Spatial variability of soil fertility in a block design .

Abstract

n the characterization of the initial state of a long-term experiment (ELD) under no-till, we hypothesized that there are horizontal and vertical fertility gradients in the topsoil. The objective of this paper is to describe and to quantify those variations in a block design wheat fertilization experiment in the field of the National University of the South (UNS), Argentina. In a slightly inclined field 4 blocks were located across the slope. Soil depth was 80 cm at the top and increased downward to 100 cm. The soil was classified as a Petrocalcic Paleustoll, thermic coarse loam. Composite samples of each block were taken at the depths of 0-6, 0-12, and 0-20 cm. They were air-dried and sieved through 2 000 µm. Determinations were: pH; electrical conductivity (CE); cation exchange capacity (CIC); exchangeable cations: calcium (Ca<sub>int</sub>), magnesium (Mg<sub>int</sub>), potassium (K<sub>int</sub>) and sodium (Na<sub>int</sub>); extractable nutrients: phosphorus (P<sub>BK</sub>), sulphur (S-SO<sub>4</sub><sup>2-</sup>), boron (B<sub>ext</sub>), iron (Fe<sub>ext</sub>), manganese (Mn<sub>ext</sub>), copper (Cu<sub>ext</sub>) and zinc (Zn<sub>ext</sub>). In samples sieved through 500 µm total organic carbon (COT) and nitrogen (N<sub>ot</sub>) were determined. Walkley & Black (W-B) was used for analysis of easily oxidizable organic carbon (COX<sub>24</sub>), lower H<sub>2</sub>SO<sub>4</sub> concentrations of 12 and 18 N were used for further determinations (COX<sub>12</sub> and COX<sub>18</sub>). Results were analyzed using descriptive statistics, t test for paired means and regressions. Soil organic carbon variables and oxidation factor for W-B were consistent with low intensity use of native grasses in the plot selected as a starting point for the ELD. The range of pH values (7.3-8.0) reflected base saturation, with ideal levels of Ca<sub>int</sub>, Mg<sub>int </sub>and K<sub>int</sub>. Nutrient availability was high for S-SO<sub>4</sub><sup>2-</sup>, B<sub>ext</sub>, Fe<sub>ext</sub> and Mn<sub>ext</sub>, moderate for Cu<sub>ext</sub> and Zn<sub>ext</sub> and low for P<sub>BK</sub>. The coefficient of variation (CV) was between 7 and 11 % for ten variables. The most stable were pH, Mg<sub>int</sub>, CIC, C:N and Cu<sub>ext</sub> (CV < 7 %). Others (COX<sub>12</sub>, N<sub>ot</sub>, Mn<sub>ext</sub> and Zn<sub>ext</sub>) showed greater variability (CV 11-20 %) and P<sub>BK</sub> (CV 33.9 %). The statistical treatment of the space variability was adapted to the block design and sampling depths. In 12 out of 20 parameters, sampling depth and/or the position of the blocks explained over 30 % of the observed variations. The changes in depth of P<sub>BK</sub>, K<sub>int</sub> and cation micronutrients (with the exception of the Cu<sub>ext</sub>) can be associated to their scarce mobility in the soil. The position in the slope with potential effects on available water and biomass production, would explain the gradients in CIC, Ca<sub>int</sub> and organic carbon. The opposite trend found for P<sub>BK</sub> is attributed to its negative non-significant relation with pH. This variability study approach can enhance monitoring of soil chemical properties in the ELD.