Nutrient Balances of New Zealand Kiwifruit (Actinidia deliciosa cv. Hayward) at High Yield Level

Abstract

The productivity of kiwifruit (Actinidia deliciosa) orchards varies widely in New Zealand. We hypothesized that such variation was attributable in part to nutrient imbalance. However, nutrient imbalance is currently diagnosed using critical nutrient ranges that are noisy and biased. Unbiased diagnosis can be performed using the isometric log ratio (ilr) technique for ad hoc balances. Our objective was to present, calibrate, and validate ilr balance standards to diagnose nutrient problems in New Zealand kiwifruit orchards. We collected leaf analytical data [nitrogen (N), sulfur (S), chloride (Cl), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), boron (B), copper (Cu), zinc (Zn), manganese (Mn), and iron (Fe)] in 433 kiwifruit orchards grown under conventional or organic farming in various agroecological zones of the North Island of New Zealand. Nutrients were arranged into sound balances illustrated by a mobile-fulcrums-buckets metaphor with balance sliders at fulcrums (domain of balances where statistics are computed) and concentrations in buckets (domain of concentrations where nutrient levels are examined relatively to each other). We developed optimum ranges of ilr balances to reach high yield levels and computed a nutrient imbalance predictor as the Mahalanobis distance. The critical Mahalanobis distance was 4.45 ± 0.13 at high yield level (>44 049 ± 334 kg ha−1), averaged from a five-fold cross-validation test; the test performance was 85 percent using the receiver operating characteristic curve, indicating that the test was informative for diagnostic purposes. After assessing relative nutrient levels in the concentrations domain, the most limiting nutrients appeared to be Cl, Mg, and Fe in poor-yield imbalanced specimens and Cl and S in organic orchards. Indeed, kiwifruit has high demand for Cl. Balance standards should be further tested in field trials.