论文部分内容阅读
This study focused on many scattered small paddy fields generated by real-time soil sensor and GIS software.The experimental site was a commercialfarm of Aguri co.Ltd.composed of 400 paddy fields of 0.1-0.4 ha in size.Within-field and between field variability was discussed.The fundamental to measure the soil properties through soil-sensor is based upon reflectance spectroscopy which provides a rapid and nondestructive method for measurement according to diffusely reflected radiation of illuminated soil.The spectra data and GIS data of each of 40 fields have been collected soil sensing by the real-time soil sensor.When the real-time soil sensor gathering continuous reflectance data from the target fields, a small number of soil samples from some of the same position were also collected for chemical analysis using conventional method in order to obtain the value of the relevant soil properties.Total nitrogen, one of the soil parameters, was focused to generate maps.After the pretreatment of data, partial least squares regression was employed to predict the soil parameter, within which the best performed one in terms of cross validation error was selected for usage in making standard curves.Prediction models composed of the value of a number of soil properties in all points where the real-time soil sensor measured was provided by the calibration of spectra data with the results of analysis of soil samples.To generate maps, the software of ArcMap was applied as a major approach to transform the calculated soil properties results into soil maps.ArcMap is a main component of ArcGIS suite of geospatial processing programs produced by ESRI of US, and it is used primarily to view, edit,create, and analyze geospatial data.In this study, two types of scale of maps,soil maps of total nitrogen of each field composed the predicted data and a coefficient of variation map of all field,were generated.Overlapping these maps created thelayered structure map visualized variability in multi-scale.Interesting results came from the map.Variability of field is depending on scale.