The use of measurements of soil mineral N in understanding the response of crops to fertilizer nitrogen in intensive cropping rotations

Abstract

The effects on succeeding crops of nitrogen in residues returned to the soil of brassica vegetable crops (Brassica oleracea) were studied on a nitrogen-retentive silt loam soil at Horticulture Research International, Kirton, Lincolnshire, UK. A sequence of four crops was started in 1988 and again in 1989. In the first sequence, two successive cauliflower crops (Brassica oleracea cv. botrytis L.), crops 1 and 2 in the first year, were followed by Brussels sprouts (Brassica oleracea cv. Gemmifera D.C.), crop 3, in the second and spring barley (Hordeum vulgare), crop 4, in the third year. The second sequence, started in spring 1989, was on an area adjacent to the first, but with spring wheat (Triticum aestivum) as crop 4. The sites followed an unfertilized, 1-year grass ley (1987) or spring barley (1988) with 73 and 107 kg N/ha soil mineral nitrogen (SMN, NH₄+NO₃) in the 0–90 cm soil profile at the start of each sequence. The marketable yield of the first cauliflowers in both sequences increased with fertilizer nitrogen up to 240 or 300 kg/ha. The response of the second cauliflower crop to fresh N declined with increasing amounts of SMN (0–90 cm) at planting, with no response when SMN exceeded c. 270 kg N/ha. Crop 3 apparently responded to fresh N in sequence 1 but not in sequence 2 even though SMN at planting ranged up to c. 400 kg N/ha. This difference in response was largely explained by the amount and distribution of SMN in the 0–30 and 30–60 cm layers when the sprouts were planted. Soil mineral N to 90 cm when the cereals were sown was only c. 100 kg N/ha, which did not reflect the large amounts of N applied to the previous crops. In spite of this, barley yield without any fresh fertilizer N did vary with the amounts of N applied to the previous crops. The results showed that SMN was a useful predictor of fertilizer response in some, but not all, situations. To use SMN more generally requires interpretation using dynamic simulation models.