GLMM和GLGM在预测CEC上准确度中等;GLMM和GLGM在预测SOC和pH上准确度较低。综合线性回归模型的调整决定系数、块金效应和全局莫兰指数,本文认为,当土壤属性与环境变量具有较低的线性回归调整决定系数(即小于5%)、土壤属性具有较弱的空间自相关性(即块金效应大于71%)和较强的局部空间变异(即全局莫兰指数小于0.09)时,GLMM和GLGM具有较高的适用性,例如本文中的黏粒。反之,GLMM和GLGM的适用性不好,例如本文中的SOC和pH。鉴于土壤空间变异的高度异质性和多尺度性,本文认为GLMM和GLGM具有较好的应用前景。但是,在未来的研究中还需要进一步研究如何提高GLMM和GLGM的模拟效率。     

基于土壤-地形关系的合理样点对土壤有机质空间预测的影响

针对土壤精细化管理体系中合理样点数及空间预测优化问题,本文将基于土壤-地形关系,探讨了不同采样方式下以局部样点数量为代表提供最优数据的可能性。以地统计学、土壤-地形关系、地理加权回归克里格(GWRK)模型为基础,经系统格网、地形单元分区和地形起伏度最佳统计单元等三种采样方式分析合理样点集的空间分布对土壤有机质空间预测精度的影响。结果表明:(1)确定地形起伏度最佳统计单元大小为10×10像元,且平原区样点分布最为密集,合理样点数为1656;(2)高程、坡向、地形位置指数、相对位置指数、地形起伏度是影响土壤有机质空间变异的主要因素,能够解释研究区内有机质含量空间变异的69.2%;(3)GWRK模型精度均比普通克里金插值(OK)精度高,且山脊、背坡、陡坡、坡脚等坡位内合理样点数分别为39、481、9、28。在样点数最多时(n=2806),GWRK精度提高幅度及样点数量对预测结果影响有限。当样点数量减少时,有机质预测值空间分布的局部变异性随样点数减少而减少。因此,不同采样方式下合理样点集明显影响有机质预测精度,但预测结果分布趋势相似,仍可完整表征土壤有机质空间分布的空间格局。     

基于协同克里格插值和地理加权回归模型的土壤属性空间预测比较

选取宜昌市红花套镇作为研究区域,研究土壤pH、有机质、有效磷、速效钾、碱解氮与土壤属性指标变量之间的关系,选择与预测变量之间具有较高相关性的变量作为辅助变量用以提高预测精度,本文试图将地理加权回归模型应用于土壤属性空间模拟中,以此与协同克里格插值的预测结果进行对照,从而比较它们的预测精度以提出更适合土壤属性预测的模型。结果表明:协同克里格插值和地理加权回归模型对土壤属性的空间模拟均有较高的预测精度,在辅助变量较多的情况下地理加权回归模型具有比协同克里格插值更为简单的算法,并且比较预测值相对误差的范围跨度和标准差以及均方根误差等方面,地理加权回归模型在土壤属性指标预测方面具有更高的预测精度,也具有更大的优势。     

空间回归分析在土壤属性预测制图中的应用

以河南省孟津县为研究区,选取坡度、高程、地面曲率和复合地形指数(CTI)作为表层土壤缓效钾含量空间预测的环境协变量,系统探讨了空间回归分析技术在土壤属性预测制图中的应用.结果表明:土壤缓效钾的空间自相关距离阈值约为10 000 m,与坡度、高程和地面曲率存在显著相关性;尽管空间回归模型的预测精度和普通回归模型相近,但前者可以更加准确地表征土壤缓效钾的空间分布格局及空间分异细节特征.     

基于协同克里格插值和地理加权回归模型的土壤属性空间预测比较

选取宜昌市红花套镇作为研究区域,研究土壤pH、有机质、有效磷、速效钾、碱解氮与土壤属性指标变量之间的关系,选择与预测变量之间具有较高相关性的变量作为辅助变量用以提高预测精度,本文试图将地理加权回归模型应用于土壤属性空间模拟中,以此与协同克里格插值的预测结果进行对照,从而比较它们的预测精度以提出更适合土壤属性预测的模型。结果表明：协同克里格插值和地理加权回归模型对土壤属性的空间模拟都有较高的预测精度,在辅助变量较多的情况下地理加权回归模型具有比协同克里格插值更为简单的算法,并且比较预测值相对误差的范围跨度和标准差以及均方根误差等方面,地理加权回归模型在土壤属性指标预测方面具有更高的预测精度,也具有更大的优势。     

基于Landsat-8数据的土壤颜色预测方法研究

选择黑龙江省292个自然风干土壤样品,室内测定土壤高光谱反射率,然后依据三刺激值法计算土壤CIE XYZ色彩空间各颜色分量,用于土壤颜色预测和验证.同时,提取各样品采集点位的Landsat-8 OLI原始反射率数据,计算归一化差值植被指数、归一化差值水体指数、归一化差值湿度指数、归一化差值不透水面指数,并据此提出建模光...     

