Regression Model for Indirect Estimation of Oil Palm Frond Area

Authors

  • M. A. Awal Associate Professor, Department of Farm Power and Machinery , Bangladesh Agricultural University, Mymensingh-2202, Bangladesh
  • Mohd. Haniff Harun Research Officer, Biological Division, Malaysian Palm Oil Board, Bangi, Malaysia
  • Wan Ishak Professor, Department of Agricultural and Biological Engineering, UPM, Malaysia
  • J. Endan Associate Professor, Department of Food and Process Engineering, UPM, Malaysia

DOI:

https://doi.org/10.61361/jambe.v5i12.95

Keywords:

Leaf area, Frond area, Regression model, Oil palm

Abstract

The fronds (leaves) play an important role in oil palm's growth and production. In oil palm, frond area is one of the most important aspects in its morphological and physiological studies and also important indicator of oil palm's future production and yield. Precise estimation tion of frond area is one of the important issues for oil palm growth analysis. Frond area is also important for analysis of canopy structure of the oil palm. Accurate, rapid, non-destructive, simple and reliable approach for determination of frond area in oil palm is essential. Destructive method and portable leaf area meter were used to develop frond area model. In this study, regression models for accurate estimation of leaflet area from simple measured leaflet length and middle width were described and two models for frond area were also developed. The regression models for leaflet were, (1) A = 75.517D-45.411, with co-relation coefficient r = 0.91 and standard error 1.62. (ii) A = 28.845 D 1.633 with co-relation coefficient r = 0.93 and standard error 0.031 and (ii) A = 0.0286 L 20142, with co- relation coefficient r = 0.93 and standard error 0.26, where A, D and L represent area (cm³), width (cm) and length (cm) of the leaflets. The developed frond area models were, (a) TFA = 0.58 PLA with co-relation coefficient r = 0.99 and standard error of estimation coefficient .007431 and (b) TFA = 0.467 PRLA with co-relation coefficient r = 0.99 and standard error of estimation coefficient 0.00536 where, TFA, PLA and PRLA represents true frond area, predicted frond area and predicted rectangular frond area in m³. Statistical analysis indicates a high degree of association (R2 = 0.99) and the low standard errors of estimation were 0.007431 and 0.00536 for model 'a' and 'b' respectively. Model 'b' was more simple and easier in terms of frond area measurement.

Downloads

Published

2009-12-31

Issue

Section

Original Research