DETERMINATION OF SPATIAL VARIABILITY OF SOME MAGNESIUM FORMS IN PHAEOZEM USING GEOSTATISTICAL METHODS

The spatial variability analysis of soil properties facilitates the prediction of their contents across various sites, useful for design of site-specific farming practices. Intrapopulation and spatial variation of soil available (Mg-A), exchangeable (Mg-E) and water-soluble (Mg-H2O) magnesium forms was evaluated in a field on Phaeozem soil, located in the village Orlinek near Mrocza, in the Province of Kujawy and Pomorze (wojewodztwo kujawsko-pomorskie), north-western Poland. Soil samples were collected from an area of 0.5 ha situated within an 80-hectare arable field. In April 2007, 50 soil samples were collected in a 10 x 10 m grid square pattern from the field cropped with winter wheat. The content of Mg forms was analyzed with descriptive statistics and the geostatistic modeling of semivariograms to plot variability maps. The Mg-A content ranged from 4.9 to 12.2 mmol kg(-1), while the Mg-E form varied from 5.6 to 16.4 mmol kg(-1). The average content of Mg-H2O was 1.27 mmol kg(-1). All the properties revealed a normal content distribution, which coincided with similar values of means, medians and significantly lower values of standard deviations (SD) than the means. Moderate variability of all the Mg forms content was confirmed by the coefficient of variation (CV%), falling within 19.4-23.6%. The spatial dependence of the Mg forms content was evaluated by the use of semivariograms and krigged maps. The parameters of variogram models, except for the Mg-A content, revealed a share of random variance (a nugget) in total variability (sill). The content of Mg-E demonstrated high spatial dependence (the nugget effect <25%), while the content of Mg-H2O and the percentage share of Mg-E in the sum of base cations (S) fell within the moderate class of spatial variability (the nugget effect between 25% and 75%). The spatial correlation of the properties studied was assessed by 11.1 to 21.3 m range. The spatial variation maps showed that the Mg-A and Mg-E contents had a similar distribution in the research area, while the Mg-H2O content and the percentage of the Mg-E in (S) presented a different pattern of variability. The maps revealed that almost 80% of the field showed a very high Mg-A content (Class I abundance), which suggested that Mg fertilizer is unnecessary.