Spatial Variability of Available Nutrients and Soil Carbon under Arable Cropping in Canterbury

Soil properties can vary spatially at a paddock to farm scale due to differences in topography, parent material, vegetation or land management. The objective of this study was to quantify the spatial variability of available nutrients and organic matter within a 10.4 ha paddock (predominantly Templeton silt loam; flat topography) at Lincoln that had a long-term history of arable cropping. Samples (0–7.5 cm depth) were collected in a grid pattern (30–35 m sampling intervals; total of 91 samples) for determination of mineral N, anaerobically mineralisable N (AMN), Olsen P, total C and total N. The data were evaluated using geostatistical as well as classical statistical methods, and a geographic information system (GIS) technique was used to produce nutrient and organic matter maps. Although the paddock had a relatively flat topography and had been managed uniformly for many years, nutrient levels exhibited substantial variability. For example, Olsen P ranged from 14 to 53 μg/g (mean 20 μg/g) and AMN ranged from 37 to 83 μg/g (mean 50 μg/g). Soil organic matter also showed significant spatial variability, with total C ranging from 19 to 31 g/kg (mean 27 g/kg). All measured variables except mineral N showed moderate to strong positional dependence. Semivariogram models showed that for Olsen P the autocorrelation distance was 700 m, compared with 184 m for AMN, and 300 m for total C. We examined the relationship between soil texture and organic matter using samples collected along two perpendicular transects within the paddock. Soil C showed a strong positive correlation (R 2 = 0.79; n = 17) with the amount of clay plus fine silt (<5 μm fraction) and a negative correlation (R 2 = 0.81) with sand content. These results suggest that textural variation was a major factor influencing within-paddock variability in soil organic matter at this study site. Introduction Soil properties can vary spatially due to a variety factors, including climate, parent material, topography, vegetation, and land management (Trangmar, 1986; West, 1989; Weijun, 2010). Soil characteristics and productivity can vary significantly over small spatial scales, e.g. within paddocks (Fulton et al., 1996; Wells et al., 2000; Gaston et al., 2001). Soil fertility assessments commonly rely on a random soil sampling protocol to obtain an average fertility value for a paddock. Thus, spatial variability is ignored and, consequently, some parts of the paddock may receive excessive fertiliser while other parts may suffer nutrient deficiency due to under-application. Geostatistics coupled with geographic information system (GIS) technologies has substantial potential to discern spatial patterns within paddocks. In this study we used geostatistics and GIS to map and interpret the spatial variability of nutrients and soil organic matter at a paddock scale. A second objective was to identify factors contributing to within-paddock variability in the measured variables, especially soil organic matter.