TY  - JOUR
T1  - Geostatistical Analyses of Soil Electrical Conductivity in a Vegetable Greenhouse Field with Different Data Sets
AU - , Yong Jiang AU - , Wei Hao AU - , Yuge Zhang AU - , Wenju Liang 
JO  - Environmental Research Journal
VL  - 2
IS  - 3
SP  - 125
EP  - 130
PY  - 2008
DA  - 2001/08/19
SN  - 1994-5396
DO  - erj.2008.125.130
UR  - https://makhillpublications.co/view-article.php?doi=erj.2008.125.130
KW  - Soil electrical conductivity
KW  -geostatistics
KW  -spatial variability
KW  -vegetable greenhouse
AB  - The study was conducted at Damintun town, Shenyang city, Liaoning Province of China, in order to analyze the spatial variability of soil Electrical Conductivity (EC) under greenhouse vegetable plantation and to compare the differences of the spatial variability using different data sets. A micro area of 8×5.7 m with 0.4×0.3  m regular rectangle grids subdivision was chosen and totally 420 points of soil EC was <I>in-situ</I> measured using a W.E.T.- sensor. The total data was named as data sets A and was divided into 3 groups that named as data sets B, C and D. The results showed that the mean, minimum, maximum, percentiles 25, median and percentiles 75 values were a little different among the four data sets. The histograms for soil EC showed that the data fitted normal distribution for each data sets. Soil EC was spatially dependent and modeled quite well with different data sets. The anisotropic semivariograms showed that the structural component of sample variance for data sets B and D was highly spatial dependence with the ratio of C/(C<SUB>0</SUB>+C) &gt;75%, while data sets A and C exhibited middle spatial dependence. The isotropic semivariograms showed that all the data sets having middle values of nugget effects and 50% of the C/(C<SUB>0</SUB>+C), indicated that the greenhouse vegetable plantation have led to sample dissimilarity in the small sampling distance within a small area. The maps obtained with kriging were quite similar with different data sets. Although, the smoothing effect existed with all the four data sets, kriging remains the best local estimator when the data is reduced. The smoothing effect of kriging can be supplemented with classical statistics. It is concluded that geostatistics combined with classical statistics is an ideal way to examine the spatial variability of soil properties in a micro- field scale. Because the soil EC was high in this study, it is suggested that measures be taken to avoid the accumulation of soil salt in greenhouses by applying fertilizers rationally according to soil fertility, vegetable varieties and fertilizer properties in the study region.
ER  -