POLYNOMIAL FUNCTIONAL MODELS TO SIMULATE CROP GROWTH IN MAIZE (ZEA-MAYS L) CULTIVARS

Polynomial and polynomial exponential models of order one to four were tried to describe crop growth in two maize cultivars viz., Deccan hyb rid and Deccan 101 of an agronomic trial conducted at the University o f Agricultural Sciences, Bangalore, on alfisols. Relative growth rate (RGR) of classical and functional methods was compared. Models perform ance was assessed based on r2, chi2, residuals sum of squares (RSS), r oot mean standard deviation (RMSD) and presence or absence of heterosc edasticity in models' predictions. In the cultivars, polynomials of or der three and four in Deccan hybrid and cubic polynomial in Deccan 101 simulated dry matter production (DMP) by 99.5 % and showed good distr ibution of error variance over the observed DMP. The RGR from the thir d and fourth degree polynomials in both the cultivars was comparable t o the classical method and explained better during post-silking period . Polynomial exponentials predicted very low RGR. Between cultivars, D eccan 101 showed higher RGR during 45-80 days after sowing and contrib uted significantly more to final DMP.