应用基于PLSR的土壤-环境模型预测土壤属性

土壤-环境模型对于正确理解土壤属性与环境因子间的关系,以及进行土壤属性预测与制图均具有重要的意义。研究区位于陕西省长武县内多年退耕还林还草沟壑区域,采集72个土壤表层样本,选择3/4的样本作为建模集,其余1/4的样本作为验证集;环境因子选择容易获取的地形因子和由遥感影像提取的植被因子和湿度因子,建立基于偏最小二乘回归(PLSR)的土壤-环境模型。结果表明:全氮、速效钾、全钾、有机质与环境因子间均有显著相关性;建立的PLSR模型可解释土壤属性的空间变异从23%(全氮)到27%(全钾);与逐步回归方法构建的模型相比,利用PLSR构建的土壤-环境模型可以更好地表征土壤属性与环境变量间的关系,拟合精度和预测精度也相对较高,说明PLSR建立的模型可以更好地应用于相似区域的土壤属性预测。     

基于HSI模型的土壤氮素空间预测不确定性的可视化表达

任何模型都无法完全准确地再现真实世界,因此各种空间预测始终面临着不确定性的困扰。从20世纪90年代开始,土壤空间预测研究中输出结果的不确定性分析与评价日益受到重视。由于不确定性信息的可视化表达有助于更加直观地揭示与展示真实世界,因此是一种增强不确定分析与决策支持之间联系的有效途径。本文首先对不确定性可视化表达的主要技术与方法进行了简要论述,进而以北京市城市边缘带样区基于地统计学的土壤表层全氮含量的空间预测作为案例研究,以相对误差表示预测结果的不确定性,应用色调-饱和度-亮度(HSI)色彩模型实现土壤属性预测结果和不确定信息的可视化同步表达。在HSI模型中,色调值被用来表示预测值,白度表示不确定性。在案例研究的全氮预测输出结果中,颜色发白的区域表示预测结果的不确定性较高,需要做进一步补充采样。实际应用过程中发现,二维图例对不确定性的可视化表达效果有一定的限制作用,仍需在二维图例完善或创建新型图例方面开展深入的研究工作。     

A multiscale soil–landform relationship in the glacial‐drift area based on digital terrain analysis and soil attributes

Understanding the linkages between structure and processes in soil landscapes involves analyses across several spatial and temporal scales. The transfer of information between scales requires the (1) identification of respective scale levels and (2) procedures for regionalization. Here, we present a multiparameter delineation of landform units and their attribution with typical Reference Soil Groups (RSG) of a landscape of NE Germany which is representative of young moraine regions. Data sources are a digital elevation model (DEM, 5 m × 5 m), a reference data set from sections of an intensively augered landscape, and expert knowledge. A conceptual digital soil map was constructed in the scale 1:5000 based on the Topographic Position Index (TPI). The methodology is applicable for multiscale analyses. Results are (1) the landform unit classified by digital terrain analysis of a DEM, (2) the attribution of RSG, and (3) the evaluation of the classification. Accuracy of the method was 57% overall, with 70% accuracy on typical erosional sites. The developed method allows identification of terrain‐related soil pattern with high spatial resolution in glacial‐drift areas. The high resolution of soil information can be used for delineation of management zones in precision farming, or as input for process studies and models requiring a translation of typological soil information into relevant soil properties (e.g., by pedotransfer functions).     

