The application of geostatistics in grain size trend analysis： A case study of eastern Beibu Gulf

There are 71 surface sediment samples collected from the eastern Beibu Gulf. The moment parameters （i.e. mean size, sorting and skewness） were obtained after applying grain size analysis. The geostatistical analysis was then applied to study the spatial autocorrelation for these parameters; while range, a parameter in the semivariogram that meters the scale of spatial autocorrelation, was estimated. The results indicated that the range for sorting coefficient was physically meaningful. The trend vectors calculated from grain size trend analysis model were consistent with the annual ocean circulation patterns and sediment transport rates according to previous studies. Therefore the range derived from the semivariogram of mean size can be used as the characteristic distance in the grain size trend analysis, which may remove the bias caused by the traditional way of basing on experiences or testing methods to get the characteristic distance. Hence the results from geostatistical analysis can also offer useful information for the determination of sediment sampling density in the future field work.

Mapping of the Water Table Levels of Unconfined Aquifers Using Two Interpolation Methods

The spatial prediction of the water table can be used for many applications related to civil works (foundations, excavations) and other urban and environmental management activities. Deterministic and geostatistical interpolation methods were used to predict the spatial distribution of water table levels (unconfined aquifers) of important geological formations of the Joao Pessoa City (capital of Paraiba State, Brazil) with dense urban occupation and high demand for new civil works. The deterministic (topo to raster) and geostatistical (ordinary kriging) interpolation methods were evaluated using a Geographic Information System (GIS)-based investigation. The water table levels were obtained from 276 boring logs of Standard Penetration Test (SPT) in situ investigation distributed over the geological formations studied (an area of 59.8 km2, covering 40 districts of the Joao Pessoa City). The Nspt values and textural characterization data are stored for levels of 1 m depth. Some boreholes located in the area investigated were not included in the interpolation processes in order to be compared with estimated values (validation of the results). Maps of the water table depths were also produced to further analyze the quality of the water table surfaces interpolated by both methods. The phreatic surface interpolations provided satisfactory results for both methods (RMSE = 1.8 m). The topo to raster method showed a slight general tendency to be less affected by local values in relation to the kriging method and also has the advantage of integrating the drainage flow system, which is a relevant aspect for spatial models of the water table levels of unconfined aquifers. The ordinary kriging (geostatistical method) provides a prediction surface and some measure of the certainty or accuracy of the predictions.

乌鲁木齐市雅玛里克山污水灌溉土壤肥力的空间变异研究

本文选取了干旱区有代表性的污水灌溉地区-乌鲁木齐市雅玛里克山，结合GIS，采用地统计学方法研究了污水灌溉区域土壤中的pH、有机质、CEC、总盐、全N、全P、全K7种土壤肥力指标的空间变异陛，并采用普通克立格法绘制了各种土壤肥力要素含量的空间分布格局图。结果表明：有机质、总盐、全N、全K的理论模型为指数模型；CEC、全P的理论模型为高斯模型；pH的理论模型为球状模型；pH、有机质、总盐、全N、全K的块金值与基台值之比都在25％以下，说明空间自相关作用强，其含量分布受结构性因素控制；CEC、全P的块金值与基台值之比在25％～75％之间，为中等强度的空间相关性，其分布是由结构性因素和随机性因素共同作用的结果。土壤肥力的相关距变化范围为246．0～2100．0m。