Spatial variability of soil boron availability and its impact on canola yield

Abstract

Soil boron (B) is influenced by factors including soil pH, texture and organic matter (OM). It was hypothesized that response of canola (Brassica napus) to B fertilization similarly may be related to these factors. The objectives of this study were to examine the spatial variability of available soil B (i.e., hot water soluble B (HWS-B» and identify soil chemical and physical factors controlling both available soil Band consequent canola yield response to applied fertilizer B. The spatial variability of HWS-B and canola response to B fertilization was assessed at two sites identified as having marginal (Carrot River) to deficient (Smeaton) levels ofHWS-B. At each site, a single 128-point linear sampling transect was established and soil samples, taken at 3-m intervals, were analyzed for various physical and chemical properties. Canola was seeded along the transects and three B treatments were imposed, i.e., an untreated control, 2 kg B ha' as surface applied Granubor'", and 0.5 kg B ha-I as foliar applied Solubor'". At crop maturity, seed yields were determined. Hot water soluble B was highly variable and was associated with, in decreasing order of importance: OM » pH > texture> inorganic C. Soil B availability increased with increasing OM. The effect of soil pH on B availability varied. At Smeaton, where acidic soils dominated, HWS-B increased with increasing soil pH whereas at Carrot River, where alkaline soils dominated, HWS-B levels typically decreased with increasing soil pH. Canola yield responses to B fertilizer were highly variable and the HWS-B test failed to predict responses. Although yield responses were weakly correlated with soil pH, correlations were inconsistent between sites and treatments. Apparently, response of canola to B fertilization was controlled by a suite of soil factors controlling B availability, and these factors varied both within and between sites. Wavelet analysis revealed that at spatial scales greater than 55 m, variability in HWS-B and OM were statistically significant. The spatial scale of the HWS-B variability suggests that successful evaluation of crop responses to B fertilization may be compromised if variability in soil B availability is not recognized and accounted for.