基于地形因子的土壤侵蚀评价指数修正 ——以陕北安塞集水区为例

地形是影响土壤流失的重要因素,小流域尺度上的土壤侵蚀评价指数(SL_(sw))对于地形的研究过于简单,因此本研究对SL_(sw)中的地形因子(G)进行修正,进而得出G因子修正后的土壤侵蚀评价指数(SL_(sw)_G)。以黄土丘陵沟壑区——安塞集水区为研究区,基于GIS技术,利用2006—2012年安塞集水区的数字高程模型(DEM)数据、水文观测站的日降雨量数据、土壤采样数据和土地利用数据,计算研究区2006—2012年逐年的SL_(sw),利用安塞集水区实测的输沙量,验证修正的SL_(sw)_G指数的有效性,并分析安塞集水区土壤侵蚀强度的时空变化特征。结果表明,修正前后土壤侵蚀情况分布基本相同,均呈现东南部地区SL_(sw)指数高,而西北部地区相对较低。在安塞集水区东南部,由于人为因素的影响,土壤保持措施相对较弱,同时土壤中黏粒和粉粒质量较轻,黏粒和粉粒湿时有明显的黏结性,降雨过程中容易被水冲刷;根据土壤采样点的采样数据表明,东南部地区土壤中黏粒和粉粒的含量较高,较容易造成土壤侵蚀。修正前后指数通过T检验得出,SL_(sw)和SL_(sw)_G具有方差齐次性,P=0.0370.05;通过与实测输沙量的比较,修正后SL_(sw)_G与实测输沙量的相关系数大于修正前SL_(sw)与实测输沙量的相关系数(0.390.36),修正后的相关性较修正前的相关性有所增加,说明修正是有效的,修正后的SL_(sw)指数能够更好地评价研究区域土壤侵蚀状况。     

基于地形属性使用回归, 克里格和人工神经网络方法估算饱和导水率空间分布

Several methods,including stepwise regression,ordinary kriging,cokriging,kriging with external drift,kriging with varying local means,regression-kriging,ordinary artificial neural networks,and kriging combined with artificial neural networks,were compared to predict spatial variation of saturated hydraulic conductivity from environmental covariates.All methods except ordinary kriging allow for inclusion of secondary variables.The secondary spatial information used was terrain attributes including elevation,slope gradient,slope aspect,profile curvature and contour curvature.A multiple jackknifing procedure was used as a validation method.Root mean square error (RMSE) and mean absolute error (MAE) were used as the validation indices,with the mean RMSE and mean MAE used to judge the prediction quality.Prediction performance by ordinary kriging was poor,indicating that prediction of saturated hydraulic conductivity can be improved by incorporating ancillary data such as terrain variables.Kriging combined with artificial neural networks performed best.These prediction models made better use of ancillary information in predicting saturated hydraulic conductivity compared with the competing models.The combination of geostatistical predictors with neural computing techniques offers more capability for incorporating ancillary information in predictive soil mapping.There is great potential for further research and development of hybrid methods for digital soil mapping.     

Mapping soil salinity and pH across an estuarine and alluvial plain using electromagnetic and digital elevation model data

Increasing pressures from agriculture and urbanization have resulted in drainage of many floodplains along the eastern Australian coastline, which are underlain by sulphidic sediments, to lower water tables and reduce soil salinity. This leads to oxidation of the sediments with a rapid decline in pH and an increase in salinity. Accurately mapping soil salinity and pH in coastal acid sulphate soil (CASS) landscapes is therefore important. One required map is the extent of highly acidic (i.e. pH < 4.5) areas, so that the application of alkaline amendments (e.g. lime) to neutralize the acid produced can be specifically targeted to the variation in pH. One approach is to use digital soil mapping (DSM) using ancillary information, such as an EM38, digital elevation models (DEM – elevation) and trend surface parameters (east and north). We used an EM38 in the horizontal (EM38h) and vertical (EM38v) modes together with elevation data to develop multiple linear regressions (MLR) for predicting EC_1:5 and pH. For pH, best results were achieved when the EM38 EC_a data were log‐transformed. By comparing MLR models using REML analysis, we found that using all ancillary data was optimal for mapping EC_1:5, whereas the best predictors for pH were north, log‐EM38v and elevation. Using residual maximum likelihood (REML), the final EC_1:5 and pH maps produced were consistent with previously defined soil landscape units, particularly CASS. The DSM approach used is amenable for mapping saline soils and identifying areas requiring the application of lime to manage acidic soil conditions in CASS landscape.     

Final Analysis of the Accuracy and Precision of Methods to Calculate the Sediment Retained by Check Dams

