Model Comparisons for Assessment of NPK Requirement of Upland Rice for Maximum Yield

Upland rice farmers in Malaysia still depend on resultant ash from burning for K and N sources. Efficient use of chemical fertilizers in upland rice needs accurate assessment of required nutrient elements. The present study was performed to determine the N, P, and K requirements of three upland rice varieties grown on idle land (Bukit Tuku soil, AQUIC KANDIUDULT) using four response models. A glasshouse experiment was conducted using 0-200 kg N ha-1 (urea, 46%N), 0-120 kg P2O5 ha-1 (TSP, 45% P2O5), and 0-150 kg K2O ha-1 (MOP, 60% K2O), each at five levels. Three upland rice varieties used in the experiment were Ageh, Kendinga and Strao. The grain yield (14% moisture content) was measured at harvest and fitted using linear (L), linear with plateau (LP), quadratic (Q), and quadratic with plateau (QP) response models. The QP proved itself as the best fitted response model for the determination of fertilizer recommendation rates for maximum yield of upland rice cultivars used. The fertilizer rates were 112 kg N ha-1, 78 kg P2O5 ha-1 and 158 kg K2O ha-1 for Ageh (QP); 138 kg N ha-1 (LP), 87 kg P2O5 ha-1 (QR), 119 kg K2O ha-1 (QP) for Kendinga; and 125 kg N ha-1 (Q), 85 kg P2O5 ha-1 (LP) and 127 kg K2O ha-1 (L) for Strao