Digital mapping of soil properties in the West of Honduras, Central America.

Digital soil property maps were generated at 30 meters resolution for the West of Honduras in order to develop the AGRI v.1 tool (Monserrate et al., 2016). AGRI (from its Spanish words AGua para RIego) is a tool that combines information about climate, relief, soils, land cover, and hydrology to identify suitable water sources for implementing small irrigation projects. The soil properties mapped were sand (%), silt (%), clay (%), texture class, field capacity (v/v), wilting point (v/v), water holding capacity (v/v), and curve numbers. A database of 1887 points from González et al. (2008) were used to generate the maps of sand, silt, and clay. This database was also used to determine field capacity, wilting point and water-holding capacity for each point by applying pedotransfer functions according to Saxton & Rawls (2006). A regression kriging approach was performed by combining 80% of point data with the terrain attributes aspect, mid-slope position, normalized height, plan and profile curvature, slope and topographic wetness index generated from a digital elevation model SRTM of 30 meters resolution. The combination of sand, silt, and clay maps resulted on texture class map. The curve number was mapped using the texture and land cover maps according to Soil Conservation Service of the United States of America (USDA-SCS, 1985). The maps performance was evaluated by the normalized root mean square error (RMSEn) expressed in percentage and using 20% of data point not used for mapping. Clay, sand, silt, field capacity, water holding capacity and wilting point presented error of 16%, 17%, 13%, 19%, 10% and 18% respectively.

为开发AGRI v.1工具（Monserrate等，2016），本研究针对洪都拉斯西部地区生成了分辨率为30米的数字土壤属性图。AGRI工具（名称取自西班牙语AGua para RIego，意为灌溉用水）是一款融合气候、地形、土壤、土地覆被及水文信息的工具，用于识别小型灌溉项目的适宜水源。本次制图涉及的土壤属性包括：砂粒含量（%）、粉粒含量（%）、黏粒含量（%）、土壤质地类别、田间持水量（体积分数，v/v）、萎蔫点（体积分数，v/v）、持水能力（体积分数，v/v）以及径流曲线数。本研究采用González等（2008）构建的包含1887个采样点的数据库，生成砂粒、粉粒及黏粒含量分布图；该数据库同时被用于结合Saxton与Rawls（2006）提出的土壤传递函数，计算每个采样点的田间持水量、萎蔫点及持水能力。本研究采用回归克里金法进行空间插值：将80%的采样点数据与由30米分辨率航天飞机雷达地形测绘任务（SRTM）数字高程模型提取的地形属性（包括坡向、坡中位置、归一化高程、平面曲率、剖面曲率、坡度及地形湿度指数）相结合。通过整合砂粒、粉粒及黏粒含量分布图，最终生成土壤质地类别分布图。参照美国农业部水土保持局（USDA-SCS，1985）发布的标准，本研究基于土壤质地及土地覆被分布图生成了径流曲线数分布图。本研究采用归一化均方根误差（RMSEn，以百分比表示），并预留20%未参与制图的采样点对制图结果进行精度验证。结果显示，黏粒、砂粒、粉粒、田间持水量、持水能力及萎蔫点的归一化均方根误差分别为16%、17%、13%、19%、10%及18%。