Knowing how much sediment check dams have trapped during their lifespan is essential to estimate their effectiveness or the sediment yields of their basins. Methods to calculate the volume of sediment trapped by check dams play an important role in the understanding of these issues. Several authors have proposed different methods to measure the volume of retained sediment, but their accuracy has, as yet, not been precisely determined and is currently a subject of debate. We compare the most frequently used methods (geometric: prism, pyramid and geometric; topographic: Digital Terrain Models (DTMs), trapezoids and sections methods) to evaluate their accuracy and precision in determining the volume of sediment retained by check dams. Our calculations are based on ten virtual check dams simulated in several gullies of Saldaña (Spain), where we determine their volumetric capacity for trapping sediment (real volume). This was made by means of an intensive topographic survey of these gullies, employing a terrestrial laser scanning system to obtain a high‐resolution digital elevation model (5 × 5 cm, ±2 mm). The results showed that topographic methods provided a very good fit to real volume with a difference of around 8%, being the sections method the most accurate. Geometric methods were less accurate, showing differences of up to 28%. Thus, the results obtained until now by geometric methods should be considered with caution. Although topographic methods were more accurate, they require many field data and much time than the others. So geometric methods can still be useful by correcting their results using our obtained percentage of variation. Knowing the accuracy of the methods before measuring is essential to obtain the most reliable results to analyse the role of check dams in controlling sediment, erosion processes and land degradation. Copyright © 2017 John Wiley & Sons, Ltd.     

流域模拟模型与GIS集成研究现状与展望

综述了流域模拟模型的现状和发展.从数据集成的角度,把流域模拟模型与地理信息系统集成方法归纳为互连式、共连式、流域模拟模型嵌入GIS和GIS嵌入流域模拟模型四种类型,讨论了各类集成方法的技术特点、应用和优缺点;并从数据、GIS技术、流域模拟模型和研究人员四个方面分析了流域模拟模型与GIS集成存在的问题;最后指出基于数字流域的综合集成方式将是流域模拟模型与GIS集成的发展方向.     

GIS在水土保持规划设计中的应用

地理信息系统(G IS)作为集计算机科学、地理学、环境科学、城市科学、空间科学、信息科学和管理科学为一体的学科,可为水土保持规划设计与管理提供强有力的技术手段支持。在介绍了G IS在水土保持规划设计中的应用背景和目标的基础上,详细介绍了数字高程模型和时态G IS这两项关键技术在水土保持规划设计中的应用。阐述了如何根据数字高程模型自动生成小班的高程、坡度和坡向,以及利用时态G IS对土地利用规划情况、规划治理措施完成情况等进行动态模拟分析的原理。     

The study on method to draw large-scale soil map in diluvial soil area in Hokkaido, using digital elevation data for a 10 m grid

(pp. 59–66)
In order to understand the detailed soil distribution of a terrace with a diluvial deposit, the method to draw a large-scale soil map was studied in Urausu Town of Sorachi district by combinig soil survey and digital elevation data for a 10 m grid.
The results are summarized as follows:

• 1) 

  From a soil survey, soils in the research area were classified into 3 soil series groups according to Classification of Cultivated Soils in Japan, Third Approximation, which were Skeletal Terrace Brown Forest Soils, Fine-textured Aquic Brown Forest Soils and Fine-textured Haplic Gray Upland Soils. In addition, Fine-textured Haplic Gray Upland Soils · were classified into 2 categories based on the abundance of gravel in the subsoil.

       
  
  

融合无人机高分辨率DOM和DSM数据语义的崩岗识别

崩岗识别是大规模的崩岗调查、治理和侵蚀机理等研究的首要任务,但以现场调查和人工解译高分辨率卫星遥感影像为主的传统方法,自动化程度低,人力、物力成本高而且效率低,不能满足大范围工作的需要。该研究借鉴遥感影像场景分类识别思路,利用视觉-地形词袋模型（Bag of Visual-Topographic Words, BoV-TW）进行崩岗区域高分辨率数字正射影像图Digital Orthophoto Map（DOM）与数字表面模型Digital Surface Model（DSM）局部特征的混合表达,并通过潜在狄利克雷分配Latent Dirichlet Allocation（LDA）的潜在语义分析融合形成低维度的高层语义表征,最后以支持向量机（Support-Vector Machine, SVM）作为监督学习训练分类器,实现崩岗的高精度快速自动识别。试验结果表明：1）LDA主题个数变化时,该方法总精度可保持在95%左右,崩岗查全率和查准率保持在80%以上,最高分别为97.22%和94.44%;2）视觉-地形词袋词汇表大小变化时,该方法总精度一直在90%以上,最高为96.10%,崩岗查全率也基本在90%以上,最高为100%,崩岗查准率随词汇表大小的增加逐渐提升,最高为85.00%;3）仅使用DOM无法较好地识别崩岗地貌特征,没有合适的特征提取和融合策略,DOM和DSM结合也无法提高崩岗识别效果。同时,该方法时间花费少,效率高,可行性强。该研究可为崩岗调查、监测、治理及侵蚀机理的定量化研究提供参考。